Musculoskeletal Disorders and Workplace Factors

A Critical Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders of the Neck, Upper Extremity, and Low Back

Edited by: Bruce P. Bernard, M.D., M.P.H.

Contributors: Vern Putz-Anderson, Ph.D. Bruce P. Bernard, M.D., M.P.H. Susan E. Burt Libby L. Cole, Ph.D. Cheryl Fairfield-Estill Lawrence J. Fine, M.D., Dr.P.H. Katharyn A. Grant, Ph.D. Christopher Gjessing Lynn Jenkins Joseph J. Hurrell Jr., Ph.D. Nancy Nelson, Ph.D. Donna Pfirman Robert Roberts Diana Stetson, Ph.D. Marie Haring-Sweeney, Ph.D. Shiro Tanaka, M.D.

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Institute for Occupational Safety and Health

July 1997 DISCLAIMER

Mention of any company name or product does not constitute endorsement by the National Institute for Occupational Safety and Health.

This document is in the public domain and may be freely copied or reprinted.

Copies of this and other NIOSH documents are available from

National Institute for Occupational Safety and Health Publications Dissemination 4676 Columbia Parkway Cincinnati, OH 45226–1998

Telephone number: 1–800–35–NIOSH (1–800–356–4674) Fax number: (513) 533–8573 E-mail: [email protected]

To receive other information about occupational safety and health problems, call 1–800–35–NIOSH (1–800–356–4674), or visit the NIOSH Home Page on the World Wide Web at http://www.cdc.gov/niosh

This publication is also available from the National Technical Information Service for $60.00 plus $4.00 handling. Please call (703) 487–4650 and ask for PB97–178628.

DHHS (NIOSH) Publication No. 97B141

ii FOREWORD

Musculoskeletal disorders (MSDs) were recognized as having occupational etiologic factors as early as the beginning of the 18th century. However, it was not until the 1970s that occupational factors were examined using epidemiologic methods, and the work-relatedness of these conditions began appearing regularly in the international scientific literature. Since then the literature has increased dramatically; more than six thousand scientific articles addressing ergonomics in the workplace have been published. Yet, the relationship between MSDs and work-related factors remains the subject of considerable debate.

Musculoskeletal Disorders and Workplace Factors: A Critical Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders of the Neck, Upper Extremity, and Low Back will provide answers to many of the questions that have arisen on this topic over the last decade. This document is the most comprehensive compilation to date of the epidemiologic research on the relation between selected MSDs and exposure to physical factors at work. On the basis of our review of the literature, NIOSH concludes that a large body of credible epidemiologic research exists that shows a consistent relationship between MSDs and certain physical factors, especially at higher exposure levels.

This document, combined with other NIOSH efforts in this area, will assist us in our continued efforts to address these inherently preventable disorders.

Linda Rosenstock, M.D., M.P.H. Director, National Institute for Occupational Safety and Health Centers for Disease Control and Prevention

iii iv NOTE TO THE READER

This second printing of Musculoskeletal Disorders and Workplace Factors: A Critical Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders of the Neck, Upper Extremity, and Low Back incorporates a number of editorial changes, including grammar, formatting, and consistency issues that were identified in the first printing. In addition, the notation of Dr. Lawrence Fine as co-editor was inadvertently omitted in the first printing and has been re-inserted.

The conclusions of the document in terms of decisions regarding the weight of the existing epidemiologic evidence for the relationship between workplace factors and musculoskeletal disorders remain unchanged. The following technical inconsistencies or errors were corrected:

Page 2-14: Text was corrected to reflect that five studies (as opposed to three) examined the relationship between force and musculoskeletal disorders of the neck.

Page 2-28: For Viikari-Juntura [1994], the “NR” entry in the Risk Indicator column was replaced with the value 3.0.

Page 2-34: Bergqvist [1995a] was changed to Bergqvist [1994]. The Risk Indicator entry for this study was changed from 4.4 to 3.7 (both noted as statistically significant), the entry for Physical Examination was changed from “Yes” to “No,” and the entry for Basis for Assessing Exposure was changed from “job titles or self-reports” to “observation or measurements.”

Page 3-3: Text was corrected to reflect that four studies (as opposed to three) met all four evaluation criteria. A description of Kilbom and Persson [1987] was moved forward in the chapter to this section and includes a clarification that health outcome in their study was based on symptoms and physical findings.

Page 3-32: The confidence interval depicted for Ohlsson [1994] was corrected to show a range from 3.5 to 5.9.

Page 3-69: Schibye et al. [1995] was added to Table 3-5.

Page 4-25: Dimberg [1989] was changed to Dimberg [1987].

v Page 5a-3: Text was corrected to reflect that nineteen studies (as opposed to fifteen) reported results on the association between repetition and carpal tunnel syndrome (CTS). Text was also corrected to reflect that five studies (as opposed to four) met the four evaluation criteria for addressing repetitiveness and CTS. A description of Osorio et al. [1994] was moved forward in the chapter to this section.

Page 5a-15: Text was corrected to reflect that eleven studies (as opposed to ten) reported results on the association between force and CTS and that four (as opposed to three) met all four evaluation criteria. Descriptions of Moore and Garg [1994] and Osorio et al. [1994] were moved forward in the chapter to this section.

Page 5a-19 : The discussion (strength of association, temporality, consistency of association, coherence of evidence, and exposure-response relationship) of force and CTS was inadvertently omitted in the first printing and has been re-inserted.

Page 5a-27: The Risk Indicator for Osorio et al. [1994] was changed from 4.6 to 6.7, and for Nathan [1992], the “No association” entry under Risk Indicator was changed to a value of 1.0.

Page 5a-29: Stetson et al. [1993] was moved to the bottom of the table, and entries for Nathan et al. [1992] and McCormack et al. [1990] were added.

Page 5a-31: This table was modified to more accurately reflect the text.

Page 5a-33: For Koskimies et al. [1990], the entry for Basis for Assessing Exposure was changed from “observation or measurements” to “job titles or self-reports.”

Page 5b-1: Text was corrected to reflect that seven studies (as opposed to eight) are referenced on Table 5b-1.

Page 5c-4: Text was corrected to reflect that five studies (as opposed to four) met three of the criteria. A brief description of Kivekäs et al. [1994] was added to this section.

A number of references were clarified, and full references for studies that were cited in the text of the first printing but were inadvertently omitted from the reference list were added.

Appendix C was added to the document to provide a concise overview of the studies reviewed relative to the evaluation criteria, risk factors addressed, and other issues.

vi CONTENTS

Foreword...... iii Note to the Reader...... v Executive Summary...... x Acknowledgments...... xvi

Chapter 1. Introduction...... 1-1 Purpose...... 1-1 Background...... 1-1 Scope and Magnitude of the Problem...... 1-1 Cost...... 1-6 Defining Health Outcomes...... 1-7 Exposure Measurements...... 1-8 Information Retrieval...... 1-9 Selection of Studies...... 1-9 Methods for Analyzing or Synthesizing Studies...... 1-10 Criteria for Causality...... 1-11 Categories Used to Classify the Evidence of Work-Relatedness...... 1-13 Summary...... 1-14 Description of Tables, Figures, and Appendices...... 1-14

Chapter 2. Neck Musculoskeletal Disorders: Evidence for Work-Relatedness...... 2-1 Summary...... 2-1 Introduction...... 2-2 Repetition...... 2-3 Force...... 2-12 Posture...... 2-16 Vibration...... 2-21 Neck or Neck/Shoulder MSDs and the Role of Confounders...... 2-21 Conclusions...... 2-22 Tables and Figures...... 2-24

vii Chapter 3. Shoulder Musculoskeletal Disorders: Evidence for Work-Relatedness...... 3-1 Summary...... 3-1 Introduction...... 3-1 Repetition...... 3-2 Force...... 3-9 Posture...... 3-14 Vibration...... 3-23 Role of Confounders...... 3-25 Conclusions...... 3-25 Tables and Figures...... 3-27

Chapter 4. Elbow Musculoskeletal Disorders (Epicondylitis): Evidence for Work-Relatedness...... 4-1 Summary...... 4-1 Introduction...... 4-2 Repetition...... 4-2 Force...... 4-6 Posture ...... 4-16 Epicondylitis and the Role of Confounders...... 4-17 Conclusions...... 4-18 Tables and Figures...... 4-20

Chapter 5. Hand/Wrist Musculoskeletal Disorders (Carpal Tunnel Syndrome, Hand/Wrist Tendinitis, and Hand-Arm Vibration Syndrome): Evidence for Work-Relatedness...... 5-1

Chapter 5a. Carpal Tunnel Syndrome...... 5a-1 Summary...... 5a-1 Introduction...... 5a-1 Outcome and Exposure Measures...... 5a-2 Repetition...... 5a-3 Force and CTS...... 5a-14 Posture and CTS...... 5a-22 Vibration and CTS...... 5a-25 Confounding and CTS...... 5a-27 Conclusions...... 5a-28 Tables and Figures...... 5a-29

Chapter 5b. Hand/Wrist Tendinitis...... 5b-1 Summary...... 5b-1

viii Introduction...... 5b-1 Repetition...... 5b-2 Force...... 5b-9 Posture...... 5b-13 Conclusions...... 5b-17 Tables and Figures...... 5b-18

Chapter 5c. Hand-Arm Vibration Syndrome...... 5c-1 Summary...... 5c-1 Introduction...... 5c-1 Evidence for the Work-Relatedness of HAVS...... 5c-2 Tables and Figures...... 5c-10

Chapter 6. Low-Back Musculoskeletal Disorders: Evidence for Work-Relatedness...... 6-1 Summary...... 6-1 Introduction...... 6-2 Heavy Physical Work...... 6-4 Lifting and Forceful Movements...... 6-13 Bending and Twisting (Awkward Postures)...... 6-21 Whole Body Vibration (WBV)...... 6-26 Static Work Postures...... 6-34 Role of Confounders...... 6-38 Tables and Figures...... 6-39

Chapter 7. Work-Related Musculoskeletal Disorders and Psychosocial Factors...... 7-1 Summary...... 7-1 Introduction...... 7-1 Psychosocial Pathways...... 7-2 Upper-Extremity Disorders (Neck, Shoulder, Elbow, Hand, and Wrist)...... 7-3 Back Disorders...... 7-7 Tables and Figures ...... 7-11

References...... R-1

Appendices ...... A-1 A. Epidemiologic Review...... A-1 B. Individual Factors Associated with Work-Related Musculoskeletal Disorders (MSDs)...... B-1 C.Summary Tables...... C-1

ix EXECUTIVE SUMMARY

The term musculoskeletal disorders (MSDs) refers to conditions that involve the nerves, tendons, muscles, and supporting structures of the body. The purpose of this NIOSH document is to examine the epidemiologic evidence of the relationship between selected MSDs of the upper extremity and the low back and exposure to physical factors at work. Specific attention is given to analyzing the weight of the evidence for the strength of the association between these disorders and work factors.

Because the relationship between exposure to physical work factors and the development and prognosis of a particular disorder may be modified by psychosocial factors, the literature about psychosocial factors and the presence of musculoskeletal symptoms or disorders is also reviewed. Understanding these associations and relating them to the cause of disease is critical for identifying exposures amenable to preventive and therapeutic interventions.

MAGNITUDE OF THE PROBLEM The only routinely collected national source of information about occupational injuries and illnesses of U.S. workers is the Annual Survey of Occupational Injuries and Illnesses conducted by the Bureau of Labor Statistics (BLS) of the U.S. Department of Labor. The survey, which BLS has conducted for the past 25 years, is a random sample of about 250,000 private sector establishments and provides estimates of workplace injuries and illnesses on the basis of information provided by employers from their OSHA Form 200 log of recordable injuries and illnesses.

For cases involving days away from work, BLS reports that in 1994 (the last year of data available at the time this report was prepared), approximately 705,800 cases (32%) were the result of overexertion or repetitive motion. Specifically, there were

C 367,424 injuries due to overexertion in lifting (65% affected the back); 93,325 injuries due to overexertion in pushing or pulling objects (52% affected the back); 68,992 injuries due to overexertion in holding, carrying, or turning objects (58% affected the back). Totaled across these three categories, 47,861 disorders affected the shoulder.

C 83,483 injuries or illnesses in other and unspecified overexertion events.

x C 92,576 injuries or illnesses due to repetitive motion, including typing or key entry, repetitive use of tools, and repetitive placing, grasping, or moving of objects other than tools. Of these injuries or illnesses, 55% affected the wrist, 7% affected the shoulder, and 6% affected the back.

Data for 1992 to 1995 indicate that injuries and illnesses requiring days away from work declined 19% for overexertion and 14% for repetitive motion. The incidence rate of overexertion (in lifting) declined from 52.1 per 10,000 workers in 1992 to 41.1 in 1995; the incidence rate for repetitive motion disorders declined from 11.8 per 10,000 workers in 1992 to 10.1 in 1995. These declines are similar to those seen for cases involving days away from work from all causes of injury and illness.

The reasons for these declines are unclear but may include: a smaller number of disorders could be occurring because of more intensive efforts to prevent them; more effective prevention and treatment programs could be reducing days away from work; employers or employees may be more reluctant to report or record disorders; or the criteria used by health care providers to diagnose these conditions could be changing.

IDENTIFICATION AND SELECTION OF STUDIES The goal of epidemiologic studies is to identify factors that are associated (positively or negatively) with the development or recurrence of adverse medical conditions. This evaluation and summary of the epidemiologic evidence focuses chiefly on disorders that affect the neck and the upper extremity, including tension neck syndrome, shoulder tendinitis, epicondylitis, carpal tunnel syndrome, and hand- arm vibration syndrome, which have been the most extensively studied in the epidemiologic literature. The document also reviews studies that have dealt with work-related back pain and that address the way work organizational and psychosocial factors influence the relationship between exposure to physical factors and work-related MSDs. The literature about disorders of the lower extremity is outside the scope of the present review.

A search strategy of bibliographic databases identified more than 2,000 studies. Because of the focus on the epidemiology literature, studies that were laboratory-based or that focused on MSDs from a biomechanical standpoint, dealt with clinical treatment of MSDs, or had other nonepidemiologic orientation were eliminated from further consideration for this document. Over 600 studies were included in the detailed review process.

METHODS FOR SYNTHESIZING STUDIES For the upper extremity studies included in this review, those which used specific diagnostic criteria, including physical examination techniques, were given greater consideration than studies that used less specific methods to define health outcomes. The review focused most strongly on observational studies whose health outcomes were based on recognized symptoms and standard methods of clinical examination. For completeness, those epidemiologic studies that based their health outcomes on reported symptoms alone were also reviewed. For the low-back studies included in this review, those which had objective exposure measurements were given greater consideration than those which used

xi self-reports or other measures. For the psychosocial section, any studies which included measurement or discussion of psychosocial factors and MSDs were included.

No single epidemiologic study will fulfill all criteria to answer the question of causality. However, results from epidemiologic studies can contribute to the evidence of causality in the relationship between workplace risk factors and MSDs. The framework for evaluating evidence for causality in this review included strength of association, consistency, temporality, exposure-response relationship, and coherence of evidence.

Using this framework, the evidence for a relationship between workplace factors and the development of MSDs from epidemiologic studies is classified into one of the following categories:

Strong evidence of work-relatedness (+++). A causal relationship is shown to be very likely between intense or long-duration exposure to the specific risk factor(s) and MSD when the epidemiologic criteria of causality are used. A positive relationship has been observed between exposure to the specific risk factor and MSD in studies in which chance, bias, and confounding factors could be ruled out with reasonable confidence in at least several studies.

Evidence of work-relatedness (++). Some convincing epidemiologic evidence shows a causal relationship when the epidemiologic criteria of causality for intense or long-duration exposure to the specific risk factor(s) and MSD are used. A positive relationship has been observed between exposure to the specific risk factor and MSD in studies in which chance, bias, and confounding factors are not the likely explanation.

Insufficient evidence of work-relatedness (+/0). The available studies are of insufficient number, quality, consistency, or statistical power to permit a conclusion regarding the presence or absence of a causal association. Some studies suggest a relationship to specific risk factors, but chance, bias, or confounding may explain the association.

Evidence of no effect of work factors (-). Adequate studies consistently show that the specific workplace risk factor(s) is not related to development of MSD.

The classification of results in this review by body part and specific risk factor is summarized in Table 1.

xii Table 1. Evidence for causal relationship between physical work factors and MSDs Strong Insufficient Evidence Body part evidence Evidence evidence of no effect Risk factor (+++) (++) (+/0) (-)

Neck and Neck/shoulder Repetition T Force T Posture T Vibration T

Shoulder Posture T Force T Repetition T Vibration T

Elbow Repetition T Force T Posture T Combination T

Hand/wrist Carpal tunnel syndrome Repetition T Force T Posture T Vibration T Combination T

Tendinitis Repetition T Force T Posture T Combination T

Hand-arm vibration syndrome Vibration T

Back T Lifting/forceful movement T Awkward posture T Heavy physical work Whole body vibration T Static work posture T

xiii CONCLUSIONS A substantial body of credible epidemiologic research provides strong evidence of an association between MSDs and certain work-related physical factors when there are high levels of exposure and especially in combination with exposure to more than one physical factor (e.g., repetitive lifting of heavy objects in extreme or awkward postures [Table 1]).

The strength of the associations reported in the various studies for specific risk factors after adjustments for other factors varies from modest to strong. The largest increases in risk are generally observed in studies with a wide range of exposure conditions and careful observation or measurement of exposures.

The consistently positive findings from a large number of cross-sectional studies, strengthened by the limited number of prospective studies, provides strong evidence (+++) for increased risk of work- related MSDs for some body parts. This evidence can be seen from the strength of the associations, lack of ambiguity in temporal relationships from the prospective studies, the consistency of the results in these studies, and adequate control or adjustment for likely confounders. For some body parts and risk factors, there is some epidemiologic evidence (++) for a causal relationship. For still other body parts and risk factors, there is either an insufficient number of studies from which to draw conclusions or the overall conclusion from the studies is equivocal. The absence of existing epidemiologic evidence should not be interpreted to mean there is no association between work factors and MSDs.

In general, there is limited detailed quantitative information about exposure-disorder relationships between risk factors and MSDs. The risk of each exposure depends on a variety of factors such as the frequency, duration, and intensity of physical workplace exposures. Most of the specific exposures associated with the strong evidence (+++) involved daily whole-shift exposure to the factors under investigation.

Individual factors may also influence the degree of risk from specific exposures. There is evidence that some individual risk factors influence the occurrence of MSDs (e.g., elevated body mass index and carpal tunnel syndrome or a history of past back pain and current episodes of low-back pain). There is little evidence, however, that these individual factors interact synergistically with physical factors. All of these disorders can also be caused by nonwork exposures. The majority of epidemiologic studies involve health outcomes that range in severity from mild (the workers reporting these disorders continue to perform their routine duties) to more severe disorders (workers are absent from the workplace for varying periods of time). The milder disorders are more common. A limited number of studies investigate the natural history of these disorders and attempt to determine whether continued exposure to physical factors alters their prognosis.

The number of jobs in which workers routinely lift heavy objects, are exposed on a daily basis to whole-body vibration, routinely perform overhead work, work with their necks in chronic flexion position, or perform repetitive forceful tasks is unknown. While these exposures do not occur in most jobs, a large number of workers may indeed work under these conditions. The BLS data indicate that

xiv the total employment is over three million in the industries with the highest incidence rates of cases involving days away from work from overexertion in lifting and repetitive motion. Within the highest risk industries, however, it is likely that the range of risk is substantial depending on the specific nature of the physical exposures experienced by workers in various occupations within that industry.

This critical review of the epidemiologic literature identified a number of specific physical exposures strongly associated with specific MSDs when exposures are intense, prolonged, and particularly when workers are exposed to several risk factors simultaneously. This scientific knowledge is being applied in preventive programs in a number of diverse work settings. While this review has summarized an impressive body of epidemiologic research, it is recognized that additional research would be quite valuable. The MSD components of the National Occupational Research Agenda efforts are principally directed toward stimulation of greater research on MSDs and occupational factors, both physical and psychosocial. Research efforts can be guided by the existing literature, reviewed here, as well as by data on the magnitude of various MSDs among U.S. workers.

xv ACKNOWLEDGMENTS

In addition to the other contributors, the following staff members of the National Institute for Occupational Safety and Health are acknowledged for their support, assistance, and advice in preparing this document:

Penny Arthur Rose Hagedorn Marty Petersen Vanessa Becks William Halperin, M.D., Sc.D. Donna Pfirman Donna Biagini Anne Hamilton Linda Plybon Jenise Brassell Denise Hill Faye Rice Karen Brewer Suzanne Hogan Cindy Riddle Carol Burnett Hongwei Hsiao, Ph.D. Kris Royer Sue Cairelli Lore Jackson Walt Ruch Dick Carlson Laurel Jones Steven Sauter, Ph.D. Shirley Carr Susan Kaelin Lucy Schoolfield Dave Case Sandy Kasper Mitch Singal, M.D. Sharon Cheesman Aileen Kiel Paul Schulte Alexander Cohen, Ph.D. Diana Kleinwachter Becky Spry Marian Coleman Nina Lalich Anne Stirnkorb Barb Cromer Leslie MacDonald Naomi Swanson, Ph.D. Judy Curless Charlene Maloney Rodger Tatken David Dankovic Diane Manning Allison Tepper, Ph.D. John Diether James McGlothlin, Ph.D. Julie Tisdale Clayton Doak Patricia McGraw Anne Votaw Karen Dragon Alma McLemore David Votaw Sue Feldmann Judy Meese Thomas Waters, Ph.D. Jerry Flesch Matthew Miller Jane Weber Larry Foster Kathleen Mitchell Joann Wess Sean Gallagher Vivian Morgan Cindy Wheeler Lytt Gardner, Ph.D. Leela Murthy Kellie Wilson Pamela Graydon Rick Niemeier Ralph Zumwalde Daniel Habes Andrea Okun

xvi We also thank the following reviewers for their thoughtful comments on an earlier draft of this document:

Gunnar B.J. Andersson, M.D., Ph.D. Robert Harrison, M.D., M.P.H. Rush-Presbyterian-St. Luke’s University of California at San Francisco Medical Center

Mohammed M. Ayoub, Ph.D., P.E., C.P.E. William S. Marras, Ph.D. Texas Tech University The Ohio State University

Sidney J. Blair, M.D., F.A.C.S. J. Steven Moore, M.D., M.P.H., C.I.H., C.P.E. Loyola Chicago University University of Texas Health Center at Tyler

Vance C. Calvez, M.S., C.P.E. Margareta Nordin, Dr. Med. Sc. The Joyce Institute New York University

Don B. Chaffin, Ph.D. Donald C. Olsen, Jr., C.S.P., C.P.E. University of Michigan ERGOSH

Jerome J. Congleton, Ph.D., P.E., C.P.E. Thomas Owens, C.I.H., P.E. Texas A&M University IBM Corporation

Thomas Cook, Ph.D., P.T. Malcolm H. Pope, Dr. Med. Sc., Ph.D. University of Iowa The University of Iowa

Theodore Courtney Laura Punnett, Sc.D. Liberty Mutual University of Massachusetts

Michael Feuerstein, Ph.D. Robert G. Radwin, Ph.D. Uniformed Services University University of Wisconsin-Madison of the Health Sciences

Eric Frumin David Rempel, M.D. Union of Needle Trades, Industrial, University of California, San Francisco and Textile Employees (UNITE)

Michael Gauf Suzanne H. Rodgers, Ph.D. CTD News Consultant in Ergonomics

Fred Gerr, M.D. C. Jivan Saran Emory University Central Missouri State University

Lawrence P. Hanrahan, Ph.D., M.S. Scott Schneider, C.I.H. Wisconsin Division of Health The Center to Protect Workers’ Rights

Barbara Silverstein, Ph.D., M.P.H., C.P.E. State of Washington Department of Labor and Industries

xvii CHAPTER 1 Introduction

PURPOSE The goal of epidemiologic studies is to identify This document examines the epidemiologic factors (such as physical, work organizational, evidence that associates selected psychosocial, individual, and sociocultural musculoskeletal disorders (MSDs) of the upper factors) that are associated positively or extremity and the low back with exposure to negatively with the development or recurrence physical factors at work. The authors have paid of adverse medical conditions. This document particular attention to analyzing the strength of addresses and evaluates the literature with the association between MSDs and work regard to these issues for work-related MSDs. factors. Because the development of an MSD may be modified by psychosocial factors, the This document reviews the epidemiologic authors have also reviewed the literature on the evidence regarding the role of physical factors relationship of these factors to the presence of in the development of MSDs for the following musculoskeletal symptoms or disorders. body areas: the neck, shoulder, elbow, Understanding these associations and relating hand/wrist, and back. The document also them to disease etiology is critical to identifying addresses the influence of work organizational workplace exposures that can be reduced or and psychosocial factors on the association of prevented. physical factors with work-related MSDs. This evaluation and summary of the epidemiologic BACKGROUND evidence focuses chiefly on disorders affecting The World Health Organization has the neck and the upper extremity—including characterized “work-related” diseases as tension neck syndrome, shoulder tendinitis, multifactorial to indicate that a number of risk epicondylitis, carpal tunnel syndrome, and factors (e.g., physical, work organizational, hand-arm vibration syndrome, which have been psychosocial, individual, and sociocultural) the most extensively studied in the contribute to causing these diseases [WHO epidemiologic literature. This document also 1985]. One important reason for the concentrates on studies that have dealt with the controversy surrounding work-related MSDs is issue of work-related back pain and sciatica. their multifactorial nature. The disagreement The literature on disorders of the lower centers on the relative importance of multiple extremities is beyond the scope of this review. and individual factors in the development of disease. The same controversy has been an SCOPE AND MAGNITUDE OF THE issue with other medical conditions such as PROBLEM certain cancers and lung disorders—both of The only routinely published, national source of which have multiple causal factors information about occupational injuries and (occupational and nonoccupational). illnesses in U.S. workers is the Annual Survey of Occupational Injuries and Illnesses (ASOII)

1-1 conducted by the Bureau of Labor Statistics number of cases decreased by 7% to 308,000 (BLS) of the U.S. Department of Labor. This reported cases; but this number still exceeds the survey is a random sample of about 250,000 number of cases in any year before 1994. private-sector establishments, but it excludes self-employed workers, farms with fewer than Because these summary data did not 11 employees, private households, and all adequately describe the nature of occupational government agencies. The ASOII provides injuries and illnesses and the related risk estimates of workplace injuries and illnesses factors, the ASOII was redesigned in 1992 to from information that employers provide to capture more detailed information about injury BLS from their OSHA Form 200 log of and illness cases requiring days away from recordable injuries and illnesses. work. This redesigned survey captures demographic information about injured workers BLS has conducted this annual survey since as well as the following characteristics of the 1972 and has thus provided basic information injury or illness: (1) the employer’s description about cases of occupational injury or illness that of the nature of the injury or illness, such as required more than first-aid (including medical sprain or carpal tunnel syndrome; (2) the part treatment, restricted work activity, or days of the body affected by the specified away from work). This information includes the condition, such as back or wrist; (3) the source total number of cases categorized on the of the injury or illness that directly produced OSHA Form 200 log as either an injury or an the disabling condition, such as a crate, heavy illness. The illness data are separated into six box, or a nursing home patient; and (4) the subcategories; the category that contains most event or exposure that describes the manner in (but not all) musculoskeletal conditions is which the injury or illness was inflicted, such as disorders associated with repeated trauma. overexertion during lifting or repetitive motion. This illness category also includes illnesses The BLS data are based on information associated with noise-induced hearing loss, provided by employers from their records of but MSDs account for the largest proportion of work-related injuries and illnesses and then these cases, especially in recent years. All back coded into these categories. disorders or injuries are placed in the single, broad injury category, which also includes all For injury and illness cases involving days away other types of injuries such as lacerations, from work, BLS reports that in 1994 (the last fractures, and burns. year for which the detailed data were complete when this report was prepared), approximately From this part of the ASOII, BLS reports that 705,800 cases (32%) resulted from in 1995, 308,000 (or 62%) of all illness cases overexertion or repetitive motion. Specifically: were due to disorders associated with repeated C 367,424 injuries were due to overexertion in trauma (excluding low-back disorders, which lifting; 65% affected the back. Another are listed as injuries). The number of repeated 93,325 injuries were due to overexertion in trauma cases increased dramatically, rising pushing or pulling objects; 52% affected the steadily from 23,800 in 1972 to 332,000 in back. In addition, 68,992 injuries were due 1994—a 14-fold increase. In 1995, the to overexertion in holding, carrying, or turning

1-2 objects; 58% affected the back. Totaled boy’s work clothes, and hats, caps, and across these three categories, 47,861 millinery; these industries also include disorders affected the shoulder. The median manufacturing sectors such as textile bags, time away from work from overexertion potato chip and similar snacks, motor vehicles, injuries was 6 days for lifting, 7 days for and meat packing plants (Table 1–2). These pushing/pulling, and industries have IRs that are more than eight 6 days for holding/carrying/turning. times the rate for all private industry.

C 83,483 injuries or illnesses occurred in other Not all workers in these high-risk industries are and unspecified overexertion events. exposed to the working conditions associated with these clearly elevated rates of illnesses and C 92,576 injuries or illnesses occurred as a injuries from overexertion and repetitive result of repetitive motion, including typing or motion; however, smaller proportions of key entry, repetitive use of tools, and workers in other industries may be similarly repetitive placing, grasping, or moving of exposed. For example, trucking and courier objects other than tools. Of these repetitive services, an industry employing over 1.6 million motion injuries, 55% affected the wrist, 7% people, had IRs for overexertion disorders that affected the shoulder, and 6% affected the were almost three times higher than the average back. The median time away from work was rate for all private industries. Thus, these 18 days as a result of injury or illness from employment estimates provide a conservative repetitive motion. approximation of the number of workers with heavy exposures to high-risk conditions. The highest incidence rates (IRs) of work- related injuries and illnesses from over- exertion The BLS data are surveillance information that occur among workers in nursing and personal might contain misclassifications of both care facilities, scheduled air transportation, and exposure and health outcomes. However, some manufacturing of travel trailers and campers. As industries have notably and consistently Table 1–1 indicates, these industries have rates elevated rates of musculoskeletal injuries and of overexertion disorders four times higher than disorders that are not likely to be attributable to the average rate for all private industry. More data collection or coding. Note that decisions than 2 million workers are employed in the about the event or exposure that resulted in an three highest-risk industries alone. However, injury or illness are associations rather than rates are not available by occupation within causal inferences. Nevertheless, they provide these industries, and not all workers within a some perspective on the magnitude of work- high-risk industry will be at equal risk of related MSDs. developing a work-related MSD.

Industries with the highest IRs of work-related injuries and illnesses from repetitive motion include a number of garment manufacturing sectors such as knit underwear mills, men’s and

1-3 Table 1-1. Private sector industries with the highest incidence rates of injuries and illnesses from overexertion resulting in days away from work, 1994

1994 annual average Incidence rate 95% confidence employment‡ (per 10,000 interval Industry* SIC code† (in thousands) workers) (rate per 10,000) Number of cases Nursing and personal care facilities 805 1,648 318.0 (286, 350) 41,884 Air transportation, scheduled 451 607 306.7 (276, 337) 16,309 Travel trailers and campers (manufacturing) 3792 22 303.7 (206, 401) 635 Food products machinery (manufacturing) 3556 24 260.1 (142, 378) 620 Bottled and canned soft drinks (manufacturing) 2086 95 255.6 (224, 287) 2,512 Beer, wine, and distilled beverages (wholesale) 518 150 254.6 (189, 321) 3,750 Coal mining 12 112 235.6 not available 2,609 Mattresses and bedsprings (manufacturing) 2515 31 233.5 (172, 295) 719 Comparison Industries: All manufacturing 2, 3 18,319 83.00 (81.4, 84.6) 151,794 All private industry§ 94,146 76.00 (75.7, 76.3) 613,251 Finance, insurance, and real estate 6 6,707 17.90 (16.5, 19.3) 11,191

Source: Bureau of Labor Statistics, U.S. Department of Labor, Annual Survey of Occupational Injuries and Illnesses, 1994 Case and Demographic Resource Tables (ftp://stats.bls.gov/pub/special.requests/ocwc/osh/c_d_data). *High rate industries were those having an incidence rate greater than three times the rate for all private industry, at the most detailed or lowest SIC level at which rates are published. Generally, manufacturing industries are published at the 4-digit code level and the remaining industries at the 3-digit level. †Standard Industrial Classification Manual, 1987 edition. ‡Annual average employment from the BLS Covered Employment and Wages (ES-202) Survey. §Excludes farms with fewer than 11 employees.

1-4 Table 1-2. Private sector industries with the highest incidence rates of injuries and illnesses from repetitive motion resulting in days away from work, 1994

1994 annual average Incidence rate 95% confidence SIC employment‡ (per 10,000 interval Industry* code† (in thousands) workers) (rate per 10,000) Number of cases Knit underwear mills (manufacturing) 2254 25 165.6 (145, 187) 370 3 House slippers (manufacturing) 3142 146.3 (92, 201) 48 Men’s and boy’s work clothes (manufacturing) 2326 42 117.2 (97, 137) 463 Textile bags (manufacturing) 2393 11 115.7 (60, 171) 117 Potato chips and similar snacks (manufacturing) 2096 35 115.2 (95, 135) 406 Motor vehicles and car bodies (manufacturing) 3711 335 113.9 (99, 129) 4,058 Hats, caps, and millinery (manufacturing) 235 21 103.9 (79, 129) 202 Meat packing plants (manufacturing) 2011 138 98.5 (76, 121) 1,402 Bras, girdles, and allied garments (manufacturing) 2342 12 96.2 (73, 119) 111 Wood products, not elsewhere classified (manufacturing) 2499 58 92.8 (69, 117) 515 Men’s and boy’s suits and coats (manufacturing) 231 40 89.1 (74, 104) 320 Electronic coils and transfers (manufacturing) 3677 17 87.0 (52, 122) 142 Men’s footwear (excluding athletic) 3143 28 84.9 (64, 106) 221 Comparison Industries: All manufacturing 2, 3 18,319 27.0 (26.4, 27.6) 49,278 All private industry§ 94,146 11.5 (11.4, 11.6) 92,576 Finance, insurance, and real estate 6 6,707 8.1 (7.4, 8.8) 5,046

Source: Bureau of Labor Statistics, U.S. Department of Labor, Annual Survey of Occupational Injuries and Illnesses, 1994 Case and Demographic Resource Tables (ftp://stats.bls.gov/pub/special.requests/ocwc/osh/c_d_data). *High rate industries were those having an incidence rate greater than three times the rate for all manufacturing workers at the most detailed or lowest SIC level at which rates are published. Generally, manufacturing industries are published at the 4-digit code level and the remaining industries at the 3-digit level. †Standard Industrial Classification Manual, 1987 edition. ‡Annual average employment from the BLS Covered Employment and Wages (ES-202) Survey. §Excludes farms with fewer than 11 employees.

1-5 The large number of work-related low-back COST injuries or illnesses reported in the BLS data is The precise cost of occupational MSDs is not consistent with the results of two representative known. Estimates vary depending on the surveillance studies in the United States and method used. A conservative estimate Ontario. In the U.S. study, about 52% of the previously published by NIOSH is back pain reports were attributed by the $13 billion annually [NIOSH 1996]. Others worker to repetitive events at work, and an have estimated the cost at $20 billion annually additional 16% were attributed to discrete, [AFL-CIO 1997]. Regardless of the estimate acute events at work; 33% were associated used, the problem is large both in health and with both types of exposures [Guo et al. 1995]. economic terms.

Although workers often consider MSDs to be Work-related MSDs are a major component of work-related, their reports of back pain do not the cost of work-related illness in the United appear to affect the reliability of their self States. The California Workers’ Compensation reports about exposure to physical work. In the Institute (a non-profit research institute) Ontario study [Liira et al. 1996], 24% of the estimates that upper-extremity MSD claims by long-term back disorders were related to workers average $21,453 each [CWCI 1993]. bending and lifting, working with vibrating Back pain is by far the most prevalent and machines, and working in awkward postures. costly MSD among U.S. industries today. Interestingly, 8% of the population were Recent analysis of the 1988 Occupational exposed to at least two of these three factors, Health Supplement of the National Health and an additional 3% were exposed to all three. Interview Survey (an ongoing household-based survey) shows that the overall prevalence of The impact of work-relatedness is self-reported back pain from repeated activities demonstrated by the elevated MSD rates for on the most recent job was 4.5%, or 4.75 certain industries in workers’ compensation million U.S. workers [Behrens et al. 1994]. The data as well as the BLS data. For example, in mean cost per case of compensable low-back the State of Washington workers’ pain was reported to be $8,321 in 1989 compensation system, the overall IR of work- [Webster and Snook 1994b]. related MSDs was 3.87/100 workers in 1992, 3.72 in 1993, and 3.52 in 1994. Work-related Webster and Snook [1994a] estimated that the MSDs in this study were defined as injuries and mean compensation cost per case of upper- illnesses involving sprains/strains, joint extremity, work-related MSD was $8,070 in inflammation, low-back pain, and nerve- 1993; the total U.S. compensable cost for compression syndromes. Four industries had upper extremity, work-related MSDs was rates at least four times the 1992–94 average $563 million in 1993. For example, the State of rate: wallboard installation (23.6/100 workers Washington averaged 44,648 work-related per year), temporary help-assembly (23.6), MSD claims, with an average total cost of roofing (19.9), and moving companies (18) $166.8 million/year for the period 1992–94. [Washington State Department of Labor and The State of Washington has a working Industries 1996]. population that is 2% that of the U.S. workforce. The compensable cost is limited to the medical expenses and indemnity costs (lost

1-6 wages). When other expenses such as the full process—such as incipient disease or residual lost wages, lost production, cost of recruiting signs of a MSD that was once clinically and training replacement workers, cost of apparent. Because of the multifactorial nature rehabilitating the affected workers, etc. are of MSDs, it has been necessary to look at a considered, the total cost to the national broad spectrum of outcome measures to assess economy becomes much greater. the effects of these factors.

DEFINING HEALTH OUTCOMES Certain authors have noted the scarcity of Work-related MSDs are defined differently in objective measures (including physical different studies; thus, it is not surprising that examination techniques) to define work-related controversy has arisen about the relative MSDs, and the lack of standardized criteria for importance of various risk factors in the defining MSD cases. Such insufficiencies etiology of these disorders. Some investigators sometimes make study comparisons difficult restrict themselves to case definitions based on [Gerr et al. 1991; Moore 1992; Frank et al. clinical pathology, some to the presence of 1995; Riihimäki 1995; Hadler 1997]. It would symptoms, some to “objectively” demonstrable be useful to have a concise pathophysiological pathological processes, and some to work definition and corresponding objective clinical disability (such as lost work-time status). test for each work-related MSD to translate the degree of tissue damage or dysfunction into an The most common health outcome has been the estimate of current or future disability and occurrence of pain, which is assumed to be the prognosis. Such definitions and tests do not yet precursor of more severe disease [Riihimäki exist. Clinically defined work-related MSDs 1995] or (as in the case of back pain) the often have no clearly delineated disorder itself. Different MSD health outcomes pathophysiological mechanisms for pathological have been assessed by investigators depending processes. In cases where some criteria exist on the particular concern or nature of the study. (such as carpal tunnel syndrome [CTS]), the The specific health outcomes studied vary standard of accuracy is relatively expensive, depending on (a) the purpose of the study, (b) elaborate, and subject to interpretation. For the composition of the study population, (c) the example, the overlap between symptoms and rarity or prevalence of the health outcome in the presence of abnormalities in nerve conduction population, (d) the need to limit specific biases, studies is not great [Stetson et al. 1993]; and (e) the decisions of the investigators. furthermore, abnormalities in nerve conduction studies cannot be reliably used to predict the Different epidemiologic measures and time future onset of CTS symptoms [Werner et al. scales have also been used to quantify MSDs in 1997]. Thus, in the interest of feasibility, groups of people (lifetime prevalence, period expense, and utility, simpler tests and less prevalence, point prevalence, IR, incidence specific case definitions may have been used in ratio, etc.). Similarly, some studies have some studies, thereby introducing some risk of included chronic cases, whereas others have misclassification for specific studied acute or subacute cases or both. Cross-sectional studies usually employ case diagnostic entities. definitions that take into account prevalent cases at different stages of the disease For upper-extremity studies in this review,

1-7 those with specific diagnostic criteria (including assessment of physical workload by the study physical examination techniques) were given subjects. greater consideration than studies that used less-specific methods to define health The accuracy of such self-assessment has been outcomes. The review focused on observational debated (both for under-estimation and over- studies whose health outcomes were based on estimation). Uhl et al. [1987] found that the constellation of recognized symptoms and workers reported performing more physical standard methods of clinical examination. For work than observational data could support. completeness, those epidemiologic studies that Armstrong et al. [1989] found that workers can based their health outcomes on reported (on average) distinguish among levels of symptoms alone were also reviewed. exposure, but workers’ ratings may not correspond with objective measurements. Therefore, this document focuses on the upper- Bernard et al. [1994] found that video display extremity MSDs that have commonly used terminal (VDT) operators (those with and those diagnostic symptoms and physical examination without symptoms of work-related MSDs) abnormality criteria. Specifically, these MSDs reported that the average time they spent typing are (1) tension-neck syndrome, (2) rotator cuff daily in the last year was twice that noted by tendinitis and impingement syndrome in the independent observers in a single work day shoulder, (3) epicondylitis in the elbow, (4) (although the 1-day observation period may CTS, have been insufficient to capture an average (5) wrist tendinitis, and (6) hand-arm vibration day of typing time). Similarly, Stubbs [1986] (HAV) syndrome. Generally, the physical found large and significant differences between examination techniques used to define these subjective and observed estimates of time spent MSD cases of the upper extremity have been working in specified postures. Fransson-Hall et similar from study to study and involve standard al. [1995], on the other hand, found that examination techniques recognized by the American Academy of Orthopedic Surgeons, workers tended to underestimate their the American College of Physicians, or the exposures to contact stress of the hand International Labor Organization compared with observation. This Musculoskeletal Task Force (thus increasing underestimation may be because workers tend the reliability of comparisons between studies). to monitor discomfort from direct contact Although physical examination techniques have pressure—not the time spent with direct not been commonly used in epidemiologic contact. Katz et al. [1996] found evidence of studies of low-back disorders, this document the validity of self-reported symptoms and also reviews those epidemiologic studies that functional status, and analysis of their data address low-back pain. yielded evidence that variability in self-reports EXPOSURE MEASUREMENTS is not influenced by potential secondary gain. Exposure measurements used in work-related MSD studies range from very crude As Riihimäki [1995] pointed out, it is difficult to assess current exposure, but it is even more measures (e.g., occupational title) to complex difficult to assess cumulative past exposure analytical techniques (e.g., spectral analysis of retrospectively. Accurate retrospective data are electrogoniometer measurements of joint usually not available; thus the exposure motions). Some studies have relied on self- assessment is often based on self-reports, and

1-8 the assessment may incur information bias. injury, repetitive strain injury, epidemiology, etiology, cumulative trauma disorders, MSDs A few studies have used observational methods (neck, tension neck syndrome, shoulder, to estimate exposures to workplace physical rotator cuff, elbow, epicondylitis, tendinitis, hazards more accurately and reliably. Because tenosynovitis, carpal tunnel, de Quervain's, studies that directly observe or assess physical nerve entrapment syndrome, vibration, back exposure factors are less likely to misclassify pain and sciatica, manual materials handling). exposure status, these studies are given greater Bibliographies of relevant articles were weight in this review. reviewed. Relevant foreign literature citations in English and included in the databases were Despite the noted limitations, occupations included in this review along with literature from classified as “high-risk” in several studies share the personal files of the contributors. This a number of workplace exposures associated search strategy identified more than 2,000 with work-related MSDs. These workplace studies. Because of the focus on the exposures occur in various combinations epidemiology literature, a number of these (singly, simultaneously, or sequentially) at studies that were laboratory-based or focused different levels for different durations. These on MSDs from a biomechanical standpoint that exposures have not been routinely broken dealt with clinical treatment of MSDs or other down into task variables and quantified, with non-epidemiologic orientations were eliminated the mechanical or physiological loads defined from further consideration for the present and measured. document. Over 600 studies were included in the detailed review process. INFORMATION RETRIEVAL This document examines scientific peer- SELECTION OF STUDIES reviewed epidemiologic journal articles, The studies that were chosen for more detailed including recent publications addressing MSD review specifically concerned the work- risk factors, conference proceedings, and relatedness of MSDs, musculoskeletal abstracts dealing with upper-extremity or back problems of the neck, upper limbs, or back, MSDs, recent textbooks, internally reviewed and/or occupational and nonoccupational risk government reports or studies conducted by factors. The following inclusion criteria were NIOSH, and other documents. Reports of used to select studies for the review: epidemiologic studies were acquired using both CD-ROM and online commercial and Population: Studies were included if the governmental databases. Searches were exposed and referent populations were well carried out on computer-based bibliographic defined. databases: Grateful Med® (which includes Medline® and Toxline®), NIOSHTIC® (a Health outcome: Studies were included if they NIOSH database), and CIS (the International involved neck, upper-extremity, and low-back Labour Organization occupational health MSDs measured by well-defined, explicit database). The search strategy included the criteria determined before the study. Studies following key terms: occupation, repetition, whose primary outcomes were clinically force, posture, vibration, cold, psychosocial, relevant diagnostic entities generally had less psychological, physiological, repetition strain misclassification and were likely to involve

1-9 more severe cases. Studies whose primary 4. The joint under discussion was subjected outcomes were the reporting of symptoms to an independent exposure assessment, generally had more misclassification of health with characterization of the independent status and a wider spectrum of severity. variable of interest (such as repetition or repetitive work). This criterion indicates Exposure: Studies were included if they whether the exposure assessment was evaluated exposure so that some inference conducted on the joint of interest and could be drawn regarding repetition, force, involved the type of exposure being extreme joint position, static loading or examined— such as repetitive work, vibration, and lifting tasks. Studies in which forceful exertion, extreme posture, or exposure was measured or observed and vibration. This criterion indicates whether recorded for the body part of concern were the exposure was measured considered superior to studies that used self- independently or in combination with reports or occupational/job titles as surrogates other types of exposures. Exposure was for exposure. also characterized by the method used to measure the level of exposure. Studies Study design: Population-based studies of that used either direct observation or MSDs, case-control studies, cross-sectional actual measurements of exposure were studies, longitudinal cohort studies, and case considered to have a more accurate series were included. exposure classification scheme, whereas studies that exclusively used job titles, METHODS FOR ANALYZING OR interviews, or questionnaire information SYNTHESIZING STUDIES were assumed to have less accurate The first step in the analytical process was to exposure information. classify the epidemiologic studies by the following criteria: During review of the studies, the greatest qualitative weight was given to studies that had 1. The participation rate was $70%. This objective exposure assessments, high criterion limits the degree of selection bias participation rates, physical examinations, and in the study. blinded assessment of health and exposure status. The chapters dealing with the different 2. The health outcome was defined by body regions—neck (including neck-shoulder), symptoms and physical examination. This shoulder, elbow, hand/wrist, and low- criterion reflects the preference of most back—summarize these characteristics for each reviewers to have health outcomes that study reviewed on the criteria table. are defined by objective criteria.

The second step of the analytical process was 3. The investigators were blinded to health to divide the studies into those with statistically or exposure status when assessing health significant associations between exposures and or exposure status. This criterion limits health outcomes and those without statistically observer bias in classifying exposure or significant associations. The associations were disease. then examined to determine whether they were

1-10 likely to be substantially influenced by viewed individually (taking into account good confounding or other selection bias (such as epidemiologic principles) but together as a survivor bias or other epidemiologic pitfalls that body of evidence for making broader might have a major influence on the interpretations about epidemiologic causality. interpretation of the findings). These include the Many investigators did not examine each risk absence of nonrespondent bias and factor separately but selected study and comparability of study and comparison groups. comparison groups based on combinations of There are also tables that summarize risk factors (such as workers in jobs involving information about confounders and high force and repetition compared with epidemiologic pitfalls for each study reviewed workers having no exposure to high force and at the end of each body region chapter. repetition).

The third step of the analytical process was to CRITERIA FOR CAUSALITY review and summarize studies with regard to No single epidemiologic study will fulfill all strength of association, consistency in criteria for causality. However, the results of association, temporal association, and many epidemiologic studies can contribute to exposure-response relationship. Each of these the evidence of causality in the relationship factors is discussed in greater detail in the next between workplace risk factors and MSDs. section (Criteria for Causality). Each study Rothman [1986] defined a cause as “an event, examined (those with negative, positive, or condition, or characteristic that plays an equivocal findings) contributed to the pool of essential role in producing an occurrence of the data for determining the strength of disease.” work-relatedness using causal inference. The exposures examined for the neck and upper This document uses the following framework of extremity were repetition, force, extreme criteria to evaluate evidence for causality. The posture, and segmental vibration. The framework was proposed by Hill [1966; 1971] exposures examined for the low back were and modified by Susser [1991] and Rothman heavy physical work, lifting, bending/twisting, [1986]. whole-body vibration, and static postures. Strength of Association Care should be taken when interpreting some The ORs and prevalence rate ratios (PRRs) study results regarding individual workplace from the reviewed studies were used to factors of repetition, force, extreme or static examine the strength of the association between postures, and vibration. As Kilbom [1994] exposure to workplace risk factors and MSDs, stated, these factors occur simultaneously or with the higher values indicating stronger during alternating tasks association. The greater the magnitude of the relative risk (RR) or the within the same work, and their effects concur and interact. A single odds ratio (OR) for an OR, the less likely the association is to be individual risk factor may not accurately reflect spurious [Cornfield et al. 1959; Bross 1966; the actual association, as not all of the studies Schlesselman 1978]. Weaker associations are derived ORs for simultaneously occurring more likely to be explained by undetected factors. Thus these studies were not only biases.

1-11 Debate is ongoing in the epidemiologic Rothman [1986] referred to specificity of effect literature about studies with small sample sizes as “useless and misleading” as a criterion for that find increased ORs or PRRs but have causality. confidence intervals (CIs) that include 1.0. The question is whether such studies simply show Temporality no significant association or can be seen as Temporality refers to documentation that the useful estimates of associated risk. cause precedes the effect in time. Prospectively Nonetheless, it is useful to identify trends across designed studies ensure that this criterion is such studies and consider whether they have strictly adhered to—that is, that exposure valuable information after taking into account precedes adverse health outcome. But cross- other epidemiologic principles. If the studies sectional studies are not designed to allow strict with and without significant findings both have adherence to this criterion because both similarly elevated ORs or PRRs, this exposure information and adverse health information is useful in estimating the overall outcome are obtained at the same point in time. level of risk associated with exposure. Even though the cross-sectional study design Consistency precludes strict establishment of cause and Consistency refers to the repeated observation effect, additional information can be used to of an association in independent studies. make reasonable assumptions that exposure Multiple studies yielding similar associations preceded the health effect—particularly when support the plausibility of a causal the relationship between physical exposures is interpretation. Finding the same association measured by observation or direct with different and valid ways of measuring measurement and by MSD-related health exposure and disease may show that the outcomes. If the exposure was directly association is not dependent on measurement measured or observed, it is also unlikely that tools. Similar studies that yield diverse results the measurement was influenced by the weaken a causal interpretation. presence or absence of the MSD in the employee. Rothman [1986] stated that it is Specificity of Effect or Association important to realize that cause and effect in an This criterion refers to the association of a epidemiologic study or epidemiologic data single risk factor with a specific health effect. cannot be evaluated without making some We have not emphasized this criterion because assumptions (explicit or implicit) about the of the different views of its utility in determining timing between exposure and disease. For causality. If this criterion is interpreted to mean example, from a cross-sectional study of that a single stressor can be related to a specific hand/wrist tendinitis and highly forceful, outcome (e.g., that forceful exertion alone can repetitive jobs, a researcher can determine be related to hand/wrist tendinitis) it becomes when exposure began from recorded work an important criterion for MSDs. However, this histories or from interviews. The researcher can criterion can be interpreted and applied too also reasonably determine the time of tendinitis simplistically. Schlesselman [1982] noted that onset by interviews. Kleinbaum et al. [1982] the concept of specificity is that is generally too said that in cross-sectional studies, risk factors simplistic and that multiple causes and effects and prognostic factors cannot be distinguished were more often the rule than the exception. empirically without additional information.

1-12 With additional information (e.g., laboratory THE EVIDENCE OF WORK- experiments or biomechanical findings), an RELATEDNESS investigator can deduce that the adverse health After assessing the quality of individual outcome followed exposure. For example, epidemiologic studies, NIOSH investigators taking other confounders into account, it is judged whether the evidence was strong unreasonable to deduce that persons with enough to relate the risk factor to the MSD. In hand/wrist tendinitis are likely to seek making this judgement, the investigators employment in jobs that require highly forceful, considered the criteria for causality. Studies repetitive exertion of the hand/wrist area. which met all four evaluation criteria were given more weight than those which met at least one Exposure-Response Relationship of the criteria. The exposure-response relationship relates disease occurrence with the intensity, The evidence of work-relatedness from frequency, or duration of an exposure (or a epidemiologic studies is classified into one of combination of these factors). For example, if the following categories: strong evidence of long-duration, forceful, repetitive work using work-relatedness (+++), evidence of work- the hands and wrists is associated with an relatedness (++), inadequate evidence of increased prevalence of hand/wrist tendinitis, work-relatedness (+/0), and evidence of no this association would tend to support a causal effect of work factors (-). interpretation. Some have challenged the importance of physical factors as causal agents, Strong Evidence of Work- but prospective studies have shown that Relatedness (+++) reduced exposures result in a decreased A causal relationship is very likely between disease [Bigos et al. 1991b]. In occupational intense and/or long duration exposure to a health, important and effective preventive specific risk factor(s) and an MSD when using actions have been initiated without prospective the epidemiologic criteria of causality. A demonstration that reduced exposure decreases positive relationship has been observed the incidence of disease. between exposure to the risk factor and the MSD in at least several studies in which Coherence of Evidence chance, bias, and confounding could be ruled Coherence of evidence means that an out with reasonable confidence. association is consistent with the natural history and biology of disease. For example, an Evidence of Work-Relatedness (++) observed association between repetitive wrist Some convincing epidemiolgic evidence exists motion and CTS (defined by nerve conduction for a causal relationship using the epidemiologic criteria) must be supported by biological criteria of causality for plausibility: repeated wrist movement can cause intense and/or long-duration exposure to a swelling of tissue in the carpal tunnel, resulting specific risk factor(s) and an MSD. A positive in injury to nerves. It is important to remember, relationship has been observed between however, that epidemiologic studies can identify exposure to the risk factor and the MSD in new associations for further study. studies in which chance, bias, and confounding are not the likely explanation. CATEGORIES USED TO CLASSIFY

1-13 Insufficient Evidence of Work- agency reports that have undergone peer Relatedness (+/0) review and are widely available. The available studies are of insufficient quality, consistency, or statistical power to permit a DESCRIPTION OF TABLES, conclusion regarding the presence or absence FIGURES, AND APPENDICES of a causal association. Some studies suggest a In each chapter on neck, shoulder, elbow, relationship to specific risk factors but chance, hand/wrist, and low back disorders, there are bias, or confounding may explain the tables summarizing the risk indicators and association. epidemiologic criteria used in examining studies relevant to each body part. For each of these Evidence of No Effect of criteria tables there are corresponding figures Work Factors (-) which depict ORs, PRRs, or IRs, along with Adequate studies consistently and strongly their associated CIs, if available. show that the specific risk factor is not related to MSDs. In a separate table for each chapter, more extensive descriptions of studies, whether or SUMMARY not they contributed to decisions regarding This document critically reviews the evidence causal inference, are provided for each body regarding work-related risk factors and their part. These tables include information from relationship to MSDs of the neck, shoulder, each study about their design, population, elbow, hand/wrist, and low back. The outcome, and exposure measures, as well as document represents a first step in assessing the reported MSD prevalence. Some studies are work-relatedness of MSDs. This step involves included in the tables that may not be examination of relevant epidemiologic mentioned in the text. These additional studies information to assess the strength of the are for information purposes only. available evidence that, under certain conditions of exposure, specific risk factors could increase Appendix A, Epidemiologic Review, is a brief the risk of MSDs or increase the likelihood of primer on occupational epidemiologic methods. impairment or disability from MSDs. The Appendix B, Individual Factors Associated second step would involve quantitative risk with Work-Related Mus-culoskeletal estimates that are beyond the purpose and Disorders (MSDs), discusses individual factors scope of this document. This review of the (age, gender, etc.) and their association with epidemiologic literature may assist national and work-related MSDs. Appendix C, Summary international authorities, academics, and policy Tables, provides a concise overview of the makers in assessing risk and studies reviewed relative to the evaluation formulating decisions about future research or criteria, risk factors addressed, and other necessary preventive measures. issues. This document does not necessarily cite all of the literature on a particular MSD. Included are articles considered relevant by NIOSH investigators and internal and external reviewers of the draft document. Only reports that have been published or accepted for publication in the openly available scientific literature have been reviewed by the authors. In certain instances, they have included government

1-14 CHAPTER 2 Neck Musculoskeletal Disorders: Evidence for Work-Relatedness

SUMMARY Over 40 epidemiologic studies have examined physical workplace factors and their relationship to neck and neck/shoulder musculoskeletal disorders (MSDs). Among these studies are those which fulfill rigorous epidemiologic criteria and appropriately address important issues so that causal inferences can be made. The majority of studies involved working groups with a combination of interacting work factors, but certain studies assessed specific work factors. Each of the studies we examined (those with negative, positive, or equivocal findings) contributed to the overall pool of data for us to use in assessing the strength of the work- relatedness using causal inference.

There is evidence for a causal relationship between highly repetitive work and neck and neck/shoulder MSDs. Most of the epidemiologic studies reviewed defined “repetitive work” for the neck as work activities which involve continuous arm or hand movements which affect the neck/shoulder musculature and generate loads on the neck/shoulder area; fewer studies examined relationships based on actual repetitive neck movements. The two studies which measured repetitive neck movements by measuring head position (using frequency and duration of movements) fulfilled the most stringent epidemiologic criteria, showing strong associations with neck/shoulder MSDs. In those studies defining repetitive work involving continuous arm or hand movements affecting the neck/shoulder, nine studies were statistically significant and had odds ratios (ORs) greater than 3.0.; eight studies fulfilled all the epidemiologic criteria except the exposure criteria, and measured repetition for the hand/wrist and not for the neck. Of these, three were statistically significant and had ORs greater than 3, five had nonsignificant ORs, all under 2.0.

There is also evidence for forceful exertion and the occurrence of neck MSDs in the epidemiologic literature. Most of the epidemiologic studies reviewed defined “forceful work” for the neck/shoulder as work activities which involve forceful arm or hand movements, which generate loads to the neck/shoulder area; no study examined a relationship based on actual forceful neck movements. Of the 17 studies addressing force as one of the exposure factors, five studies found statistically significant associations, but did not derive ORs; two studies found ORs greater than 3.0, seven studies from 1 to 3.0, and two studies with ORs less than 1.0. Many of the studies relating measured force (as workload, etc.) to MSDs are in the biomechanical and ergonomic literature.

There is strong evidence that working groups with high levels of static contraction, prolonged static loads, or extreme working postures involving the neck/shoulder muscles are at increased risk for neck/shoulder MSDs. Consistently high ORs were found (twelve statistically significant studies with ORs over 3.0) providing evidence linking tension-neck syndrome with static postures or static loads.

The epidemiologic data were insufficient to provide support for the relationship of vibration to neck disorders. At this time, further studies must be done before a decision regarding causal inference is made. The few prospective studies which have included interventions to decrease workplace exposures that include decreasing repetitive work and less extreme working postures showed a decrease in the incidence of neck MSDs and an improvement in symptoms among affected workers. The data on intervention provide additional evidence that these disorders are related to workplace risk factors.

2-1 INTRODUCTION symptoms has ranged from approximately 50% Studies from the United States have generally (Silverstein et al. [1987]; Blåder et al. [1991]; classified neck disorders separately from Bernard et al. [1993]; Hales et al. 1994]) to shoulder disorders when evaluating work- about 85% (Andersen and Gaardboe related risk factors. Scandinavian studies [1993b]). Forty-seven of the listed studies examining work-related factors, on the other referenced included physical examination hand, have often combined neck and shoulder findings in their health outcome assessment MSDs into one health outcome variable. This criteria. was based on the concept that several muscles act on both the shoulder girdle and the upper Many of the neck and neck/shoulder MSD spine together. We have divided our reviews of studies referenced in the tables were part of the neck and shoulder MSDs into two larger studies that inquired about chapters: Chapter 2 addresses neck and musculoskeletal symptoms and physical findings neck/shoulder MSDs and Chapter 3 addresses in multiple body sites. In most of these studies, shoulder MSDs. there were no separate ergonomic exposure observations or measurements made that Our discussion of the evidence for work- pertained to the neck region (e.g., there were relatedness of the neck will include criteria no neck posture observations, neck angle Tables 2-1 through 2-6 and Figures 2-1 measurements, neck work-load assessment, through 2-6. Shoulder MSDs will be discussed trapezius electromyographic testing, etc.). In in the next chapter. these studies, the primary interest and measurement strategies focused on the hand Epidemiologic studies have defined neck and wrist region (e.g., Kuorinka and Koskinen MSDs in one of two ways: (a) by symptoms [1979]; Ohlsson et al. [1989]; Hales et al. occurring in the neck (usually with regard to a [1989]; Kiken et al. [1990]; Baron et al. specific duration, frequency, or intensity), or (b) [1991]). In the studies, workers were by using both symptoms and physical categorized only by hand/wrist exposures. examination findings. Hand/wrist categorization will not reflect exposures of the neck region (or other The prevalence of reported MSDs is generally musculoskeletal sites). For example, workers lower when they are defined using both who may have frequent and rapid awkward symptoms and physical examination results than postures of the neck but less frequent or when defined using symptoms alone. For extreme postures of the hand and wrist region example, the prevalence rate of tension neck may be misclassified as low risk if classification syndrome (TNS) among male industrial depends only on hand/wrist exposure. In workers in the United States was reported to general, we have given these studies less weight be 4.9% from interview data and 1.4% when because of a significant potential for case definitions included physical exam findings misclassification. [Hagberg and Wegman 1987]. The percent of work-related MSD cases defined by physical The text of this section on neck and examination findings to those defined solely by neck/shoulder MSDs is organized by work-

2-2 related exposure factor. The discussion within during a cycle), or each factor is organized according to the criteria for evaluating evidence for work- relatedness in epidemiologic studies using the (2) repeated arm or shoulder motions that strength of association, the consistency of generate loads to the neck/shoulder area (e.g., association, temporal relationships, exposure- trapezius muscle). Most of the studies that response relationship, and coherence of examined repetition or repetitive work as a evidence. Conclusions are presented with potential risk factor for neck or neck/shoulder respect to neck and neck/shoulder MSDs as a MSDs had several concurrent or interacting single disorder for each exposure factor. physical workplace factors that were being Summary information relevant to the criteria evaluated. Therefore, repetitive work was not used to evaluate study quality is presented in necessarily considered the primary exposure Tables 2-1 through factor but was considered along with the other 2-6. A more extensive summary, which work factors. includes information on health outcome, covariates, and exposure measures, is Studies Reporting on the Association presented at the end of this chapter. of Repetition as a Work Factor for Neck and Neck/Shoulder Studies Included in Neck MSDs MSDs Tables Either the risk factor “repetition” or “repetitive Forty-six epidemiologic studies dealing with work” was included in 26 studies as a factor neck MSDs and 23 dealing with neck/shoulder for selection of the study population in their MSDs appear in the summary tables. Of the examination of neck and neck/shoulder MSDs studies, 38 were cross-sectional, 2 were case- in the workplace. However, only a handful of control studies, and 6 were prospective studies. these studies examined repetitive movements of Among all the studies pertaining to the neck or the neck. Few of these studies observed or neck/shoulder area, 35 had participation rates measured: (a) the frequency or duration of of over 70%, 3 had less than 70%, and 8 did tasks pertaining to the neck, (b) the ratio of not report their participation rates. work-time-to-recovery time for neck or neck/shoulder involvement, or (c) the percentage of the workday spent on repetitive REPETITION activities involving the neck. Instead, studies Definition of Repetition for Neck and tended to compare and contrast the Neck/Shoulder MSDs prevalences of neck symptoms and/or physical For our review of the neck or neck/shoulder findings in workers in occupations requiring a region, we chose those epidemiologic studies combination of forceful, repetitive movements that examined repetition or repetitive work and extreme postures of the upper extremities activities and MSDs. Studies generally address (mainly of the hand/wrist) to workers in repetition as cyclical work activities that occupations without those requirements. involved either: (1) repetitive neck movements (e.g., the frequency of different head positions Twenty studies that mentioned repetitive work

2-3 or repetitive movements found a measurements of critical statistically significant positive association angles (15E and 30E) of flexion of the neck. between repetition and neck or neck/shoulder Two independent readers determined MSDs; 6 others had non-significant findings frequency, duration, and critical angles of (Tables 2-1 and 2-2, Figures 2-1 and 2-2). In movement for each variable by taking the terms of magnitude of the association, 11 average of the two readings. Weekly working studies had ORs greater than 3.0, 11 had ORs time, work rotation, patterns of breaks, and between 1.0 and 3.0, and none had an OR less individual performance rate (piece rate) were than 1.0. Four studies did not report their recorded and used in the analysis. The study results in terms of ORs or Prevalence Rate controlled for age, gender (only females were Ratio (PRRs), although all of these found included), and psychosocial variables significant associations (p<0.05). (“tendency for stress” and “worry”).

Studies Meeting the Four Evaluation Criteria The other study that fulfilled the four criteria Of the 27 investigations (see Tables 2-1 and 2- concerned a 3-year prospective study written 2), 2 fulfilled all four evaluation criteria outlined up in a series of articles by Kilbom et al. earlier in the introduction section [Ohlsson et al. [1986], Kilbom and Persson [1987], and 1995; Jonsson et al. 1988]. Only the Ohlsson Jonsson et al. 1988]. Female electronic study reported ORs. The investigations workers in highly repetitive tasks with static assessed repetitive work as an independent postural loads to the neck and shoulder areas variable in terms of frequency and duration of were followed over a 3-year period. In the neck movements. second year, some of the employees had workplace interventions that decreased the In the cross-sectional study by Ohlsson et al. number of repetitive tasks involving extreme [1995], female industrial assembly-line workers neck and shoulder postures, while others exposed to repetitive tasks with short (<30 continued to work at unaltered tasks. Three seconds) cycles were compared to 2 referent separate physical exams were carried out at groups: 68 former assembly workers and 64 yearly intervals, the first one initially assessing other workers with no repetitive exposure at tenderness on palpation and pain or restriction their current jobs. Industrial workers had to with active and passive movements. Ergonomic perform tasks with a posture requiring an assessments occurred at the outset of the study intermittently flexed neck and elevated arms, and included video analysis of postures and which were abducted intermittently. Workers movements of the head, shoulder, and upper and referents reported neck/shoulder arm. The evaluation recorded work-cycle time symptom(s) and had physical exams performed and number of cycles per hour; time at rest for by a single examiner. The examiner was blinded the arm, shoulder, and head; total number of to exposure status but not completely to group rest periods; and average and total duration per status. Ergonomic exposure assessment was work cycle and hour. (The method was extensive. It included videotaping, observation, designed to study short-cycle repetitive work and analysis of postures, including under visual control.) The mean number of

2-4 neck forward flexions included only occasional sitting tasks, >20E per hour was 728 (standard deviation caretaking work, surveillance of machinery, or [s.d.] 365) in the initial 96 workers. The assembling of bigger and heavier equipment. participation rate of the study was 72% after 3 The article documenting the last phase of the years; the investigators analyzed several cohort study by Jonsson et al. [1988] did not variables separately for dropouts and found no specifically address the neck but broadened the significant differences with regards to medical health outcome definition to include the status, physiologic capacity, working technique, neck/shoulder area and the rest of the upper or work history. The investigators performed extremity using “cervicobrachial region” as the step-wise logistic regression with deterioration health outcome of interest. A significant of disorders or remaining healthy in the different association between deterioration of health locations (neck and neck/shoulder) as the two status of the cervicobrachial region between dependent variables. Age, muscle strength, job Year II and Year III of the study and “work satisfaction, and high productivity were cycle, total time” at the p<0.05 level was found included in the logistic regression analyses of (ORs were not given). these studies. Video analysis and observation were used to assess repetitive exposure on all Studies Meeting at Least One of the Four subjects, using work cycle time, number of Criteria—Strength of Association cycles per hour, as well as number of neck Of the studies that found significant ORs over flexions per hour as criteria. Work cycle time 3.0 but did not mention or fulfill all of the varied between 4.6 and 9.1 min, with a mean criteria, almost all focused on working groups value of 6.6 min. with a combination of repetitive and forceful work and compared them to either population Strength of Association for referents or groups in occupations with lower Repetition exposure. Almost all were cross-sectional In the Ohlsson et al. [1995] study, the OR for surveys. These studies used health outcomes the association between repetitive work related from symptom surveys and self-reported to the neck and any neck/shoulder diagnoses workplace exposure (no direct observations) was 4.6; for a diagnosis of tension neck and either compared symptomatic workers syndrome, it was 3.6. (neck MSD cases) to asymptomatic workers in the same workforce (e.g., Yu and Wong For the cohort study carried out by Kilbom et [1996]; Bergqvist et al. [1995a]; Schibye et al. al. [1986], at the 2-year followup, the number [1995]; Hünting et al. [1981]) or in other of neck flexions per hour appeared as a strong occupations (e.g., Liss et al. [1995]; Andersen predictor for deterioration to severe disorders and Gaardboe [1993b]; Milerad and Ekenvall of the neck. Improvement to a “healthy status” [1990]; Onishi et al. [1976]). Onishi et al. classification from [1976] found significant differences in Year I to Year II was seen with reallocating neck/shoulder MSDs (OR 3.8) between workers to more varied work tasks (which groups involved in repetitive upper limb required a reorganization of monotonous and operations and office workers. They found repetitive work tasks). The new tasks were workers involved in repetitive activity had 10% characterized as more dynamic and varied and to 30% maximum voluntary contraction (MVC)

2-5 of the trapezius muscle. They concluded that disorders among checkers was 2.0 (95% habitual neck or shoulder muscle fatigue is confidence interval [CI] 0.6–6.7), in a model that included age, hobbies, second jobs, caused by repetitive tasks that result in systemic disease, and obesity. localized tenderness and may be a precursor to chronic MSDs. Bergqvist et al. [1995a] carried out a study comparing office workers using video display Andersen and Gaardboe [1993a] used a terminals (VDTs) to those who did not. A cross-sectional design to compare sewing physiotherapist’s diagnosis of tension-neck machine operators with a random sample of syndrome was used to define a case. Exposure women from the general population of the same assessment was based on both self-reports and region. A neck case required a strict the investigators’ observation of work postures, predetermined symptom and physical movements, and measurements of heights of examination definition. Exposure was assessed work-station equipment in conjunction with the through observation and categorization of jobs, user. Statistical modeling included several based on the authors’ experience and individual factors, organizational factors, and judgements. However, the main interest for ergonomic factors. For “tension neck” exposure assessment was duration of exposure syndrome, no factor related to repetitive work as a sewing machine operator. Statistical was found to be significantly related. modeling controlled for age, having children, not doing leisure exercise, smoking, and Blåder et al. [1991] surveyed 199 sewing socioeconomic status found a significant trend machine operators from 4 plants. Of the 155 for “neck/shoulder syndrome” in relation to who reported shoulder or neck pain, 131 were years of exposure as a sewing machine examined. Exposure assessment was by operator, with ORs from 3.2 to 36.74. The OR questionnaire and addressed employment for the lowest exposure category, 0-7 years, duration and hours per week. Authors stated was not statistically significant, although the that the study involved a control group and higher exposure levels were. For this study, the took into account psychosocial factors, but the exposure classification scheme does not allow results were not included in the article. Both separation of the effects of repetition from employment duration and working more than those of force, and there was no precise 30 hours per week were found to be measure of repetitiveness. statistically significant at the p<0.05 levels. For this study, the exposure as duration of work Baron et al. [1991] studied neck MSDs in 124 (per week and per years) does not allow grocery store checkers and 157 other grocery separation of the effects of repetition from store workers who were not checkers. The those of force. There was no direct measure of neck MSD case definition met predetermined repetitiveness. symptom and physical exam criteria. Physical examinations had higher participation rates Ekberg et al. [1994] carried out a case-control among the checkers (85%) than among the study involving cases from a semi-rural referents (55%). Telephone interviews to non- community in southern Sweden who had checkers resulted in questionnaire completion consulted a community physician for MSDs of by 85% of the non-checkers. The OR for neck the neck, shoulder, arm, or upper thorax.

2-6 Cases had to have been ill immediately prior to (data entry to “conversational” VDT use), physician visit and Hünting et al. [1981] used a case definition requiring symptoms and physical exams and an have been on sick leave less than 4 weeks. extensive exposure assessment using Cases were excluded for trauma, infectious questionnaire, observation, and measurements causes, accident, malignancy, rheumatic of workstations, and body posture disease, abuse, or pregnancy. Controls were measurements using a prescribed method. Data randomly selected from the Swedish insurance entry terminal users, whose tasks required registry. Exposure was obtained by more extensive repetitive work than traditional questionnaire. The analysis showed that for office workers, found an OR of 9.9 with the neck disorders with precise repetitive comparison. There were no adjustments for movements the OR was 3.8 for medium confounders in this analysis. exposure and 15.6 for high exposure comparing jobs with low force and low Kamwendo et al. [1991] compared 420 repetition. Gender, immigrant status, work medical secretaries with frequent, significant pace, and current smoking were also analyzed neck pain to those with few episodes based on in the logistic model. questionnaire responses. Exposure was also questionnaire based. The analysis was Ekberg et al. [1995] surveyed 637 Swedish controlled for age and length of employment. A residents for the presence of neck symptoms in surrogate for repetitive work consisted of hours the past six months. Exposure was based on sitting or working with office machines with high questionnaire responses. Twenty questionnaire exposure equal to 5 hrs or more/day. items on physical work conditions were factor analyzed. Age, smoking, exercise habits, and Kiken et al. [1990] also studied poultry family situation with preschool children were workers at two plants with exposure to highly not significantly associated with symptoms. forceful, highly repetitive jobs and compared Repetitive movements demanding precision them to other poultry workers with less was found to have an OR of 1.2 for neck pain. exposure. Neck case definition required symptoms and physical examination findings Hales and Fine [1989] compared 89 female that met predetermined criteria. Exposure workers in 7 high exposure jobs to 25 female assessment was based on hand/wrist poultry workers in low exposure jobs assessment of forceful and repetitive jobs. No employed in poultry processing. Neck case assessment of neck repetition was performed. definition required symptoms and physical Job turnover was around 50% at plant 1 and examination findings that met predetermined 70% at plant 2 making survivor bias a strong criteria. Exposure assessment was based on possibility. hand/wrist assessment of forceful and repetitive jobs. No assessment of neck repetition was Kuorinka and Koskinen [1979] studied performed. Twelve percent of workers in high occupational rheumatic diseases and upper limb risk jobs versus none in low risk jobs were found strain among 93 scissor makers and compared to have neck MSDS. them to the same group of department store assistants (n=143) that Luopajärvi et al. [1979] In a study of VDT users in a range of jobs used as a comparison group. Temporary

2-7 workers and Milerad and Ekenvall [1990] compared the self-reported neck and neck/shoulder those with recent trauma were excluded from symptoms between dentists and pharmacists. the scissor makers group. Exposure assessment Dentists had been considered the high risk included videotape analysis of scissor maker group because of awkward postures and tasks, however exposure assessed for the hand repetitive use of small handtools. Exposure was and wrist region and not the neck. No formal based on self-reports. The authors examined exposure assessment was conducted on the several covariates and stratified by gender for shop assistants. Health assessment involved an their analysis. No difference between groups in interview and physical examination by a leisure time, smoking, systemic disease, and physiotherapist following a standard protocol. exposure to vibration. Diagnoses of tension neck syndrome were determined using predetermined criteria [Waris Ohlsson et al. [1989] studied 148 electrical et al. 1979]. In problem cases, orthopedic and equipment and automobile assemblers, physiatric teams determined case status. It is 76 former female assembly workers who quit unclear whether cashiers were excluded from within 4 years and compared these two groups the comparison group in this study as they were to 60 randomly sampled females from the in the Luopajärvi et al. [1979] study. The study general population. A case was determined by group was 99% female. questionnaire; exposure was based on job categorization and questionnaire responses. Luopajärvi et al. [1979] compared the Repetitive exposure was based upon the prevalence of neck/shoulder disorders among number of items completed per hour. The work 152 female assembly line packers in a food pace was divided into four classes: (1) Slow: production factory to 133 female shop <100 items/hr; (2) Medium: 100 to 199 assistants in a department store. Exposure to items/hr; (3) Fast: 200 to 700 items/hr; repetitive work, awkward hand/arm postures, (4) Very Fast: >700 items/hour. The OR and static work was assessed by observation increased with increasing work pace, except at and videotape analysis of factory workers. No very high paces, where there was a decrease. formal exposure assessment was conducted on This was attributed to “selective quitting of the department store workers; their job tasks subjects with complaints, only the healthiest were described as variable. Cashiers were being left in the assembly work.” excluded, presumably because their work was repetitive. No formal assessment occurred for Onishi et al. [1976] compared several groups neck/shoulder repetition. The health assessment of workers with varying exposure to repetitive consisted of interviews and physical tasks. Health outcome was based on symptoms examinations conducted by a physiotherapist, of shoulder stiffness, dullness, pain, numbness; and diagnoses of tension neck syndrome were pressure measured by strain transducer at later determined by medical specialists using which a subject felt pain; and a physical exam. these findings and predetermined criteria (95% Observation and measurements of some job CI 2.63–6.49). Age, hobbies, and housework tasks, including some measures of repetition, were considered in the analysis. were performed then job categorization was done. Based on job

2-8 physical exam findings in the case definition of categorization and job analysis, and taking into neck and neck/shoulder MSDs when account shift length, activities, number of comparing workers bagging pears versus breaks, repetitive movements of the hands, arm apples. Exposure was again based on manipulations, and length of employment, there measurements of job tasks by a representative was not a difference between workers with worker. tenderness threshold above 1.5 kg/cm² and those below with respect to age, height, weight, Schibye et al. [1995] followed up 303 sewing skinfold thickness, grip strength, upper arm machine operators at nine factories representing abduction strength, and back muscle strength. different technology levels who completed a questionnaire in 1985. In April 1991, 241 of Punnett et al. [1985] compared neck/shoulder 279 traced workers responded to the same MSDs based on symptom reporting alone in 1985 questionnaire. Operators still working 162 women garment workers and 76 women were compared to those who moved to other hospital workers such as nurses, laboratory employment in 1991. Exposure was assessed technicians, and laundry workers. There was a through a questionnaire asking type of machine low participation rate among the hospital operated, work organization factors, workers. Eighty-six percent of the garment workplace design factors, units produced per workers were sewing machine operators and day, the payment system, and the duration of finishers (sewing and trimming by hand). The employment as a sewing machine operator. sewing machine operators were described as Although the authors state that the analysis did using highly repetitive, low force wrist and not show that neck symptoms among workers finger motions, while the finishers had shoulder who had worked as a sewing machine operator and elbow motions as well. The exposed to be significantly related to exposure, exposure garment workers likely had more repetitive time, or age, there was a significant drop-out jobs than most of the hospital workers. The rate of those above 35 years. neck/shoulder cases were found to lift both the “typical” and “heaviest” loads with greater Rossignol et al. [1987] chose 38 random sites frequency than non-cases. from Massachusetts workers with more than 50 employees, and selected Sakakibara et al. [1995] found among orchard 191 workers from computer and data workers that neck shoulder MSDs based on processing services, and public utilities symptom and physical findings were and the Commonwealth Government. Subjects significantly higher when performing pear were selected after the bagging than when apple bagging. Exposure observation of the worksite. A self- was based on measurements of specific angles administered questionnaire case definition was of the neck and shoulder and job tasks in a used for neck MSD. Exposure was also based representative worker. ORs were not derived upon self-reports of number of hours worked in this study. Confounders were not checked each day with a keyboard machine with a for in this study. VDT. Analysis controlled for the

Sakakibara et al. [1987] did not include following confounding factors: age, cigarette

2-9 smoking, industry, and educational VDT between symptoms and repetitive work. The training. Ekberg studies specifically asked about “precise repetitive movements” in their Yu and Wong [1996] chose to compare 90 questionnaire and controlled for confounders data entry, data processing, and computer and effect modifiers (age, gender, having pre- programmers from an International Bank in school children) in their analyses. Milerad and Hong Kong and 61 infrequent users of VDTs. Ekenvall [1990] compared dentists and Both neck MSD case definition and exposure pharmacists, stratified by gender, and found no assessment were based on symptom data. association between neck or neck/shoulder Analysis controlled for “age and gender, and MSDs with metabolic disease, smoking, leisure other covariates” (as stated in the paper). For time, exposure, or vibration. Significant ORs of frequent VDT use an OR of 28.9 was found. 2.0 to 2.6. for neck MSDs were reported for dentists compared to pharmacists. Kuorinka and Koskinen [1979] found a significant difference in neck MSDs between Of those studies reporting no significant scissor makers (an occupation chosen for study association between repetition and neck or because of its assembly-line repetitive hand neck/shoulder MSDs, none included exposure tasks) and shop assistants (non-stereotypic, assessment or observations of the neck or non-repetitive jobs) with an OR of 4.1. In the neck/shoulder area that were both objective same study, comparing the different and independent of the hand/wrist. Several of stereotypic, repetitive jobs in scissor-making, these studies [Baron et al. 1991; Kiken et al. those in short-cycled tasks (2–9.5 sec) had no 1990; Hales et al. 1989; Ohlsson et al. 1989; significantly different prevalence of neck Luopajärvi et al. 1979] categorized workers disorders than workers in longer-cycled tasks into high and low exposure groups based (7.3–26 sec) (OR 1.6, 95% CI 0.7– 3.8). It is strictly on hand/wrist exposure and not arm, important to note that both the longer-cycled shoulder, or neck exposure. All of these studies tasks and short-cycled tasks in Kuorinka’s reported ORs below 2.0. study would have been classified as “highly repetitive” in most other ergonomic studies In the study of VDT users by Bergqvist et al. [Silverstein et al. 1987; Chiang et al. 1993; [1995a], exposure was based on self-reports Viikari-Juntura et al. 1991a; Kurppa et al. of “the presence of repeated work movements” 1991]. When comparing two groups in which for all work tasks and not specifically focused the level of repetitive exposure may not differ on the neck or neck/shoulder area. They found by much (in this case, where both groups have no significant association with neck/shoulder highly repetitive tasks), it is unlikely that one will MSDs when the variable “repeated work find a significant difference because there is not movements” was analyzed in the logistic model enough variance between the exposures. alone, but found a significant relationship with a combination of variables: (1) workers wearing Three studies [Ekberg et al. 1994, 1995; glasses, (2) who reported VDT use, and (3) Milerad and Ekenvall 1990] used health VDT use for more than 20 hours/week. In this outcomes and exposure assessments based on case, it was the combination of variables at self-reports and found significant associations higher levels of exposure (VDT use more than

2-10 20 hours per week) that was found to be with less than 6 months (or even longer) of job statistically significant. experience, thereby omitting from their study workers who may have developed their MSDs Temporal Relationship—Repetition prior to working at the job of interest, or who and Neck/Shoulder MSDs had experienced discomfort or fatigue due to Of the prospective studies of neck MSDs that new activities or a “break-in period” at work. It can be used to establish a temporal relationship is reasonable to assume that in those studies, between exposure to repetitive work and neck given the exclusions required by the case or neck/shoulder disorders, the study by definitions, the onset of exposure was prior to Jonsson et al. [1988] fulfills all the four study the onset of neck/shoulder MSDs in the criteria. Jonsson’s study was a followup of the majority of participants. cohort studied by Kilbom et al. [1986], electronic workers who entered the study without MSDs. Exposure assessment pertaining Consistency in Association for specifically to the neck/shoulder area was Repetition and Neck/Shoulder MSDs completed three times over 3 years. In the studies fulfilling the four criteria [Ohlsson et al. 1995; Jonsson et al. 1988; Kilbom et al. In the longitudinal study by Ohara et al. [1976], 1986], significantly positive associations the authors attributed the increase in neck between neck MSDs and repetitive work were symptoms in cash register operators to the found. Many more studies involved workers in introduction of new electronic cash registers repetitive work from a range of industries placed at unsuitable heights. They noted an (VDT workers, dentists, electronic assembly, increase in repetitiveness and an increase in sewing machine operators, etc.), comparing awkward and static postures by cash register symptom prevalences to those in less repetitive operators using the new registers. The authors jobs. There was also significant association reported a relationship between static loading between neck and neck/shoulder MSDs and and MSDs and found that a subsequent jobs with repetitive tasks, with ORs between reduction in exposure to static loading resulted 1.6 and 5.9 [Onishi et al. 1976; Kuorinka and in less worker disability (sick leave). Koskinen 1979; Rossignol et al. 1987; Vihma et al. 1982; Kamwendo et al. 1991; Andersen Although temporality cannot be obtained from and Gaardboe et al. 1993b; Ekberg et al. cross-sectional studies, several studies 1994, 1995; Schibye et al. 1995] indicating attempted to insure that disorders developed that workers exposed to higher levels of work following the exposure being studied. In certain risk factors have greater rates of neck and studies [Baron et al. 1991; Kiken et al. 1990; neck/shoulder symptoms. None of the studies Hales et al. 1994; Hoekstra et al. 1994], the that failed to find significant associations carried health outcome definition excluded persons out exposure assessment of the neck or reporting symptoms prior to the job or neck/shoulder. reporting acute injury thought to be unrelated to work, insuring that exposure preceded MSD Coherence of Evidence for occurrence. Other studies excluded participants Repetition

2-11 Studies outside the epidemiologic literature give exposure assessments for their analyses and did supportive evidence that repetitive work is not conduct specific neck, shoulder, or upper related to neck/shoulder disorders. Stevens et extremity (apart from hand/wrist) exposure al. [1966] found that the neck injuries among assessment. (Only one of the studies finding fork-lift truck drivers were from repetitive, significant associations did so using hand/wrist extreme head rotations needed for the exposure assessment.) The possibility of operation of fork lift trucks and introduced the misclassification affecting the results must be a sideways-sitting driver forklift. Eklund et al. consideration. [1994] reported following up on a “sideways- sitting” forklift (in an unpublished study); these FORCE drivers experienced neck pain three times as Definition of Force for Neck and often as other drivers on traditional Neck/Shoulder MSDs forklifts—indicating that moderate head rotations during long periods of time can be For our review, we included studies that more risky than short term and extensive head examined force or forceful work or heavy loads rotations. Nicholas [1990] reported in his to the neck and neck/shoulder, or described discussion on pathophysiologic mechanisms of exposure as strenuous work involving the upper sports injuries that a low-load force with high extremity that generates loads to the trapezius repetition results in a gradual deterioration of muscles. Most of the studies that examined tissue strength from strain to fatigue to force or forceful work as a risk factor for deformation, with prefailure symptoms, such as neck/shoulder had several concurrent or pain on use, a common clinical sign of early interacting physical work load factors. inflammation from overuse. Force has generally been defined as: (1) either Exposure-Response Relationship for externally as a load or internally as a force on a Repetition body structure, or (2) a force magnitude expressed in newtons or pounds or as a There were no studies reviewed that showed a proportion of an individual’s strength capacity, clear dose-response relationship between that is, of a person’s MVC, usually measured repetition and neck and neck/shoulder MSDs. by EMG. Most studies that have dealt with force loading of the neck or stress generated on Conclusions Regarding Repetition the neck structures are from biomechanical The association between neck or studies performed in the laboratory. These neck/shoulder MSDs and repetitive work studies are not included in this document. In the epidemiologic studies reviewed, force is usually was found to be statistically significant in 19 estimated by either questionnaire, studies using different epidemiologic biomechanical models, in terms of weight lifted, approaches and under different circumstances electromyographic activity, or the variable, “ of exposure. Twenty-seven studies found ORs heavy physical workload.” above one; of these, 13 were above 3.0. Almost all the studies (6 of 8) with non- Seventeen studies reported results on the significant associations used hand/wrist association between force or forceful work (in

2-12 combination with repetition) and neck and maximum voluntary contraction (MVC) were neck/shoulder MSDs. Of the 17 studies of calculated. Post-intervention (which involved force and neck MSDs, 11 found a statistically changes to the workstation, tools, and significant positive association between force organization of work)—see Table 2-4 at the and neck or neck/shoulder MSDs; six others end of the chapter for further explanation, the had non-significant findings. In terms of mean static trapezius load in assemblers was magnitude of the association, two studies had reduced from 4.3% MVC to 1.4%, the mean ORs greater than 3.0, seven were between 1.0 static trapezius load in VDT users reduced and 3.0, and two were less than 1.0. Six from 2.7% MVC to 1.6% MVC (post- studies did not report their results in terms of intervention). Sick leave also decreased ORs or prevalence rate ratios (PRRs) but considerably. Because so many interventions reported that the findings were statistically were involved in this study, it is not clear to significant at the p<0.05 level. what intervention changes the decrease in sick- leave per man-labor years might be attributed. Studies Meeting the Four Criteria for Bjelle et al. [1981] compared 13 workers of an Force and Neck/Shoulder MSDs industrial plant consecutively seen at a health There were no studies that met the four clinic with acute, nontraumatic shoulder-neck epidemiologic evaluation criteria for forceful pain not due to causative disease or exertion of the neck. malformation compared to 26 controls, matched on age, gender and place of work. Studies Not Meeting the Four Criteria for Force and Neck/Shoulder MSDs In another cohort study, Veiersted and Åaras [1994] carried out a cohort study of four Westgaard [1994] followed 30 female groups, 15 female assembly workers making chocolate manufacturing workers, 17 of whom telephone exchanges, 27 female VDT users, 25 contracted trapezius myalgia within 6 to female data entry operators, and 29 male VDT 51 weeks compared to those workers who did users. Case definition for neck MSD was not. Diagnosis was based on both symptoms based on self-reports. However, and physical exam. There were prospective musculoskeletal sick leave per man-labor years interviews every 10 weeks to detect symptoms was also used as an endpoint. For force of muscle pain. Daily “pain diaries” were also estimate the load on the kept by subjects. trapezius was measured by electromyography Exposure assessment consisted of measured static muscle tension recorded by EMG. (EMG). Interviews concerning exposure at work were also conducted prospectively every 10 weeks Quantification of the muscle load was done by for 1 year. Only 55% of the subjects were ranking the interval estimate (0.1 s) to produce retained during the full study; however, the an amplitude probability distribution function. ‘drop-outs’ were follow-up subjects and had Both the total duration and number of periods no significant differences in static muscle tension per minute when muscle activity was below 1% compared to the participants.

2-13 Viikari-Juntura et al. [1994] , the third about 10 to 30% of the maximum contraction longitudinal study discussed under force and of the trapezius. This level, 10 to 30% of the neck and neck/shoulder MSDs, used maximum contraction, was found by Tanii et al. questionnaire to assess neck symptoms and [1972] to induce static fatigue significant based exposure on job category, comparing enough to produce electromyographic changes. 688 machine operators, 553 carpenters, and Hales et al. [1989] and Kuorinka and 591 office workers. For the initial evaluation, Koskinen [1979] reported statistically observation of work sites were performed. In significant ORs (1.6 and 4.1, respectively) for multivariate analysis occupation, age, and the association between neck MSDs and high current smoking were significant predictors in levels of force combined with high levels of change from no neck trouble to severe neck repetition estimated for the hand/wrist areas. trouble (ORs were not given for logistic There were no separate force measurements model.) for the neck area. Both studies controlled for age, gender, and length of employment in the Wells et al. [1983] evaluated letter carriers with current job. Two of the four studies that used an increased load on the shoulder from a estimated hand and wrist exposure mailbag. Letter carriers were compared to gas measurement combinations of force and meter readers (without heavy loads) and postal repetition (but carried out no neck, shoulder, or clerks. A telephone survey was used to obtain upper extremity exposure measurements) found both symptoms and exposure. This analysis non-significant associations between neck was adjusted for age, number of years on the MSDs and force/repetition exposure [Baron et job, quetelet (body mass) ratio and previous al. 1991; Kiken et al. 1990]. work experience. Temporal Relationship—Force and Of the studies in the tables, five (that did not Neck/Shoulder MSDs fulfill all the inclusion criteria) examined the risk See temporal relationship above in Repetition factor, force, either as trapezius muscle load and Neck/Shoulder MSDs. (using EMG), or as forceful work in combination with other risk factors [Aåras Consistency in Association for Force 1994; Wells et al. 1983; Onishi et al. 1976; and Neck/Shoulder MSDs Andersen and Gaardboe 1993a; Punnett Both Kilbom et al. [1986] in their cross- 1991]. Wells et al. [1983] found a significant sectional study and Jonsson et al. [1988] in difference (p<0.05) in reported neck pain between letter carriers and postal clerks and their follow-up cohort studies found that attributed it to weight from carrying heavy mail bags on shoulder straps. In the Wells study, confounding due to age, number of years on the “time spent in physically heavy work before the job, previous work experience, or quetelet present employment” appeared as a strong risk ratios was ruled out. As noted above, Onishi et factor for deterioration of health of the al. [1976] reported that the operations studied neck/shoulder area (specifically, the health required continuous contraction of the trapezius outcome was for the cervicobrachial region in muscle to sustain the arms, estimated to be the Jonsson study). Jonsson et al. [1988] noted that the physical demands of the previous jobs

2-14 had only been assessed at the initial interview Other laboratory studies have examined muscle and constituted a subjective estimate. damage that may arise during static muscle However, the relationship was strengthened by contractions used to maintain static postures. the consistency of findings in the prospective Hägg et al. [1990] proposed that while and cross-sectional studies. maintaining static postures (that have low force levels), the same low-threshold motor units are Coherence of Evidence for Force contracted repeatedly for prolonged periods, There is coherence with the biological during which time they work close to their mechanisms proposed by Hagberg [1984] for maximal capacity. This may lead to injury of occupational muscle-related disorders, such as these units, despite the fact that the total tension neck syndrome. The first mechanism workload is low. This hypothesis was recently concerns stress on the trapezius and supported by a longitudinal study by Veiersted surrounding muscles of the neck from heavy et al. [1993] who investigated the number of physical exertion that causes rupture of the rest-pauses during muscle fiber activity using muscle’s z-discs, and an outflow of metabolites EMG recording from neck and shoulder from the muscle fibers, and activation of pain muscles. Among subjects performing machine- receptors through edema or other mechanisms. paced repetitive packing work, those with This temporary high, local stress involving symptoms had fewer rest-pauses (0.9 versus eccentric contractions in the shoulders 8.4 per minute) and a tendency toward shorter improves with time through a re-orientation of total duration of rest-pauses in the muscle fiber collagen in the muscles. This mechanism is activity of their trapezius muscle when offered as an explanation for MSDs in workers compared with those without symptoms. These unaccustomed to the work. The second mechanisms of decreased blood flow, mechanism is from local decreased blood flow increased metabolite concentration, and (ischemia), as seen in assembly workers whose prolonged activation of certain small units at tasks involved dynamic, frequent contractions near maximum capacity may explain the chronic above 10 to 20% of the MVC and few rest myofascial shoulder pain seen in workers breaks. Reduced blood flow was found to be performing repetitive assembly work with static correlated with myalgia (muscle pain) and loading of the trapezius muscles [Hagberg and ragged red fibers in 17 patients with chronic Kvarnström 1984; Larsson et al. 1988]. myalgia thought to be associated with static load during repetitive assembly work [Larsson Exposure-Response Relationship et al. 1990]. The third pathophysiologic for Force mechanism for muscle pain deals with energy Åaras [1994] reported that by reducing static metabolism disturbance, caused by long-term muscle loading (an indication of force static contractions of the muscles. Supporting measurement) through equipment changes this theory was a study finding a correlation among VDT users, as well as improving between muscle tension and plasma myoglobin workplace organization, he was able to among patients with regional muscle tenderness decrease the prevalence of neck pain, decrease and pain [Dammeskiold-Samsøe et al. 1982]. the number of sick days taken, and cause a significant reduction in trapezius load measured

2-15 by EMG in VDT operators. prevalences of neck symptoms and/or physical findings in workers in occupations or tasks Conclusions Regarding Force requiring some combination of forceful, There is evidence for forceful exertion and repetitive movements, and extreme or static neck MSDs in the epidemiologic literature. postures of the upper extremity, and compared Most of the epidemiologic studies reviewed them to workers in occupations without those defined “forceful work” for the neck/shoulder requirements. as work activities that involve forceful arm or hand movements that, in turn, generate the Twenty-seven studies that considered extreme loads to the neck/shoulder area; no study or static posture found a statistically significant examined a relationship based on actual positive association between posture and neck forceful neck movements. Of the 17 studies or neck/shoulder MSDs; three had non- addressing force as one of the exposure significant findings (Table factors, 5 found statistically significant 2-1. Overall, in terms of magnitude of the associations but did not derive ORs; 2 found association, looking at both significant and non- ORs greater than 3.0, 7 found ORs from 1 to significant findings, 13 studies had estimations 3.0, and 2 studies showed ORs less than 1.0. of risk (ORs or PRRs) greater than 3.0, 9 had Many of the studies regarding measured force risk estimates between 1 (as workload, etc.) and MSDs are in the and 3, and none had an estimate less than 1.0. biomechanical and ergonomic literature. Eleven studies did not report their results in terms of ORs or PRRs; of these, all but one POSTURE found a significant relationship.

Definition of Posture for Neck and Studies Meeting the Four Evaluation Criteria Neck/Shoulder MSDs Of the 31 studies evaluating neck postures and We included those articles that mentioned neck neck MSDs, the four investigations mentioned or head postures, adverse or extreme head or above [Ohlsson et al. 1995; Jonsson et al. neck postures, or static postures of the head 1988; Kilbom and Persson 1987; Kilbom et al. and/or neck. 1986] fulfilled the four evaluation criteria. Three of these studies [Jonsson et al. 1988; Kilbom et al. 1986; Kilbom and Persson 1987], dealt with the same cohort; female electronics workers Studies Reporting on Posture as a Work Factor for Neck and Neck/Shoulder Musculoskeletal Disorders followed for 3 successive years. These studies We included 31 studies of the association found significant association between posture between extreme or static posture and neck variables and neck MSDs; however, none used and neck/shoulder MSDs, including TNS. methods that reported ORs. Studies usually focused on the different

2-16 Studies Not Meeting the Four Criteria for on questionnaire responses regarding heavy Posture and Neck/Shoulder MSDs lifting, monotonous or assembly line work, Bernard et al. [1993] carried out a cross- sitting, uncomfortable work postures (bending sectional study of 894 newspaper employees and twisting), and vibration. The psychosocial using a questionnaire survey for case definition work environment was also studied; the based on frequency, duration, and intensity of analysis was stratified for age and gender. symptoms in the neck. Exposure was based upon both questionnaire and job analysis. Time As part of a longitudinal study, Viikari-Juntura spent on the telephone was associated with an et al. [1994] studied 154 subjects from increased prevalence of neck MSDs, with a Helsinki, Finland that originally entered the slightly elevated OR of 1.4. Analysis was study in 1955, and had repeated cross- controlled for age, gender, height, psychosocial sectional exams from 1961 to 1963. During factors, and medical conditions. that time, 1084 subjects underwent cross- sectional examination. In 1985, a questionnaire Kukkonen et al. [1983] compared 104 data was sent to all subjects; 801 (74%) responded. entry operators with 57 female workers in Of the respondents, 180 lived in the Helsinki varying office tasks. Neck MSD was based on area. It was from this group that 162 pre-determined symptom and physical exam. responded. Eight were excluded due to Exposure was based on observation of illnesses. Outcome was based on questionnaire posture, movements and working techniques, data for this study — because of small number assessment of equipment, interview with of abnormal physical findings, the physical workers and supervisors. An intervention exam was eliminated from analysis. Exposure consisting of adjustment of office furniture and was also based on survey, asking the amount of equipment was carried out. The study group work with hands overhead, work in forward was given a short course of basic training on bent position, and work in twisted or bent pertinent aspects of ergonomics. Four lessons position. This analysis was controlled for on relaxation was given by means of exercises. physical and creative hobbies, with no There was no controlling of confounders. There interactions seen. was a significant decrease in tension neck In a cross-sectional study of machine syndrome among the cases involved in the operators, carpenters were compared to office intervention compared to those workers who workers by Tola et al. [1988], who used a had no change. postal questionnaire to obtain both health outcome and exposure information. Analysis Linton and Kamwendo [1989] surveyed used “occupation” to examine relationships. 22,180 employees undergoing screening Pain Drawing Diagrams were used to examinations at their occupational health care distinguish body areas. For the logistic service in Sweden. Neck cases defined from regression model a 12 month prevalence of questionnaire responses as those persons neck and shoulder symptoms on 8 days or reporting “yes” to having seen a health care more was used. The logistic regression models professional for neck pain in the last year. were adjusted for years working in an Cases were compared to “non-cases” defined occupation and age. by outcome (neck pain). Exposure was based

2-17 Welch et al. [1995] examined 39 electricians at forward flexion, upper arm elevation, and a screening convention using surveys to collect cervicobrachial disorders weakened information on symptoms and exposures. The (compared with the results that Kilbom et al. questionnaire included questions concerning the [1986] had found), but that the results still frequency of tasks performed, including the remained statistically significant in some of the percent of time spent hanging duct work. The multifactorial analyses (no numerical results analysis did not control for confounders except were reported). The most important finding, for length of employment. according to the authors, was that reallocation to more varied work tasks was a strong Strength of Association for Posture predictor of improvement over the second Ohlsson et al.'s [1995] study, discussed year. This change would have decreased static previously, compared female industrial workers loading and increased the dynamic pattern of performing repetitive tasks to referents without movements of the workers. such exposure and found significant associations (p<0.05) between (1) neck and Of those studies not fulfilling the four criteria, neck/shoulder diagnoses with time spent in results regarding extreme or static posture were neck flexion, with critical angles greater than similar to those of the studies which did fulfill 15E; and (2) neck/shoulder diagnoses and time them. Sakakibara et al. [1995] found a spent with upper arm abduction greater than significant difference in the prevalence of neck 60E. MSDs when they examined orchard workers who picked and bagged pears and two months Kilbom et al. [1986], in the initial paper later picked and bagged apples. Exposure was concerning the electronic workers, reported assessed by job analysis and posture two findings: (1) that the more dynamic the measurements of two representative workers. working technique, the fewer neck symptoms Arm and neck elevation was significantly experienced by electronic workers; and (2) that greater for bagging pears (more than 90E for the greater the average time per work cycle 75% of the time) than for bagging apples (less spent in neck flexion, the greater the association than 40% of the time). The same authors found with symptoms in the neck and neck/shoulder similar results in 1987 when only the symptoms angle. A statistically significant association of orchard workers were studied. They found (p<0.05) was also obtained from the job significant a positive association between analysis variables describing neck forward posture and neck MSDs, reporting histograms flexion and upper arm elevation and neck and (not ORs) in their article. neck/shoulder disorders. Jonsson et al. [1988], in the follow-up study, performed an analysis Although they did not mention the participation that grouped the different parts of the neck and rates in their methods, Aåras [1994], Veiersted upper extremity into a health outcome labeled and Westgaard [1994], and Bjelle et al. [1981] “cervicobrachial disorder” (unlike the cross- found significant relationships between postures sectional study by Kilbom et al. [1986] that and neck MSDs (they fulfilled the other three used “neck” and shoulder”). They found that criteria). Veiersted and Westgaard [1994] the relationships between MSDs and neck found an association between “perceived

2-18 strenuous postures” and neck MSDs (OR 7.2), of neck MSDs for scissor makers (chosen for but found that these perceived postures were their stereotypic, repetitive work using extreme not reflected in any of the conventional EMG postures) compared to shop assistants, parameters (static, median or peak loads) although no quantitative measurements or measured in the participants. One explanation observations of neck posture were reported. for these results may be information bias, if the One study by Hünting et al. [1981] showed a data concerning perceived strenuous posture fairly strong association (OR 4.9) with are from questionnaires. Another explanation constrained postures and neck MSDs in those may be that EMG testing results reflect workers having neck flexion of more than 56E parameters for a single day, whereas symptoms and an OR of 9.9 from the comparison of were asked about concerning the entire groups. Several articles with significant posture previous year. and neck MSD associations dealt with comparisons of workers in occupations chosen Several studies that carried out no independent for higher observed combinations of exposure assessment of ergonomic factors, but relied on factors and compared them to workers with fewer observed exposure stressors: Viikari- self-reported exposure found significant relationships between posture variables and Juntura et al. [1994], OR 3.9 to 4.2; Milerad neck disorders. Ekberg et al. [1994] found an and Ekenvall [1990], OR 2.6; and Wells et al. [1983], OR 2.57. OR of 4.8 for the variable “work with lifted arms,” and an OR of 3.6 for “uncomfortable sitting position” and neck MSDs. Hales et al. For those studies that did not find a significant relationship, 2 out of the 3 did not carry out [1994] found that “use of bifocals” (OR 3.8) in VDT users was significantly associated with observation or measurement (ergonomic assessment) of the neck or upper extremity neck MSDs; this variable was interpreted to be a surrogate for neck posture, as bifocals postures. Ferguson [1976] stated that seven require either neck flexion or extension for eye body dimensions were measured in the telephonists studied, but that neither discomfort accommodation when viewing a VDT screen. Bernard et al. [1994] reported that as workers’ nor aching were linked with any of these body time spent on the telephone increased, so did postures. The article does not mention the body postures that were measured. Ferguson’s the ORs for neck symptoms, and interpreted this variable as a surrogate for static posture conclusion, that “physical complaints in requiring neck deviation to cradle the telephone telephonists are probably due to static load on joints and muscles occasioned by the fixed receiver. Holmström et al. [1992] found that the odds of workers with neck MSDs reporting forward bent position determined by visual, working with hands above their shoulders for auditory greater than 4 hrs/day compared with those reporting less than 1 hr/day was 2.0, a and manipulative tasks.” Ferguson's data are contrary to the conclusions presented. These statistically significant finding. Bergqvist et al. [1995a] reported an OR of 4.4 for workers conclusions may then only be speculative. using highly placed keyboards in their logistic modeling of neck MSDs. Kuorinka and Koskinen [1979] found an increased OR (4.1)

2-19 Temporality for Extreme or Static Coherence of Evidence for Extreme Postures Or Static Postures The prospective study by Veiersted and See section above under Coherence of Westgaard [1994] followed the development Evidence for Force. of trapezius myalgia among 30 female chocolate manufacturing workers. Seventeen Exposure-Response Relationship for workers developed the MSD within 6 to 51 Specific or Static Postures weeks of starting work. Perceived strenuous The study by Ohara et al. [1976], mentioned postures on the assembly line were found to earlier, not only portrayed the multifactorial contribute to the disorders. Although retention nature of neck and shoulder MSDs, but of subjects was low (55%), the authors found documented that an increase in specific and that the “drop-outs” did not differ in exposure static postures by cash register operators using estimates and symptom reporting from those new registers placed on unsuitable counter retained in the study. The prospective study of heights increased symptoms in neck MSDs. Viikari-Juntura et al. [1994] used self-reported symptoms and exposure defined by Several studies have suggested an occupational status to find a temporal exposure-response effect between increased relationship between the development of severe level or duration of exposure and an increase in and persistent severe neck pain and jobs number of cases of neck MSDs. Burt et al. involving dynamic work, static posture, and [1990], in their investigation at a major urban whole body vibration, as compared to office newspaper, found that an increase in the self- work. reported percentage of time spent typing at VDT keyboards was associated with a Consistency in Association for moderate increase in neck symptoms. (Job Extreme or Static Postures and analysis found a significant relationship between Neck/Shoulder MSDs independent observation of time spent typing Of the 31 studies we reviewed reporting results and self-reported time) Keyboard time was on the association between specific or static considered by the authors to be a surrogate for posture and neck and neck/shoulder MSDs, 27 time spent with the neck held in static postures found statistically significant associations. There with arms unsupported. Rossignol et al. [1987] were many different studies reporting ORs of found that the prevalence of neck symptoms greater than 3.0 with CIs above 1, indicating among 1,545 clerical workers increased with that the effects were not explained by chance. the number of hours per day using VDTs. Consistent associations were also found in Knave et al. [1985] found that, among VDT those studies dealing with specific postures and operators, total daily working hours and time neck MSDs across many industries, from fish spent at the VDT screen were significant risk workers [Ohlsson et al. 1995] to fruit pickers factors for neck pain. Andersen and Gaardboe [Sakakibara et al. 1995], to assembly line [1993a,b] found an exposure-response workers [Jonsson et al. 1988], to garment relationship between persistent neck pain and workers [Vihma et al. 1982; Andersen and years of being a sewing machine operator, Gaardboe 1993a,b]. controlling for age.

2-20 performed in this study, and vibration was likely Conclusions Regarding Extreme or to be confounded by neck twisting and static Static Postures loads. Overall, the strength of the association (OR ranging from about 1.6 [Vihma et al. 1982] to Conclusions—Vibration and Neck or 7 [Veiersted and Westgaard 1994], dropping Neck/Shoulder MSDs the outliers) between specific postures and We conclude that there is insufficient neck MSDs was similar between studies using evidence to support an association between the most restrictive criteria and carrying out a vibration and neck or neck/shoulder MSDs prospective design and those that used based on strength-of-association criteria. Too symptom-based health outcome or self- few studies of neck or neck/shoulder MSDs reported exposures to static or specific have examined the relationship between postures and cross-sectional methods. We exposure to vibration and to draw any conclude that there is strong evidence for conclusions about their relationship. support of an association between static or specific postures and neck and neck/ shoulder NECK OR NECK/SHOULDER MSDs MSDs based on strength of association criteria. AND THE ROLE OF CONFOUNDERS A positive relationship has been observed As in many MSDs, prevalence of neck and between exposure to this risk factor and neck neck/shoulder disorders tends to increase with or neck/shoulder MSDs in studies where age. Therefore, it is important that studies take chance, bias, and confounding can be ruled out into account when examining the strength of with reasonable confidence. occupational versus non-occupational factors. Age and gender were the primary potential VIBRATION confounders that investigators addressed in many of the studies on neck and neck/shoulder No study of neck MSDs met the four criteria to MSDs (The tables at the end of the chapter list address strength of association between summaries of each of the articles and include vibration and neck MSDs and only one of the which particular covariates or confounders reviewed studies in the tables mentioned neck were considered.) These were either dealt with MSDs and vibration. Viikari-Juntura et al. by logistic regression modeling, as in the case [1994] selected study groups for their of age (e.g., Andersen and Gaardboe [1993a]; longitudinal study based on different work Rossignol et al. [1987]; Tola et al. [1988]; exposures. Machine operators exposed to Ohlsson et al. [1989]; Baron et al. [1991]), static work and whole-body vibration were through matching of case subjects and referents compared to carpenters exposed to dynamic (e.g., Vihma et al. [1982]), or through study of physical work and presumably no vibration to a single gender (e.g., Luopajärvi et al. [1979]; see whether occupational status was related to neck MSDs. Results found that the OR for Hünting et al. [1994]), or stratifying by gender progressing from no neck pain to moderate to [Sakakibara et al. 1995]. Most studies severe neck trouble was from 3.9 to 4.2; for performed univariate analysis prior to logistic operators compared to carpenters; a significant regression to consider factors which needed to difference. No vibration measurements were be introduced into the logistic models as

2-21 confounders or covariates. There is also reasonable evidence for a causal relationship between highly repetitive work and Almost all the studies we reviewed accounted neck and neck/shoulder MSDs. Most of the for the confounders of age and gender. Many epidemiologic studies reviewed defined of the studies controlled for leisure exercises “repetitive work” for the neck as work [Andersen and Gaardboe [1993a,b] smoking activities which involve continuous arm or hand (Linton [1990]; Milerad and Ekenwall [1990]; movements which affect the neck/shoulder Bergqvist et al. [1995a,b]; Viikari-Juntura et al. musculature and generate loads to the [1994]), medical conditions [Bernard et al. neck/shoulder area; fewer studies examined [1994]; Hales et al. [1994]). Reviewing the relationships based on actual repetitive neck methods and results of these studies, the movements. The two studies which measured confounding factors do not account for the repetitive neck movements by head position consistent relationship that is found with the (using frequency and duration of movements), work-related factors. and fulfilled the four criteria, found strong associations with neck/shoulder MSDs. In CONCLUSIONS those studies defining repetitive work as Interpreting association for individual continuous arm or hand movements affecting workplace factors is difficult, as most the neck/shoulder, nine studies found epidemiologic studies of MSDs used statistically significant ORs greater than 3.0. populations selected because of multiple factors Eight studies fulfilled all the criteria except for (such as forceful exertion and repetitive tasks). objective exposure assessment and measured Unlike laboratory experiments, one cannot repetition for the hand/wrist, not the neck. Of isolate exposure factors, nor alter some factors these, three had statistically significant ORs while keeping others constant to insure greater than 3, and five had non-significant accuracy in examining, recording, and ORs, all under 2.0. interpreting results. However, one can examine the body of epidemiologic evidence and infer There is reasonable evidence for forceful relationships. There have been over 40 exertion and neck MSD found in the epidemiologic studies which have examined epidemiologic literature. Most of the work factors and their relationship to neck and epidemiologic studies reviewed defined neck/shoulder MSDs. Many studies identified “forceful work” for the neck/shoulder as work individuals in heavier industrial occupations and activities which involve forceful arm or hand compared them to workers in light industry or movements which generate the loads to the office environments. Other studies identified a neck/shoulder area; no study examined a symptomatic group of workers, or those with relationships based on actual forceful neck symptoms and physical exam abnormalities, movements. Of the 17 studies and compared them to asymptomatic workers at the same worksite, or to population addressing force as one of the exposure referents, and looked for differences in factors, five studies found statistically significant exposure. These approaches, although quite associations but did not derive ORs; two different, by and large have chosen to focus on studies found ORs greater than 3.0, seven similar workplace risk factors. These include studies from 1 to 3.0, and 2 studies with ORs repetition, forceful exertions, and constrained less than 1.0. or static postures, usually found in combination.

2-22 There is strong evidence that working groups have included interventions to decrease with high levels of static contraction, prolonged workplace risk factor exposures, including static loads, or extreme working postures decreasing repetitive work and less extreme involving the neck/shoulder muscles are at working postures, have shown a decrease in increased risk for neck/shoulder MSDs. incidence of neck MSDs, and an improvement Consistently high ORs (12 studies found in symptoms among affected workers. These statistically significant ORs over 3.0) for tension data provide additional evidence that these neck syndrome associated with static postures disorders are related to work factors. or static loads have been found. However, cumulative exposure-response data The epidemiologic data are insufficient to is lacking, although VDT studies using document relationship of vibration and neck surrogate exposure variables suggests a disorders. The few prospective studies which relationship.

2-23 Table 2-1. Epidemiologic criteria used to examine studies of neck MSDs associated with repetition

Investigator Risk indicator blinded to case (OR, PRR, IR Participatio Physical and/or exposure Basis for assessing neck Study (first author and or p-value)*,† n rate $70% examinatio status exposure to repetition year) n

Met all four criteria:

Ohlsson 1995 3.6† Yes Yes Yes Observation or measurements

Met at least one criterion:

Andersen 1993b 6.8† Yes Yes Yes Job titles or self-reports

Baron 1991 2.0 No Yes Yes Job titles or self-reports

Bergqvist 1995b 6.9† Yes Yes Yes Job titles or self-reports

Hales 1989 1.6 Yes Yes Yes Job titles or self-reports

Kamwendo 1991 1.65† Yes No NR‡ Job titles or self-reports

Kiken 1990 1.3 Yes Yes Yes Job titles or self-reports

Knave 1985 NR† Yes No NR Job titles or self-reports

Kuorinka 1979 4.1† Yes Yes NR Job titles or self-reports

Luopajärvi 1979 1.6 Yes Yes Yes Job titles or self-reports

Onishi 1976 3.8† NR Yes NR Observation or measurements

Sakakibara 1987 NR† Yes No NR Job titles or self-reports

Schibye 1995 3.3† Yes No NR Job titles or self-reports

Yu 1996 28.9† Yes No NR Job titles or self-reports

Met none of the criteria:

Liss 1995 1.7† No No No Job titles or self-reports

Ohlsson 1989 1.9 NR No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

2-24

Table 2-2. Epidemiologic criteria used to examine studies of neck/shoulder MSDs associated with repetition

Investigator Risk indicator blinded to Basis for assessing (OR, PRR, IR Participation Physical case and/or neck/shoulder exposure Study (first author and or p-value)*,† rate $70% examinatio exposure to repetition year) n status

Met all four criteria:

Jonsson 1988 NR†,‡ Yes Yes Yes Observation or measurements

Ohlsson 1995 4.6† Yes Yes Yes Observation or measurements

Met at least one criterion:

Andersen 1993a 4.6† Yes No Yes Job titles or self-reports

Bergqvist 1995a 3.6 Yes No Yes Observation or measurements

Blåder 1991 NR† Yes Yes No Job titles or self-reports

Ekberg 1994 15.6† Yes No NR Job titles or self-reports

Ekberg 1995 1.2† Yes No NR Job titles or self-reports

Hünting 1981 9.9† NR Yes NR Observation or measurements

Milerad 1990 2.1† Yes No NR Job titles or self-reports

Punnett 1991 1.8 Yes No NR Observation or measurements

Rossignol 1987 1.8–4.6† Yes No NR Job titles or self-reports

Vihma 1982 1.6† NR No NR Observation or measurements

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

2-26

Table 2-3. Epidemiologic criteria used to examine studies of neck MSDs associated with force

Investigator Risk indicator blinded to case (OR, PRR, IR Participatio Physical and/or Basis for assessing neck Study (first author and year) or p-value)*,† n rate $70% examinatio exposure status exposure to force n

Met at least one criterion:

Baron 1991 2.0 No Yes Yes Job titles or self-reports

Hales 1989 1.6 Yes Yes Yes Job titles or self-reports

Kiken 1990 1.3 Yes Yes Yes Job titles or self-reports

Kuorinka 1979 4.1† Yes Yes NR‡ Job titles or self-reports

Luopajärvi 1979 1.6 Yes Yes Yes Job titles or self-reports

Veiersted 1994 6.7† No Yes NR Observation or measurements

Viikari-Juntura 1994 3.0† Yes No Yes Job titles or self-reports

Wells 1983 2.57† Yes No NR Job titles or self-reports

Met none of the criteria:

Liss 1995 1.7† No No No Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported.

2-28

Table 2-4. Epidemiologic criteria used to examine studies of neck/shoulder MSDs associated with force

Risk Investigator Basis for assessing indicator Participatio Physical blinded to case neck/shoulder exposure Study (first author and year) (OR, PRR, IR n rate $70% examinatio and/or exposure to force or p-value)*,† n status

Met at least one criterion:

Åaras 1994 NR†,‡ NR No NR Observation or measurements

Andersen 1993a 3.2 Yes No Yes Job titles or self-reports

Bjelle 1981 NR† NR Yes Yes Observation or measurements

Jonsson 1988 NR† Yes Yes Yes Job titles or self-reports

Punnett 1991 0.9 (females) Yes No NR Observation or measurements 1.8 (males)

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

2-30

Table 2-5. Epidemiologic criteria used to examine studies of neck MSDs associated with posture

Investigator Risk blinded to case indicator Participation Physical and/or Basis for assessing neck Study (first author and year) (OR, PRR, IR rate $70% examination exposure exposure to posture or p-value)*,† status

Met at least one criterion:

Bernard 1994 1.4† Yes No Yes Job titles or self-reports

Ferguson 1976 NR‡ Yes No No Observation or measurements

Hales 1994 3.8† Yes Yes Yes Job titles or self-reports

Kamwendo 1991 1.65† Yes No NR Job titles or self-reports

Kukkonen 1983 3.6† NR Yes Yes Job titles or self-reports

Kuorinka 1979 4.1† Yes Yes NR Job titles or self-reports

Linton 1990 3.5† Yes No NR Job titles or self-reports

Onishi 1976 3.8† NR Yes NR Observation or measurements

Sakakibara 1987 NR† Yes No NR Observation or measurements

Sakakibara 1995 1.5 Yes Yes NR Observation or measurements

Veiersted 1994 7.2† No Yes NR Observation or measurements

Viikari-Juntura 1994 3.9–4.2† Yes No§ Yes Job titles or self-reports

Welch 1995 7.5 Yes No No Job titles or self-reports

Wells 1983 2.57† Yes No NR Job titles or self-reports

Yu 1996 784.4† Yes No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on posture alone (i.e., posture plus force, repetition, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported. §Physical examinations were not analyzed because there were too few cases.

2-32

Table 2-6. Epidemiologic criteria used to examine studies of neck/shoulder MSDs associated with posture

Investigator Risk indicator blinded to case Basis for assessing (OR, PRR, IR Participation Physical and/or neck/shoulder exposure Study (first author and year) or p-value)*,† rate $70% examination exposure to posture status

Met all four criteria:

Jonsson 1988 NR†,‡ Yes Yes Yes Observation or measurements

Kilbom 1986 NR† Yes Yes Yes Observation or measurements

Ohlsson 1995 NR† Yes Yes Yes Observation or measurements

Met at least one criterion:

Åaras 1994 NR† NR No NR Observation or measurements

Bergqvist 1995a 4.4† Yes No Yes Observation or measurements

Bjelle 1981 NR† NR Yes Yes Observation or measurements

Blåder 1991 NR† Yes Yes No Job titles or self-reports

Ekberg 1994 4.8†, Yes No NR Job titles or self-reports 3.6†

Holmström 1992 2.0† Yes No Yes Job titles or self-reports

Hünting 1981 9.9† NR Yes NR Observation or measurements

Milerad 1990 2.6† Yes No NR Job titles or self-reports

Rossignol 1987 1.8, Yes No NR Job titles or self-reports 4.0†, 4.6†

Ryan 1988 NR† Yes No Yes Observation or measurements

Tola 1988 1.8† Yes No NR Job titles or self-reports

Vihma 1982 1.6† NR No NR Observation or measurements

Viikari-Juntura 1991a 1.5 Yes Yes§ NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on posture alone (i.e., posture plus force, repetition, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

2-34

Table 2–7. Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD Prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen and Cross- 701 female sewing machine Outcome: Case of chronic pain 26.2% General SMO compared Participation rate: 78.2%. Gaardboe sectional operators (SMO), compared was defined as continuous pain population: to: (1) General 1993a to 781 females from the lasting for a month or more after 9.9% population: Examiners blinded to general population of the beginning work and pain for $30 Internal OR=3.2 control/subject status. region and internal referent days within the past year. referent (2) Internal 2.3-4.5 group of 89 females from group: referent group: Controlled for age, having children, the garment industry. Exposure: Job categorization 6.7% OR=4.9 not doing leisure exercise, smoking based on “authors’ experiences 2.0-12.8 socioeconomic status. as occupational health Logistic Model: physicians” and involved crude Years as SMO: Age-matched exposure groups assessment of exposure level 0 to 7 years: and controls. and exposure repetitveness. 1.9 Jobs involving high 8 to 15 years: 1.3-2.9 Logistic regression limited to a repetitiveness (several 3.8 combined neck/shoulder case times/min) and low or high force, >15 years: 2.3-6.4 definition. and jobs with medium 5.0 repetitiveness (many times/hr) 2.9-8.7 No difference in education, marital combined with high force were Age $ 40 status, number of children. classified as high exposed jobs; years: 1.5 jobs with medium repetitiveness 1.1-2.2 Poor correlation between and low force and jobs with Children (>0): degenerative X-ray neck changes more variation and high force 1.3 and cervical syndrome. were classified as medium 0.8-2.0 exposed. Job titles such as Exercise: 0.9 Most frequent diagnosis among teachers, self-employed, trained 0.6-1.3 study group was “cervicobrachial nurses, and the academic Socioeconomic fibromyalgia” significant for test of professions were “low status: 1.29 trend with exposure time in years. exposed.” 0.7-2.3 Smoking: 1.39 Chronic neck pain and palpatory 0.99-1.9 findings: Sensitivity: 0.85; Current Specificity: 0.93. Exposure: 1.3 0.9-1.9

(Continued)

2-36

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen and Cross- From a historical cohort of Outcome: Measured by health Referents: 0 to 7 years: Participation rate: 78.2%; logistic Gaardboe sectional 424 sewing machine interview and exam of the neck, OR=1 2.3 0.5-11 regression limited to a combined 1993b operators, 120 were shoulder and arm. Case of neck/shoulder case definition. randomly selected and 82 chronic pain was defined as 8 to 15 years: exposed workers were continuous pain lasting for a 6.8 1.6-28.5 Age-matched exposure groups categorized by number of month or more after beginning and controls. years of employment: 0-7 work and pain for at least >15 years: years, 8-15 years and 30 days within the past year. 16.7 4.1-67.5 Examiners blinded to greater than 15 years. Physical examination: Restricted control/subject status. These were compared to a movements in the cervical spine Age at least 40 referent group of 25 and either palpatory tenderness years: 1.9 O.9-4.1 Controlled for age, having children, auxiliary nurses and home in cervical segments or not doing leisure exercise, helpers. A total of 107 irradiating pain or tingling at Children >0 smoking, socioeconomic status. subjects participated. maximum movements or positive years: 0.5 0.1-1.7 foraminal test. Poor correlation between Exercise: 1.4 0.6-2.96 degenerative X-ray neck changes Exposure: Exposure and cervical syndrome. categorization broken down Smoking: 1.5 0.7-3.3 according to current Most frequent diagnosis among occupational status by job title. Current high study group was “cervicobrachial Classification into exposure exposure: 1.6 0.7-3.6 fibromyalgia” significant for test of groups based on author’s trend with exposure time in years. experiences as occupational health physicians and involved Chronic neck pain vs. palpatory crude assessment of exposure findings: Sensitivity: 0.85; level and exposure Specificity: 0.93. repetitiveness. High exposure jobs: Involved high repetition/high force or high repetition/low force or medium repetition/high force. Medium exposure jobs involved medium repetition/low force and low repetition and high force. Low exposure jobs were low repetition/low force.

(Continued)

2-37

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Baron et al. Cross- 124 grocery checkers Outcome: Based on symptom 16% 5% Odds of neck Participation rate: 85% checkers; 1991 sectional using laser scanners (119 questionnaire and physical pain, 55% non-checkers in field study. females, 5 males) exam. Case defined as having checkers vs. Following telephone survey 91% compared to 157 grocery positive symptoms and a positive non-checkers: checkers and 85% non-checkers. non-checkers (56 females, physical exam. Symptoms must OR=2 0.6-6.7 101 males); excluded 18 have begun after employment at Examiners blinded to worker’s job workers in meat, fish, and supermarket of employment and and health status. deli departments, workers in current job; lasted one week under 18 and pregnant or occurred once a month during Adjusted for duration of work, age, workers. the past year; no history of hobbies, systemic disease obesity. acute injury to part of body in question. Total repetitions/hr ranged from 1,432 to 1,782 for right hand and Exposure: Based on job 882 to 1,260 for left hand. categorization. Estimates of repetition and average and peak Average forces for cashiers were forces of hand and wrist based low and peak forces medium. on observed and videotaped Force was not analyzed in the postures, weight of scanned models. items, and subjective assessment of exertion. Multiple awkward postures of all upper extremities recorded but not Specific neck assessment was analyzed in models. not done. Statistically significant increase in neck MSD with increase in years “checking.”

(Continued)

2-38

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bergqvist et Cross- Office workers using Outcome: Neck discomfort— any Neck: Asympto- Tension neck Participation rate: 92% of 353 al. 1995a sectional VDTs, (n=260), 198 discomfort over the last 12 61.5% matic syndrome: office workers. females; symptomatic months; intense neck Female: workers Females no cases compared to non- discomfort— as above, if 63% children: Adjusted for age and gender. cases. occurred in last 7 days and Male: 57% OR= 2.0 0.7-5.6 interfered with work. Factors included in analysis: Age, Females with gender, smoking, children at home, Outcome: Physiotherapist's TNS: 22% children: negative affectivity, tiredness- diagnosis of: (1) tension neck Female: OR=6.4 1.9-21.5 related stress reaction, stomach- syndrome (TNS): ache/pain in 25% related stress reaction, use of the neck; feeling of tiredness Male: 13% Limited rest spectacles, peer contacts, rest and stiffness in neck; possible break: OR=7.4 breaks, work task flexibility, headache; pain during 3.1-17.4 overtime, static work position, non- movements; muscular Cervical Too highly use of lower arm support, hand in tenderness; (2) cervical diagnosis: place non-neutral posture, repeated diagnoses—ache/pain in neck 23% keyboard: movements with risk of tiredness, and arm; headache; decreased Female: OR=4.4 height differences mobility due to cervical pain 25% 1.1-17.6 keyboard/elbow, high visual angle during isometric contraction; Male: 20% Cervical to VDT, glare on VDT. often root symptoms such as Diagnoses: numbness or parathesias. Age >40 Found that “frequent overtime” OR=2.7 protective for cervical diagnoses Exposure: Based on 1.0-7.2 OR=0.48 (0.23, 0.99). observation— static work Spectacles: posture, nonuse of lower arm OR=4.0 Examiner and workplace support, hand in non-neutral 1.3-12.5 investigators blinded to case and position, insufficient leg space at Static Posture: exposure status. table, repeated movements with OR=5.1 risk of tiredness, specular glare 0.6-42.5 There are problems with present on VDT. Measured: Spectral glare: interpreting results because of Height difference of VDT OR=1.9 multiple comparisons and multiple keyboard-elbow, high visual 0.9-4.2 models. angle to VDT. Stomach reactions: Not all significant findings OR=3.9 presented in paper. 2.0-7.7 Tiredness: 1.9 1.0-3.5

(Continued)

2-39

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bergqvist et Cross- 322 office workers; VDT Outcome: Neck discomfort—any Neck Current VDT Participation rate: 76%. al. 1995b sectional users compared to non- discomfort over the last 12 discomfort: work: VDT users. months; intense neck/shoulder 60% OR=1.4 0.8-2.4 Adjusted for age and gender. 52% interactive, discomfort—as above, if 29% data entry, 19% non- occurred in last 7 days and Intensive neck discomfort VDT users. interfered with work. Intense Intense associated with VDT work over 20 neck neck hr and having stomach reactions Outcome: Physiotherapist's discomfort: discomfort: often and repetitive movements: diagnosis of tension neck 7.4% OR=0.5 0.2-1.8 OR=3.9 (1.1, 13.8). syndrome (TNS)—ache/pain in the neck; feeling of tiredness Tension Tension Originally 535 workers queried in and stiffness in neck; possible neck neck 1981. Of those, 182 had left the headache; pain during syndrome: syndrome: workplace (quit, retired, etc.). movements; muscular 21% OR=1.0 0.5-1.9 Possible bias from “healthy worker tenderness. effect.” TNS Diagnosis: Exposure: Based on self- <20 hr/week Covariates considered: Children at reporting of VDT use. VDT VDT: 1.2 0.4-3.7 home, smoking, negative users categorized into data affectivity, stomach-related stress entry or interactive VDT users. >20 hr/week reactions, tiredness-related stress VDT: 0.7 0.3-1.5 reactions. Organizational factors considered: limited or excessive TNS diagnosis peer contacts, limited rest break with bifocal or opportunity, limited work task progressive flexibility, frequent overtime. glasses at VDT work and $20 For cervical diagnoses: Excess OR hr/week VDT suggested for combined work duration: occurrence of VDT work of OR=6.9 1.1-42.1 >20 hr/week and specular glare on the VDT screen.

(Continued)

2-40

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Bernard et al. Cross- Of a total population of Outcome: Health data and 26% (case) ÕÕ Females: Participation rate: 93%. 1994 sectional 3,000 workers in the psychosocial information were OR=2.1 1.4-2.4 editorial, circulation collected using a self- Cases with Examiners blinded to case and classified advertising and administered questionnaire. daily neck Number of hr exposure status. accounting departments, Definition: Presence of pain, pain: 22% spent on 1,050 were randomly numbness, tingling, aching, deadline/week Analysis controlled for selected for study and 973 stiffness or burning in the neck (30 to 39 hr vs. confounders, age, gender, height, participated. Those fulfilling occurring $ once a month or 0 to 10 hr) psychosocial factors, medical case definition compared to 7 days continuously within the OR=1.7 1.4-3.0 conditions. those workers not fulfilling past year, reported as definition. moderately severe. The Work variance Psychosocial scales analyzed by symptom must have begun (continually splitting the responses into during the current job. Workers changing work quartiles, then comparing the 75% with previous nonoccupational load; response score to the 25% injuries to the relevant area occasionally response score for deriving the were excluded. vs. often) ORs in each scale. OR=1.7 1.2-2.5 Exposure: Based on In sub-analysis of jobs having observation of work activity Time spent on comparable number of males and involving keyboard work, work the telephone females. Only number of hr spent pace, posture, during a typical (4 to 6 hr vs. 0 on deadline/week and perceived day of a sample of 40 workers to 2 hr): lack of importance for ergonomic with and 40 workers without OR=1.4 1.0-1.8 issues by management were symptoms. Exposure to work significant. organization and psychosocial Perceived lack factors based on questionnaire of importance responses. for ergonomic issues by management: OR=1.9 1.4-2.4

(Continued)

2-41

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ferguson Cross- 418 telephonists Outcome: Symptoms by Tele- Chi sq=11.01 Participation rate: 95%. 1976 sectional interviewed questionnaire. Neck ache phonists: (df=2), p<0.005 categorized on 3-point Uncomfort- Although author states the discomfort scale: (1) very able or following: “Discomfort, aching, and comfortable, (2) barely very other symptoms are common, comfortable, and (3) uncomfort- important but usually neglected uncomfortable, very able neck problems in telephonists which uncomfortable. ache =26% could be ameliorated by ergonomic job and equipment,” the results of Exposure: Personal and social his study did not support his attributes and attitudes to conclusion. aspects of the work and the equipment were obtained by Neither discomfort nor aching was questionnaire. Seven body linked to any of the body postures dimensions were measured, and observed. standing posture was categorized by observation Height and weight were not related against a grid according to to discomfort or aching. predetermined criteria. Multiple correlations not helpful in identifying combinations of personal, equipment, environmental or other variables predictive of aching and discomfort.

(Continued)

2-42

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hales and Cross- Of 96 female workers Outcome: Period prevalence— Period Period Outcome: Neck Participation rate: 93%. Fine 1989 sectional employed in 7 high symptoms in last 12 months by prevalence: prevalence: symptoms: exposure jobs in poultry questionnaire. Case defined as: 21% 13% RR=1.64 Adjustment for age and duration of processing: 89 were Pain, aching, stiffness, 0.4-3.19 employment. compared to 23 of 25 numbness, tingling or burning in Point Point Outcome: Neck female workers in low the neck and symptoms began prevalence: prevalence: symptoms and Examiner blinded to case and exposure jobs. after employment at the plant; 12% 0% physical: exposure status. were not due to a previous OR injury or trauma to the joint; indeterminate Exposure based on repetitive and lasted >8 hr; and occurred 4 or because of “0" forceful hand/wrist motions and more times in the past year. cell not neck exposure assessment.

Point prevalence: Determined by Estimated OR 80% of workers involved in job physical exam of the neck using by adding 1 to rotation program. standard diagnostic. Tension each cell in neck syndrome: Palpable muscle crude 2 X 2 No information collected on non- tightness, hardening or pain $ 3 table: 3.69 work related risk factors. (on 8 point scale) on passive or resisted neck flexion or rotation. 0.4-164 Cervical root syndrome: Pain $ (on 8 point scale) radiating from neck to one or both arms with numbness in the hand criteria. Case must also fulfill symptom definition.

Exposure: Observation and walk-throughs; jobs categorized as high exposure and low exposure based on estimates of force and repetition of hand maneuvers.

(Continued)

2-43

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hales et al. Cross- Telecommunication workers Outcome: Self-administered 9% ÕÕ Lack of Participation rate: 93%. 1994 sectional (n=518, 416 females, 117 questionnaire and standardized decision males) in 3 offices, physical exam (PE). Case making Physician examiner blinded to employed > 6 months. defined as: Pain, aching, opportunities: worker case status. stiffness, burning, numbness or OR=4.2 2.1-8.6 Logistic analysis adjusted for "Cases" fulfilling neck tingling lasting >1 week or demographics, work practices, work-related MSD definition >12 times a year; no previous Use of work organization, individual compared to non-cases. traumatic injury to neck; bifocals: factors; electronic performance occurring after employment on OR=3.8 1.5-9.4 monitoring; DAO keystrokes; current job within the last year Denver DAO keystrokes/day. and positive PE—moderate to Lack of a ORs for psychosocial variables worst pain experienced with productivity represent risk at scores one tension neck or cervical root standard: standard deviation above mean syndrome. OR=3.5 1.5-8.3 score compared to risk at scores one SD below mean. Exposure: Assessed by Fear of being questionnaire and observation; replaced by Because of readjustments and changes of workstations during number of keystrokes/day; no computers: study period, measurements of exposure questions were OR=3.0 1.5-6.1 VDT workstations considered specifically aimed at the neck unreliable and excluded from region. High analyses. information Physical workstation and processing Number of hr spent in hobbies and recreational activities not postural measurements were demands: significant. taken but not analyzed in OR=3.0 1.4-6.2 models. Although keystrokes/day found not Job requiring a significant, data available was for variety of workers typing an average of 8 tasks: words/min over 8-hr period. OR=2.9 1.5-5.8 97% of participants used VDT $6 hr so not enough variance to Increasing evaluate hr of typing. work procedure: Over 70 variables analyzed in OR=2.4 1.1-5.5 models may have multiple comparison problem.

(Continued)

2-44

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hunting et al. Cross- 308 of 400 apprentice and Outcome: Three-symptom 16% ÕÕ ÕÕ ÕÕ Participation rate: 75%. 1994 sectional journeymen, electricians definitions used; most restrictive from one labor union includes neck symptoms 3% with 1 to 3 years 98% of participants were male. participated. occurring $once/month or lasting medical worked: OR=1 >1 week during past year, and visits, Stratified by most experienced vs. no previous traumatic injury to missed 4 to 5 years least experienced electrician, by site. work, or worked: years worked, by age group, light duty OR=1.3 current work as an electrician. Exposure: Questionnaire dealing with lifting activities, working 6 to 10 years Analysis of specific work factors overhead, working with hand worked: (repetition, force, extreme posture, tools. OR=1.6 vibration, or combinations of risk factors) not analyzed in this paper >10 years which dealt with prevalence of worked: symptoms among electricians. OR=1.3

(Continued)

2-45

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kamwendo et Cross- 420 medical secretaries; Outcome: Questionnaire using 6 63% ÕÕ OR for work Participation rate: 96%. al. 1991 sectional compared those frequently point scale ranging from “very period with office having neck pain to those often” to “almost never” and prevalence. machines 5 hr Neck symptoms associated with a less frequently having pain. Nordic Questionnaire. Definition or more/day: "poorly experienced psychosocial of neck MSD: Discomfort, ache, 33% point 1.65 1.02-2.67 work environment.” or pain during previous year; prevalence. whether they had pain in last 7 Working >5 Age, length of employment days, whether pain prevented 15% with years: OR=1.6 significantly related to neck pain. them from doing daily duties. constant 0.9-2.8 10 questions on psychosocial neck pain. Sitting 5 or Questionnaire included work environment included. more hr/day: psychosocial scales, length of OR=1.9 employment, part-time or full-time Exposure: Based on 0.86-2.6 work, average hr sitting working questionnaire. Low exposure with machines. was regarded as 1 to 4 hr sitting or working with office machines, Ability to influence work, a friendly high exposure was regarded as spirit of cooperation between co- 5 to 8 hr. workers, being given too much to do significantly positively associated with neck pain.

(Continued)

2-46

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kiken et al. Cross- 294 poultry processors. Outcome: Period prevalence— Plant #1: Plant #1: Participation rate: 98%. 1990 sectional Plant #1 (n=174) based on questionnaire. Case— (High (Low Plant #2 (n=120) pain, aching, stiffness, burning, exposure) exposure) OR= Analysis stratified by gender and numbness or tingling in the neck, Any symp- Any symp- age. began after employment at the toms: 34% toms: 16% 2.2 0.9-5.0 plant; not due to previous Period Higher exposure jobs (HE) were accident or injury outside work; Period preva- located in the receiving, lasted >8 hr and occurred 4 or prevalence: lence: 3% evisceration, whole bird grading, more times in the past year. 9% 2.9 0.4-21.4 cut up and deboning departments. Point Lower exposure jobs (LE) were Point prevalence: Based on Point preva- located in the maintenance, symptom and physical exam prevalence: lence: 3% sanitation, quality assurance and using standard diagnostic 4% 1.3 0.2-11 clerical departments. criteria. Case must fulfill symptom definition listed above. Plant #2: Examiners blinded to case and Plant #2: (Low expo- exposure status. Exposure: Observation and (High sure) Any walkthrough; jobs categorized exposure) symptoms: OR= 30% of workers in job rotation as high exposure and low Any 11% program may influence exposure based on observed symptoms: 3.9 1.5-10.2 associations. force and repetition of hand 42% maneuvers. Period Annual turnover rate -50% at plant Period prevalence: 1 and 70% at plant 2; making prevalence: 3% survivor bias a strong possibility. 5% 1.8 0.2-15.2 Point Point prevalence: prevalence: 0% 1% ÕÕ ÕÕ

(Continued)

2-47

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Knave et al. Cross- 400 VDT operators from Outcome: Questionnaire— Results Results ÕÕ ÕÕ Participation rate: Initially exposed 1985 sectional 4 industries using VDTs >4 symptom questionnaire based on estimated estimated 97%; referent 100%; Phase IV hr/day; compared to frequency and intensity scores: from from Typing hr exposed 84% referents 84%. 157 office employees negligible=1, slight=2, histogram: histogram: significantly Cases and referents matched on without VDT work at the pronounced=3. related to neck age and gender. same industries. Rt. side of Rt. side of symptoms. Exposure: Based on self- neck: 5% neck: 5% Musculoskeletal complaints assessment “hrs of typing.” A Dose-response grouped in analysis; because of special gaze direction instrument Lt. side of Lt. side of relationship large number of comparisons, recorded time spent looking at neck: 20% neck: 0% found between some without a prior hypotheses, reliable conclusions limited to VDT screen. Observation was registered p<0.001. conducted but not included in work duration analysis. and musculo- Significant difference between skeletal females and males in reported neck complaints. symptoms. No statistical difference between cases and referents in discomfort scores, but “tendency towards higher discomfort scores for shoulder, neck, and back among the exposed group.” No difference in cases and referents in whether work was “interesting” or they had a “positive attitude” towards work. Age, smoking, educational status, and drinking did not correlate with symptoms. Females reported more symptoms than males in both referent and case groups.

‘Registered’ total work hr associated with musculoskeletal symptoms p<0.05.

(Continued)

2-48

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kukkonen et Cross- 104 female data entry Outcome: Questionnaire— Data entry 28% 2.3 1.1-4.6 Participation rate: Not reported. al. 1983 sectional/ workers. 60 data entry stiffness and pain in the neck groups: Inter- operators (noted as “study and shoulder region, frequency 47% Examiners blinded to case status. vention group”) were grouped with of symptoms and localization. 44 data entry operators Physical exam (PE): A clinical No adjustment for confounders. who worked at another functional examination Tension Tension bank and were compared performed by a physiotherapist. neck neck Examiner blinded to case status. with 57 female workers in syndrome syndrome varying office tasks. Exposure: Observation of in study in data Average duration of employment posture, movements and group pre- entry 3.5 years. working techniques, interven- comparison assessment of characteristics tion: 54% group pre- Intervention consisted of: of desk, chair, equipment, interven- Adjustment of desk, chairs, data interview with foremen and tion: 43% processing equipment individually workers to get determination of to suit each worker, who was physical, mental, and social Tension Tension instructed to carry out adjustments environment at workplace. neck neck herself. Document holders were Foremen and workers were syndrome syndrome added. The study group was interviewed so that the in study in data given a short course of basic organization of work and the group post- entry training on pertinent aspects of physical, mental, and social interven- comparison ergonomics. Four lessons on environment at the workplace tion: 16% group post- relaxation was given by means of could be determined. interven- exercises. tion: 45% Physiotherapy was given to workers for whom the doctor prescribed—17 from the study group and none from the first reference group had treatments.

(Continued)

2-49

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kuorinka and Cross- 93 scissor makers, (n=90 Outcome: Symptoms and 61% 28% Scissor makers Participation rate: 81%. Koskinen sectional females, 3 males) physical examination—two vs. referents: 99% female study group, no 1979 compared with 113 female tender spots symptoms of neck OR=4.1 2.3-7.5 significant age difference. department store shop stiffness and fatigue/ weakness Used Waris [1979] criteria for assistants from and/or palpable hardenings + examination which called for Luopäjarvi’s 1979 study. muscle tenderness in neck Short cycle blinding of examiners, otherwise it movements. Physiotherapist tasks vs. long- was not mentioned. Excluded those with examined workers, diagnoses cycle tasks No association between tension seropositive rheumatic were from predetermined criteria and tension neck syndrome and: (1) age, (2) affections as well as [Waris 1979]. In problem cases neck duration of employment, and cashiers. orthopedic and physiatric teams syndrome: (3) weight/height2. handled cases. OR=1.64 0.7-3.8 Total workload for the number of pieces handled in one year Exposure: Based on job significantly associated with analysis from work history of tension neck syndrome previous year from production Although authors state no and salary forms. Conducted relationship between short cycled record review of hr and longer cycled tasks; both groups of tasks would be worked/task, production classified as highly repetitive using statistics, absences: used only Kilbom, Silverstein’s and other cases where 80% of hr cross- criteria. Lack of variance in checked (n=76). Work methods comparison groups. for each type of station Authors noted: “earlier unpublished analyzed. Stations classified questionnaire pertaining to according to dominance of activities outside factory — extra inspection or manipulation of work, hobbies, did not indicate scissors, and length of cycle correlations with work...” using observation and video- Found that “diseases” seem to taping. Observations made accumulate in same individuals. looking at hand/wrist force, Physical workload was low. repetition and hand grasp. A slight trend towards tension Calculated index for wrist neck being more common in deviation. manipulation tasks than in —Work methods for each work inspection but not statistically station analyzed: Cycle time. significant. —Total workload during investigation/year recorded individually as pieces handled.

(Continued)

2-50

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Linton 1990 Cross- 22,180 employees Outcome: Cases defined from 18% had ÕÕ Monotonous Participation rate: Authors had sectional undergoing screening questionnaire responses as seen health work and poor access to all workers’ records; examinations at their those persons reporting “yes” to care psychosocial 85% of working population has occupational health care having seen a health care profes- environment: occupational health care services. service in Sweden. 85% of professional for neck pain in the sional for OR = 3.6 2.8-4.6 the Swedish workforce is last year. neck pain Analysis stratified for age, gender. covered by health care Lifting and services. Exposure: Based on 31% had poor Lifestyle factors asked: Exercise, questionnaire responses— experi- psychosocial eating, smoking, alcohol Cases compared to “non- questions asked regarding enced neck environment: consumption. cases” defined by outcome. heavy lifting, monotonous or pain OR=2.7 2.0-3.6 Groups selected a priori assembly line work, sitting, On univariate analysis, heavy which would represent uncomfortable work postures Uncomfortable lifting, monotonous work, exposure as well as little or (bending or twisting), vibration. posture and uncomfortable posture, and no exposure for Psychosocial work environment: poor vibration had elevated ORs. Sitting psychosocial variables. Work content, workload, social psychosocial did not. support. environment: OR=3.5 2.7-4.5 On univariate analysis, eating regularly and smoking had elevated ORs. Alcohol and exercise did not.

Authors caution direct comparison of ergonomic and psychosocial variable’s ORs. The scales were not consistent for the different factors measured.

(Continued)

2-51

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Liss et al. Cross- 1,066 of 2,142 dental Outcome: Mailed survey, case 43% 30% 1.7 1.1-2.6 Participation rate: 50% from both 1995 sectional hygienists from Ontario definitions based on Nordic groups. Canada Dental Hygienists Questionnaire, percent reporting Association compared to neck symptoms >7 days in past Study population >99% female. referent group, 154 of 305 12 months. Had to modify dental assistants who do their work or No association with duration of not scale teeth. Exposure: Based on mailed were unable to employment. survey and self-reported work at some answers—length of practice, point, Not controlled for confounders. days/week worked, (hygienists patients/day, patients with compared to Very low response rate, heavy calculus, percent of time dental confounders not considered, study trunk in rotated position relative assistants): has methodologic problems which to lower body, instruments used, OR=2.4 1.1-5.4 influence interpretation of results. hr of typing/week, type of practice.

(Continued)

2-52

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Luopäjarvi et Cross- Assembly line workers Outcome: Tension neck 37% 28% TNS: OR=1.56 Participation rate: 84%. al. 1979 sectional (n=152 females) compared syndrome (TNS): Neck stiffness 0.9-2.7 to shop assistants in a and fatigue/weakness and two Had seen a Workers excluded from department store tender spots and/or palpable doctor for neck participation for previous trauma, (n=133 females). hardenings + muscle tenderness symptoms: arthritis and other pathology. in neck movements. OR=4.38 Cashiers excluded from 2.1-9.24 No difference in mean ages comparison group. Exposure: Observation, video between exposed and referents. analysis, and interviews used to assess exposure to repetitive Examined only females. arm work, static muscle work affecting neck/shoulder area. Factory opened only short time so no association between duration of employment and MSDs possible.

Social background, hobbies, amount of housework not significant.

(Continued)

2-53

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Milerad and Cross- 99 dentists randomly Outcome: Based on telephone Pharma- Participation rate: 99%. Ekenvall 1990 sectional selected from Stockholm questionnaire. Neck symptoms at 54% cists: 26% 2.1 1.4-3.1 dentist registry who any time before the interview Analysis stratified by gender. practiced $ 10 years ("lifetime prevalence"). Further Male: 45% Male: 18% 2.6 1.2-5.0 compared to analyzed according to Nordic No difference in leisure time 100 pharmacists selected questionnaire as to duration Female: Female: exposure, smoking, systemic from all pharmacists in during last 12 months and during 63% 32% 2.0 1.3-3.1 disease, exposure to vibration. Stockholm. last 7 days, effect on work performance and leisure Symptoms increased with age in activities, and sick leave. female dentists only.

Exposure: Based on Duration of employment highly questionnaire. Exposures correlated with age: included: (1) abduction of arm dentists (r=0.84), pharmacists particularly in sit-down dentistry; (r=0.89). (2) work hrs/day; and (3) static postures. No relation between symptoms and duration of employment.

Equal problems dominant and nondominant sides.

Genders “equally prone to develop neck symptoms when subjected to equal work-related musculoskeletal strain.”

No analysis of exposure factors. Only discussion of “probable reasons” for high risk using work positions, flexing neck.

(Continued)

2-54

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson et al. Cross- Industrial Workers Outcome: Pain in the last Tension Tension Tension neck Participation rate: Current 1995 sectional (n=82 females) exposed to 7 days and physical exam (PE) neck: 40% neck: 13% syndrome workers: 96%; past workers: repetitive tasks with short diagnosing tension neck (industrial 86%; referents: 100%. cycles mostly far <30 sec, syndrome, cervical syndrome. workers Controlled for age. usually with a flexed neck compared to and arms elevated and Tension neck: Tightness of referents): No exposure information available abducted intermittently; 68 muscles, tender spots in the OR=3.6 1.5-8.8 to examiners, “not possible to completely blind the examiners.” former workers (mean muscles. Cervical syndrome: Cervical Cervical employment time 21 years) Limited neck movement, radiating syndrome: syndrome: Questionnaire included individual who had left the factory pain provoked by test 1% 0% Õ Õ factors, work/environment, during the seven years movements, decreased symptoms, psychosocial scales. before the study; these sensibility in hands/fingers; Muscle strength measured by workers were compared to muscle weakness of upper limb. (maximum voluntary capacity) at 64 referents with no elevation, abduction, and outward repetitive exposure at their Exposure: Videotaping and rotation of both arms measured by current jobs. observation. Analysis of dynamometer. postures, flexion of neck (critical Videotape analysis revealed angles 15E and 30E). 74 considerable variation in posture workers videotaped $10 min even within groups performing from back and sides. Average similar assembling tasks. counts of two independent Logistic models replacing repetitive readers for frequencies, work with videotape variables duration, and critical angles of found muscular tension tendency movement used. and neck flexion movements Repetitive industrial work tasks significantly associated with divided into 3 groups: (1) fairly neck/shoulder diagnoses. mobile work, (2) assembling or Inverse relationship between pressing items, and (3) sorting, duration of industrial work and polishing and packing items MSDs, largest OR employed <10 Weekly working time, work years. rotation, patterns of breaks, Assembly group has high OR (6.7) individual performance rate with regard to neck/shoulder MSD (piece rate). Only exposure compared to referents. readings from right arm were used. Significant association between time spent in neck flexion positions < 60E.

(Continued)

2-55

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson et al. Cross- Electrical equipment and Outcome: Determined by Pain in last Pain in last Participation rate: Not reported. 1989 sectional automobile assemblers questionnaire—any neck pain, 12 months: 12 months: (n=148), 76 former female neck pain affecting work ability, 39% 32% 1.9 0.9-3.7 For younger females, increase in assembly workers who quit and neck pain in the last 7 days pain occurred with increased within 4 years compared to and the last 12 months. Work Work duration of employment. 60 randomly sampled inability in inability in female from general Exposure: Based on job last 12 last 12 OR increased with increasing work population. categorization and months: months: 7% pace, except for very high paces, questionnaire—number of items 13% 2.8 0.9-8.8 which there was a decrease. completed/hr. Pain in last Pain in last 7 days: Logistic models checked for Work pace divided into four 7 days: 17% interaction and controlled for age. classes: (1) slow: <100 21% 1.9 0.7-3.6 items/hr; (2) medium: 100 to 199 Study group consisted of females items/hr; (3) fast: 200 to 700 only. items/hr; (4) very fast: >700 items/hr. Significant association between symptoms and duration of employment much stronger for workers <35 years old than workers >35 years old.

(Continued)

2-56

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Onishi et al. Cross- The following were Outcome: Based on Group I: Participation rate: Not reported. 1976 sectional compared to 101 female (1) symptoms of neck stiffness, 29% office workers: dullness, pain, numbness; (2) Body weight, weight skin fold pressure (<1.5 kv/cm²) measured by strain transducer Group II: thickness, muscle strength and grip Film rolling workers: 127 at which subject felt pain; 39% strength obtained. (females). (3) physical exam: range of motion, tests, nerve Group III: Body height and weight Subjects categorized as: compression tenderness. 23% differences not statistically significant. Exposure: Observation of job Group I: Without symptoms tasks, then job categorization. of cervico- No difference between workers brachial disorder. Film rollers wind 1 roll of 35 mm with tenderness threshold above film every 2.5 to 5 sec over 7.5 1.5 kg/cm² and those below with Group II: Subjective hr/day. respect to age, height, weight, skin symptoms in the neck, fold thickness, grip strength, upper shoulder, or upper limbs. Loading of trapezius was arm abduction strength, back examined in two workers during work activities by muscle strength. Group III: Symptoms and electromyography. clinical signs. Authors noted that continuous loading of the trapezius seems characteristic to repetitive operations where the upper limbs are used.

(Continued)

2-57

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ryan and Cross- Data process operators Outcome: Symptoms (pain, Shoulder: ÕÕ Not reported ÕÕ Participation rate: 99%. Bampton 1988 sectional (n=143). Group with ache, sore, hurts, numb, 44% highest scores (n=41) swollen, etc.) occurring symptom Interviewers blinded to designated "cases," $3 times/week with no physical only questionnaire responses. compared to lowest scores exam signs or $ weekly with (n=28). physical exam signs of muscle Neck: 43% No adjustment for confounders; tenderness present; diagnosed symptoms cases for analysis were those “myalgia” as diffuse muscle pain only with either neck, shoulder, or and tenderness. lower arm scales having higher Neck/ symptom scores compared to Exposure: Ergonomic shoulder those with low scores. assessment measuring angles symptoms and distances of each operator occurring $ Cases had higher visual glare seated at his/her workstation. 3 times index, feeling there was Wrist extension, ulnar deviation, weekly insufficient time for rest breaks, elbow angle, shoulder with no more boredom, more work stress, abduction, and shoulder flexion signs or and needed to push themselves >3 were measured. Also weekly times/week; lower peer cohesion, measured: person and furniture with signs: autonomy, clarity. Higher staff fit, eye-copy and eye-keyboard 44% support and work pressure. fit, elbow-keyboard height difference, popliteal-chair height Significant differences in those difference, and copy placement. trained in adjustment of their chairs.

No differences for height, weight, age, marital and parental status, handedness, time in current job, time spent keying or typing, whether this was their first job, length of training time.

Significant difference in smaller mean elbow angle and shoulder flexion of the left arm, and smaller eye-copy distance.

(Continued)

2-58

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Sakakibara et Cross- Orchard workers (n=48, Outcome: Shoulder pain Estimated Estimated ÕÕ Participation rate: 77%. al. 1987 sectional 20 males and 20 females). described as the presence of from from stiffness and pain daily. histograms histograms Stratified by gender. Compared symptoms after Pears: Apples: completion of thinning of Exposure: Observation of jobs. General fatigue, gastric pears, bagging of pears Angles of flexion of the shoulder Rt. side: Rt. side: disturbances, appetite loss and and bagging of apples and extension of the neck on 20% 9% p<0.05 headache showed no difference in (covering fruit with paper one subject were measured Lt. side: Lt. side: 9% frequency between tasks. bags while on the trees). every 25 min during a whole day 20% p<0.01 doing each task. No observation Exposure data based on Internal comparison using was made on neck repetition. measurement of one worker may same study population. not be generalized to others. Farmers worked approximately 8 hr/day for 10.6 to 13.6 days The angle of forward flexion in the each year bagging or thinning shoulder and that of extension in pears and bagging apples. the neck was statisticallly significantly positively correlated (r=0.88, p#0.01). The proportion of workers with >90E forward shoulder flexion was significantly higher for thinning out pears and bagging pears than for bagging apples.

The authors presumed that the symptoms of dizziness and tinnitus may be associated with the cochlear-vestibular symptoms of vertebral insufficiency due to continuous extension of the head.

Results presented in paper in histograms.

(Continued)

2-59

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Sakakibara et Cross- Of 65 female Japanese Questionnaire: Stiffness and Pear Apple Participation rate: 80%. al. 1995 sectional farmers. 52 completed the pain in neck region. Symptoms bagging bagging questionnaire and physical in past 12 months for $one day, Examiners not blinded to case exam in late June for or symptoms in past 12 months status due to design of study. bagging pears and late July for $8 days. Neck Neck Workers for bagging apples. pain=40% pain=25% bagging pears Same population examined two Exam: Pain in motion of the neck with neck pain times. 2nd exam occurred one joint such as flexion/extension, vs. apple month after first. These results lateral bending, and rotation. bagging used in analyses for comparison of with neck pain, two tasks. Exposure: Observation of tasks p<0.05 and measurements of Stiffness and pain during apple representative workers (only bagging may have been pain that two workers measured) . Neck pain Neck pain Workers was a residual of pear bagging in joint in joint bagging pears operations. Angle of arm elevation during motion: motion: with pain in bagging was measured in one 55.8% 36.5% joint motion vs. Number of fruit bagged/day was subject. controls apple bagging significantly more in pear bagging with pain in than in apple bagging. joint motion: PRR=1.5 Exposure measurements only 0.99-2.35 obtained on 2 workers and generalized to all workers.

(Continued)

2-60

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Schibye et al. Cohort Follow-up of 303 sewing Outcome: Nordic Neck Developing Participation rate, 1985: 94%. 1995 machine operators at nine Questionnaire— discomfort, symptoms neck symptom factories representing ache, or pain in the neck during in previous improvement in Participation rate, 1991: 86%. different technology levels the previous year; whether they year for 1991 among All participants were female. who completed had neck pain in last 7 days, and employees operators 77 of 241 workers still operated a questionnaire in 1985. whether pain prevented them maintaining compared to sewing machine in 1991. from doing daily duties. a piece- other 82 workers had another job in In April 1991, 241 of 279 work employment 1991. Among those 35 years or traced workers responded Exposure: Assessed by groups of group below, 77% had left job; among to same 1985 questions regarding type of <100 OR=0.85 0.29-2.4 those above 35 years, 57% left questionnaire. machine operated, work units/day: job. organization, workplace design, 36% 20% reported musculoskeletal Operators still working units produced/day, payment symptoms as the reason for were compared to those system, and duration of Neck Neck symptom leaving job. who moved to other employment as a sewing symptoms improvement in No significant changes in employment in 1991. machine operator. in previous other prevalences among those year for employment employed as sewing machine employees group vs. operators from 1985 to 1991; maintaining operator group: significant decrease in those who a piece- 12 month changed employment. work symptoms: As many as 50% of respondents groups of OR=3.3 reported a change in the response 100 to 125 1.4-7.7 to positive or negative symptoms units/day: from 1985 to 1991. 53% Operators always working at the 7 day same machines showed Neck symptoms: significantly higher neck symptoms symptoms OR=3.9 compared to those working at in previous 1.3-11.9 different machines year for Although the authors state that the employees analysis did not show the maintaining development of neck (or shoulder) a piece- symptoms among workers who work had worked as a sewing machine groups of operator to be significantly related to exposure, exposure time, or >125 age, there was a significant drop- units/day: out rate of those above 35 years. 61%

(Continued)

2-61

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Veiersted and Cohort 30 female chocolate manu- Outcome: Trapezius 56% ÕÕ Perceived Participation rate: 55%. Westgaard facturing workers. 17 who myalgia—neck and shoulder pain strenuous Drop-out rate may limit generaliz- ability of results although drop-outs 1994 contracted trapezius lasting >2 weeks of a degree postures: did not differ in exposure estimates myalgia within 6 to making it difficult to continue OR=7.2 2.1-25.3 and complaints. 51 weeks compared to work. At least one tender or those workers without. trigger point present. Physical Excluded subjects with: (1) no similar occupation during last 5 Prospective interviews every environment: years; (2) known musculoskeletal 10 weeks to detect symptoms of OR=0.9 0.5-1.7 disorder predisposing for myalgia; muscle pain. Daily “pain diaries” (3) neck or shoulder pain sufficient kept by subjects. Psychosocial to initiate medical visit, (4) if factors: employed <26 weeks. Exposure: Static muscle tension OR=3.3 0.8-14.2 Several anthropometric, non-work- during work was between 1 and related, general health, personality, 2% of maximal voluntary activity Perceived psychosocial, and previous employment variables included in of the trapezius muscles strenuous initial interview and follow-ups. recorded by electromyographic previous work: measurements of trapezius OR=6.7 1.6-28.5 Subjects on a fixed-wage system. muscle in earlier study. Work was mainly machine-paced. Interviews conducted Nine of 17 with trapezius myalgia prospectively every 10 weeks had sick leave after medical concerning exposure at work consultation. for 1 year. No difference in general health status, anthropometric measures. None of the models showed any effect of the “physical environ- ment.” Parameters which in- cluded exposure to draft, vibration (floor or machine), or noise. Observation time was con- siderably shorter for workers who contracted neck pain compared to status used in analysis. Non- patients had more opportunities to report a positive answer. The perceived strenuous postures were not reflected in any of the conventional EMG parameters (static, median or peak loads).

(Continued)

2-62

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Viikari-Juntura Cohort 688 machine operators and Outcome: Neck trouble, 12 month ÕÕ Carpenters vs. Participation rate: 81% machine et al. 1994 longitud- 553 carpenters compared categorized on 5 point scale prevalence office workers: operators; 79% carpenters; inal; to 591 office workers. All ("not any" to "daily"). for severe No neck pain to 89% office workers. 2 quest- male. neck pain moderate: ionnaires Exposure: Based on job for OR=1.6 1.0-2.5 Adjusted for occupation, smoking, 3 years category. Machine operators— 1984/1987 and physical exercise, age, apart static work with whole body No neck pain to duration or current occupation. vibration, carpenters—dynamic Machine severe: physical work, office operators: OR=1.6 0.8-3.0 2% had retired. workers—sedentary work. For 28/40% initial evaluation, observation of Persistently In multivariate analysis; work sites were performed. Carpenters: severe: “occupation” was only significant 25/32% OR=3.0 1.4-6.4 predictor in change from no neck trouble to moderate neck trouble. Office Machine workers: operators vs. Twisting or bending trunk not a 9/12% office workers: significant predictor of neck pain.

No neck pain to In multivariate analysis: moderate: occupation, age, and current OR=1.8 1.1-2.8 smoking were significant predictors in change from no neck No neck pain to trouble to severe neck trouble. severe: OR=3.9 2.3-6.9 Interaction between age and occupation not significant. Persistently severe: Job satisfaction not associated OR=4.2 2.0-9.0 with neck trouble and other predictors.

(Continued)

2-63

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Welch et al. Cross- 39 of 47 sheet metal Outcome: Symptom survey; 21% Compari- Percent time Participation rate: 83%. 1995 sectional workers attending a pain, aching, stiffness, burning, son group hanging duct: screening for occupational numbness or tingling in neck with no OR=7.5 0.8-68 Smoking cigarettes, average lung disease. Cases $ once/month, or lasting > one symptoms number of years working not found compared to those without week, no history of previous to be significantly different symptoms. traumatic injury. Symptoms between symptomatic and began after working as a sheet asymptomatic; other confounders metal worker and prior to (age, gender) not mentioned. retirement. Average length of employment in Exposure: Questionnaire survey trade: 33 years. obtaining types of tasks performed, tools used, Pilot study. frequency of task performance. Hanging duct work dichotomized Hrs/week using hand tools, into > and <40% of time worked. percent of time in the shop vs. time in the field not significant.

Duration of employment not included in article.

(Continued)

2-64

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Wells et al. Cross- 196 male letter carriers Outcome: Telephone interview All letter Postal All letter Participation rate: 99% among 1983 sectional compared to 203 male case status based on current carriers: clerks: carriers vs. letter carriers, 92% meter readers, meter readers and postal pain; frequency, severity, 12% 5% clerks and 97% postal clerks. clerks. interference with work, etc.; readers: score of 20 required to be a Meter OR=2.57 1.13-6.2 No significant difference in 104 letter carriers had case—more points given to neck readers: schooling and marital status. weight increased from 25 and shoulder problems that Letter 7% Letter carriers to 35 lbs. in the year prior to interfered with routine daily carriers with increased Comparison group (gas meter the study. activities. with weight vs. readers) used because of similar increased clerks: “walking rate” without carrying Exposure: Based on job weight: OR=2.63 0.9-8.8 weight compared to letter carriers. category; based on self- 12% Postal clerks neither walk nor carry reported information on weight weight. carried, previous work involving Letter Letter carriers lifting and work-related injuries. carriers with no weight More weight given to scoring neck with no increase vs. and shoulder. Outcome influenced weight clerks: results when ranking of body increase: OR=2.87 0.9-9.8 MSDs though would not influence 12% group comparisons.

Adjusted for age, number of years on the job, Quetelet ratio and previous work experience.

Study limited to males.

Letter carriers with increased bag weight walked on average 5.24 hr; those with no change in bag weight walked 4.83 hr.

Letter bag straps usually carried on the shoulder.

(Continued)

2-65

Table 2–7 (Continued). Epidemiologic studies evaluating work-related neck musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Yu and Wong Cross- 151 VDT users from an Outcome: Questionnaire survey 31.4% Frequent users Participation rate: 80%. Ages 1996 sectional international bank in Hong used to collect information on of VDTs vs. ranged from 18 to 41 years, 74% Kong; of these 90 were discomfort or ache during work infrequent between 21 to 30 years. data entry, data after starting the current job. users: processing, computer p=0.0025 Analysis controlled for “age and programmers; 61 infrequent Exposure: Questionnaire survey gender, and other covariates.” users of VDTs. on “undesirable postures” Logistic model including frequent bending of the for neck pain Queried about personal particulars, back and inclining the neck inclining neck job nature and characteristics, forwards. at work: 33.2- working posture, general health OR=784.4 18,630 conditions.

Fixed keyboard Males with significantly longer height: mean VDT working experience OR=90.1 7.6-1056 compared to females (5 vs. 2.7 years). Frequent VDT use: Non-workplace factors not OR=28.9 2.8-291.8 examined.

Female gender: OR=1.6 0.35-6.8

Age (years): OR=1.2 1.02-1.5

2-66

Table 2–8. Epidemiologic studies evaluating work-related neck/shoulder disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Åaras 1994 Prospective 15 female assembly Outcome: Assembly Workers: Number of Duration of Participation rate: Not reported. workers making musculoskeletal sick leave/man- musculoskeletal sick-leave/man- telephone exchanges. labor years; pre- and post- diagnoses: pre- labor year Study designed to evaluate if there is intervention. intervention, (days) a relationship between trapezius load 27 female VDT users. 1967 to 1974: 52 and incidence of MSD. Data Entry and VDT Users: (30.6%) Median sick 25 female data entry Survey: Pain intensity for the days pre- Other intervening variables that may operators. neck and shoulder region Number of intervention: have reduced symptoms or sick according to Nordic musculoskeletal 22.9 4.4-50.8 leave were not discussed. 29 male VDT users. questionnaire. diagnoses post- intervention, Median sick Mean static trapezius load in Exposure: Load on trapezius as 1975 to 1982: 35 days post- assemblers was reduced from 4.3% measured by EMG. (14.3%) intervention: 1.8 MVC to 1.4% (post-intervention); Quantification of the muscle load 0-34.4 mean static trapezius load in VDT done by ranking the interval Shoulder pain users reduced from 2.7% MVC to estimate (0.1 s) to produce an intensity: 1.6% MVC (post-intervention). amplitude probability distribution 3.4 function. Both total duration and 2.3-4.4 The mean intensity and duration of number of periods/min. when 2.2 neck pain showed no significant muscle activity was below 1% 1.3-3.3 reduction after intervention in the MVC were calculated. data dialogue females.

Intervention: Replacing workstands with fixed heights to workplaces easily adjustable for both sitting and standing. Hand tools were counter- balanced and adjustable arm rests introduced. For VDT operators, tables and chairs adjusted to give more relaxed position of the shoulders, operators given more work surface for keyboard and mouse, and distances between operators and screen/documents adjusted.

(Continued)

2-67

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen Cross- 701 female sewing Outcome: Case of chronic neck 34.2% General Sewing Participation rate: 78.2%. and sectional machine operators, pain was defined as continuous population: machine Gaardboe compared to pain lasting for a month or more 12.9% operators Examiners blinded to case status. 1993a 781 females from the after beginning work and pain for compared to: (1) General general population of $ 30 days within the past year. Internal population: Respondents excluded if had the region and internal referent OR=3.5 2.6-4.7 previous trauma to neck, shoulder, or referent group of Exposure: Categorization broken group: 10.1% (2) Internal arms or had inflammatory disease at 89 females from the down according to current referent group time of response. garment industry. occupational status by job title. OR=4.6 2.2-10.2 Classification into exposure Odds ratios adjusted for age, having Logistic model groups based on author’s Years as children, not doing exercise, experiences as occupational sewing socioeconomic status, smoking, and health physicians and involved machine current neck/shoulder exposure. crude assessment of exposure operator (0 to 7 level and exposure years): Age-matched exposure groups and repetitiveness. High exposure OR=3.17 0.6-16.1 controls. jobs were those involving high (8 to 15 years): OR=11.2 repetition/high force or high (>15 years): 2.4-52.3 Presented study as “general survey repetition/low force or medium OR=36.7 of health in the garment industry” to repetition/high force. Medium 7.1-189 minimize information bias. exposure jobs were those Age >40 years: involving medium repetition/low OR= 1.96 force and low repetition and high Current high 0.8-5 force. Low exposure jobs were exposure (-/+): low repetition/low force. OR=0.32 For the analysis, “length of Children (>0): 0.1-1 employment as a sewing OR =0.35 machine operator” was considered the variable of 0.1-1.9 interest, the rest were confounders. Exercise (-/+): OR=1.28 0.5-3.4 Smoking (=/-): OR=2.3 0.9-6.1

(Continued)

2-68

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen Cross- From a historical cohort Outcome: Measured by health 50.9% 46.2% Referents: Participation rate: 78.2%. and sectional of 424 sewing machine interview and exam of the neck, OR=1 Gaardboe operators, 82 were shoulder and arm. Case of Tension neck Logistic regression limited to a 1993b randomly selected and chronic pain was defined as syndrome: 40% 0 to 7 years: combined neck/shoulder case categorized by number continuous pain lasting for a OR=2.3 0.5-11 definition. of years of month or more after beginning Cervical employment: 0 to work and pain for $ 30 days Syndrome: 20% 8 to 15 years: Age-matched exposure groups and 7 years, 8 to 15 years within the past year. Physical OR=6.8 1.6-28.5 controls. and greater than 15 examination: Restricted years. These were movements in the cervical spine >15 years: Examiners blinded to control/subject compared to a referent and either palpatory tenderness OR=16.7 4.1-67.5 status. group composed of in cervical segments or 21, 25 and 36 irradiating pain or tingling at Age $ 40 Controlled for age, having children, operators from each maximum movements or positive years: OR=1.9 O.9-4.1 not doing leisure exercise, smoking, group and 25 of foraminal test. socioeconomic status. 55 auxiliary nurses and Children >0 home helpers who Exposure: Exposure categoriza- years: Poor correlation between participated in the tion broken down according to OR= 0.5 0.1-1.7 degenerative X-ray neck changes study. current occupational status by and cervical syndrome. job title. Classification into Exercise: exposure groups based on OR=1.4 0.6-2.96 Most frequent diagnosis among study author’s experiences as occupa- group was “cervicobrachial tional health physicians and Smoking: fibromyalgia” significant for test of involved crude assessment of OR=1.5 0.7-3.3 trend with exposure time in years. exposure level and exposure repetitiveness. High exposure Current high Chronic neck pain vs. palpatory jobs: Involved high repetition/ exposure: findings: Sensitivity: 0.85; high force or high repetition/ low OR=1.6 0.7-3.6 Specificity: 0.93. force or medium repetition/ high force. Medium exposure jobs involved medium repetition/ low force and low repetition and high force. Low exposure jobs were low repetition/low force.

(Continued)

2-69

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bergqvist Cross- 260 office workers Outcome: Neck/shoulder Neck/shoulder: Intensive Participation rate: 92% of 353 office et al. 1995a sectional using VDTs, (198 discomfort: Any discomfort over 61.5% neck/shoulder workers, of which 260 were VDT females); symptomatic the last 12 months; intense neck Female: 63% discomfort: users. cases compared to discomfort: As above, if Male: 57% stressful non-cases. occurred in last 7 days and stomach Adjusted for age and gender. interfered with work. reactions: OR=5.4 1.6-17.6 Examiner and workplace Physiotherapist's diagnosis of investigators blinded to case and (1) Tension neck syndrome: Repeated work exposure status. Ache/pain in the neck; feeling of movements: tiredness and stiffness in neck; OR=3.6 0.4-29.6 Factors included in analysis: Age, possible headache; pain during gender, smoking, children at home, movements; muscular Too highly negative affectivity, tiredness-related tenderness; (2) Cervical placed VDT: stress reaction, stomach-related diagnoses: Ache/pain in neck OR=4.4 0.9-60.3 stress reaction, use of spectacles, and arm; headache; decreased peer contacts, rest breaks, work mobility due to cervical pain task flexibility, overtime, static work during isometric contraction; position, non-use of lower arm often root symptoms such as support, hand in non-neutral posture, numbness or parathesias. repeated movements with risk of tiredness, height differences Exposure: Based on observation keyboard/elbow, high visual angle to an ergonomic evaluation using VDTs, glare on VDTs. data on each individual’s most common work situations: Static There are problems with interpreting work posture, nonuse of lower results because of multiple arm support, hand in non-neutral comparisons and multiple models. position, insufficient leg space at table, repeated movements with Not all significant findings presented risk of tiredness, specular glare in paper. present on VDT. Measured: Height difference of VDT keyboard-elbow, High visual angle to VDT.

(Continued)

2-70

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bergqvist Cross- 322 office workers Outcome: Neck/shoulder Neck/shoulder Neck/shoulder Participation rate: 92% questionnaire; et al. 1995b sectional from 7 Stockholm discomforts: Any discomfort discomfort: 60% discomfort: 91% physiotherapy exam; companies; VDT users over the last 12 months; intense Current VDT 82% workplace exam. compared to non-VDT neck/shoulder discomfort: As work vs. no users 52% interactive, above, if occurred in last 7 days VDT work: Examiner and workplace 29% data entry, and interfered with work. OR=1.4 0.8-2.4 investigators blinded to case and 19% non-VDT users. exposure status. Physiotherapist's diagnosis of For tension neck syndrome: accumulated Intensive neck/shoulder discomfort Ache/pain in the neck; feeling of VDT work > 5 was associated with VDT work over tiredness and stiffness in neck; PY²: OR=1.3 0.7-2.5 20 hr and having “stomach reactions” possible headache; pain during Intense often and repetitive movements. movements; muscular neck/shoulder Intense OR=3.9 (1.1-13.8). tenderness. discomfort: 7.4% neck/shoulder discomfort: Originally 535 workers queried in Exposure: Video display terminal Current VDT 1981, of those 182 had left the use: Based on self-reporting of work vs. no workplace (quit, retired, VDT use. VDT users VDT work: etc.)–possible bias from “Healthy categorized into data entry or OR=0.5 0.2-1.8 Worker Effect.” interactive VDT users. For Covariates considered: Children at Ergonomic Factors: Same as accumulated home, smoking, negative affectivity, Bergqvist 1995a. VDT work >5 stomach-related stress reactions, PY²: OR=0.8 0.3-2.5 tiredness-related stress reactions; organizational factors considered limited or excessive peer contacts, limited rest break opportunity, limited work task flexibility, frequent overtime.

For cervical diagnoses: Excess OR suggested for combined occurrence of VDT work of >20 hr/wk and specular glare on the VDT screen.

(Continued)

2-71

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bjelle et al. Case- 13 workers of Outcome: Physician diagnosed 6 with tendinitis Controls Cases had Participation rate: Not reported. 1981 control industrial plant neck/shoulder pain. without significantly consecutively seen at tendinitis longer duration Investigators completed the video health clinic with acute, Exposure: Anthropometric and and higher analyses blinded to case status. nontraumatic isometric muscle strength were frequency of neck/shoulder pain not tested with strain gauge abduction or Anthropometric data, age no due to causative instruments. Patients asked to forward flexion difference between cases and disease or perform their maximal efforts. than controls, controls. malformation compared Measurements made for the 2.5/min. to 26 controls. following contractions: Shoulder (p<0.001). Isometric strength test: Controls Matched on age, elevation at the acromion, significantly stronger in 6 of 14 tests gender, and place of abduction and forward flexion of Cases had but probably influenced by pain work. the shoulder joints at neutral significantly inhibition in cases. position, and semi-pronated. higher shoulder loads than No significant difference in cycle time Grip strength measured by controls. (9 vs. 12 min). vigorimeter. Median number Video recording of arm of sick-leave movements at work. Shoulder days loads estimated from videos. significantly Consisted of measuring the different duration and frequency of between cases shoulder abduction or forward and controls flexion of >60E. (p<001).

Electromyography measurement of shoulder load during assembly work on 3 patients and 2 healthy volunteers. Muscular load level determination made by computer analysis of myoelectric amplitude.

(Continued)

2-72

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Blåder et al. Cross- Of 224 sewing Outcome: Survey: Shoulder or Muscle Õ Age p <0.05 Participation rate: 89% for 1991 sectional machine operators neck pain in past 12 months. tenderness: questionnaire, 87% for physical from 4 plants, 199 Acromioclavicula Nationality non- exam. completed a symptom Exam: Tenderness on palpation, r joint: 15% significant survey. Of 155 who range of motion, pain during Only those with symptoms given reported shoulder or motion or isometric muscle Biceps tendon: Employment p <0.05 physical exam. Physicians and neck pain in the past contraction, active and passive 35% duration physiotherapist not blinded to 12 months, 131 were range of motion was measured symptom status. examined. by use of a goniometer. Decreased ROM: Working >30 Diagnoses were not made during 30% hr/wk p <0.05 High rate of turnover in plant. the examinations, but test forms were later analyzed by criteria Acromioclavi- Authors state that study involved from Waris [1979]. cular: 5% control group taking into account psychosocial factors, but results not Exposure: From questionnaire: included in this article. employment duration, hr/wk. Questionnaire included information on Plants selected by background, family situation, representatives of Swedish employment, job conditions, health. Labour Union familiar with work sites with similar loads. Physical exam occurred 1 to 3 months after questionnaire.

In 3 consecutive years 147 sewing machine operators left this work in the factories. 48% answered follow- up questionnaire. (17% left because of neck problems contributing to decision to leave work.)

(Continued)

2-73

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Ekberg et al. Case- Study population were Outcome: Self-administered Õ Õ 90% CI used Participation Rate: 73%. 1994 control aged 18 to 59 years, questionnaire; a modified version in this paper had to have yearly of the Nordic questionnaire Female gender: Logistic analysis adjusted for age, incomes of SEK 45,000 asking about musculoskeletal OR=15.5 3.4-71 gender, smoking, having preschool and not been on sick symptoms in the past 6 months. Immigrant: children. OR=28.3 3.1-257 leave for more than Questionnaire included 2 months in past 6 background factors, age, Current smoker: Age and having preschool children OR=8.2 2.3-29 months, not employed gender, ethnic background, were not significant factors. Repetitive in large rubber industry family situation, smoking habits, Precision in area. and exercise. Movements: Low: Ambiguity of work role, demands on OR=1 Med: OR=3.8 0.7-20 attention and work content also “Cases” had consulted Exposure: Assessed by High: OR=15.6 2.2-113 statistically significant. a community physician questionnaire; seven Light Lifting: for musculoskeletal determinants were: Low: OR=1.0 Med + High: disorders of the neck, uncomfortable sitting position, OR=49.7 9.0-273 shoulder, arm, or upper uncomfortable standing position, Lifted arms: thorax during the study physically demanding work, light Low: OR=1.0 period from semi-rural lifting (less than 6 kg), repetitive Med: OR=5.9 0.9-37 community in southern movements demanding precision, High: OR=3.7 0.4-30 Sweden. Cases had work with lifted arms, and Work Pace: Low: OR=1 to have been ill monotonous work position. Med: OR=7.6 1.6-36 immediately prior to Rating scales were based on Rushed: OR=10.7 2.2-52 physician visit and average duration of hours per ORs for controls have been on sick day of each item of exposure. with MSD symptoms in both leave at most less than neck and shoulder and other body 4 weeks. No trauma, 52 items on psychosocial work parts: infectious cause, conditions reduced to 8 factors Repetitive accident, malignancy, by factor analysis: psychological Precision rheumatic disease, work climate, quality of work Movements: OR=7.5 abuse, or pregnancy. content, work pace, demands on 2.4-23 attention, work planning, job Light lifting: OR=13.6 Controls were security, job constraints, and 4.8-39 randomly selected from work role ambiguity. Lifted arms: OR=4.8 Swedish insurance 1.3-18 registry. Uncomfortable sitting positions: OR=3.6 1.4-9.3

(Continued)

2-74

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Ekberg et al. Cross- 637 of 900 residents Outcome: Based on modified Symptoms neck: Õ Gender: Participation rate: 73%. 1995 sectional between the ages of Nordic questionnaire; case Male: 33% OR=1.3 1.1-1.5 18 to 59 years, with an defined as the presence of Female: 53% Symptom responses in neck and average yearly income symptoms during the past Immigrant shoulder correlated (r=0.56) and of $ $8000 U.S. dollars. 6 months. Shoulder: Status: collapsed into one variable for the Male: 35% OR=1.3 1.0-1.6 analyses. Exposure: 20 questionnaire Female: 40% items on physical work Repetitive Age, smoking, exercise habits, family conditions which were factor movements situation with preschool children not analyzed. Self-reported demanding significantly associated with perception of physical work precision: symptoms. environment factors considered: OR=1.2 1.0-1.3 Uncomfortable sitting or standing Social work climate, demands on position; physically demanding High work attention, work planning, job security work; light lifting; repetitive pace: OR=1.2 1.0-1.3 and job constraints not significantly movements demanding precision; associated with symptoms. work with lifted arms, Low work monotonous work position. content lack of stimulation and Questionnaire on work variation: organization, work content and OR=1.3 1.1-1.5 relations in the work situation. Work role ambiguity: OR=1.2 1.0-1.3

(Continued)

2-75

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Holmström Cross- Of 2500 construction Outcome: Self-reported history Hands above Õ Participation rate: 71%. et al. 1992 sectional workers randomly of musculoskeletal problems was shoulder selected from 4,159 obtained through a mail survey. <1 hr/day 1.1 0.8-1.5 Neck/shoulder pain related to active members of 1 to 4 hr/day 1.5 1.2-1.9 increasing age, smoking, weight inactivity during free time, height trade union registry of Case of “neck and shoulder pain” >4 hr/day under 185 cm. the south of Sweden, defined as: Pain, ache, 1,773 (71%) discomfort from the Hands at waist 2.0 1.4- 2.7 Controlled for age, physical factors. participated. This neck/shoulder are experienced <1 hr/day /1 to 4 group was sometimes often or very often hr/day Dose-response relationship for represented by all during the past 12 months. >4 hr/day working with hands above shoulder 1.0 0.7-1.3 level. construction trades 1.1 0.9-1.3 except painters, Case of “considerable neck and Stooping Stress index showed a dose- electricians and shoulder pain” defined as neck <1 hr/day response. Stress questions plasterers. All and/or shoulder trouble with 1 to 4 hr/day pertained to rushing, job pressure, participants must “severe” or “very severe” >4 hr/day 1.2 0.8-1.6 and inability to relax. have worked in the functional impairment. past 6 months, Kneeling Psychosocial factors strongly associated with neck and/or shoulder including short periods Exposure: Data on physical <1 hr/day 1.0 0.8-1.3 trouble and neck and shoulder pain of sick leave or workload, psychosocial factors 1 to 4 hr/day 1.4 1.1-1.8 when age and physical factors kept unemployment. and individual and employment >4 hr/day constant in logistic models for related factors obtained from mail 1.5 1.1-2.1 psychosocial pre-rate ratio, “high” survey. Sitting level compared with “low” level for <1 hr/day 1.4 1.1-1.8 considerable neck pain; the following 1.4 1.1-1.8 psychosocial scales were 1 to 4 hr/day significant: >4 hr/day 1.5 1.1-2.1 Qualitative demands: 1.4 (1.0-2.0) Quantitative demands: 3.0 (2.1-4.0) Roofers 0.6 0.3-1.0 Solitary work: 1.5 (1.2-1.8) Plumbers 1.6 0.9-2.7 Anxiety (health): 3.2 (2.5- 4.0) Floor Psychosomatic: 5.0 (3.6-6.9) Machines/ Tools. Psychological: 4.7 (3.6-6.0) Stress: 3.4 (2.6-4.2)

0.7 0.4-1.2 The following were not significant: Discretion, support, under- 1.6 Õ stimulation, anxiety (work), job 1.5 Õ satisfaction, quality of life. 1.3 Õ 1.1 Õ

(Continued)

2-76

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hünting et al. Cross- VDT users: 53 data Outcome: Questionnaire: Medical findings Medical Medical Participation rate: Not reported. 1981 sectional entry; 109 Symptoms of pain, stiffness in shoulder and findings in findings: conversational VDT fatigue, cramps, numbness, neck: shoulder and No adjustment for age and gender. users; 78 typists; tremor scaled as: Daily, neck: compared to occasionally, seldom, never; Blinding of examiners not mentioned 55 “traditional office Conversational Traditional Conversational in paper. workers” not using Medical Exam: Included an VDT users: 28% office terminal VDT VDTs or typewriters. anamnesis and palpation of workers: users vs. trad. Medical findings in neck and shoulder painful pressure points and 11% office workers: significant in data entry workers for tendons and tendon insertion OR=1.35 0.6-3.1 head inclination greater than 56E vs. points in the shoulders, arms, <56E. Not significant in and hands. Typewriter: Typewriter vs. conversational terminal workers or 35% trad. office typewriters. Exposure: (1) Questionnaire, workers: (2) Observation and OR=3.18 1.3-2.6 Medical findings in neck and shoulder measurements of work-station, significant for typists with head and (3) Body posture measured Data Entry Data entry rotation greater than 20E compared to using method described by terminal VDT terminal users <20E. Hünting et al. 1980b. users: 38% vs. trad. Office workers: The lower the table and keyboard OR=9.9 3.7-26.9 heights, the more frequently pains in the shoulder, neck, and arms. No document holders used. Authors concluded the higher the table, the higher the documents, the better the posture of the head and trunk. Increased neck/shoulder findings occurred with increased turning of the head or head inclination.

Job satisfaction, relationship with colleagues, superiors, decision making abilities, use of skills not significantly different among groups.

(Continued)

2-77

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Jonsson Cohort Electronics Workers Outcome: Three separate Severe neck Severe neck Predictors of Participation rate: 72%. et al. 1988 (n=69 female) out of physical exams at yearly disorders: disorders: change of initial 96 workers. intervals (one initially) assessing After 1 year: 11% initially health status Predictors of deterioration were tenderness on palpation, pain or 24% from 2nd to 3rd previously physically heavy jobs, restriction with active and examination: high productivity (after 1 year), and passive movements; symptoms 22% at 2nd previous sick leave. in previous 12 months with exam Palpation regard to character, frequency, tenderness, Predictors of improvement were duration, localization, and relation At 3rd exam, 38 neck/ shoulder reallocation, physical activity in spare to work or other physical subjects angle: OR=1.6 time, and high productivity (after 2 activities. Analyzed if score on reallocated to years). any symptom of 2 or greater than varied tasks had Shoulder on a 4 point scale; “severe” improved (16% elevated, % of Predictors of remaining healthy were symptom score = 4. of these had work-cycle: work without elevating the shoulders severe OR= 1.04 and satisfaction with work tasks. Carried out at outset of study: symptoms) MVC of forearm flexors, Satisfaction Subjects reallocated to new tasks shoulder strength, handgrip, 26% with with work characterized as more dynamic and heart rate using a bicycle unchanged colleagues: varied: Non-sitting, no inspection of ergometer and rating of working OR=25 small details on printed circuit boards, perceived exertion. conditions standing and walking, occasionally deteriorated Satisfaction sitting, caretaking work, surveillance Exposure: Computerized via two further with work of machinery, assembling of bigger video recordings (rear and side), tasks: OR=24.5 and heavier equipment. real time; obtained frequency and duration of working postures and movements, neck flexion greater than 20E.

(Continued)

2-78

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Kilbom et al. Cross- 106 of 138 female Outcome: Three separate MSD symptoms Õ Logistic Participation rate: 77%. The authors 1986 sectional assemblers in two physical exams at yearly in the neck/ Regression followed up on the non-participants electronic intervals (one initially) assessing shoulder using a model (all and found no significant differences Kilbom and manufacturing tenderness on palpation, pain or 4 point severity variables from participants. Persson companies agreed to restriction with active and scale: significant at 1987 participate; 10 passive movements; symptoms the p<0.05 No relation between maximal static excluded because of in previous 12 months with None: 78% level) strength and symptoms. symptoms in past 12 regard to character, frequency, months. 96 underwent duration, localization, and relation Slight: 8% Headache Examiner blinded to case status. medical, physiological, to work or other physical and ergonomic activities. Analyzed if score on Moderate: 7% Average Questions included spare time evaluation. any symptom of 2 or >on a 4 time/work cycle physical activities, hobbies, point scale; “severe” symptom Severe: 3% with upper arm perceived psychosocial stress at score = 4. 0-30E abducted work, work satisfaction, number of breaks, rest pauses. Exposure: Carried out at outset Average of study: MVC of forearm time/work cycle Clinically diagnoses found were flexors, shoulder strength, in neck flexion largely myofascial symptoms. handgrip, heart rate using a bicycle ergometer and rating of Excessive Headache, sleep problems, dizziness perceived exertion. Included general fatigue showed a weak positive correlation. video analysis of postures and at end of movements of the head, shoulder working day Age, years of employment, and upper arm including productivity, muscle strength were durations and frequencies. not related to symptoms. Recorded work cycle time and number of cycles/hr, time at rest There was large inter-worker for the arm, shoulder and head, variation in working posture and rest periods, and average and working techniques. total duration/work cycle and hr. The mean number of neck The more dynamic working forward flexions >20E/hr was technique, the less symptoms in the 728 (s.d. 365) in the initial 96 neck and neck/shoulder symptoms. workers. Authors note: “a strong positive relationship to disorders was obtained with VIRA variables describing neck forward flexion and upper arm elevation.” See Jonsson et al. 1988 for follow- up.

(Continued)

2-79

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Linton and Cross- 420 of 438 medical Outcome: 3-point scale Shoulder pain Õ Those Participation rate: 96%. Kamwendo sectional secretaries and office collapsed from 6-point frequency frequency frequently 1989 personnel at a scale ranging from “almost having neck 75% sat >5 hr/day. Swedish hospital. never” to “almost always” having Very often: and shoulder neck or shoulder discomfort; and 16.9% pain vs. those 43% worked with office machines Those reporting Nordic Musculoskeletal Pain less frequently each day. frequently having neck Questionnaire. Sometime wk: having pain: and shoulder pain 3.8% Psychosocial scale scored: 10 to 20 (1 to 3) compared to Exposure: 10-question Poor Work as good environment. 20 to 40 as those less frequently standardized form on the Sometimes a wk: Content: poor environment. having pain (4 to 6) psychological work environment 4.8% OR= 2.5 1.3-4.9 points). with 1 to 4 categorical scales. Authors noted that: (1) Secretaries Overall score and indexes on Sometimes days: Lack of Social exposed to high work demands work content, psychologic work 13.8% Support: periodically, (2) they also felt helpless demand and social support at OR=1.6 0.9-2.8 to change the work environment, and work. Sometimes 1 that (3) internal conflict within day: 28.6% departments may have affected Duties included daily use of responses. typewriter, VDT, plus mail Never: 32.1% telephone and appointment duties.

(Continued)

2-80

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Maeda 1982 Cross- 119 accounting Outcome: Based on Partial Participation rate: Not reported. sectional machine operators questionnaire responses of pain correlation aged 17 to 29 years in and stiffness in the right and left coefficient Examiners blinded to case status: a post-check office. sides of the neck and shoulder between head Not reported. based on frequency of “almost neck tilt and every day, occasionally, and factor score Constrained tilted head posture was never or seldom” during the 1 to 5, associated with neck/shoulder previous several wk. Scores controlling for stiffness. were factor analyzed. other angles “A and C”, age, Exposure: Anthropometric and length of parameters relevant to the job p<0.05 service 0.25 tasks were measured on 51 operators who showed large or small factor scores.

(Continued)

2-81

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Milerad and Cross- 99 dentists randomly Outcome: Based on telephone All dentists: Participation rate: 99%. Ekenvall sectional selected from questionnaire: Neck symptoms Neck and 1990 Stockholm dentist at any time before the interview Shoulder: 36% 17% 2.1 1.3-3.0 Analysis stratified by gender. registry who practiced ("lifetime prevalence"). Further $ 10 years compared analyzed according to Nordic Neck and No difference in leisure time, smoking, to 100 pharmacists questionnaire as to duration Shoulder and systemic disease, exposure to selected from all during last 12 months and during Arm: 16% 3% 5.4 1.6-17.9 vibration. pharmacists in last 7 days, effect on work Stockholm. performance, leisure activities, Symptoms increased with age in and sick leave. female dentists only. Exposure: Questionnaire Duration of employment highly included: (1) abduction of arm correlated with age (r=0.84, 0.89). particularly in sit-down dentistry, (2) work hr/day, (3) static No relation between symptoms and postures. duration of employment. Equal problems dominant and non- dominant sides.

(Continued)

2-82

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohara et al. Cross- For cross-sectional Outcome: Assessed by Cash register Office NR Participation rate: for prospective 1976 sectional study: 399 cash standard health inventory and operators machine study = 100%. register operators medical examination (used operators and pro- compared with clinical classification according to and other Participation rate: for cross-sectional spective 99 office machine the committee on cervicobrachial Interventions did workers study, unable to calculate from data operators and disorders of the Japan not result in (clerks and presented. 410 other workers Association of Industrial Health, reduced muscle saleswomen (clerks and in Table 3 in the paper). fatigue of the ) Unknown whether examiners blinded saleswomen). All neck, shoulders, to case status. female. Periodic physical exam and upper back performed twice a year from brought on Interventions did not reduce For prospective study: 1973. Primary exams performed presumably by complaints in the shoulder region, but 56 workers employed on 371 operators. 130 (35%) the continuous did improve symptoms in the arms, <7 months had testing received detailed exams. lifting of the hands, fingers, low back, and legs. pre- and post- upper limbs. The lack of improvement in the intervention using Exposure: To repetitive shoulder region was stated to be due questionnaire and movements relocating to the use of the same narrow check physical exam. merchandise across counter and stands, unsuitable counter height, bagging, involved muscle activity and necessity of continuous lifting of 86 operators, newly of the fingers, hands, and arms; the upper limbs. hired after extreme and sustained postures. interventions, also had Operators hired after the evaluation after Interventions: (1) a 2-operator interventions and then examined after 10 months of working. system, 1 working the register, 10 months had less Grade I, one packing articles, changing II , or III occupational cervicobrachial roles every hr; (2) continuous disorders in examination than those operating time <60 min; max. hired before intervention. working hr/day 4.5 hr; (3) 15-min resting period every Only 14.5% with >3 years hr; (4) electronic cash registers employment at worksite. with light touch keyboard substituted for half of previously Narrow work space and counter used height not adjusted for height of worker. mechanical cash registers.

(Continued)

2-83

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson Cross- Industrial Workers Outcome: Pain in the last 7 days Industrial Referents: All Participation rate: Current workers: et al. 1995 sectional (n=82 females) and PE diagnosing tension neck workers: 50% 16% neck/shoulder 96% Past workers: 86%; exposed to repetitive syndrome, cervical syndrome. clinical Referents: 100%. tasks with short cycles diagnoses mostly far <30 sec., Tension neck: Tightness of (industrial No exposure information available to usually with a flexed muscles, tender spots in the workers examiners, “not possible to neck and arms muscles. Cervical syndrome: compared to completely blind the examiners.” elevated and abducted Limited neck movement, radiating referents): intermittently; 68 former pain provoked by test OR=2.7 1.2-6.3 Questionnaire included individual workers (mean movements, decreased factors, work/environment, employment time 21 sensibility in hands/fingers; Logistic Model: symptoms, psychosocial scales. years) who had left muscle weakness of upper limb. Repetitive work the factory during the 7 vs. none: Videotape analysis revealed years before the Muscle strength measured by OR=4.6 1.9-12 considerable variation in posture study; these workers MVC at elevation, abduction, and even within groups performing similar were compared to 64 outward rotation of both arms Age (57 vs. assembling tasks. referents with no measured by dynamometer. 37): OR=1.9 1.0-3.5 repetitive exposure at Logistic models replacing repetitive their current jobs Exposure: Videotaping and Muscular work with videotape variables found (female residents of a observation. Analysis of tension muscular tension tendency and neck nearby town currently postures, flexion of neck (critical tendency: flexion movements significantly employed as customer angles 15E and 30E). 74 (score 4.5 vs. associated with neck/shoulder service, ordering and workers videotaped $10 min. 1) : OR=2.3 1.3-4.9 diagnoses. price marking in from back and sides. Average supermarkets, as counts of two independent Stress/worry Inverse relationship between duration office workers (no readers for frequencies, tendency: of industrial work and MSDs, largest constant computer duration, and critical angles of OR=1.9 1.1-3.5 OR in those employed <10 years. work) or as kitchen movement used. workers. Assembly group had high OR (6.7) Repetitive industrial work tasks with regard to neck/shoulder MSD divided into 3 groups: (1) Fairly compared to referents. mobile work; (2) Assembling or pressing items; and (3) sorting, Significant association between time polishing and packing items. spent in neck flexion positions <60E. Weekly working time, work rotation, patterns of breaks, individual performance rate (piece rate). Only exposure readings from right arm were used.

(Continued)

2-84

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Punnett et al. Cross- 254 of 275 (92%) Outcome: Based on self- Overall Õ Male: 1.8 1.0-3.2 Participation rate: 92%. 1991 sectional meatcutters and reported symptom survey. Prevalence Female: 0.9 0.5-1.9 wrappers who Cases were defined if they met Neck/Shoulder: Stratified by gender and age. attended health and the following: $ 20 episodes in 53% safety training classes. the previous year or usual Neck/shoulder disorders associated duration of $ one wk; reported with external duration of static Workers fulfilling date of pain onset after postures (>5 sec.) or lifting $ 5 lbs. outcome case employment in the retail meat while abducting, flexing or extending definition (cases) were industry; no history of systemic the shoulder. compared to non- disease related to soft tissue cases; also compared pain; and, no history of acute Neck/shoulder pain did not vary by to the U.S. industrial injury. job category. population. Exposure: Based on interview 98% of respondents performed lifting and authors observations. tasks at work. “They judged lifting an average load/day was 41 (+ 23) lb Exposure: Repetitive and lifted 33 times and carried 9 feet. strenuous activities (it was not Heaviest load = 71 (+ 31 lb), lifted stated whether this was for 11 times and carried 9 feet/lift.” specific area or involved neck Listing an average load with a 40 to and all upper extremity areas) for 50% standard deviation can be 0.5 to 8 hr/day in refrigerated misleading. areas. Neck/shoulder cases lifted both the Cutters cut an average “typical” and “heaviest” loads with 121 (+ 278) large pieces of greater frequency than non-cases. meat/day filled 701 (+ 830 boats). Association was found for extended Wrappers filled duration of and lifting weight in 374 (+ 602 boats/day). Wrapped abduction/flexion and extension of 1,299 (+ 1,365 boats and the shoulder. weighed 1,399 boats).

(Continued)

2-85

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Rossignol Cross- 191 Computer and data Outcome: Self-administered ½ to 3 hr of VDT No VDT use Up to 3 hr of Participation rate: In 6 industry et al. 1987 sectional processing services, questionnaire case defined as: use/day (n=31): (n=28): VDT use groups 67 to 100%. public utilities of Neck pain, stiffness, or soreness 39% 25% compared to 0 Massachusetts State occurring almost always or hr of use: Participation rate: For individual Department, at 38 work missed work due to neck pain, OR=1.8 0.5-6.8 clerical workers; 94 to 99%. sites selected at stiffness or soreness. random from 4 to 6 hr of VDT 4 to 6 hr of VDT Assessed magnitude of confounding Massachusetts Exposure: Self-reports of use/day (n=28): use compared by age, cigarette smoking, industry, employers of >50 number of hr worked each day 57% to 0 hr of use: educational VDT training. workers. with a keyboard machine with a OR=4.0 VDT. Subjects selected after 1.1-14.8 Study presented to participants as a 28 of the 191 did not observation of worksite. >7 hr of VDT “general health” survey (as opposed use a computer. 7 or more hr of use compared to an occupationally related survey) VDT use/day to 0 hr of use: to avoid observation bias. (n=104): 61% OR=4.6 1.7-13.2

(Continued)

2-86

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ryan and Cross- 143 data process Outcome: Based on symptoms Shoulder: 44% Comparison More non- Participation rate: 99%. Bampton sectional operators; using a 0 to occurring three or more times/wk symptom only group had cases trained in 1988 10 point scale, the with no physical exam signs, or symptom adjustment of Interviewers blinded to questionnaire group with symptom $ weekly symptoms with Neck: 43% scores <2. chairs p<0.05 responses. scores of 8 or above physical exam signs of muscle symptoms only (n=41) were tenderness or hardening Cases with Height, weight, sex, age, marital designated "cases," present. Neck/shoulder higher scores status, parental status evaluated and and were compared to symptoms of visual not found to be confounders. group with symptom Cases were selected by having occurring $ 3 discomfort p<0.05 scores of 2 or less a combination of symptoms in the times weekly Handedness, time spent in current (n=28). lower arm and shoulder/neck with no signs or Cases felt there job, time spent altogether keying or area meeting a summary score weekly with was not typing work, training in adjustment of of eight or more. These cases signs: 44% enough time for keyboard and desk evaluated in two were compared to a comparison rest breaks groups and no significant differences group with a score of 2 or less. compared to found. non-cases p<0.05 Exposure: Ergonomic Psychosocial and work environment assessment measuring angles Cases had scales included pertaining to job and distances of each operator more boredom, satisfaction as well as the Work seated at his/her workstation more work Environment Scale [R. Moos 1974]. performed; Questionnaire stress, and responses to: Time spent in needed to push Authors diagnosed “myalgia” as current job, time spent altogether themselves >3 diffuse muscle pain and tenderness. keying or typing work, training in times/wk; lower the adjustment of their chair, peer cohesion, desk, or keyboard. autonomy, clarity in the authority structure. Higher staff support and work pressure. p<0.05

(Continued)

2-87

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Tola et al. Cross- 828 Machine Outcome: Postal questionnaire Daily symptoms: Daily Machine vs. Participation rate: 74% machine 1988 sectional operators; 658 on neck or shoulder symptoms symptoms: office: operators, 67% carpenters, 67% carpenters; frequency in last year, and machine OR=1.7 1.5-2.0 office workers. compared to 657 office influence on work methods, daily operators: 11% office Carpenter vs. Adjusted for years in occupation, workers; All male, duties and activities or leisure carpenters: 8% workers: 2% office: age. Interaction terms tested for, ages 25 to 49 years. time hobbies. Pain Drawing OR=1.4 1.1-1.6 none found. Diagram used to distinguish body Change work Change work areas. For logistic regression methods: methods: Machine vs. Education, general health, and leisure model 12 month prevalence of carpenter: time activities, car driving included in neck and shoulder symptoms on machine office OR=1.3 1.1-1.4 analysis. 8 days or more. operators: 19% workers: Use of twisted Study restricted to males aged 25 to carpenters: 10% or bent 49 years. Exposure: Exposure based on 21% postures during occupation: Machine operators work Education status (“$ some vocational known to be exposed to static Little: OR=1.0 1.0-1.5 school” compared to “no > some loading due to prolonged sitting Moderate: courses”) statistically significant for OR=1.2 1.4-1.9 machine operators’ and carpenters’ and low-frequency whole body Rather much: reporting of symptoms. vibration, fast work pace, and OR=1.6 1.5-2.2 upper trunk twisting. Carpenters Very much: exposed to dynamic physical OR=1.8 work with varying postures and Working in a loads, static loading of draft: neck/shoulder-arm, and male No: OR=1.0 1.0-1.3 office workers, of whom only Yes: OR=1.1 40% were performing routine office tasks. Job satisfac- tion Very good: OR=1.0 1.0-1.3 Rather good: OR=1.1 Moderate or poor: OR=1.2 1.1-1.4 Age (years) 25 to 29: OR=1.0 30 to 34: OR=1.2 1.0-1.5 35 to 39: OR=1.3 1.1-1.6 40 to 44: OR=1.5 1.3-1.8 45 to 49: OR=1.6 1.4-1.9

(Continued)

2-88

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Vihma et al. Cross- 40 Sewing machine Outcome: Neck or shoulder Sewing machine Seam- Participation rate: Not reported. 1982 sectional operators with short complaints defined by operators with stresses work cycles compared questionnaire: Recurrent pain or neck/shoulder with neck/ Random selection of participants. to 20 seamstresses. aching in present work (during or complaints: 98% shoulder after work). complaints: Cases and referent group matched 60% PRR = 1.6 1.1-2.3 for age and duration of employment. Exposure: Observation and interview; hr continuously sitting, Sewing machine operators found to standing time, survey of work have significantly greater static work postures, length of work cycle. compared to seamstresses. Sewing machine operator cycle time was 30 to 60 sec. in duration. Seamstresses had longer cycle.

(Continued)

2-89

Table 2–8 (Continued). Epidemiologic studies evaluating work-related neck/shoulder disorders MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Viikari- Cohort 154 subjects (72 Outcome: Based on 10% of female Õ Female: Participation rate: 90%. Juntura et al. female, 82 male) from Questionnaire data: Ache, pain, and 2% of male Severe Controlled for physical and creative 1991a Helsinki, Finland. stiffness, numbness in their reported severe neck/shoulder hobbies, no interactions seen. Subjects were part of neck/shoulder in last 12 months. radicular neck symptoms vs. a longitudinal study Visual analogue scale of pain no symptoms Alexithymia Because of low numbers, males population that started intensity, disability. Severe neck (low verbal were not included in analysis. in Finland in 1955; and disability: Pain for >7 days in last 21% of female productivity) from 1961 to 1963. 12 months and mean disability and 2% of male (continuous): Subjects comprised of mostly high During that time, 1084 index $ 15. reported any OR=1.02 0.97-1.1 subjects underwent type of severe socioeconomic status who reported cross-sectional Physical exam (P.E.): Two tests neck/shoulder Social confi- light physical workloads. examination. In 1985, a for cervical nerve root pain dence (mode- questionnaire was involvement, neck compression rate fears vs. Data collection in 1955 to 1963: sent to all subjects; test, shoulder abduction test. no fears): Intelligence, alexithymia, social 801 (74%) responded. Because of small number of OR=0.04 0.0-4.5 confidence, hobbies, motor Of the respondents, abnormal physical findings, the (much fear vs. development, verbal development, 180 lived in the Helsinki P.E. was eliminated from analysis no fears): level of education of parents, type of area. It was from this OR=1.4 0.05-42.2 income of family. group that Exposure: Questionnaire: Type of income 162 responded. Eight Amount of work with hands (monthly Data collection in 1985: were excluded due to overhead, work in forward bent salary): OR=0.5 Questionnaire on family relationships, illnesses. The position, work in twisted or bent 0.05-5.2 socioeconomic status, work history, proportions of the position. Sense of characteristics of present work, job highest income levels coherence in the sample (continuous): satisfaction, mental resources. exceeded the Finnish OR=0.95 population. 0.9-0.99 Data collection in 1986 to 1987: Twisted or bent Questionnaire: Physical torso characteristics of work, amount of (>3 hr/day vs. physical exercise, illnesses, trauma. <1 hr/day: OR= 0.9 Measurements taken in adolescence, >3 hr/day vs.<1 0.8-10.0 such as intelligence, alexithymia, hr/day social confidence, hobbies and socioeconomic status of the family Sitting in a showed no consistent association forward with neck/shoulder symptoms in posture 1-3 hr/day vs. adulthood. <1hr/day: OR=10.7 >3 .4-291 hr/day vs. <1 hr/day: OR=1.5 0.07,29.6

2-90 CHAPTER 3 Shoulder Musculoskeletal Disorders: Evidence for Work-Relatedness

SUMMARY There are over 20 epidemiologic studies that have examined workplace factors and their relationship to shoulder musculoskeletal disorders (MSDs). These studies generally compared workers in jobs with higher levels of exposure to workers with lower levels of exposure, following observation or measurement of job characteristics. Using epidemiologic criteria to examine these studies, and taking into account issues of confounding, bias, and strengths and limitations of the studies, we conclude the following:

There is evidence for a positive association between highly repetitive work and shoulder MSDs. The evidence has important limitations. Only three studies specifically address the health outcome of shoulder tendinitis and these studies involve combined exposure to repetition with awkward shoulder postures or static shoulder loads. The other six studies with significant positive associations dealt primarily with symptoms. There is insufficient evidence for a positive association between force and shoulder MSDs based on currently available epidemiologic studies. There is evidence for a relationship between repeated or sustained shoulder postures with greater than 60 degrees of flexion or abduction and shoulder MSDs. There is evidence for both shoulder tendinitis and nonspecific shoulder pain. The evidence for specific shoulder postures is strongest where there is combined exposure to several physical factors like holding a tool while working overhead. The association was positive and consistent in the six studies that used diagnosed cases of shoulder tendinitis, or a constellation of symptoms and physical findings consistent with tendinitis, as the health outcome. Only one [Schibye et al. 1995] of the thirteen studies failed to find a positive association with exposure and symptoms or a specific shoulder disorder. This is consistent with the evidence that is found in the biomechanical, physiological, and psychosocial literature.

There is insufficient evidence for a positive association between vibration and shoulder MSDs based on currently available epidemiologic studies.

INTRODUCTION Shoulder MSDs and their relationship to work work-relatedness, especially showing an risk factors have been reviewed by several increased risk for overhead and repetitive authors [Hagberg and Wegman 1987; work. Kuorinka and Forcier 1995; Sommerich et al. 1993; Winkel and Westgaard 1992]. Hagberg The focus of this review is to assess evidence and Wegman [1987] attributed a majority of for a relationship between shoulder tendinitis shoulder problems occurring in a variety of and workplace exposures to the following: occupations to workplace exposure. Kuorinka awkward postures, forceful exertions, repetitive and Forcier [1995] looked specifically at exertions, and segmental vibration. Also shoulder tendinitis and stated that the included are studies relevant to shoulder epidemiologic literature is “most convincing” disorders—as defined by a combination of regarding symptoms and physical examination findings or by symptoms alone, but not specifically defined as tendinitis—and those studies for which

3-1 the health outcome combined neck and considered the primary exposure factor, shoulder disorders, but where the exposure particularly independent of posture. Some was likely to have been specific to the shoulder. studies indirectly inferred shoulder repetition by Chapter 2 discusses studies involving neck- characterizing hand, wrist, and forearm shoulder disorders where assessment of movements. exposure was likely specific to the neck region. Studies Reporting on the Association Pertinent information about the 39 reviewed of Repetition and Shoulder MSDs studies is presented in several ways. Detailed Three of the reviewed studies reported results descriptions of the studies are provided in on the association between repetition and Table 3-5. The text of this section on shoulders shoulder tendinitis [English et al. 1995; Ohlsson is organized by exposure risk factor. The et al. 1994, 1995]. For all three studies, some discussion within each risk factor is organized or all of the results were for associations with a according to criteria presented on Pages 1-1 to combined exposure to repetition and awkward 1-10 of the Introduction. Conclusions are posture. Six additional studies reported results presented with respect to the specific MSD of on the association between repetition and non- concern, shoulder tendinitis. specific shoulder disorders [Sakakibara et al. 1995], non-specific shoulder symptoms REPETITION [Andersen and Gaardboe 1993a; Ohlsson et al. 1989], combined neck-shoulder disorders Definition of Repetition for Shoulder [Bjelle et al. 1981; Chiang et al. 1993] or MSDs combined neck-shoulder symptoms [Kilbom et Studies that addressed the physical factor of al. 1986; Kilbom and Persson 1987]. repetition and its relation to shoulder MSDs were included in this review. Studies usually Studies Meeting the Four Evaluation Criteria defined repetition, or repetitive work, for the Four studies met all four of the criteria [Chiang shoulder as work activities that involved et al. 1993; Kilbom et al. 1986; Ohlsson et al. cyclical flexion, extension, abduction, or 1994, 1995] (Table 3-1, Figure 3-1). Chiang rotation of the shoulder joint. Repetitiveness et al. [1993] studied workers in the fish was defined in four different ways in the processing industry in Taiwan. The health reviewed studies: (1) the observed frequency of outcome of “shoulder girdle pain” was defined movements past pre-defined angles of shoulder as self-assessed symptoms of pain in the neck, flexion or abduction, (2) the number of pieces shoulder or upper arms, and signs of muscle handled per time unit, (3) short cycle tender points or palpable hardenings upon time/repeated tasks within cycle, and (4) a physical examination. Pain referred from a descriptive characterization of repetitive work nerve root or other spinal source was included or repetitive arm movements. Some of the in the case definition. The force requirements of studies that examined repetition as a risk factor the jobs were estimated by surface for shoulder MSDs had several concurrent or interacting physical work load factors. Therefore, repetitive work should not be

3-2 electromyographs (EMGs) in the forearm flexor In the first of the Ohlsson et al. studies, a cross- muscles. This is not a direct measure of sectional study, women in the fish industry were shoulder muscle activity. There may be no compared to a control population of women relationship between the level of activity in the employed in municipal workplaces in the same forearm and shoulder girdle muscles. Three towns [Ohlsson et al. 1994]. Diagnoses of categories, based on both force and shoulder disorders (e.g., tendinitis, repetitiveness, were used as the exposure acromioclavicular syndrome, frozen shoulder) outcome: Group I (low force, low were made on the basis of symptoms repetitiveness), Group II (high force or high determined by interview and a physical exam. repetitiveness), and Group III (high force and Exposure evaluation of each work task held by high repetitiveness). Force was also evaluated the fish industry population was evaluated with independently in multivariate analyses. ergonomic workplace analysis (EWA). Ten different factors were rated on a scale from 1 to Kilbom et al. [1986] performed a prospective 5 and the combined ratings were used as a study in which female employees in the profile of the work task. Based on this profile, electronics manufacturing industry were the authors reported that fish industry work was observed for a 2-year period. The health found to be “highly repetitive” and to include outcome in the neck, shoulder, or arm regions “poor work postures.” was based on symptoms and physical findings. Symptom severity was coded on the basis of its Ohlsson et al. [1995] compared a group of character, frequency, and/or duration. Changes women who performed industrial assembly in severity status at follow-up evaluations were work to a referent group of women from a used as the dependent variables in multiple nearby town who were employed in jobs regression analyses. Neck, shoulder, and upper characterized as having varied and mobile work arm posture was determined by VIRA. tasks. One examiner assessed signs and Although the health outcome combined symptoms. The examiner was blinded to symptoms from different body regions, specific exposure information, but not knowledge of biomechanical theory can be completely blinded to factory worker versus used to identify significant predictors related to referent group status. Shoulder tendinitis the shoulder symptom severity. included supraspinatus, infraspinatus, and bicipital tendinitis. Another health outcome For the two Ohlsson et al. [1994, 1995] combined neck and shoulder disorders (tension studies, the authors reported that the examiners neck, cervical syndrome, thoracic outlet could not be completely blinded to exposed syndrome, frozen shoulder, tendinitis, versus referent status, but that a standard acromioclavicular syndrome). In a descriptive protocol was followed and observer bias was assessment, it was reported that the work tasks likely to have been minimal. As examiners were in the study group involved repetitive arm blinded to objective exposure measures, movements with static muscular work of the analyses testing associations between neck- shoulder disorders and specific postures would not have been biased [Ohlsson et al. 1995].

3-3 neck and shoulder muscles. The percentage of outcome was defined by self-assessed time spent in specific upper arm postures was symptoms of shoulder stiffness and pain and a determined from videotaped observation of 74 physical examination for muscle tenderness and (out of 82) workers. The average result from joint pain on movement. Whether the examining two independent videotape analyses was used. physician was aware of the prior hypothesis Posture category demarcations included 0, 30, regarding differing exposures between the two and 60 degrees for arm elevation, and 30, 60, tasks (bagging pears versus bagging apples) and 90 degrees for arm abduction. was not stated. Exposure was based on self- report of the number of hours per day spent Studies Not Meeting the Four Evaluation bagging, the number of pears or apples bagged Criteria per day, and the total number of days spent Bjelle et al. [1981] compared cases with acute, bagging each fruit. One worker was observed non-traumatic shoulder-neck pain to age- and for 3 hours while performing each bagging job, sex-matched, paired controls. To determine with repeated goniometric measures of exposure, each case and control was filmed shoulder forward flexion angles done each and a biomechanical analysis was performed to minute. While there was no difference in the determine the frequency and duration of total number of days or number of hours per shoulder abduction or forward flexion > than day spent bagging each fruit, significantly more 60 degrees. pears than apples were bagged per day. The proportion of time spent with the angle of In the study by English et al. [1995], cases shoulder forward flexion greater than 90 were determined by medical diagnosis and degrees was significantly larger when bagging controls were selected from patients evaluated pears (75%) than when bagging apples (41%). at specified orthopedic clinics. For statistical analyses, all diagnoses were grouped by One study did not meet any of the criteria. In a anatomical site. The diagnoses for shoulder cross-sectional study by Ohlsson et al. [1989], cases were rotator cuff injury, rupture of long the exposed population was factory employees head of biceps, shoulder capsulitis, and who produced and assembled plastic symptomatic acromioclavicular arthritis. It is components. Work exposure was assumed that shoulder tendinitis is included in characterized as “repetitive arm and hand this group. Exposure measures were movements in constrained work postures.” The determined by a standardized interview referent population was composed of women conducted by an interviewer who was randomly sampled from the general population “unaware of the case-control status of the in a nearby area. The health outcome was individual wherever this was possible.” determined by self-reported symptoms of shoulder pain during the previous seven days. In a study by Sakakibara et al. [1995], the The exposure measure was the self-reported health status of a group of women farm number of items completed per hour. The range workers was assessed during the performance was from less than 100 items completed per of two different tasks, with a hour (slow category) to more than 700 items 1-month interval between the tasks. The health

3-4 per hour (very fast category). Self-reporting shoulder MSDs, the major limitation of this was believed to be accurate because workers study was that the exposure assessment was were paid by the piece. not specific to movement at the shoulder joint and may therefore have either over- or Strength of Association: underestimated repetition at the shoulder. In Repetition and Shoulder MSDs some cases the exposure assessment may have Using the data presented in the study by been a measure of repetitive upper arm Ohlsson et al. [1994], for supraspinatus, movements, but it may also have been a infraspinatus, or bicipital tendinitis the odds measure of repetitive hand and distal upper ratio (OR) for working in the fish industry extremity activity occurring in the context of a (repetitive work, poor posture) was calculated static load on the shoulder muscles. as 3.03 (95% CI 2.5–7.2). For shoulder tendinitis alone, the PRR was calculated as 3.5 For the shoulder diagnoses used to form their (95% CI 2.0–5.9). For clinical diagnoses of the group of cases, English et al. [1995] found an neck and shoulder, the OR for working in the association with repeated shoulder rotation with fish industry versus the referent population was an elevated arm (OR 2.30, p< 0.05). They also 3.2 (95% CI 2.0–5.3). found what appeared to be a protective effect associated with elbow flexion (OR 0.4, 95% Using data presented in the study by Ohlsson et CI 0.2–0.8). This effect was greatest at low al. [1995] for supraspinatus, infraspinatus, or amounts of daily cumulative exposure to elbow bicipital tendinitis, the OR for being an flexion; the protective effect decreased (RR assembly worker (repetitive arm movements increased) as the number of hours of total daily with static load on shoulders) versus the elbow flexion increased. In a laboratory study referent population was 4.2 (95% CI of shoulder muscle activity in relation to 1.35–13.2). For neck-shoulder disorders, the different combinations of shoulder and elbow OR for being an assembly worker versus the joint postures (a total of 21 different postures), referent group was 5.0 (95% CI 2.2–11.0). Herberts et al. [1984] found that humeral rotation and elbow flexion had insignificant Using multiple logistic regression analysis with effects on shoulder muscle activity. However, age, gender, and force as covariates, Chiang et the postures tested by that study were al. [1993] found that highly repetitive upper stationary, whereas the associations reported extremity movements were associated with by English et al. [1995] appear to be related to shoulder girdle pain (OR 1.6, 95% CI repetitive movements. 1.1–2.5). When tested in the same model with force and repetition, the interaction term for For symptoms of shoulder pain within the force and repetition was also significant (OR previous 7 days, the OR for assembly workers 1.4, 95% CI 1.0–2.0). Several factors could versus the referent group was 3.4 (95% CI have resulted in an underestimation of the 1.6–7.1) [Ohlsson et al. 1989]. A significantly strength of association: no requirement that higher proportion of the farm workers studied symptoms had begun on current job means that by Sakakibara et al. [1995] some symptomatic workers may have transferred to lower risk jobs. Relative to

3-5 had signs of shoulder muscle tenderness while Repetition Characterized as Short Cycle Time bagging pears than while bagging apples. There Chiang et al. [1993] found a significant was no way to analyze the relative contribution association between a very short or repetitive to risk of repetitive shoulder exertions cycle (<30 seconds or >50% spent repeating (increased number of pears picked per day) same task) and shoulder girdle pain. and awkward posture (greater portion of each day spent with extreme forward flexion when Repetition Characterized Descriptively picking pears). Three studies by Ohlsson et al. found a significantly higher proportion of shoulder Consistency of Association MSDs in exposed populations with work Repetitiveness was defined in four different characterized as involving repetitive arm and ways in the reviewed studies: (1) the observed hand movements than in referent populations frequency of movements past pre-defined [Ohlsson et al. 1989, 1994, 1995]. angles of shoulder flexion or abduction, (2) the number of pieces handled per time unit, (3) Repetition Combined with Static Shoulder short cycle time/repeated tasks within cycle, Load and (4) a descriptive characterization of Except for the study by Sakakibara et al. repetitive work or repetitive arm movements. [1995], in which the increased number of pears bagged per day was associated with an Repetition Characterized as Frequency of increased proportion of the work day spent Movements Past Pre-Defined Shoulder with extreme shoulder flexion, the studies using Angles measures of piece work or repetitive arm Bjelle et al. [1981] and Ohlsson et al. [1995] movements as the exposure outcome did not found a significant positive association between specify which joints or body regions the prevalence of neck-shoulder disorders and participated in the repetitive action. Ohlsson et the frequency of upper arm movements past 60 al. [1995] described the assembly work degrees of flexion or abduction. English et al. performed by the exposed population as [1995] found a significant association between combining repetitive arm movements with a diagnosed cases of shoulder disorders and static shoulder load. It is possible that the repeated shoulder rotation with an elevated arm association between piece work, short cycles, posture. or repetitive hand-arm movements and shoulder disorders reported by the other Repetition Characterized as the Number of authors is related to a sustained, static load on Pieces Handled per Time Unit the shoulder muscles as the upper arm is A significant positive association was found stabilized in a posture of mild to severe flexion between both nonspecific shoulder symptoms or abduction, while repetitive movements are [Ohlsson et al. 1989] and nonspecific shoulder performed by the hand-wrist-forearm. disorders [Sakakibara et al. 1995] and the number of pieces handled per hour or per day.

3-6 Temporal Relationship associated (p<0.05) with both the number of In the prospective study by Kilbom et al. 1986; arm elevation movements from less than to Kilbom and Persson 1987; and Jonsson et al. greater than 60 degrees and the number of arm 1988 the number of shoulder elevations per abduction movements from less than to greater hour was a strong predictor for a change to than 60 degrees. Bjelle et al. [1981] found that severe status at the 1- and 2-year follow-up the frequency of shoulder abduction or forward evaluations. Although the change in status flexion (past 60 degrees) was significantly included problems in the neck and arm, as well greater (p<0.005) for cases with neck-shoulder as the shoulder, it is reasonable to assume that disorders than for controls. repetitive shoulder elevations would have had the greatest effect on disorders of the shoulder. In the study of assembly workers by Ohlsson et al. [1989], the number of pieces completed per Several studies with a cross-sectional design hour was categorized as follows: slow: <100, used techniques to determine whether the medium: 100–299, fast: 300–699, very fast: health outcome of interest had occurred since, >700. In this study, the ORs are shown in a or was present during, exposure to figure, rather than reported in the text. hypothesized risk factor(s) of interest. Case Compared with the slow-paced group, the definitions which required a positive physical odds for symptoms of shoulder pain is examination finding [Chiang et al. 1993; approximately seven times that for those Ohlsson et al. 1994, 1995] or where symptoms workers in the medium-paced group and had occurred within the recent past [Chiang et approximately nine times that for those in the al. 1993; Ohlsson et al. 1989, 1994] were fast-June 26, 1997 pace group. While adjusting designed to focus on disorders most likely to for age and length of employment, the OR for have been caused or aggravated by current shoulder pain was significantly higher for the work exposures. medium- and fast-paced groups than for the slow-paced group (p=0.0006). The OR for the Exposure-Response Relationship very fast-paced group compared to the slow- Chiang et al. [1993] found a significant paced group was between 1.0 and 2.0 and was increasing trend in the prevalence of shoulder not significantly different from the slow-paced girdle pain from Group I (low force, low group. The authors hypothesized that repetitiveness) to Group III (high force, high symptomatic workers may have self-selected repetitiveness). However, the health outcome out of the very fast paced jobs or that other was not specific to shoulder disorders, and the unknown factors may have mitigated the effects exposure categories combine increasing of work pace. repetitiveness—as defined by either less than a 30-second cycle time or a repeated task within When comparing fish industry workers to the the job cycle—and increasing forearm flexor reference population, Ohlsson et al. [1994] muscle activity. Ohlsson et al. [1995] found that found that among those workers younger than neck and shoulder disorders among assembly age 45, the ORs for disorders of the neck and workers were significantly shoulders were significantly elevated and

3-7 increased with duration of employment [0–5 Some of the significant associations reported years, OR 3.2 (95% CI 1.5–7.0); >5 years, may have been related to exposure to repetitive OR 10 (95% CI 4.5–24)]. In their study of work in the distal upper extremity while the assembly workers, Ohlsson et al. [1989] found shoulder and upper arm were maintained in a a statistically significant increase in the odds for static posture [Chiang et al. 1993; Ohlsson et pain in the shoulder with duration of al. 1989, 1994, 1995]. Winkel and Westgaard employment (p=0.03) which was dependent on [1992] have pointed out that, “It is not possible age. The increase with duration of employment to use the arm/hand without stabilizing the had a steeper slope for younger (<35 years) rotator cuff girdle and the glenohumeral joint. assembly workers than for the older subgroup Therefore, work tasks with a demand of (i.e., among those workers employed for short continuous arm movements generate load durations, older women had more symptoms, patterns with a static load component.” and among those workers employed for long durations, younger women had more The finding that the supra- and infraspinatus symptoms). This was thought to be a reflection muscles were particularly prone to fatigue when of both survivor bias as well as the possibility subjects performed overhead work led that older new hires may have experienced a Herberts et al. [1984] to hypothesize that the relatively more rapid onset of symptomatic rotator cuff muscles may develop high problems than do younger women. intramuscular pressures at relatively low contraction levels. These high intramuscular Coherence of Evidence pressures could lead to an impairment of Repetitive movements of the upper extremity intramuscular circulation, which could involving flexion or abduction of the contribute to the early onset of fatigue. glenohumeral joint would increase the Intramuscular pressure increases with the frequency of effects such as fatigue and tendon muscle contraction level, and impaired circulation disruption hypothesized to occur as circulation has been demonstrated at levels of a result of such postures. These effects could contraction as low as 10–20 percent of be magnified by the addition of a hand-held maximal voluntary contraction (MVC). load. Repetition may also be solely related to [Hagberg 1984]. the development of tendinitis. In a laboratory study, Hagberg [1981] induced acute shoulder The increased pressure in rotator cuff muscles tendinitis in female subjects performing and increased pressure on the supraspinatus repetitive shoulder elevations for one hour. Six tendon may trigger two different events that are female students, ages 18–29, all developed both related to impaired microcirculation. The shoulder tenderness (two with tendinitis) when impaired microcirculation in the tendon may exposed to 15 shoulder flexions (from 0 to 90 also result from tension within the tendon degrees) per minute for 60 minutes while produced by forceful muscle contractions holding up to 3.1 kg (6.4 lb) of weight. [Rathburn and Macnab 1970]. An inflammatory infiltrate with increased

3-8 vascularity and edema within the rotator cuff method, and shoulder MSDs or symptoms tendons, especially the supraspinatus tendon were found in all studies. Of the eight studies in may be a result of or a contributor to the which the effect of repetition was examined, process. If the inflammation process is three studies found ORs above 3.0 [Ohlsson et sufficiently intense, then shoulder tendinitis may al. 1989, 1994, 1995] and three studies found occur. If the process is less intense, and more ORs from 1.0 to 3.0 [Chiang et al. 1993; chronic, then it may contribute to a English et al. 1995; Sakakibara et al. 1995]. degenerative process in the tendons of the The remaining studies were prospective studies rotator cuff. In the muscles of the rotator cuff, [Jonsson et al. 1988; Kilbom and Persson the impaired microcirculation may lead to small 1987] or studies that reported risk indicators areas of cell death. A reasonable hypothesis is other than OR [Bjelle et al. 1981]. that repeated or sustained episodes of muscle ischemia result in localized cell death and In none of these studies is it likely that age, the persistent inflammation. most important personal characteristic associated with shoulder tendinitis and other Neither of these proposed models for shoulder shoulder disorders, or nonoccupational factors muscle pain or tendinitis suggest that all muscle such as sports activities, caring for young activity is potentially harmful. Both muscles and children, or hobbies explained these tendons are strengthened by repeated activity if associations. There is evidence of a relationship there is sufficient recovery time. However, the between shoulder tendinitis and highly repetitive models present plausible mechanisms by which work. work tasks with substantial shoulder abduction could contribute both to shoulder pain and FORCE tendinitis. Definition of Force for Shoulder There is evidence of a relationship between MSDs shoulder tendinitis and highly repetitive work. Studies that examined force or forceful work or However, there are several limitations to the heavy loads to the shoulder, or described evidence. In the three studies for which the exposure as strenuous work involving the health outcome was shoulder tendinitis, the shoulder abduction, flexion, extension, or exposure combined repetition with awkward rotation that could generate loads to the shoulder posture and/or a static shoulder load shoulder region were also included. Most of the [English et al. 1995; Ohlsson et al. 1994, studies that examined force or forceful work as 1995]. Five out of the eight studies reviewed a risk factor for shoulder symptoms or tendinitis used either nonspecific shoulder disorders, had several concurrent or interacting physical nonspecific shoulder symptoms or combined work load factors. However, there is still a neck-shoulder disorders as the health outcome. need to summarize present knowledge about the relationships between forceful work and Despite the limitations of the evidence, shoulder MSDs. This section summarizes that significant and positive relationships between knowledge, while acknowledging that other repetitiveness, regardless of the measurement factors can modify the response.

3-9 Neck-shoulder disorders are discussed in experience and knowledge of the jobs were Chapter 2. used to assign job titles to exposure categories based on crude assessments of force and Studies Reporting on the Association repetitiveness. High exposure was of Force and Shoulder Tendinitis characterized as a combination of high There are five studies which reported results on repetitiveness (activity repeated several times the association between force and adverse per minute) and low or high force, or medium shoulder health outcomes (Table 3–2, Figure repetitiveness (activity repeated many times per 3–2). The epidemiologic studies that addressed hour) and high force. Medium exposure was forceful work and shoulder MSDs tended to characterized as medium repetitiveness and low compare working groups by classifying them force, or low repetitiveness (jobs with more into broad categories based on an estimated variation) and high force. Those in teaching, amount of resistance or force of exertion and a academic, self-employed, or nursing combination of estimated rate of repetition professions were classified as low exposure. [Andersen and Gaardboe 1993a; Chiang et al. The exposure classification scheme in this study 1993] or in terms of overall load [Herberts et does not allow separation of the effects of force al. 1984; Stenlund et al. 1992; Wells et al. from those of repetition. More sewing machine 1983]. operators than referents were considered to have high exposure (41% versus 15%), but Studies Meeting the Four Evaluation Criteria more in the referent population were Chiang et al. [1993] studied workers in the fish considered to be in the medium exposure group processing industry. (This study was described (44% versus 22%). Because the outcome of in detail in the section on shoulder MSDs and interest was duration of historical exposure, repetition.) Chiang et al. [1993] did not report current exposure was included as an an exposure specific to the shoulder. independent variable in multivariate regression analyses. Studies Not Meeting the Four Evaluation Criteria Herberts et al. [1984] added to the 1981 study Andersen and Gaardboe [1993a] performed a by comparing the prevalence of supraspinatus cross-sectional study in which a cohort of tendinitis between plate-workers and office sewing machine operators was compared to a clerks. Tendinitis in welders was determined by random sample of women in the general a combination of self-reported symptoms and population of the same region. Chronic positive physical examination findings. The only shoulder pain was defined as a having information given regarding plate-work is that it experienced a continuous pain episode lasting is dynamic in character. It is presumed that more than 1 month and either daily pain or pain plate-workers handled heavy loads more lasting more than 30 days in the same location frequently than office clerks. within the previous year (per self-administered questionnaire). In order to compare the current In a cross-sectional study, the prevalence of exposure of sewing machine operators and osteoarthrosis in the acromioclavicular joint, those in the control group, the authors’

3-10 as determined by radiography, was compared current work as a sewing machine operator among three groups of workers in the was associated with chronic shoulder pain (OR construction industry [Stenlund et al. 1992]. 1.72, 95% CI 1.17–2.55). Using multiple The three groups were bricklayers, rock logistic regression analysis with age, gender, blasters, and construction foremen. The and repetitiveness as covariates, Chiang et al. foremen did not perform manual work [1993] found that high force exertions currently, or in the past, and were considered measured in the forearm were associated with the control population. A standardized shoulder girdle pain (OR 1.8, 95% CI interview was used to determine exposure 1.2–2.5). When tested in the same model with factors, including job title and the sum of loads force and repetition, the interaction term for lifted during all working years (expressed in force times repetition was also significant (OR tonnes). Analyses were performed separately 1.4, 95% CI 1.0–2.0). Two factors could have for right and left sides. resulted in an underestimation of the strength of association: (1) no requirement that symptoms In a study of letter carriers, Wells et al. [1983] have started on current job meant that some evaluated the effect of a load carried on the symptomatic workers may have transferred to shoulder. Letter carriers, who carry a load and lower risk jobs, and (2) no matching of health status and exposure status by side (left, right, or walk, were compared to gas meter readers both) may have caused non-differential (who walk without carrying a load) and postal misclassification. For supraspinatus tendinitis, clerks. Utilizing information from telephone Herberts et al. [1984] calculated a prevalence interviews, points were assigned to symptom rate ratio (PRR) for plate-workers versus office characteristics such as frequency, length of clerks of 16.2 (90% CI 10.9–21.5) “under the episodes, and interference with work ability. assumption that missing data had the same Case definition required a report of recurrent characteristics as those considered.” The shoulder pain with greater than 20 points. A absence of specific exposure information was a subset of letter carriers had experienced an major limitation of this study. increased load during the previous year. (The Postal Service had increased maximum weight The age-adjusted OR associated with carried from 25 to 35 pounds, but not all osteoarthritis of the acromioclavicular locations had implemented this change.) joint was 2.16 (95% CI 1.14–4.09) (right side) and 2.56 (95% CI 1.33–4.93) Strength of Association—Force (left side) for manual construction workers and Shoulder MSDs versus foremen [Stenlund et al. 1992]. Because The studies are presented in alphabetical order there was a lower participation rate among bricklayers and blasters, self-selection into the in Table 3-2. Results of studies where ORs, or study because of having symptoms could have other measures of association, were specifically resulted in overestimation of the strength of associated with a measure of exposure, are association. While some of the items handled presented in the section on Exposure-Response required a bilateral lift (e.g., jackhammer), Relationship. other loads may have been specific to the right or left hand. Because the Andersen and Gaardboe [1993a] found that

3-11 exposure measure did not separate load by disorders between letter carriers with and sides, non-differential misclassification may without the increased load was related to have caused underestimation of the strength of current exposure rather than past peak association. exposures or cumulative duration. Although baseline symptom status in the group with the Consistency of Association: increased load could not be obtained, there Force and Shoulder MSDs was no significant difference in the prevalence Despite different outcome and exposure of shoulder problems between the two groups measures, all of the studies had positive when results were adjusted for the amount of associations. Each study used a different case weight currently carried. Therefore, the definition, ranging from relatively mild difference in symptom prevalence was likely symptoms to radiographic evidence of related to the load increase rather than prior osteoarthritis, and a different measure of differences in symptom status. The cross- exposure. Chiang et al. [1993] used EMG sectional studies are consistent with exposure measures of forearm flexor muscle activity. occurring before the onset of the shoulder Wells et al. [1983] evaluated the effect of a MSDs. direct load on the shoulder. Stenlund et al. [1992] used an estimate of the cumulative, Exposure-Response Relationship lifetime load carried. Andersen and Gaardboe When sewing machine operators were [1993a] compared sewing machine operators compared with an external control population, to a referent population. However, positive and there was a trend of increasing ORs for chronic significant associations were found, regardless shoulder pain with increasing duration of work of the measure of health outcome or exposure. as a sewing machine operator [Andersen and Gaardboe 1993a]. The OR for 0–7 years was Temporal Relationship: Force and 1.38 (95% CI 0.86–2.39), for 8–15 years it Shoulder MSDs was 3.86 (95% CI 2.29–6.50), and for >15 All of the studies of forceful exertions used a years it was 10.25 (95% CI 5.85–17.94), cross-sectional study design. To increase the while controlling for other factors including age likelihood that shoulder symptoms were caused and current exposure. or aggravated by current exposure, Chiang et al. [1993] required that symptoms had Chiang et al. [1993] found a significant occurred within the previous 30 days. increasing trend in the prevalence of shoulder girdle pain from Group I (low force, low Wells et al. [1983] used several analytical repetitiveness) to Group III (high force, high methods to increase confidence in a relationship repetitiveness). However, the health outcome is between carrying the increased load and having not specific for shoulder tendinitis and the shoulder disorders. The use of age, the number exposure categories combine increasing force, of years on the job, and previous heavy work as measured in the forearm flexor muscles, and experience as covariates when performing increasing repetitiveness. analysis of covariance helped ensure that the difference in the proportion of shoulder

3-12 In the study of bricklayers and blasters, and that muscular fatigue will occur at EMG levels acromioclavicular osteoarthritis, Stenlund et al. as low as 5% of maximal voluntary contraction [1992] found that for the left side, ORs (MVC) if sustained for 1 hour. Other studies increased with the level of lifetime load lifted. have demonstrated that when the period of For a lifetime load of 710–24,999 tonnes muscle contraction is extended to more than an versus less than 710 tonnes, the left side OR hour, the endurance limit of force may be as was 7.29 (95% CI 2.49–21.34), and for low as 8% MVC [Jonsson 1988]. Workers greater than 25,000 tonnes versus less than 710 performing repetitive work with the hands and tonnes, the left side OR was 10.34 (95% CI wrists, while maintaining static upper arm 3.10–34.46). elevation may experience fatigue even at low load levels. Jonsson [1988] reported that many For severe, but not disabling, shoulder pain, the constrained work situations are characterized OR for letter carriers versus postal clerks was by static load levels near or exceeding 5% 3.6 (95% CI 1.8–7.8) [Wells et al. 1983]. For MVC, even when characterized by a fairly low those letter carriers who had experienced a mean muscular load. weightload increase within the previous year, versus postal clerks, the OR was 5.7 (95% CI Because the five studies reviewed had a 2.1–17.8). Furthermore, letter carriers who had considerable diversity of exposure assessment experienced the weightload increase had approaches and health outcomes, there is significantly more shoulder problems than those insufficient epidemiologic evidence to conclude whose bag weight had not been increased. If that forceful exertions are associated with letter carriers tend to keep the mail-bag strap rotator cuff or bicipital tendinitis. The one study on one shoulder, the fact that the side of the that used shoulder tendinitis as the health load was not matched with the side of the outcome reported a strong association related shoulder problem could have resulted in non- to job category (OR for plate-workers versus differential misclassification and an clerks: 16.2 (95% CI 10.9–21.5), but did not underestimation of the strength of association. describe or measure specific exposure risk However, some of the health effects may have factors [Herberts et al. 1984]. One of the been related to activation of contralateral reviewed studies did present evidence for an muscles involved in stabilizing the shoulder association between acromioclavicular girdle [Winkel and Westgaard 1992]. osteoarthrosis and cumulative, lifetime load on the shoulder muscles [Stenlund et al. 1992]. Coherence of Evidence Another study reported a significant association High shoulder muscle force requirements can between severe shoulder pain and a direct cause increased muscle contraction activity, shoulder load [Wells et al. 1983]. which may lead to an increase in both muscle fatigue and tendon tension, and may possibly impair microcirculation as well.

Force may also be related to a static load on shoulder muscles. Sjøgaard et al. [1988] found

3-13 POSTURE Ohlsson et al. 1995] or combined neck- shoulder symptoms [Kilbom and Persson Definition of Awkward Posture for 1987]. Shoulder MSDs For the shoulder, a relaxed, neutral posture is Studies Meeting the Four Evaluation Criteria one in which the arm hangs straight down by Four studies met all four of the evaluation the side of the torso. As the arm is flexed, criteria. abducted, or extended, the included angle between the torso and the upper arm increases. Using a prospective study design, Jonsson et al. In one study, postures in which the included [1988] assessed the health and exposure status angle was equal to or greater than 45 degrees of 69 electronics manufacturing plant required substantial supraspinatus muscle employees at the beginning of the study and activity, while deltoid muscle activity underwent after one and two years. Employees who a pronounced increase as the angle of shoulder dropped out before completion of the study flexion or abduction increased from 45 to 90 were compared to those who fully participated; degrees [Herberts et al. 1984]. As the arm is there was no significant difference in medical elevated, the space between the humeral head status, working technique, or work history. and the acromion narrows such that mechanical Employees who had upper extremity disorders pressure on the supraspinatus tendon is greatest resulting in a physician visit or sick leave were between 60 and 120 degrees of arm elevation excluded from the initial study group. The [Levitz and Iannotti 1995]. While there is a dependent variables related to health status continuum of severity from an included angle of were of two types: a change in symptom 30 degrees to a maximally abducted arm, severity and being symptom free. Symptom postures with shoulder abduction or flexion past status was assessed by interview and a physical 60 degrees are considered awkward. examination by a physiotherapist. The symptoms severity index compiled data from Studies Reporting on the Association the five body regions combined and was not of Awkward Postures and Shoulder specific for the shoulder region. Because the MSDs exposure was determined by direct observation Six of the reviewed studies reported results on for each individual, and clearly separated the association between awkward postures and ergonomic risk factors by body region, it was shoulder tendinitis [Baron et al. 1991; Bjelle et still possible to evaluate associations likely to al. 1979; English et al. 1995; Herberts et al. specifically involve the shoulder. 1981; Ohlsson et al. 1994, 1995] (Table 3-3, Figure 3-3). Seven additional studies reported Kilbom and Persson [1987] and Kilbom results on the association between awkward et al. [1986] performed a study in postures and non-specific shoulder disorders which female employees in the [Sakakibara et al. 1995], non-specific shoulder electronics manufacturing industry were symptoms [Hoekstra et al. 1994; Milerad and observed for a 2-year period. The health Ekenvall 1990; Schibye et al. 1995] combined outcome of fatigue, ache, or pain neck-shoulder disorders [Bjelle et al. 1981; Jonsson et al. 1988;

3-14 in the neck, shoulder, or arm regions was based different factors were rated on a scale from 1 to on symptoms information. Symptom severity 5 and the combined ratings were used as a was coded on the basis of its character, profile of the work task. Based on this profile, frequency, and/or duration. Changes in severity the authors reported that fish industry work was status at follow-up evaluations were used as the found to be “highly repetitive” and include dependent variables in multiple regression “poor work postures.” analyses. Neck, shoulder, and upper arm posture was determined by computerized Ohlsson et al. [1995] compared a group of analysis (VIRA) of videotapes of individuals. women who performed industrial assembly Although the health outcome combined work to a referent group of women from a symptoms from different body regions, nearby town who were employed in jobs knowledge of biomechanical theory can be characterized as having varied and mobile work used to identify significant predictors related to tasks. One examiner assessed signs and the shoulder symptom severity. symptoms. The examiner was blinded to specific exposure information, but not Two of the reviewed studies in which tendinitis completely blinded to factory worker versus was the health outcome are Ohlsson et al. referent group status. Shoulder tendinitis [1994, 1995]. For both studies, the authors included supraspinatus, infraspinatus, and reported that the examiners could not be bicipital tendinitis. Another health outcome completely blinded to exposed versus referent combined neck and shoulder disorders (tension status, but that a standard protocol was neck, cervical syndrome, thoracic outlet followed and observer bias was likely to have syndrome, frozen shoulder, tendinitis, and been minimal. Because examiners were blinded acromioclavicular syndrome). In a descriptive to objective exposure measures, analyses assessment, it was reported that the work tasks testing associations between neck-shoulder in the study group involved repetitive arm disorders and specific postures would not have movements with static muscular work of the been biased [Ohlsson et al. 1995]. neck and shoulder muscles. The percentage of time spent in specific upper arm postures was In a cross-sectional study, women in the fish determined from videotaped observations of 74 industry were compared to a control population (out of 82) workers. The average result from of women employed in municipal workplaces in two independent videotape analyses was used. the same towns [Ohlsson et al. 1994]. Posture category demarcations included 0, 30, Diagnoses of shoulder disorders (e.g., and 60 degrees for arm elevation, and 30, 60, tendinitis, acromioclavicular syndrome, frozen and 90 degrees for arm abduction. shoulder) were made on the basis of symptoms determined by interview and a physical exam. Studies Not Meeting the Four Evaluation Exposure evaluation of each work task held by Criteria the fish industry population was evaluated with Summaries of studies that specifically evaluated ergonomic workplace analysis (EWA). Ten associations with shoulder tendinitis are presented next [Baron et al. 1991; Bjelle et al. 1979, 1981;

3-15 English et al. 1995; Herberts et al. 1981]. working as a checker. Summaries of other studies are presented in alphabetical order. Bjelle et al. [1979] compared cases with persistent shoulder pain to controls employed In the study by Baron et al. [1991], grocery as manual workers. After an extensive medical store workers who performed the job of evaluation, a diagnosis of bicipital and/or checker were compared to a non-checker supraspinate tendinitis was made for a majority group that performed a variety of other jobs (12/17) of the cases. Physical workload was (e.g., general stocking, working in the produce categorized in relation to sitting or standing section, the bakery, salad bar, pharmacy, and posture, weight lifting, and carrying. The work courtesy counter). There was a low height of the hands was categorized based on participation rate among non-checkers (55%), position relative to the acromion height, per which could have resulted in an underestimation individual. Placement of workers into exposure of the OR for checkers if symptomatic non- categories was determined by the combined checkers were more likely to participate than efforts of each study participant and a those non-checkers without symptoms. The physician. authors evaluated this possibility by performing a sufficient number of telephone interviews with Bjelle et al. [1981] compared cases with acute, non-participants to raise the non-checker non-traumatic shoulder-neck pain to age- and participation rate for interviews to 85%. The sex-matched, paired controls. An extensive OR for shoulder symptoms among the full physical examination was performed and participant population was similar to the OR for workers with inflammatory rheumatoid diseases the full participant plus telephone interview were excluded. To determine exposure, each population. The case definition was shoulder case and control was filmed and a symptoms lasting at least one week or biomechanical analysis was performed to occurring at least once per month during the determine the duration and frequency of previous year that began while the worker was shoulder abduction or forward flexion greater performing her current job and positive physical than 60 degrees. examination findings consistent with a shoulder tendinitis. Detailed descriptions of the checker In a study by English et al. [1995], cases jobs were presented based on both on-site and determined by medical diagnosis, and controls videotape analyses of a few representative were selected from patients evaluated at workers per workstation. No videotaping of specified orthopedic clinics. For statistical non-checkers was performed. Shoulder flexion analyses, all diagnoses were grouped by and/or abduction ($90 degrees) was observed anatomical site. The diagnoses for shoulder during a variety of different tasks performed by cases included rotator cuff injury, rupture of the the checkers. The exposure measures used in long head of the biceps, shoulder capsulitis, and statistical analyses were: (1) checker versus symptomatic acromioclavicular arthritis. It is non-checker and, (2) for exposure-response assumed that shoulder tendinitis was included in assessment among checkers, the total number this group. Exposure measures were of months and the number of hours per week determined by a standardized interview

3-16 conducted by an interviewer who was, In a prospective study by Sakakibara et al. “unaware of the case-control status of the [1995], the health status of a group of women individual wherever this was possible.” farm workers was assessed during the performance of two different tasks, with a 1- In a study by Herberts et al. [1981], the month interval between the tasks. The health prevalence of supraspinatus tendinitis was outcome was defined by self-assessed compared between welders and office symptoms of shoulder stiffness and pain and a workers. Tendinitis cases were based on a physical examination for muscle tenderness and combination of symptoms reported on a nurse- joint pain on movement. Whether the examining administered questionnaire and a positive physician was aware of the prior hypothesis physical examination done by a physiotherapist. regarding differing exposures between the two For welders, an “experienced physiotherapist” tasks (bagging pears versus bagging apples) rated work-load on the shoulder as low, high, was not stated. Exposure was based on self- or very high; no description of the classification report of the number of hours per day spent scheme was given. bagging, the number of pears or apples bagged per day, and the total number of days spent Hoekstra et al. [1994] evaluated government bagging each fruit. One worker was observed office workers at two locations. The case for 3 hours while performing each bagging job, definition for shoulder symptoms was with repeated goniometer measures of shoulder symptoms that began after starting current job, forward flexion angles done each minute. While lasting greater than one week, or occurring at there was no difference in the total number of least once per month during the past year with days or number of hours per day spent bagging an intensity greater than two on a five point each fruit, significantly more pears than apples scale, and no preceding acute, non- were bagged per day. The proportion of time occupational injury. A self-administered spent with the angle of shoulder forward flexion questionnaire was used to determine exposure greater than 90 degrees was significantly larger to factors such as “perceived adequacy of when bagging pears (75%) than when bagging adjustment of video display terminal (VDT).” apples (41%). Walk-through ergonomic evaluations of factors such as workstation surface height and furniture Schibye et al. [1995] performed a prospective adjustability were used to provide descriptive study of a population of sewing machine differences between the two office locations. operators in which the change in self-reported shoulder symptom status was compared with Milerad and Ekenvall [1990] compared the those sewing machine operators who continued prevalence of self-reported, non-specific to work and those operators that moved into shoulder symptoms between dentists and other occupations (e.g., shop assistant, health pharmacists. Dentistry, as a profession, was care worker, and fishing industry worker). described as work “with the arms abducted and unsupported” whereas, pharmacists had “physically light and varied work.”

3-17 Strength of Association—Awkward OR for being a checker versus a non-checker Posture and Shoulder MSDs was 3.9 (95% CI 1.4–11.0). Because non- Results are presented in the section on checkers also performed work requiring Exposure-Response Relationship (Table 3-3, awkward postures, the reported OR may Figure 3-3) for studies where ORs, or other underestimate the risk for checkers. Short measures of association, were specifically stature (# 5'2") was associated with an associated with a measure of exposure. elevated, but not statistically significant, OR for shoulder disorders (2.1, 95% CI 0.7–6.9). Using data presented in the study by Ohlsson et Because work-station height was fixed, it is al. [1994], for supraspinatus, infraspinatus, or likely that short stature workers experienced bicipital tendinitis, the PRR for working in the more frequent and/or more severe episodes of fish industry (repetitive work, poor posture) shoulder flexion and/or abduction. versus the referent population was calculated as 3.03 (95% CI 2.0–4.6). For shoulder tendinitis The OR for work performed at or above alone, the PRR was calculated as 3.5 (95% CI acromion height (i.e., hands above the 2.0–5.9). In the same study, the authors also shoulder) versus work performed below interviewed a large group of former fish acromion height was 10.6 (95% CI 2.3–54.9) industry employees and found that a quarter of [Bjelle et al. 1979]. In this study, all cases were those workers who left employment had done patients who had been examined by the same so because of problems with their neck or physician. Placement of cases and controls into upper limbs. This proportion increased with age exposure categories was performed by each and also occurred after a shorter duration of subject in collaboration with a physician who employment among the oldest workers. This “had personal knowledge of the work involved evidence of a survivor bias highlights the in each case.” Whether or not the physician importance of controlling for age. Higher risks who performed the clinical examinations is the were found for the workers less than 45 years same person as the physician involved in old and these risks may be a more accurate exposure classification is not stated. If this was assessment of the true risk. the same person, a potential bias towards assigning cases to higher exposure categories Using data presented in the study by Ohlsson et could have resulted in overestimation of the al. [1995], for supraspinatus, infraspinatus, or strength of association. However, two other bicipital tendinitis, the OR for being an factors could have resulted in an assembly worker (repetitive arm movements underestimation of the strength of association. with static load on shoulders) versus the The exposure outcome was based on current referent population was 4.2 (95% CI work load without any stated restriction that 1.35–13.2). For neck-shoulder disorders, the cases’ symptoms had started on their current OR for being an assembly worker versus the job. If some of the cases, defined as having referent group was 5.0 (95% CI 2.2–11.0). problems non-responsive to therapy lasting longer than 3 months, had transferred to a For shoulder disorders consistent with lower risk job, the strength of association tendinitis, Baron et al. [1991] found that the

3-18 may have been underestimated. Location of the two locations provided a plausible explanation disorder and exposure were not matched by for this finding. At the higher risk location, the side (left, right, or both) and this would have workstation surface was too high to serve as a caused non-differential misclassification, keyboard support, there were nonadjustable resulting in some underestimation of the strength chairs, and it was observed that “nonadjustable of association. furniture universally promoted undesirable postures (i.e. elevated arms, hunched English et al. [1995] found that the risk of shoulders).” Having shoulder symptoms was having a medically diagnosed shoulder also positively associated with using a non- condition was increased by repeated shoulder optimally adjusted desk height (OR 5.1, 95% rotation with an elevated arm (OR 2.30, CI 1.7–15.5) and a non-optimally adjusted p<0.05). Non-differential misclassification due VDT screen (OR 3.9, 95% CI 1.4–11.5). to a combination of complicated exposure Because exposure was self-reported without definitions using a questionnaire, and the fact any indication of whether or not study that analyses did not relate health outcomes and participants had received education regarding exposure on a temporal basis, or by left/right good VDT workstation design, the phrase, side, may have caused an under-estimate of the “non-optimally adjusted,” may have had various strength of association. meanings to the study participants. This could have caused non-differential misclassification of For supraspinatus tendinitis, Herberts et al. exposure and an under-estimation of the [1981] found that the PRR for welders strength of association. On the other hand, a (characterized as using awkward postures to possible reporting bias related to self perform overhead work) versus clerks was assessment of both symptoms and exposure 18.3. However, in determining this PRR, the could have resulted in an overestimation of the authors performed extrapolation based on an strength of the association. A plausible assumption that, “the drop-out group does not explanation for the association between deviate from the examined group,” without any shoulder symptoms and these workstation data to support this assumption. To determine a design factors is that the non-optimally adjusted more reliable indicator of risk, unextrapolated workstation components forced the employees data presented in the study were used to to abduct the upper arms and/or hunch the calculate a crude OR=8.3 (95% CI 0.63–432). shoulders. The office clerks were older than the welders, so that confounding by age may have caused an For shoulder symptoms without concomitant under-estimation of the strength of association. neck symptoms, Milerad and Ekenvall [1990] found that the OR for being a dentist (work In a study of teleservice employees, there was with both arms abducted) versus being a an association between reporting shoulder pharmacist was 3.8 (95% CI 1.2– 10.3). As symptoms and working at one location versus with most cross-sectional studies, the survivor another location; the OR was 4.0 (95% CI bias may have resulted in 1.2–13.1) [Hoekstra et al. 1994]. Descriptive differences between workstation design at the

3-19 underreporting of the strength of exposure. shoulder symptoms and awkward shoulder Conversely, the exposed group may have had posture. Awkward postures were consistently better recall of self-reported symptoms with a described as overhead work, arm elevation, resultant overestimation of the OR. and specific postures relative to degrees of upper arm flexion or abduction. This In the study of farm workers by Sakakibara et association was found in cross-sectional, case- al. [1995], the point prevalence of muscular control, and prospective studies among a great tenderness in the shoulder regions (per physical variety of types of work performed. examination) was significantly higher when performing pear bagging (48%) than when Temporal Relationship performing apple bagging (29%). The It is important to determine whether symptoms proportion of time spent with the shoulder in or MSDs occur as a consequence of work- forward flexion greater than 90 degrees was related exposures. This can be done most significantly larger when bagging pears (75%) clearly with a prospective study design. than when bagging apples (41%). Whether or not there was a recovery period between pear In the study by Jonsson et al. [1988], the and apple bagging is not stated. If there was percent of the work cycle spent with the insufficient recovery after pear bagging, shoulder elevated was negatively associated persistent muscle tenderness or increased with remaining healthy (symptom free). susceptibility may have caused underestimation Because workers with pre-existing shoulder of the difference in shoulder conditions were excluded from study disorder prevalence between these two work ta participation, the onset of new symptoms may sks. have been associated with the daily and/or cumulative duration of exposure to elevated With the exception of the study by English et al. shoulder postures. In the study by Kilbom and [1995], in which the strength of association may Persson [1987], three of the work exposure have been underestimated, for the studies in variables that were strong predictors for a which the health outcome was shoulder change to severe status at the 1- and/or 2-year tendinitis [Baron et al. 1991; Bjelle et al. 1979; follow-up evaluations were related to shoulder Herberts et al. 1981; Ohlsson et al. 1994, posture: (1) percent of work cycle time with 1995], the magnitude of association was strong. arm abduction greater than 30 degrees, (2) ORs ranged from 2.0 to 10.6. In none of these percent of work cycle time with arm abduction studies is it likely that nonoccupational factors greater than 60 degrees, and (3) percent of such as sports activities or personal work cycle time with arm extension. characteristics such as age explain these associations. A few studies utilized techniques to improve the ability to detect possible relationships Consistency of Association All but one of the reviewed studies relevant to posture and shoulder disorders found a positive association between shoulder disorders or

3-20 despite a cross-sectional study design. The this association was significant despite very little case definition used by Baron et al. [1991] variability in exposure to arm abduction greater required that symptoms began while the worker than 60 degrees. While the analysis among was on the currently held job. Bjelle et al. assembly workers was performed without [1979] filmed and analyzed the job held at the controlling for age, there is no evidence to time the worker/case became symptomatic. suggest that older workers were more likely to The results of the prospective studies are be on high exposure jobs, and therefore a similar to the cross-sectional studies. There is substantial bias is unlikely. no evidence that shoulder disorders predicted the onset of exposure. When comparing fish industry workers to the reference population, Ohlsson et al. [1994] Exposure-Response Relationship found that among those workers younger than The level of an exposure can be described in 45 years, the ORs for disorders of the neck two different ways. It may be related to the and shoulders were significant and increased amount of exposure over a relatively short time with duration of employment (0–5 years, OR period, such as a day or week, or it may be 3.2; 95% CI 1.5–7.0) (>5 years, OR 10; 95% related to cumulative or life-time exposure over CI 4.5–24). Ohlsson et al. [1995] found a a number of years. Studies that tested decreasing trend when they compared OR after associations related to daily or weekly variation stratifying the factory workers by employment in exposure are presented first, followed by duration (<10 years, OR 9.6; 10–19 years, OR studies that evaluated cumulative exposure by 4.4 and $20 years: 3.8). Given the cross- using independent variables, such as duration of sectional study design, this finding could be an employment or estimated lifetime exposure. artifact caused by the survivor bias (i.e., workers with disorders left, while symptom- Four studies have some evidence of exposure- free ‘survivors’ stayed; see Table 3-5). The response relationships. Baron et al. [1991] assumption of a survivor bias is based on the found a significantly larger OR for shoulder finding that 28% of a group of former assembly disorders among employees working greater workers reported pain in the musculoskeletal than 25 hours/wk as a checker compared to system as their reason for leaving employment those working less than 20 hours/wk. Bjelle et at the factory. In the study by Schibye et al. al. [1981] found that the duration of hours [1995], improvement in shoulder symptoms worked per day with the shoulder flexed or among those who were no longer sewing abducted >60 degrees was significantly higher machine operators appeared greater at follow- (p<0.025) for cases with neck-shoulder up, but was not significant. The fact that many disorders than for controls. Ohlsson et al. of those who left sewing jobs moved into [1995] found that neck and shoulder disorders industries such as health care and fishing, where among assembly workers were significantly awkward postures and high force loads may associated (p<0.05) with the percent of time occur, might explain why a large difference spent with the shoulder abducted or elevated between sewing machine operators and non- >60 degrees. Although it is more difficult to detect associations with homogenous exposure,

3-21 sewing machine operators was absent. These is evidence for both shoulder tendinitis and four studies provide some support for the nonspecific shoulder pain. The evidence for relationship between shoulder abduction and increased risk of MSDs due to specific shoulder MSDs. shoulder postures is strongest when there is a combination of exposures to several physical Coherence of Evidence factors such as force and repetitive work. An Discussions of the probable influence of example of this combination would be holding a workplace exposure factors in the tool while working overhead. The strength of pathophysiology of localized muscle fatigue, association was positive and consistent in the myalgia, and tendinitis have been presented by six studies that used diagnosed cases of a number of authors [Bjelle et al. 1981; shoulder tendinitis, or a combination of Hagberg 1984; Herberts and Kadefors 1976; symptoms and physical findings consistent with Herberts et al. 1984; Levitz and Iannotti 1995]. tendinitis, as the health outcome [Baron et al. Posture is important: when the arm is raised or 1991; Bjelle et al. 1979; English et al. 1995; abducted, the muscle activity in supraspinatus Herberts et al. 1981; Ohlsson et al. 1994, and other muscles increases, and the 1995]. Only one [Schibye et al. 1995] of the supraspinatus tendon comes in contact with the thirteen studies failed to find a positive undersurface of the acromion. The mechanical association with exposure and symptoms or a pressure on the tendon from the acromion is specific shoulder disorder. However, in this greatest between 60 and 120 degrees of arm study discontinuing employment as a sewing elevation. [Levitz and Iannotti 1995]. The machine operator was associated with a degree of upper arm elevation is also important reduction in neck and shoulder symptoms. in the onset and intensity of localized muscle While most of the studies that considered fatigue in the trapezius, deltoid, and rotator cuff specific shoulder postures as an exposure muscles. [Hagberg 1981; Herberts and variable were cross-sectional, the two Kadefors 1976; Herberts et al. 1984]. In a prospective studies found that the percent of laboratory study, EMG signals from these work cycle spent with the shoulder elevated muscles were analyzed. The supraspinatus [Jonsson et al. 1988] or abducted [Kilbom et muscle was found to be highly active at $45 al. 1986; Kilbom and Persson 1987] predicted degrees of abduction. The deltoid muscle change to more severe neck and shoulder underwent a pronounced increase in activity as disorders. While there is insufficient evidence to shoulder flexion or abduction increased from 45 develop a quantitative exposure-disorder to 90 degrees [Herberts et al. 1984]. The relationship, three studies reported a significant earlier sections on Coherence of Evidence also association with shoulder flexion or abduction discussed the rate of fatigue and role of greater than 60 degrees [Bjelle et al. 1981; impaired micro-circulation in shoulder tendinitis. Kilbom and Persson 1987; Ohlsson et al. 1995]. Among the studies for which shoulder Overall, there is epidemiologic evidence for a tendinitis was the health outcome, the largest relationship between repeated or sustained ORs were associated with work above shoulder postures with more than 60 degrees of acromion height [Bjelle et al. 1979; flexion or abduction and shoulder MSDs. There

3-22 Herberts et al. 1981]. These results are non-significant. consistent with the current models for the pathophysiology of shoulder tendinitis and In the study by Stenlund et al. [1993], the same stressful shoulder muscle activities. In none of population of bricklayers, rock blasters, and these studies does “age,” an important personal foremen described in Stenlund et al. [1992] characteristic associated with shoulder were evaluated to determine whether signs of tendinitis, explain the positive results. Most of tendinitis or muscle attachment inflammation in the studies controlled for a variety of the shoulders were related to lifetime work confounders, such as occupational sports load, years of manual work, lifetime exposure activities in their analyses. In summary, there is to vibration, or job title. The case definition for evidence that repeated or sustained shoulder “signs of shoulder tendinitis” was pronounced abduction or flexion is associated with shoulder (i.e., grade 3 out of 3) pain upon palpation of tendinitis, and the evidence is stronger for highly the muscle attachment or pronounced pain in repetitive, forceful work. response to isometric contraction of any of the rotator cuff muscles or the biceps muscle. The VIBRATION case definition of “clinical entity of tendinitis” Three of the studies evaluated exposure to low- was “signs of shoulder tendinitis” plus the frequency vibration found in industrial settings subject’s report of shoulder pain during the past (Table 3-4, Figure 3-4). Because of the small year. Using multivariate models that included number of studies, the full outline used for the age and hours spent in arm intensive sports sections on repetition, force, and posture will activities, a significant association with not be repeated here. The study by Stenlund et cumulative vibration exposure was found when al. [1992] is summarized in the section on it was tested in isolation from the other force. Vibration exposure occurred in one of exposure variables. For “clinical entity of the three job categories: rock blaster. The tendinitis” the OR for the left side was 1.86 exposure outcome, lifetime exposure to (95% CI 1.00–3.44) and the OR for the right vibration expressed in hours, was determined side was 2.49 (95% CI 1.06–5.87). from a weighted summary of the number of For “signs of shoulder tendinitis” the OR self-reported hours using specific tools. for the left side was 1.66 (95% CI 1.06–2.61) However, because the rock blaster job and the OR for the right side category was also the only one where workers was 1.84 (95% CI 1.10–3.07). When performed heavy lifts several times per day, the cumulative vibration exposure was tested authors concluded that, “vibration exposure is in the same model with cumulative lifting load, indivisible from static load and heavy lifting in significant associations were not found for the present data.” When both cumulative lifting either variable. Several factors could have exposure and cumulative vibration exposure resulted in an underestimation of the strength of were included in the same multivariate model of association: (1) bricklayers or rock blasters an association with acromioclavicular with tendinitis may have been more likely to osteoarthrosis, the OR for lifting and right- side leave their jobs than foremen, (2) subjects may osteoarthrosis remained significant have had difficulty recalling exposure while the weaker ORs for vibration became throughout their

3-23 lifetimes, (3) the inability to separate exposure weighted according to International Standards by left and right sides. These factors may have Organization (ISO) standards. Tool vibration caused nondifferential misclassification. Most profiles and time-work studies of riveters and important is the authors’ observation that controls were used to determine daily vibration vibration exposure occurred through the used exposure for each group. For riveters, on the of hand-held, heavy tools (e.g., jack-hammers) basis of daily tool operating time, the equivalent and thus is intertwined with exposure to a static frequency-weighted acceleration for a period of load on the shoulders (from stabilizing the upper 4 hours was 2.8 m s -2. For controls, it was 1.0 extremity while using the tool) as well as being m s -2. Using a multiple logistic regression associated with the heavy lifting tasks model that included age, there was a weak performed by rock blasters. association between shoulder symptoms and the number of years riveting (0.05# p<0.10). In a cross-sectional study by Burdorf and When the age-adjusted ORs for riveters Monster [1991], riveters and control subjects compared to controls were plotted by the in an aircraft company were investigated for duration (in years, from 0 to 20) of riveting, the vibration exposure and self-reported symptoms slope for shoulder symptoms was very gradual, of pain or stiffness in the shoulder. Riveters with ORs ranging from 1.0 to 2.0. While the were exposed to hand-arm vibration from results of the analysis of non-respondents working with hand drills, riveting hammers, described above suggest that the strength of bucking bars, and grinders. Controls were association may have been underestimated, the manual workers selected from the machine reported associations are weak and it is shop, maintenance, and welding departments in unlikely that the response bias would have the same factory. In order to focus on the effect resulted in a large increase in the magnitude of of vibration alone, a walk-through survey was association. performed to confirm that there were “no striking differences in dynamic and static joint There is insufficient evidence for an association loads during normal working activities.” between shoulder tendinitis Participation was 76% among riveters and 64% and exposure to segmental vibration. In among controls. An analysis of non- four separate evaluations, stratified by “signs of respondents revealed that controls with health tendinitis” (positive physical examination complaints were more likely to have findings), “clinical entity participated than those without, while riveters of tendinitis” (signs plus symptoms), left and with health complaints were less likely to have right side, Stenlund et al. [1993] participated. The health outcome, determined found an association between shoulder by a self-administered questionnaire, was tendinitis and vibration exposure to segmental shoulder pain or stiffness occurring for at least a vibration; the range of ORs few hours during the prior year. Only subjects was from (OR for right side 1.66, 95% CI who reported having no symptoms before 1.06–2.61) (OR for left side 1.84, 95% CI starting their present work were included in 1.10–3.07). However, work with vibration logistic regression analyses. The vibration exposure also placed a large, static load on transmitted by hand-tools was measured and

3-24 shoulder muscles so that the effects of forceful characteristics. Using epidemiologic criteria to shoulder muscle exertions could not be examine these studies, and taking into account separated from vibration. issues of confounding, bias, and strengths and limitations of the studies, we conclude the ROLE OF CONFOUNDERS following: Shoulder MSDs are multifactorial in origin and may be associated with both occupational and There is evidence for a positive association non-occupational factors. The relative between highly repetitive work and shoulder contributions of these covariates may be MSDs. The evidence has important limitations. specific to particular disorders. For example, Only three studies specifically addressed the the confounders for non-specific shoulder pain health outcome of shoulder tendinitis and these may differ from those for shoulder tendinitis. studies investigated combined exposure to Two of the most important confounders or repetition with awkward shoulder postures or effect modifiers for shoulder tendinitis are age static shoulder loads. The other six studies with and sport activities. Most of the shoulder significant positive associations dealt primarily studies considered the effects of age in their with symptoms. There is insufficient evidence analysis. Some studies considered sport for a positive association between force and activities [Baron et al. 1991; Stenlund et al. shoulder MSDs based on currently available 1993; Jonsson et al. 1988; Kilbom et al. epidemiologic studies. There is epidemiologic 1986]. Some studies also used multivariate evidence for a relationship between repeated methods to simultaneously adjust for several or sustained shoulder postures with greater than confounders or effect modifiers. For example, 60 degrees of flexion or abduction and shoulder Ohlsson et al. [1995] found that for MSDs. There is evidence for both shoulder shoulder/neck diagnoses, repetitive work was tendinitis and nonspecific shoulder pain. The the strongest predictor 4.6 (95% evidence for specific shoulder postures is CI 1.9–12); age, muscle tension, and strongest where there is combined exposure to stress/worry tendency were also significant several physical factors like holding a tool while predictors. It is unlikely that the majority of the working overhead. The strength of association positive associations between physical was positive and consistent in the six studies exposures and shoulder MSDs are due to the that used diagnosed cases of shoulder tendinitis, effects of non-work confounders. or a combination of symptoms and physical findings consistent with tendinitis, as the health CONCLUSIONS outcome. Only one [Schibye et al. 1995] of the There are over 20 epidemiologic studies that thirteen studies failed to find a positive have examined workplace factors and their association with exposure and a specific relationship to shoulders (MSDs). These shoulder disorder or symptoms of a shoulder studies generally compared workers in jobs disorder. with higher levels of exposure to workers with lower levels of exposure, following observation or measurement of job

3-25 This is consistent with the evidence that is found There is insufficient evidence for a positive in the biomechanical, physiological, and association between vibration and shoulder psychosocial literature. MSDs based on currently available epidemiologic studies.

3-26 Table 3-1. Epidemiologic criteria used to examine studies of shoulder MSDs associated with repetition

Investigator Risk indicator blinded to case (OR, PRR, IR, Participatio Physical and/or Basis for assessing Study (first author and or p-value)*,† n rate $70% examinatio exposure shoulder year) n status exposure to repetition

Met all four criteria:

Chiang 1993 1.6† Yes Yes Yes Observation or measurements

Kilbom 1986, 1987 NR†,‡ Yes Yes Yes Observation or measurements

Ohlsson 1994 3.5† Yes Yes Yes Observation or measurements

Ohlsson 1995 5.0† Yes Yes Yes Observation or measurements

Met at least one criterion:

Bjelle 1981 NR† NR Yes Yes Observation or measurements

English 1995 2.3†,§ Yes Yes Yes Job titles or self-reports

Sakakibara 1995 1.7† Yes Yes NR Job titles or self-reports

Met none of the criteria:

Ohlsson 1989 3.4† NR No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported. §Repeated shoulder rotation with elevated arm.

3-27

Table 3-2. Epidemiologic criteria used to examine studies of shoulder MSDs associated with force

Investigator Risk indicator blinded to (OR, PRR, IR or Participatio Physical case and/or Basis for assessing Study (first author and p-value)*,† n rate $70% examination exposure shoulder exposure to year) status force

Met all four criteria:

Chiang 1993 1.8† Yes Yes Yes Observation or measurements

Met at least one criterion:

Andersen 1993a 1.38–10.25† Yes No Yes Job titles or self-reports

Herberts 1981, 1984 15–18† NR‡ Yes NR Job titles or self-reports

Stenlund 1992 2.2–4.0† Yes Yes Yes Job titles or self-reports

Wells 1983 5.7† Yes No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported.

3-29

Table 3-3. Epidemiologic criteria used to examine studies of shoulder MSDs associated with posture

Investigator Risk blinded to indicator Participatio Physical case and/or Basis for assessing Study (first author and (OR, PRR, IR, n rate $70% examination exposure shoulder year) or status exposure to posture p-value)*,†

Met all four criteria:

Jonsson 1988 NR†,‡ Yes Yes Yes Observation or measurements

Kilbom 1986, 1987 NR† Yes Yes Yes Observation or measurements

Ohlsson 1994 3.5† Yes Yes Yes Observation or measurements

Ohlsson 1995 5.0† Yes Yes Yes Observation or measurements

Met at least one criterion:

Baron 1991 3.9† No Yes Yes Observation or measurements

Bjelle 1979 10.6† NR Yes No Observation or measurements

Bjelle 1981 NR† NR Yes Yes Observation or measurements

English 1995 2.3†,§ Yes Yes Yes Job titles or self-reports

Herberts 1981 8.3 NR Yes NR Job titles or self-reports

Hoekstra 1994 5.1† Yes No Yes Job titles or self-reports

Milerad 1990 2.4† Yes No NR Job titles or self-reports

Sakakibara 1995 NR† Yes Yes NR Observation or measurements

Schibye 1995 NR Yes No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on posture alone (i.e., posture plus force, repetition, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported. §Repeated shoulder rotation with elevated arm (p< 0.05 level, most of study used 0.01 level).

3-31

Table 3-4. Epidemiologic criteria used to examine studies of shoulder MSDs associated with vibration

Investigator Risk blinded to case indicator Participatio Physical and/or Basis for assessing Study (first author and (OR, PRR, IR, n rate $70% examination exposure shoulder exposure to year) or p-value)*† status vibration

Met at least one criterion:

Burdorf 1991 1.5 No No NR‡ Observation or measurements

Stenlund 1992 2.2–3.1† Yes Yes Yes Self-reports, weight of tools

Stenlund 1993 1.7–1.8† Yes Yes Yes Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on vibration alone (i.e., vibration plus force, posture, or repetition). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported.

3-33

Table 3-5. Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen Cross- 424 female sewing Outcome: Case of chronic Shoulder pain: Participation rate: 78.2%. and sectional machine operators shoulder pain was defined as Sewing Gaardboe (SMO), compared to continuous pain lasting for a machine Examiners blinded to case status. 1993a 781 females from the month or more after beginning operators, general population of the work and pain for at least 25.2% 8.5% 3.21 1.68-7.39 Respondents excluded if had region and internal 30 days within the past year. previous trauma to neck, shoulder, referent group of Years of or arms or had inflammatory disease 89 females from the Exposure: Categorization broken exposure: at time of response. garment industry. down according to current 0-7=12.3% 1.56 0.76-3.75 occupational status by job title. ORs adjusted for age, having Classification into exposure 8-15=33.7% 4.28 2.14-10.0 children, not doing exercise, groups based on author’s socioeconomic status, smoking, and experiences as occupational >15=57.1% 7.27 3.82-16.3 current neck/shoulder exposure. health physicians and involved crude assessment of exposure Age-matched exposure groups and level and exposure controls. repetitiveness. High exposure jobs were those involving high Presented study as “general survey repetition/high force or high of health in the garment industry” to repetition/low force or medium minimize information bias. repetition/high force. Medium exposure jobs were those involving medium repetition/low force and low repetition and high force. Low exposure jobs were low repetition/low force.

For the analysis, “length of employment as a sewing machine operator” was considered the variable of interest, the rest were confounders.

(Continued)

3-35

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen Cross- From a historical cohort Outcome: Measured by health Rotator cuff Participation rate: 78.2%; logistic and sectional of 424 sewing machine interview and exam of the neck, syndrome: regression limited to a combined Gaardboe operators, 120 were shoulder and arm. Case of neck/shoulder case definition. 1993b randomly selected and chronic pain was defined as Number of Controls: 1 Chi sq for 82 exposed workers continuous pain lasting for a workers by trend=9.51, Age-matched exposure groups and were categorized by month or more after beginning exposure time p<0.01 controls. number of years of work and pain for at least in years: employment: 0 to 30 days within the past year. 0-7: 1; Examiners blinded to control/subject 7 years, 8 to 15 years Physical examination: Restricted 8-15: 6; status. and greater than 15 movements in the cervical spine >15: 11 years. These were and either palpatory tenderness Controlled for age, having children, compared to a referent in cervical segments or not doing leisure exercise, smoking, group of 25 auxiliary irradiating pain or tingling at socioeconomic status. nurses and home maximum movements or positive helpers. A total of 107 foraminal test. Poor correlation between subjects participated. degenerative X-ray neck changes Exposure: Exposure and cervical syndrome. categorization broken down according to current Most frequent diagnosis among occupational status by job title. study group was “cervicobrachial Classification into exposure fibromyalgia” significant for test of groups based on author’s trend with exposure time in years. experiences as occupational health physicians and involved Chronic neck pain vs. palpatory crude assessment of exposure findings: Sensitivity: 0.85; level and exposure Specificity: 0.93. repetitiveness. High exposure jobs: Involved high repetition/high force or high repetition/low force or medium repetition/high force. Medium exposure jobs involved medium repetition/low force and low repetition and high force. Low exposure jobs were low repetition/low force.

(Continued)

3-36

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Baron et al. Cross- 124 Grocery checkers Outcome: Based on symptom Checkers: Other Checkers vs. Participation rate: 85% checkers; 55% 1991 sectional using laser scanners (119 questionnaire and physical exam. 15% grocery others: non-checkers in field study. females, 5 males) (1) Rotator cuff syndrome—pain workers: OR=3.9 1.4-11.0 Following telephone survey 91% compared to 157 other with resisted abduction or deltoid 7% checkers and 85% non-checkers. grocery workers (56 palpation (2) Bicipital Checkers Checkers Examiners blinded to worker’s job and females, 101 males). tendinitis—pain on Yergason’s using using Excluded 18 workers in maneuver. Case defined as scanners: scanners vs. health-status. meat, fish, and deli having positive symptoms in 34% others: Logistic regression model adjusted for departments, workers shoulder and a positive physical OR=8.6 1.0-72.2 duration of work. No difference in under 18, and pregnant exam of a particular body part. groups between age, gender, and workers. Symptoms must have begun after Checkers Other Checkers hobbies so that these were not employment at the supermarket 5'2" or less in grocery <5'2" vs. controlled for. and in the current job; lasted one height: 21% workers other grocery week or occurred once a month 5'2" or workers Number of hr worked/week as a during the past year; and where less in <5'2": checker statistically significantly there was no history of acute height: OR=2.1 0.7- 6.9 related to shoulder disorders for injury to body part in question. 13% workers checking >25-hr/ /week (OR=3.5, p<0.05) Exposure: Job category and (OR estimated from figure). estimates of repetitive and Total repetitions/hr ranged from 1,432 average and peak forces based to 1,782 for right hand and 882 to on observed and videotaped 1,260 for left hand. postures, weight of scanned items, and subjective assessment Average forces were low and peak of exertion. forces medium. Multiple awkward postures recorded for upper extremities among cashiers. No statistical significance associated between duration of employment as a checker and shoulder MSDs.

(Continued)

3-37

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Bergenudd Cross- 574 of 830 survey Outcome: Based on symptom Prevalence Õ Õ Participation rate: 69%. et al. 1988 sectional respondents participated in survey: Occurrence of shoulder of Unknown whether examiners blinded a health exam. pain lasting $24 hr during the last occupational to case status. month and physical exam (joint workload in Analysis stratified by gender. In 1983, 1,070 residents of motion, tenderness on palpation subjects with Only 9% of workers included in study Malmö, Sweden, of supraspinatus, biceps, shoulder pain were in the Heavy Physical Demands responded to questions on tendons and acromioclavicular Jobs category, compared to 49% in shoulder pain in a health joint). Heavy work: Light category and 42% in moderate survey as part of a 11% category. Only 1% of females were longitudinal study begun in Exposure: Based on job in Heavy Physical Demand 1,938 of 1,542 residents. classification; classified as: Light Moderate Jobs category. physical demands (white collar) work: 49% Sick leave due to shoulder pain was =275; Moderate physical restricted to males in jobs with demands (nurses, light Light work: moderate or heavy physical demands industry)=237; Heavy (blue collar, 40% (p<0.05) (data not shown in article). e.g., carpenters, bricklayers)=50. At one year follow-up, 61 (77%) of 79 subjects with shoulder pain re- examined. 35 had continued shoulder pain. Misclassification of work categories a possibility: Likely no observation of job tasks performed.. No differences in overall physical demands of jobs among subjects with shoulder pain compared to those without shoulder pain, but females with signs of supraspi-natus tendinitis more often had jobs with physical demands. Authors state that shoulder pain may be related to intelligence in males in this study; “more talented” males had less shoulder joint symptoms. We question author’s conclusions. Females showed significant association with shoulder pain and dissatis-faction. No association with relation to family or friends or level of life success. Author states both groups of females rated their life success low, and subjects with shoulder pain did not rate level of success differently.

(Continued)

3-38

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Bernard Cross- Of a total population of Outcome: Health data and 17% (case) Õ Female: Participation rate: 93%. et al. 1994 sectional 3,000 workers in the psychosocial information were OR=2.2 1.5-3.3 editorial, circulation, collected using a self- 3% (case Examiners blinded to case and classified advertising, and administered questionnaire. with daily Perceived exposure status. accounting departments, Definition: Presence of pain, pain) lack of 1,050 were randomly numbness, tingling, aching, decision For calculation of the ORs of the selected for study and stiffness or burning in the making psychosocial scales, the responses 973 participated; shoulder occurring $once a participation: were divided into quartiles, then the 894 responded to the month or 7 days continuously OR=1.6 1.2-2.1 75th percentile was compared to 25th shoulder questions. within the past year, reported as percentile. moderately severe. The symptom Years at the Cases fulfilling shoulder must have begun during the newspaper: Model adjusted for race, age, gender, definition compared to non- current job. Workers with OR=1.4 1.2-1.8 height, psychosocial factors, medical cases. previous injuries to the relevant conditions. area were excluded. Perceived increased job Age, height, hr typing away from Exposure: Based on observation pressure: work, other medical conditions were of work activity involving OR=1.5 1.0-2.2 not found to be significant. keyboard work, work pace, posture, during a typical day of a In a sub-analysis of jobs with sample of 40 workers with comparable number of males and symptoms and 40 workers females, there were no significant without symptoms. Exposure to factors related to shoulder MSDs. work organization and psychosocial factors based on questionnaire responses.

(Continued)

3-39

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bjelle et al. Case- 17 cases of shoulder Outcome: Cases were non- With work at With work Participation rate: Not reported. 1979 control tendinitis from a population responsive to analgesics, non- or above at or of male industrial workers steroidal anti-inflammatory shoulders: above Matched for age, gender and who were patients at an agents, physiotherapy, and 65% shoulders: workshop. occupational health center. outcome measured by exam. 15% 10.6 2.3-54.9 These 17 were chosen Case defined as shoulder pain Three of the 20 were diagnosed with from 20 consecutive male lasting >3 months with no inflammatory rheumatoid diseases not patients from 6 industries resolution post-treatment. previously diagnosed, 17 had no and had been suffering inflammatory rheumatic disease. from pain over a period of Exposure: Defined as work with >3 months in one or both hands at or above shoulder level. Mean age (53 years) of cases shoulders. 3 classes work performed: (A) significantly older than other workers with hands below shoulder or (37.6 years). 34 non-cases were acromion height, (B) at or above matched for age and acromion 3 to 8 times/day (<1/hr Myopathic signs not found on EMG or workshop. plus for duration >1 min) (C) $8 muscle biopsies. Muscle enzymes times at or above acromion (creatine phosphokinase and/or ($1/hr. plus duration >1 min). aldolase) were elevated in 6 cases. Exposure assessed by interview and physician observation and Present and previous employment, knowledge of work. physical workload not different between cases and referents. Electromyographs on 15 cases. Work performed with hands above Open muscle biopsies on acromion height significantly greater 11 cases. for cases than referents.

2-year follow-up showed that only 8 cases working in the same or less heavy types of work, 7 of these had slight shoulder complaints.

(Continued)

3-40

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bjelle et al. Case- 20 workers of industrial Outcome: Physician evaluated all 6 with right No Cases had Participation rate: Not reported. 1981 control plant consecutively seen at patients with acute non-traumatic shoulder Controls significantly health clinic with acute, shoulder-neck pains referred to tendinitis: with longer Video analyses were done blinded to the outpaient clinic of the nontraumatic shoulder-neck rheumatology department. Each 46% tendinitis: duration and case status. pain. patient had to undergo an 0% higher extensive clinical examination, frequency of No significant difference between Of these, 13 were not due including local anaesthesia for the abduction or cases and controls in anthropometry. to causative disease or definition of pain location. forward malformation. These Exploratory puncture of the flexion than Isometric strength test: controls 13 were compared to glenohumeral joint was performed controls, significantly stronger in 6 of 14 tests in patients with tenderness over 26 controls, matched on the joint. p<0.001. but probably influenced by pain age, gender and place of inhibition in cases. work. Exposure: Anthropometric and Cases had Isometric muscle strength were significantly No significant difference in cycle time tested with strain gauge higher (9 vs. 12 min) between cases and instruments. Patients asked to shoulder controls. perform their max-mal efforts. Measurements made for the loads than following contractions: shoulder controls. The supraspinatus muscle showed a elevation at the acromion, significant change of the mean power abduction and forward flexion of frequency (p<0.05) towards lower the shoulder joints at neutral Median levels, indicating a fatiguing process position and semipronated. Grip number of for four of the five investigated strength measured by sick-leave assemblers during work. vigorimeter. days Video recording of arm significantly movements at work. Shoulder different loads estimated from videos. between Consisted of measuring the cases and duration and frequency of controls shoulder abduction or forward flexion of >60°. (p<0.01). EMG measurement of shoulder load during assembly work on 3 patients and 2 healthy volunteers. Muscular load level determination made by computer analysis of myo-electric amplitude.

(Continued)

3-41

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Burdorf and Cross- 194 riveters exposed to Outcome: Standardized Nordic 31% 20% 1.5 Participation rate: Riveters=76%, Monster sectional vibration compared to 194 questionnaire, pain or stiffness. controls=64%. 1991 workers in the same plant with little or no exposure to Exposure: Employed >12 months, Examiners blinded to exposure or vibration. not exposed to hand/arm case status: Not reported. vibration. Confounders controlled for included Observation, time-work studies, height, weight, and smoking habits. measurements of vibrating tools. Age and height significantly different No shoulder measurements. between groups.

Occupational history treated as Years of riveting work associated dichotomous variable with “1” for with pain or stiffness in shoulder heavy physical work. (0.05#p#0.10).

Follow-up of nonrespondants showed no difference in age or work experience. Sick leave significantly different.

(Continued)

3-42

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Burt et al. Cross- 836 Newspaper employees Outcome: Based on symptom Time spent Participation rate: 81%. (Authors 1990 sectional in the Editorial Department questionnaire. Case defined as typing: 50% 42% Õ Õ note that those out on assignment or ill and selected jobs in the pain, aching, stiffness, burning, or on vacation counted as non- Advertising, Circulation, numbness or tingling in shoulder Typing Typing Speed: participants.) Data Processing, and lasting >1 week or occurring one Speed: Finance Departments from time/month in the past year. Slow: 6% Moderate: 2.6 Number of workers in number of non- 4 company locations, Symptoms must have begun on Moderate: Fast: 4.1 typing jobs not reported. (460 female and 376 male). current job; no previous accident 11% 1.1-5.9 or acute injury to the joint, no Fast: 15% Pre-existing 1.8-9.4 Reporters characterized by high Cases compared to non- related systemic disease. Arthritis: periodic demands (deadlines) cases. OR=2.3 although they had high control and job Exposure: Based on satisfaction. questionnaire and job sampling. Dissatisfied 1.2-4.4 Exposure variables included work with job: Job analysis found significant time spent typing on computer; OR=2.3 correlation (r=0.56) between reported typing speed; keyboard type; hr average typing time/day and observed worked/week; workload; number 1.2-4.3 8 hr period of typing (p<0.0001). of years worked. Length of employment and symptoms in shoulder not significant.

(Continued)

3-43

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Chiang et al. Cross- 207 fish processing Outcome: Shoulder girdle pain as Prevalence Prevalenc Repetitive Participation rate: Not quantified; 1993 sectional workers, 67 males and defined by Anderson (1984) (the of Physician- e of movement of however, authors stated that “all of 140 females, divided in painful condition of the shoulder observed Physician- the upper limb the workers who entered the fish 3 groups: with limitation of movement, Disorders: observed (Rep): processing industry before June 1990 which may occur in association Disorders: OR=1.6 1.1-2.5 and were employed there full-time (I) Low force, low with tension neck or merge with were part of the cohort.” Of the repetition (comparison pain in the suprascapular or Group II: 37% Group I: Sustained 232 employees who agreed to group, n=61); upper dorsal regions). Symptoms (male 31%; 10% forceful participate, 207 met study criteria. in these regions occurring in last female 39%) (male 9% movement of Examiners blinded to exposure status. (II) High force or high 30 days and physical exam female the upper limb (“Workers examined in random repetition (n=118); findings of $two tender points or Group III: 10%) (force): sequence to prevent observer bias.”) palpable hardenings which may 50% OR=1.8 1.2-2.5 Workers with hypertension, diabetes, (III) High force and high either be caused or aggravated (male 50% history of traumatic injuries to upper repetition (n=28). by work conditions. female 50%) Rep times limbs, arthritis, collagen disease force: excluded from study group. Exposure: Assessed by OR=1.4 1.0-2.0 Eight plants used in study. Authors observation and recording of reported “no plant effect". tasks and biomechanical Age: Case definition based on physician movements of three workers OR=1.0 0.9-1.1 diagnosis not significantly different each representing one of 3 study from definition based on symptoms in groups. Highly repetitive jobs Gender: Groups II : 37% vs. 44% or Group III: with cycle time k=<30 sec or OR=1.1 0.7-1.7 50% vs. 50%. Group I about 2/3 the >50% of cycle time performing prevalence (10% vs. 15%). the same fundamental cycles. Dose-response for physician Hand force estimate from EMG observed shoulder girdle pain among recordings of forearm flexor three exposure groups. muscles. Classification of Dose-response for physician workers into 3 groups according observed shoulder girdle pain by to the ergonomic risks of the gender in three exposure groups. shoulders and upper limbs: Group Logistic model controlled for age and I: Low repetition and low force; gender. Group II: Low repetition or low Significant trend found for duration of force; Group II: High repetition employment and exposure group in and high force. workers <12 months, 12 to 60 months, but not in workers employed >60 months.

(Continued)

3-44

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments English Case- Cases: n=580; 174 males Outcome: Based on standard Frequency of Per 5 years of Participation rate: 96%. et al. 1995 control and 406 females with diagnosis for rotator cuff injury; shoulder age: 1.4 1.2-1.5, diagnosed soft tissue rupture of the long head of problems p<0.01 Administered questionnaire blinded to conditions of the upper limb biceps, shoulder capsulitis, case status. at 2 orthopedic clinics; symptomatic acromioclavicular Rotator cuff: For elbow ages 16 to 65 years. arthritis. 8.3% Õ flexion: 0.4 0.2-0.8, Controlled for age, height, gender, p<0.01 weight, whether MSD was due to an Controls: 996 controls; Exposure: Based on self- Rupture of Per hr of total accident, study center. 558 males and 438 females reported risk factors at work for long head of daily elbow attending the same clinics; musculoskeletal disorders biceps: 0.3% flexion: 1.1 0.9-1.2, Total daily exposure to elbow flexion diagnosed with conditions concentrating on detailed Õ p<0.01 did not contribute to shoulder injury. other than diseases of the components of movements and Shoulder upper limb, cervical or activities at work: awkward capsulitis: Repeated Risks highest for female hairdressers. thoracic spine; ages 16 to postures, grip types, wrist 3.6% shoulder 65 years. motions, lifting, shoulder Õ rotation with Not “Repetitive” defined as a frequency of postures, static postures, hand Symptomatic elevated arm: reported >once/min of 14 specific movements. tool use, and job category. acromiocla- RR=2.3 p<0.05 vicular Sporting activities, hobbies; average Questionnaire obtained arthritis: Wrist rotation Not hr of driving/week; whether claim for information on repetitive 0.2% at low rates: reported compensation made were analyzed in movements of the upper limb: Õ RR=0.18 p<0.05 models. Shoulder flexion, shoulder rotation with elevated arm, Wrist rotation Jobs with pinching between thumb keeping the whole arm raised >1 with and forefinger protective against min, shoulder rotation with elbow increasing shoulder disorders. May reflect hand flexed. rates: Not movement and exertion with no RR=2.02/30 reported shoulder movement or exertion. reps/min. p<0.05 Small number of subjects/group limits power to detect significant differences.

(Continued)

3-45

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Flodmark Cross- 58 industrial workers Outcome: Questionnaire survey Symptoms in Symptoms Participation rate: 87%. and Aase sectional making ventilation shafts using Nordic questionnaire for past 12 in past 12 1992 (51 males and 7 females) symptoms as to duration during months: 40% months: Aim of the study was to further compared to symptom last 12 months and during last 23% 2.2 1.4-4.4 investigate relationship between Type prevalence in 170 blue- 7 days, effect on work A behavior and musculoskeletal collar workers in Örebro, performance and leisure symptoms. Sweden. activities, and sick leave. Type A behavior assessed by Bortner The Bortner Score for Type A Compared workers with questionnaire. behavior significantly higher for those symptoms to those with shoulder symptoms than those workers without symptoms Exposure: No objective without. for risk factor analysis. measurements. No difference in headache, tiredness, sleeping, irritation, lack of concentration or problems with eyes, nose, stomach, skin.

Authors suggest that Type A persons more likely to ignore symptoms to minimize their potential effect on work capacity.

(Continued)

3-46

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hales and Cross- Of 96 female workers Outcome: By questionnaire: Any 43% (low Participation rate: 91%. Fine 1989 sectional employed in 7 high Period Prevalence: Symptoms in symptom of exposure exposure jobs in poultry last 12 months. Case defined as: the shoulder: group) 1.2 0.7-2.0 Examiner blinded to case and processing, 89 were pain, aching, stiffness, 49% (high exposure status. compared to 23 of numbness, tingling or burning in exposure 25 female workers in low the shoulder, and symptoms group) Analysis adjusted for age and exposure jobs. began after employment at the duration of employment. plant; were not due to a previous Period injury or trauma to the joint; lasted prevalence Although shoulder MSDs surveyed by >8 hr; and, occurred 4 or more for shoulder questionnaire, exposure assessment times in the past year. case: 19% 4% 3.8 0.6-22.8 was based on hand/wrist exposure, so that risk for shoulder may not be Point Prevalence: Determined by Point accurate. physical exam of the upper prevalence extremity using standard for shoulder High exposure departments: Breast diagnostic criteria case must also case: 7% 4% 0.9 0.1-7.3 trim, thigh debone, leg cut/disjoint, fulfill symptom definition (listed tender cut, knuckle cut, breast, above). knuckle cut, thigh fat trim.

Exposure: Observation and Lower exposure departments: Breast, walk-through; jobs categorized thigh, or quality control inspectors. as High exposure and Low exposure based on estimated hand force and hand repetition, not shoulder exposure.

(Continued)

3-47

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Hales et al. Cross- 533 Telecommunication Outcome: Self-administered Rotator cuff Fear of Participation rate: 93%. 1994 sectional workers (416 females and questionnaire and standard tendinitis: replacement 117 males) in 3 offices, physical examination; case 6% (n=513) by computers: Physician examiner blinded to worker employed $6 months. defined as: pain, aching, 1.5 case study. stiffness, burning, numbness or Bicipital 1.1-2.0 "Cases" fulfilling shoulder tingling >1 week or >12 times a tendinitis: Number of Logistic analysis adjusted for WRMSD definition year; no previous traumatic injury less than 1% times arising demographics, work practices, work compared to non-cases. to the area; occurring after (n=516) from chair: organization, individual factors; employment on current job within 1.9 1.2-3.2 electronic performance monitoring; the last year and positive physical Overall DAO keystrokes; Denver DAO exam: moderate to worst pain shoulder: keystrokes/day. experienced with positive 6% physical finding of the ORs for psychosocial variables symptomatic joint. represent risk at scores one standard deviation above mean score Exposure: Work practices and compared to risk at scores one SD work organization assessed by below mean. questionnaire and observation; number of keystrokes/day. Because of readjustments and changes of workstations during study Physical workstation and postural period, measurements of VDT measurements obtained but not workstations considered unreliable used in final analyses. and excluded from analyses.

Number of hr spent in hobbies and recreational activities not significant.

Although keystrokes/day was found to not be significant, data available was for workers typing an average of 8 words/min over 8-hr period.

97% of participants used VDT $6 hr/day, so not enough variance to evaluate hr of typing.

Over 70 variables analyzed in models may have multiple comparison bias.

(Continued)

3-48

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Herberts Cross- 131 male shipyard welders Outcome: Positive answers to Supraspi- Shoulder Prevalence Participation rate: Not reported. et al. 1981 sectional with >5 years of work questions about repeated natus Pain rate ratio experience compared to 57 occurrences of shoulder pain tendinitis Prevalenc (PRR) of Incidence estimated to be 15 to 20% a male office clerks. All during work; shoulder stiffness (ST) results e from shoulder pain year. workers participated in the that affected work and of 23 question- results from shipyard’s medical program weakness in shoulder that welders naire: questionnaire, Welders with and without tendinitis which offered medical affected work or weakness or called back 1.8% welders vs. were age-matched. exams every 5 years. numbness in arm or hand and for clinical office participation in a follow up exam. follow-up workers: We question the methods used to exams: PRR=15.2 2.1-108 approximate the prevalence of Clinical examination with joint 16 welders (90% CI) shoulder tendinitis. Authors stated range of motion, active and had PRR from that they took into account the missing passive and simultaneous pain supraspi- estimated data in the investigation and assumed analysis, rating of gross power in natus prevalence that the drop-out group did not deviate flexion, abduction and rotation, tendinitis. (“propor- from the examined group, so they rating of tenderness to palpation. tionation” of used “proportionation” to obtain the cases) number of cases of supraspinatus Exposure: Estimation of workload Shoulder reported in tendinitis cases in the welders for with assessment of the Pain reports article: calculations of prevalence rate ratios; workplace into 3 groups very from the PRR=18.3 14.7-22.1 number of supraspinatus tendinitis high, high or low. Static loading question- (90% CI) cases increased from 16 to 24. while holding tools; awkward naire: 27% postures; shoulder level or Number of years active welding, overhead work. shoulder load, and welding years showed no significant difference. However, a sample size of 11 matched pairs may not have enough power to detect a difference.

Turnover of shipyard welders mentioned at 33%.

Shoulder tendinitis was not found to be associated with increasing age.

(Continued)

3-49

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Herberts Cross- 131 male shipyard welders Nurse-administered symptom Question- Question- Participation rate: Not reported. et al. 1984 sectional and 188 plate workers questionnaire: Case defined as naire results, naire compared to 57 male office pain, weakness, stiffness in shoulder pain results, Not mentioned whether examiners clerks. Welders and shoulder excluding effects of the supra- shoulder blinded to case or exposure status. plateworkers chosen had originating from neck, plus clinical spinatus pain of the >5 years of job experience. exam with tenderness, range of tendinitis supraspi- Controls were matched for age and motion gross power measured by type natus gender. 23 symptomatic welders, dynamometer. Welders: tendinitis 27% type: PRR=18.3 13.7-22.1 30 symptomatic plate Plateworkers with shoulder pain Office (90% CI) workers compared to 18 Exposure: Observation of jobs; Plate- averaged 6 years older than welders workers: worker: asymptomatic welders and workers compared by use of job 2% PRR=16.2 10.9- with shoulder pain. 30 plate workers by clinical title; EMG measurements of 32% 21.5 exam. muscles of shoulder region. Supraspi- (90% CI) EMG analysis using fine monopolar natus wire electrodes showed that in work Age-matched pairs: Electromyographic analysis of the tendinitis where the hand was positioned 11 welders; shoulder muscle load completed results of 23 overhead, the intramuscular pressure 15 plateworkers. on 9 volunteers to study the welders in the supraspinatus muscle had influence of hand tool mass and called back extremely high pressure levels arm posture. for clinical compared to pressure levels in other follow-up skeletal muscles. exams: 16 welders Turnover rate of welders was 30%; had supra- may be explanation for lack of spinatus association with duration. tendinitis Supraspi- Welding seen as static work; natus plateworking dynamic work. tendinitis results of 30 plate- workers called back for clinical follow-up exams: 15 plateworkers had supra- spinatus tendinitis

(Continued)

3-50

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Hoekstra Cross- 108 of 114 teleservice Outcome: Self administered Center A: Õ Õ Participation rate: 95%. et al. sectional representatives working at questionnaire. Case defined as 13% 1994 2 Government the presence of pain, numbness, Representatives perceived little administration centers: A tingling, aching, stiffness or Center B: control over actions of others; little and B. burning in the shoulder, and no 44% 4.0 1.2-13.1 participation in decision making; little previous injury; symptoms began freedom to regulate own activities. after starting the job; lasting >1 Non-optimally week or occurred once a month adjusted Perception that workload was high within the past year; reported as desk height and variable. “moderate” or greater on a 5-point work scale. 5.1 1.7-15.5 Analysis controlled for gender and Non-optimally location and interactions checked. Exposure: Observation of work adjusted stations, measurement and screen Variables considered in logistic model evaluation of work station; included location, age, seniority, hr observation of postures. 3.9 1.4-11.5 spent typing at VDT, hr on the phone, 3 chair variables, and perceived adequacy of: (1) chair adjustment, VDT screen, (2) keyboard adjustment, VDT screen, (3) desk adjustment; job control, workload variability.

Center B location had nonadjustable work stations and mostly nonadjustable chairs causing elevated arms, hunched shoulders and other undesirable postures.

Linear regression also performed on psychosocial variables in separate models for health outcomes of job dissatisfaction and mental and physical exhaustion (not for shoulder MSDs).

Did not include non-work-related variables in analyses.

(Continued)

3-51

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Hughes et Cross- 104 male aluminum smelter Outcome: Symptoms occurring in 14.9% with Õ Model based Participation rate: carbon al. 1997 sectional workers: 62 carbon the shoulder >once per month or positive on MSD setters: 65%; crane operators: 56%; setters, 36 crand lasting longer than 1 week in the symptoms defined by carbon plant: 33%. operators, 9 carbon plant previous year, no acute or and physical symptoms workers. There were 14 traumatic onset; occurrence exam. and physical Examiners blinded to exposure and workers who were not since working at the plant, no exam health status: Not reported. from selected jobs and systemic disease. Physical 24% had Age: OR=0.93 0.8-1.0 were excluded. examination: Active, passive, symptoms in Good health: Analysis controlled for age, smoking and resisted motions, pinch and the elbow- OR=0.35 0.1-0.87 status, sports and/or hobbies. grip strength, 128 Hz vibration forearm in sensitivity, two-point the previous Low decision Psychosocial data collected discrimination. Psychosocial week. latitude: individually; physical factors based on scales from questionnaire based OR=4.0 0.8-19 estimates of each job. on Theorell and Karasek Job Years of Stress Questionnaire, and on forearm twist: Job risk factors entered into the model Work Apgar questionnaire used. OR=46 for hand/wrist included (1) the 3.8-550 number of years of handling >2.7 Exposure: For carbon setters Model based kgs./hand, (2) push/pull, (3) lift/carry, and crane operators (non- on MSD (4) pinching, (5) wrist repetitive jobs) and modified job- defined by flexion/extension, 60 ulnar deviation, surveillance checklist method symptoms and (7) forearm twisting. was used. Job task analysis Age: OR=0.96 used a formula based on the 0.8-0.98 Health interview included information relative frequency of occurrence Smoker: about metabolic diseases, acute of posture during tasks. OR=0.41 0.1-1.4 traumatic injuries, smoking, hobbies. Low decision latitude: Low participation rate limits OR=4.5 1.3-16 interpretation. High Job demand: OR=3.0 0.7-13 Years forearm twist: 92 7.3-4

(Continued)

3-52

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ignatius Cross- 1,917 of 3,248 male postal Outcome: history of symptoms Recurrent Participation rate: 59% et al. 1993 sectional employees completed an and severity of recurrent joint joint pain: interviewer-administered pain as defined by Wells et al. 55.1% 38.4% 1.8 1.5 -2.2 Severe shoulder pain associated with questionnaire; 1,081 were [1983]. age, work experience, bag weight letter delivery postmen Severe joint and walking time. compared to 836 other Exposure: work factors related to pain: 12.0% 6.2% 2.2 1.5-3.1 postal workers. weight of letter bags, distance Bags usually carried on one shoulder. walked each day, use of transporting tools.

Postmen carry/day an average load of 45 lbs; walked 4.5 km plus 1,300 steps for 3.7 hr/day.

(Continued)

3-53

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Jonsson Prospec- Electronics Workers Outcome: Three separate Severe At 3rd exam Participation rate: 72% of original et al. 1988 tive (n=69 females) out of initial physical exams at yearly intervals shoulder during 3rd group had 3 exams one year apart. 96 workers. (one initially) assessing disorders: year of longi- 80% had 1st and 3rd year exams. tenderness on palpation, pain or tudinal study: Questionnaire included spare time (See Kilbom et al. 1986 restriction with active and 22% at 2nd Initially: 38 subjects physical activity, hobbies, perceived for initial study.) passive movements; symptoms in exam 11% of reallocated to psychological stress at work, work previous 12 months with regard subjects varied tasks satisfaction, number of breaks, rest pauses. to character, frequency, duration, had had improved localization, and relation to work shoulder (16% of these Most of physiologic and ergonomic or other physical activities. MSDs had severe evaluations conducted only at outset Analyzed if score on any symptoms of study. symptom of $2, on a 4 point initially) Low muscle strength not a risk factor scale; “severe” symptom score significance for subsequent symptoms. equals 4. at p<0.05 Relative time spent with shoulder elevated negatively related to Exposure: Carried out at outset Those with “remaining healthy” after both 1 and 2 of study: Maximum voluntary After 1 year; 20% with unchanged years. isometric contraction (MVC) of 24% unchan- working tasks Muscular strength and endurance not forearm flexors, shoulder ged deteriorated related to improvement nor remaining strength, handgrip, heart rate working further (26%). healthy. using a bicycle ergometer and conditions At 2nd and 3rd examination, there rating of perceived exertion. was a strong negative relationship Videotaping performed for the between “remaining healthy” and analysis of working postures and satisfaction with colleagues. movements. Predictors of remaining healthy were work without elevating the shoulders Reallocation tasks: and satisfaction with work tasks. Non sitting; no inspection of small No mention of examiner being blinded details on printed circuit boards; to case status. standing and walking, Predictors of deterioration were occasionally sitting; caretaker previously physically heavy jobs, high work; surveillance of machinery; productivity (after 1 year), and and assembling bigger and previous sick leave. heavier equipment. Predictors of improvement were reallocation, physical activity in spare time, and high productivity (after 2 years).

(Continued)

3-54

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kiken et al. Cross- 294 Poultry Processors at Outcome: Period prevalence Plant #1: Participation rate: 98%. 1990 sectional 2 plants. symptom in last 12 months by Any Plant #1=174 questionnaire. Case: Pain, symptom for Examiners blinded to case and Plant #2=120 aching, stiffness, burning, shoulder exposure status. numbness or tingling in the case: 46% 28% 1.6 0.9-2.9 shoulder, began after employment Analysis stratified for gender and at the plant; not due to previous Period age. accident or injury outside work; prevalence: lasted >8 hr and occurred 4 or 13% 3% 4.0 0.6-29 Higher exposure jobs (HE) were more times in the past year. located in the receiving, evisceration, Point whole bird grading, cut up and Point prevalence determined by prevalence deboning departments. Lower physical exam. Rotator cuff for shoulder exposure jobs (LE) were located in defined as pain $3 on a 0 to 8 case: 3% 0% Indeterminate Õ the maintenance, sanitation, quality scale on active and resisted assurance and clerical departments. shoulder abduction. Case must Plant #2: fulfill symptom definition (listed Any 30% of workers involved in a job above). symptom for rotation program may have influenced shoulder associations made. Exposure: Determined by case: 50% 30% 1.7 0.8 -3.3 observation; level of exposure Annual turnover rate close to 50% at was based on exposure to Period plant 1 and 70% at plant 2 making repetitive and forceful hand prevalence: survivor bias a strong possibility -- motions, not shoulder. 14% 5% 2.8 0.4-19.6 leading to underestimation of associations. Exposure measurements Point estimated for the hand and prevalence wrist region and NOT the for shoulder shoulder area. case: 3% 0% Indeterminate Õ

(Continued)

3-55

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kilbom et al. Cross- 106 of 138 female Outcome: Three separate MSD Logistic Participation rate: 77%. 1986 sectional assemblers in two physical exams at yearly intervals symptoms in Regression electronic manufacturing (one initially) assessing the shoulder model (all See Jonsson et al. 1988 for follow-up. companies agreed to tenderness on palpation, pain or using a four variables participate; 10 excluded restriction with active and point severity significant at No relation between maximal static because of symptoms in passive movements; symptoms in scale: the p<0.05 strength and symptoms. past 12 months. previous 12 months with regard level). 96 underwent medical, to character, frequency, duration, None: 84% Examiner blinded to case status. physiological, and localization, and relation to work Shorter ergonomic evaluation. or other physical activities. Slight: 5% stature Questions included spare time Analyzed if score on any physical activities, hobbies, perceived (See Jonsson et al. 1988, symptom of $2, on a 4 point Moderate: Years of psychosocial stress at work, work earlier in this table, for scale; “severe” symptom score 7% employment in satisfaction, number of breaks, rest follow-up.) equals 4. electronics. pauses. Severe: 3% Exposure: Carried out at outset Fewer total 59% had no symptoms or only slight of study: Maximum voluntary number of ones. There were no cases of isometric contraction (MVC) of upper arm shoulder tendinitis. forearm flexors, shoulder flexions/hr. strength, handgrip, heart rate Age showed a weak positive using a bicycle ergometer and Greater correlation. rating of perceived exertion. percentage of Videotaping during the work cycle Years of employment, productivity, representative part of working time with muscle strength were not related to day from rear and side. Upper upper arm symptoms. arm studied at rest and in 0 to abducted 0 to 30E, 30 to 60E, 60 to 90E, in 30E. There was large inter-worker extension and >90E abduction. variation in working posture and The shoulder recorded as resting working techniques. or elevated; also frequency of changes in posture between The authors followed up on the non- different angular sectors/hr, participants and found no significant duration of postures. Work cycle differences from participants. time and number of cycles/hr, time at rest for arm, shoulder, The more dynamic working technique, head. the less symptoms.

(Continued)

3-56

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kvarnström Cross- 112 cases of prolonged Outcome: Shoulder cases Die casting Participation rate: Not reported. 1983b sectional shoulder disorders fulfilled the following: symptoms machine identified in a workplace of from shoulder was the main operators Examiners not blinded to exposure, and 11,000 employees. The reason for inability to work, off (involved but selection based on diagnosis of total number of employees work longer than 4 weeks, heavy work shoulder MSD. Case- was approximately half fatigue in one of both shoulders, with repetitive All 112 shoulder disorders occurred in control factory workers and half pain in shoulder brought on by movements of laborers; none in office workers. office workers. work and aching at rest were the RR for Swedish workers: 0.46; RR for present, and Clinical examination shoulders): immigrants: 3.1. demonstrated tenderness of the RR=5.4 All cases except one were paid piece Case more than control shoulder muscles, especially rate. study: muscularis trapezius, levator Plastic “Young persons significantly less ill scapulae, and/or infraspinatus workers: than middled-aged.” Controls chosen at random and/or tenderness at the tendon RR=2.2 The following questionnaire from factory workers, insertions of the rotator cuff responses were significantly different matched for age and muscles. Spray between cases and controls: Group gender. painters: piece rate, shift work, heavy work, Muscle strength in shoulder RR=3.7 monotonous, stressful, detrimental to assessed with regards to four health, heavy lifting, and unsuitable functions Surface working conditions. 9 cases and 1 treatment control cited poor relationship with Exposure: (1) Information operators: supervisor. obtained through interview: RR=4.7 No difference in environmental organization of work, physical condition, job content. work load, physical environment, Assembly line Cases more likely to be married, have psychosocial work environment, workers: ill spouses, have children at home, social and ethnic conditions, RR=5.2 work alternating shifts than controls. (2) detailed work history. Factors Ergonomic Work history showed no difference 0,1, or 2 given to different types experts’ between points for cases and of work depending on the evaluation: controls (see exposure column). workload borne by the shoulder. cases had Muscle strength bilaterally significantly This factor multiplied by number significantly lower in cases in four functions. of years spent at job, and more mono- products were added, (3) 2 tonous and company engineers graded the repetitive degree of monotony and work than repetitiveness in each job held by controls. cases and controls.

(Continued)

3-57

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments McCormack Cross- Manufacturing workers: Outcome: Questionnaire and Packaging/ Participation rate: 91%. et al. 1990 sectional packaging or folding physical examination initially by folding non-office workers (41 males, 328 nurse screening; if employee workers: workers: Examiners not blinded to exposure females); sewing workers answered affirmative to question 2.7% 2.1% 1.3 0.5-3.8 status (information obtained from (28 males, 534 females); regarding symptoms in upper personal communication). boarding workers (19 extremity and/or had any positive Sewing males, 277 females) physical findings, then had workers: 11 Physician examiners; inter- compared to knitting physician examination. The term 2.5% 2.1% 1.2 0.5-2.7 examiner potential problem workers (203 males, 149 "shoulder condition" used to acknowledged by authors. females); non-office define abnormalities of shoulder; Boarding workers (204 males, 264 consisted of bursitis, bicipital workers: Questionnaire asked types of jobs, females) compared with tendinitis and impingement 2.4% 2.1% 1.1 0.4-2.9 length of time on job, production rate, knitting workers syndrome. nature and type of upper extremity (203 males, 149 females). Knitting complaint and general health history. These groups were Exposure: Based on observation workers: compared to a referent of job activities; only the boarding 1.1% 2.1% 1.3 0.5-3.1 Age, sex, race, job category and group consisting of non- workers had activities requiring years of employment not statistically office workers maintaining reaching overhead (from significant with "shoulder conditions." machinery, involved in personal communication with first transportation, or worked author). Patients with objective diagnostic as cleaners and sweepers. shoulder findings: Of 45 cases None of the referent group diagnosed: 25 graded as “mild”, used rapid repetitive 19 graded as “moderate; 1 graded as movements comparable to severe. the employees in the other job categories. 21, 25 and 36 operators from each group and 25 of 55 auxiliary nurses and home helpers (controls) participated in the study.

(Continued)

3-58

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Milerad and Cross- 99 Dentists randomly Outcome: Based on telephone Male: 36% Participation rate: 99%. Ekenvall sectional selected from Stockholm questionnaire: Shoulder Female: 67% 15% 2.4 1.0 -5.4 1990 dentist registry who symptoms at any time before the 28% 2.4 1.5-3.7 Stratified analysis by gender. practiced $10 years interview "lifetime prevalence." Neck and compared to Further analyzed according to shoulder: No difference in leisure time 100 pharmacists selected Nordic questionnaire as to 36% exposure, smoking, systemic disease, from all pharmacists in duration during last 12 months 17% 2.1 1.3-3.0 exposure to vibration. Stockholm. and during last 7 days, effect on Neck and work performance and leisure shoulder and Symptoms increased with age in activities, and sick leave. upper arm: female dentists only. 16% Exposure: Questionnaire 3% 5.4 1.6-17.9 Duration of employment highly included: (1) abduction of arm, correlated with age (r=0.84, 0.89). particularly in sit-down dentistry, (2) static postures, (3) work No relation between symptoms and hr/day. duration of employment.

Equal problems dominant and nondominant sides.

(Continued)

3-59

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohara et al. Cross- For cross-sectional study: Outcome: Assessed by standard Shoulder Shoulder Participation rate: for prospective 1976 sectional 399 cash register health inventory and medical stiffness: stiffness : study = 100%. and Pro- operators compared with examination (used clinical spective 99 office machine classification according to the Cashiers: Office Participation rate: for cross-sectional operators and 410 other committee on cervicobrachial 81% Workers: study, not reported. workers (clerks and disorders of the Japan 72% 1.7 1.0-2.8 saleswomen). All female. Association of Industrial Health, in Unknown whether examiners blinded Table 3 in the paper). Shoulder Shoulder to case status. For prospective study: dullness dullness 56 workers employed <7 Periodic physical exam performed and pain: and pain: Interventions did not reduce months had testing pre- twice a year from 1973. Primary complaints in the shoulder region, but and post-intervention using exams performed on 371 Cashiers: Other did improve symptoms in the arms, questionnaire and physical operators. 130 (35%) received 49% workers: hands, fingers, low back, and legs. exam. detailed exams. 68% 2.0 1.4-2.8 The lack of improvement in the shoulder region was stated to be due 86 operators, newly hired Exposure: To repetitive to the use of the same narrow check after interventions, also movements relocating Office stands, unsuitable counter height, and had evaluation after merchandise across counter and workers: necessity of continuous lifting of the 10 months of working. bagging, involved muscle activity 30% 2.2 1.4-3.5 upper limbs. of the fingers, hands, and arms; extreme and sustained postures. Operators hired after the interventions and then examined after 10 months Interventions: (1) a 2-operator had less Grade I, II , or III occupational system, 1 working the register, cervicobrachial disorders in one packing articles, changing examination than those hired before roles every hr; (2) continuous intervention. operating time <60 min; max. working hr/day 4.5 hr; Only 14.5% with >3 years (3) 15- min resting period every employment at worksite. hr; (4) electronic cash registers with light touch keyboard Narrow work space and counter substituted for half of previously height not adjusted for height of used mechanical cash registers. worker.

(Continued)

3-60

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson Cross- Electrical equipment and Outcome: Based on Shoulder Participation rate: Not reported. et al. 1989 sectional automobile assemblers questionnaire: Any shoulder pain in (n=148), former female pain, shoulder pain affecting previous 12 Significant association for shoulder assembly workers who work ability, and shoulder pain in months: 55% symptoms and medium and fast pace quit within 4 years (n=76) the last 7 days. 45% 2.0 1.1-4.0 compared to slow pace but not very compared to randomly Shoulder fast pace. sampled females from Exposure: Based on job pain in general population (n=60). category. previous 7 Significant association with duration days: 38% of employment (p=0.03), but much 18% 3.4 1.6-7.1 stronger for workers <35 years than Work in workers >35 years. auxiliary previous 12 Significant interaction between age months: 21% and employment.

10% 2.4 1.0-5.8 Older females employed for shorter periods had more symptoms than younger ones.

(Continued)

3-61

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson Cross- Exposed Group: 206 of Outcome: Defined by criteria Frozen Participation rate: 83%. et al. 1994 sectional 247 females working in from questionnaire and physical shoulder: 2% No exposure information available to 13 fish processing plants examination: standard diagnosis 0.5% 4.1 0.5-37 examiners, however, it was not participated. of frozen shoulder, Supraspi- possible to completely blind the natus supraspinatus tendinitis, study/referent group status. tendinitis: All activities (trimming of cod, packing 322 females who left infraspinatus tendinitis, bicipital 15% employment in the fish tendinitis acromioclavicular 5% 3.4 1.6-7.2 fish and herring filleting) were found processing industry in the syndrome. to be highly repetitive with poor Infraspinatus working postures and fast 10 years prior to the study. tendinitis: movements by standardized Exposure: Assessed by 12% “ergonomic workplace analysis” Comparison group: All 208 questionnaire (length of 3% 4.7 1.4-15.2 (EWA) methods; very few pauses in females employed in the employment; psychosocial Bicipital the work cycle; tasks not varied. same towns as the factors, physical factors) and by tendinitis: 10% Sports activities were highly exposed; 71 were observational methods 4% 2.4 1.1-5.4 associated with shoulder tendinitis employed in day nurseries; (Ergonomic Workplace Analysis) Acromiocla- (OR=4, 9) in multiple logistic 92 in offices; 42 caretakers and NIOSH guidelines for lifting. vicular regression analysis. of elderly; 3 gardeners. Analyzed 10 items: work site, syndrome: In the control group, prevalences of general physical activity, lifting, 17% upper limb disorders increased work postures and movements, 6% 3.1 1.6-6.0 substantially with age. Among the exposed, the prevalence remained job content, job restrictiveness, almost constant with age. worker communication, difficulty of decision making, repetitiveness Excess prevalence for exposed females most pronounced for females of the work, and attentiveness. PRR of <45 years. There was a pronounced shoulder dose-response for disorders of the 74 workers videotaped $10 min. disorders: neck or shoulders vs. duration of from the back and sides. 2.95 2.2-4.0 exposure in the industry. No such Average counts of two PRR for associations seen in group >45 years. independent readers for Authors explained as perhaps due to suprapi- the “healthy worker effect,” but, it frequencies, duration and critical natus, would be more accurate to describe it angles of movement used. infraspinatus as “survivor bias.” and bicipial tendinitis: 3.03 Psychosocial work environment, 2.0-4.6 stress and worry factors, tendencies PRR for towards muscular tension differed suprapinatus significantly between exposed and and controls. infraspinatus tendinitis alone: 3.5 2.0-5.9

(Continued)

3-62

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson Cross- Industrial Workers Outcome: Measured by physical 50% (n=82) 16% 5.0 2.2-11.0 Participation rate: current workers: et al. 1995 Sectional (n=82 females) exposed to exam and questionnaire. (n=64) 96%; past workers: 86%; repetitive tasks with short Frozen shoulder: Limited out- referents: 100%. cycles mostly far <30 sec, ward rotation and abduction. Employment Questionnaire included individual usually with a flexed neck duration: factors, work/environment, Infraspinatus, supraspinatus symptoms. and arms elevated and tendinitis: Local tenderness over <10 years No exposure information available to abducted intermittently; tender insertion, pain with (n=19): 53% 68 former workers (mean 9.6 2.8-33.0 examiners, however, it was not resisted abduction. possible to completely blind the employment time 21 years) Bicipital tendinitis: Pain with 10 to 19 study/referent group status. who had left the factory years resisted elevation of arm, resisted Psychosocial scales assessed: during the 7 years before flexion of elbow. (n=25): 48% control over one’s work, stimulation, the study; these workers Acromicoclavicular syndrome: psychological climate, work strain, were compared to 64 >20 years 4.4 1.5-13.0 Pain with horizontal adduction fellowship at work and social network referents with no repetitive (n=38): 50% at work. Age, stress/worry exposure at their current and/or outward rotation of arm. tendency, subjective muscular tension jobs (female residents of a Exposure: Videotaping and tendency, social network outside of nearby town currently observation. Analysis of 3.8 1.4-10.0 work, psychosomatic symptoms. employed as customer elevation of the arm: 0E, 30E, 60E, Age and employment status (repetitive service, ordering and price and for abduction 30E, 60E, 90E. vs. referent) controlled for in logistic model. marking in supermarkets, 74 workers videotaped $10 min. as office workers (no from back and sides. Average For continuous variables, OR are for constant computer work) counts of two inde-pendent 75th vs. 25th percentiles. or as kitchen workers. readers for frequencies, duration, Videotape analysis revealed and critical angles of movement considerable variation in posture even within groups performing similar used. assembling tasks. Repetitive industrial work tasks Logistic models replacing repetitive divided into 3 groups: (a) fairly work with videotape variables found mobile work, (b) assembling or muscular tension tendency and neck pressing items, and © sorting, flexion movements significantly polishing and packing items associated with neck/shoulder diagnoses. Weekly working time, work Significant association between time rotation, patterns of breaks, spent with upper arm abducted >60° individual performance rate (piece and neck/shoulder diagnoses. rate). Only exposure readings from right arm were used. Muscle strength (maximum voluntary capacity) measured by hand dynamometer at elevation,

(Continued)

3-63

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Onishi et al. Cross- Female industrial workers: Outcome: Based on Shoulder Participation rate: Not reported. 1976 sectional 42 reservationists; (1) symptoms of shoulder Tenderness: 95 fluorescent lamp stiffness, dullness, pain, Unknown whether examiners blinded assemblers; numbness; (2) pressure (<1.5 Reserva- Office to case status. 109 photographic film kv/cm2) measured by strain tionists: workers rollers; 46 teachers of transducer at which subject felt assemblers: (n=101): Body height, weight skin fold handicapped children; pain. (3) physical exam: range 70% 48% 1.1 0.6-1.9 thickness and muscle strength, grip 101 office workers. of motion, tests, nerve strength, obtained. compression tenderness. Film rollers: 84% 6.0 3.0-12.2 Body height and weight differences Exposure: Observation of job not significant. tasks, then job categorization. Teachers: 58% 1.6 0.7-3.3 Significant difference between body Reservations; Key 15,000 to fat in reservationists and office 20,000 strokes/day or more on Shoulder workers. busy days 2 to 3 times/week. Stiffness: Significant difference in grip strength Assemblers inspect lamps once Reservatio- in teachers and nurses compared every 3.5 to 4.5 sec; all work nists with office workers. 12 hr/day. (N=45): 56.6% 34.7% 2.5 1.1-5.6 Those with habitual shoulder stiffness Film rollers wind 1 roll of 35mm had lower threshold of local film every 2.5 to 5 sec over 7.5 Assemblers tenderness than those without hr/day. (N=94): stiffness. 66.6% 3.7 2.0-7.0 Prolonged contraction of No difference between workers with trapezius noted in 2 film rollers. Film Rollers tenderness threshold above (N=127): 1.5 Kb/cm2 and those below with Teachers and nurses daily care 59.1% 2.7 1.5-4.9 respect to age, height, weight, skin of disabled children e.g., lifting. fold thickness, grip strength, upper Teachers arm abduction strength, back muscle Office workers: Record keeping, (N=52): strength. copying, etc. 65.4% 2.1 0.9-4.6

(Continued)

3-64

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Punnett Cross- 162 female garment Outcome: Self-administered Garment Hospital Shoulder Participation rate: 97% (garment et al. 1985 sectional workers, 85% were questionnaire about pain and workers: employees MSDs in workers), 40% (hospital workers). employed as sewing standardized physical exam. 19.6% 8.8% Garment machine operators and workers vs. Analysis stratified for number of sewing and trimming by Cases defined as the presences Hospital years employed, decade of age, hand. of persistent shoulder pain employees: native language. (lasted for most days for one OR= 2.2 1.0-4.9 Comparison: 76 of 190 full month or more within the past Age and length of employment not a or part-time workers on year); were not associated with Shoulder predictor of risk of shoulder MSDs. day shift in a hospital who previous injury, and, began after MSDs in worked as nurses or aids; first employment in garment Straight stitch Prevalence of pain not associated lab techs or therapists; manufacturing or hospital workers vs. with years of employment in garment food service workers. employment. Key questions Hospital workers. based on the arthritis supplement employees: Employees typing >4 hr/day questionnaire of NHANES. OR=3.9 p#0.05 Non-English speakers significantly excluded from comparison less likely to report pain (RR 0.6 group. Exposure: Self-administered Shoulder p<0.05). questionnaire; number of years in MSDs in Top the industry, job category, stitch Native English speakers significantly previous work history. workers vs. older than non-native English Hospital speakers (p<0.03). employees OR=5.0 p#0.05 Logistic regression model found garment work and language significantly related to shoulder pain.

(Continued)

3-65

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Rossignol Cross- 191 computer and data Outcome: Self-administered 0.5 to 3 hr of Compari- Up to 3 hr of Participation rate: in six industry et al. 1987 sectional processing services, public questionnaire case defined as: VDT use/day son group VDT use groups 67 to 100%. utilities of Massachusetts “Almost always experienced” (n=31): 35% (with no compared to State Department, 28 of shoulder pain, stiffness or computer 0 hr of use. Participation rate: for individual whom did not use a soreness or missed work due to use) OR=2.5 0.7-10.8 clerical workers: 94 to 99%. keyboard with a VDT. shoulder pain, stiffness or 4 to 6 hr of (n=28): soreness. VDT use/day 18% “Assessed magnitude of confounding Centers selected at random (n=28): 48% 4 to 6 hr of by age, cigarette smoking, industry, from 38 work sites with Exposure: Self-reported number VDT use educational VDT training.” >50 employees. of hr/day working on a keyboard >7 hr of VDT compared to with a VDT. Subjects selected use/day 0 hr of use: The study was presented as “General after observation of work sites. (n=104): OR=4.0 1.0-16.9 health survey to avoid observation 51% bias.” >7 hr of VDT use compared to 0 hr of use: OR=4.8 1.6-17.2

(Continued)

3-66

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Sakakibara Cross- 48 Orchard workers Outcome: Shoulder pain Workers Workers Workers Participation rate: 77%. et al. 1987 sectional (20 males and 20 females). described as the presence of thinning bagging thinning pears stiffness and pain daily. pears apples: vs. workers Stratified by gender. Compared symptoms after (estimated 21% bagging completion of thinning of Exposure: Observation of jobs. from apples: General fatigue, gastric disturbances, pears, bagging of pears Angles of flexion of the shoulder histograms): OR=2.2 1.2-4.1 appetite loss and headache showed and bagging of apples on one subject were measured 46% no difference in frequency between (covering fruit with paper every 25 min. during a whole day tasks. bags while on the trees). doing each task. Workers Workers bagging pears Stiffness and pain in shoulders Internal comparison using Farmers worked approximately 8 bagging vs. bagging significantly higher from thinning and same study population. hr/day for 10.6 to 13.6 days each pears apples: bagging pears than apples which year bagging or thinning pears (estimated OR=1.4 0.7-2.8 authors attributed to working posture and bagging apples. Median from of elevated arms and neck extension. shoulder flexion was 110E to histograms): 119E for thinning pears and 29% Exposure data based on bagging pears; 30E bagging measurement of one worker may not apples. be generalized to others.

The proportion of workers with >90E forward shoulder flexion was significantly higher for thinning out pears and bagging pears than for bagging apples.

(Continued)

3-67

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Sakakibara Cross- Of 65 female Japanese Questionnaire: Stiffness and pain Pear bagging Apple Participation rate: 80%. et al. 1995 sectional farmers. 52 completed the in shoulder region. Symptoms in bagging questionnaire and physical past 12 months for $one day, or Examiners not blinded to case status exam in late June for symptoms in past 12 months for Muscle Muscle Workers due to design of study. bagging pears and late July $8 days. tenderness: tender- bagging pears for bagging apples. 48.1% ness: with muscle Same population examined two times. Exam: Muscular tenderness in 28.8% tenderness 2nd exam occurred one month after shoulder region; maximal grasping vs. apple first. These results used in analyses power measured by bagging for comparison of two tasks. dynamometer and back muscle with muscle power by myosphenometer. tenderness: Stiffness and pain during apple OR=1.7 1.1-2.9 bagging may have been pain that was Exposure: Observation of tasks a residual of pear bagging operations. and measurements of Pain in joint Pain in joint Workers representative workers (only two motion: motion: bagging pears Number of fruit bagged/day was workers measured). 23.1% 21.2% with pain in significantly more in pear bagging than controls joint motion in apple bagging. Angle of arm elevation during vs. apple bagging was measured in one bagging with Exposure measurements only subject. pain in joint obtained on 2 workers and motion: generalized to all workers. Angle of forward flexion of OR=1.1 0.53-2.3 shoulder for bagging pears was 110 to 139o. 75% of angles were above 90o. For bagging apples the angle of forward flexion was 0 to 140o; 41% of the angles were >90o.

(Continued)

3-68

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Schibye et Pro- Follow-up of 303 sewing Outcome: Cases defined by the Workers Õ Õ Õ Participation Rate in 1985: 94%. al. 1995 spective machine operators at nine Nordic questionnaire for who Participation Rate in 1991: 86%. factories representing symptoms as to duration during delivered or All participants were female. different technology levels last 12 months and during last 7 collected who completed days, effect on work their own 77 of 241 workers still operated a questionnaire in 1985. performance and leisure materials: sewing machine in 1991. activities, and sick leave. 18% In April 1991, 241 of shoulder 82 workers had another job in 1991. 279 traced workers Exposure: Assessed by symptoms; Among those 35 years or younger, responded to same questions regarding type of the rest 33% 77% had left their jobs; among those questionnaire. machine operated, work above 35 years, 57% had left their organization, workplace design, jobs. units produced/day, and payment system, time of employment as a 20% reported musculoskeletal sewing machine operator. symptoms as the reason for leaving job.

No significant changes in prevalences among those employed as sewing machine operators from 1985 to 1991; significant decrease in those who changed employment.

(Continued)

3-69

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Stenlund Cross- 55 of 75 rockblastors, 54 of Outcome: Based on a grading of Bricklayers Foremen Participation rate: 80%. et al. 1992 sectional 75 bricklayers randomly acromioclavicular joints of Rt side: Classification of X-rays achieved with selected from union shoulders. 59.3% 36.7% 2.2 1.0-4.7 blinding of investigators to age, name records and 98 of 110 Grade 0 = normal Lt side: or exposure status. foremen selected from Grade 1 = minimal changes 40.7% 23.4% 1.8 0.8-3.9 foremen employed in large Grade 2 = moderate changes Study looked at manual work and construction firms. Grade 3 = severe osteoarthritis Rockblasters Foremen exposure to vibration and relationship Grade 4 = joint destroyed Rt side: to osteoarthritis in acromioclavicular 61.8% 36.7% 2.1 0.9-4.6 joint using shoulder x-rays. Exposure: Based on self- Lt side: Logistic regression models adjusted reported estimates of loads lifted, 56.4% 23.4% 4.0 1.8-9.2 for age, smoking, dexterity, checked hr of exposure to vibration, job for interactions. title, and years of employment. Years of The weights of tools also manual work Questionnaire included questions obtained. >28 years vs. about smoking, dexterity, ethnicity, citizenship. <10 years Bricklayers lifted a mean of Rt side: 2.9 1.2-7.4 Risks were elevated as length of 29,439 tonnes; Rockblasters, a Lt side: 2.5 1.0-5.9 employment increased and as mean of 33,210 tonnes; Foremen, exposure to vibration and amount a mean of 2,261 tonnes. 10 to 28 lifted increased. years vs. X-ray grades 2 and 3 for analysis. <10 years Rt side: 1.1 1.1-4.7 Smoking significantly associated with Lt side: 2.3 1.0-5.3 osteoarthritis of right shoulder (OR=2, 2.4) but not left side. Significance Load lifted found, but is it meaningful? 725,000 vs. Left handedness significantly 710 tonnes associated with osteoarthritis of left Rt side: 3.2 1.1-9.2 side (OR=2.5). Lt side:10.3 3.1-34.5 The age adjusted odds ratio for osteoarthrosis in the right Vibration acromioclavicular joint for brick layers 725,000 hr and rock blasters as compared with vs <9001 hr foremen, was 2.16 on the right side Rt side: 2.2 1.0-4.6 95%CI(1.14-4.09), and was 2.56 95% Lt side: 3.1 1.4-6.9 CI (1.33-4.93).

(Continued)

3-70

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Stenlund Cross- 55 of 75 rockblasters and Outcome: Based on Bricklayers Foremen Participation rate: 80%. et al. 1993 sectional 54 of 75 bricklayers questionnaire of previous injuries Rt. side: selected randomly from and diseases of musculoskeletal 11.1%; 8.2% 0.4 0.2-1.3 Examiners blinded to exposure status union records, and 98 of system and previous shoulder Lt. side: or job title. 110 foremen randomly pain, and physical exam. 14.8% 17.1% Õ Õ selected from foremen Unconditional multiple regression employed in large Case defined as “Signs of Rockblasters analysis adjusted for age, construction companies. shoulder tendinitis” as palpable Rt. side: handedness, smoking and sport pain of the muscle attachment or 32.7% 8.2% 1.7 0.7-4 activities. In all models left and right pronounced pain reaction to Lt. side: sides calculated separately. isometric contraction in any of the 40.0% 17.1% 3.3 1.2-9.3 4 rotator cuff muscles or biceps Vibration related to shoulder tendinitis muscles. Clinical Entity although confounded by static loads Load and lifting. ”Clinical entity of tendinitis” Rt. side: 1.0 0.5-2.2 defined as pain during the last Lt. side: 1.6 0.6-4.1 Interactions tested for. year, pronounced pain reaction to palpation or isometric contraction. Vibration The study looked at manual work and Rt. side: 1.9 1.0-3.4 exposure to vibration and their Exposure: Based on self- Lt. side: 2.5 1.1-5.9 relationship to signs of tendinitis of the reported estimates of load lifted, shoulder. hr of exposure to vibration, job Manual Work title and years of employment. Rt. side: 0.9 0.5-1.8 Exposure-response found where Lt. side: 2.3 0.9-6.3 comparison of high vibration exposure Load defined as 0 to 709 tonnes, compared to low exposure. 710 to 25,999 tonnes, >25,999 Signs of tonnes vibration defined as hr of Tendinitis exposure: 0 to 8,999, 9000 to Load 255,199, >255,999 hr to each tool Rt. side: 1.0 0.6-1.8 multiplied by factor corresponding Lt. side: 1.8 0.9-3.4 to vibration energy. Years of manual work: 0 to 9, 10 to 28, Vibration >28 years. Rt. side: 1.7 1.1-2.6 Lt. side: 1.8 1.1-3.1

Manual Work Rt side: 1.1 0.7-1.8 Lt side: 1.9 1.0-3.4

(Continued)

3-71

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Sweeney Cross- 105 of 164 sign language Outcome: Symptom questionnaire Symptom >20 hr Participation rate: 64%. et al. 1994 sectional interpreters for the deaf, and physical exam: case: 22% signing, who attended a compared Examiner blinded to exposure status. professional conference of Symptom case defined as the Symptom to sign language interpreters. presence of pain, aching, case with <10 hr/we Generalizability of results to other sign stiffness, burning, numbness or moderate to ek 2.5 0.8- 8.2 language interpreters is limited. tingling in the shoulder lasting severe $ one week or once/month within shoulder the past 12 months; no previous discomfort: injury and symptoms occurred 50% after becoming a sign-language interpreter. Positive Õ Õ Õ symptom + Symptom-exam case: Defined as positive the presence of symptoms and a exam: 1% positive exam for the shoulder.

Exposure: Based on questionnaire (years of employment as a sign language interpreter; numbers of hrs/week engaged in signing).

(Continued)

3-72

Table 3-5 (Continued). Epidemiologic studies evaluating work-related shoulder musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Wells et al. Cross- Of 199 letter carriers, Outcome: Telephone interview All letter Postal Participation rate: 99% among letter 1983 sectional 196 were compared to based on current pain; carriers: clerks: carriers, 92% meter readers, 97% 76 of 79 meter readers and frequency, severity, interference 18% 5% 3.6 1.8-7.8 postal clerks. 127 of 131 postal clerks. with work, etc; score of 20 required to be a case. More Letter Schooling and marital status asked. points given to neck and shoulder carriers: Postal problems that interfered with increased clerks: Symptoms alone used for MSD routine daily activities. weight: 23% 5% 5.7 2.1-17.8 definition.

Exposure: Based on job Letter Comparison group (gas meter category; based on self-reported carriers: readers) used because of similar information on weight carried, no weight Postal “walking rate” without carrying weight previous work involving lifting and increase: clerks: compared to letter carriers. Postal work-related injuries. 13% 5% 3.3 1.1-11.1 clerks neither walk nor carry weight.

During analysis, more weight was given to scoring neck and shoulder than other body regions. Outcome influenced results when ranking of body MSDs, though, would not influence group comparison.

Adjusted for age, number of years on the job, quetlet ratio and previous work experience.

104 letter carriers had bag weight increased from 25 to 35 lbs in the year prior to the study.

Letter carriers with increased bag weight walked on average 5.24 hr; those with no change in bag weight walked 4.83 hr.

Letter bags usually carried on the shoulder.

3-73 CHAPTER 4 Elbow Musculoskeletal Disorders (Epicondylitis): Evidence for Work-Relatedness

SUMMARY Over 20 epidemiologic studies have examined physical workplace factors and their relationship to epicondylitis. The majority of studies involved study populations exposed to some combination of work factors, but among these studies were also those that assessed specific work factors. Each of the studies examined (those with negative, positive, or equivocal findings) contributed to the overall pool of data to make our decision on the strength of work-relatedness. Using epidemiologic criteria to examine these studies, and taking into account issues of confounding, bias, and strengths and limitations of the studies, we conclude the following:

There is insufficient evidence for support of an association between repetitive work and elbow musculoskeletal disorders (MSDs) based on currently available epidemiologic data. No studies having repetitive work as the dominant exposure factor met the four epidemiologic criteria.

There is evidence for the association with forceful work and epicondylitis. Studies that base exposure assessment on quantitative or semiquantitative data tended to show a stronger relationship for epicondylitis and force. Eight studies fulfilling at least one criteria showed statistically significant relationships.

There is insufficient evidence to draw conclusions about the relationship of postural factors alone and epicondylitis at this time.

There is strong evidence for a relationship between exposure to a combination of risk factors (e.g., force and repetition, force and posture) and epicondylitis. Based on the epidemiologic studies reviewed above, especially those with some quantitative evaluation of the risk factors, the evidence is clear that an exposure to a combination of exposures, especially at higher exposure levels (as can be seen in, for example, meatpacking or construction work) increases risk for epicondylitis. The one prospective study which had a combination of exposure factors had a particularly high incidence rate (IR=6.7), and illustrated a temporal relationship between physical exposure factors and epicondylitis.

The strong evidence for a combination of factors is consistent with evidence found in the sports and biomechanical literature. Studies outside the field of epidemiology also suggest that forceful and repetitive contraction of the elbow flexors or extensors (which can be caused by flexion and extension of the wrist) increases the risk of epicondylitis.

Epidemiologic surveillance data, both nationally and internationally, have consistently reported that the highest incidence of epicondylitis occurs in occupations and job tasks which are manually intensive and require high work demands in dynamic environments—for example, in mechanics, butchers, construction workers, and boilermakers.

Epicondylar tenderness has also been found to be associated with a combination of higher levels of forceful exertions, repetition, and extreme postures of the elbow. This distinction may not be a true demarcation of different disease processes, but part of a continuum. Some data indicate that a high

4-1 percentage of individuals with severe elbow pain are not able to do their jobs, and they have a higher rate of sick leave than individuals with other upper extremity disorders.

INTRODUCTION cohort. Those studies using symptom and Epicondylitis is an uncommon disorder, with the physical examination findings to define overall prevalence in the general population epicondylitis used consistent criteria— reported to be from 1% to 5% [Allender almost all studies using physical examination for 1974]. There are fewer epidemiologic studies diagnosis required pain with palpation of the addressing workplace risk factors for elbow epicondylar area and pain at the elbow with MSDs than for other MSDs. Most of these resisted movement of the wrist. However, studies compare the prevalence of epicondylitis studies using a definition based on symptom in workers in jobs known to have highly data alone used various criteria, some based on repetitive, forceful tasks (such as meat frequency and duration of symptoms [Burt et processing) to workers in less repetitive, al.1990; Hoekstra et al. 1994; Fishbein et al. forceful work (such as office jobs); the majority 1988] others based on elbow symptoms of these studies were not designed to identify preventing work activities [Ohlsson et al. individual workplace risk factors. 1989].

The text of this section on epicondylitis is REPETITION organized by work-related exposure factor. Definition of Repetition for Elbow The discussion within each factor is organized MSDs according to the criteria for evaluating evidence For our review, we chose studies that for work-relatedness in epidemiologic studies addressed the physical factor of repetition and using the strength of association, the its relation to elbow MSDs, especially those consistency of association, temporal studies that focused on epicondylitis. Studies relationships, exposure-response relationship, usually defined repetition, or repetitive work, and coherence of evidence. Conclusions are for the elbow as work activities that involved presented with respect to epicondylitis for each (1) cyclical flexion and extension of the elbow exposure factor. Summary information relevant or (2) cyclical pronation, supination, extension, to the criteria used to evaluate study quality is and flexion of the wrist that generates loads to presented in Tables 4-1 to 4-4. A more the elbow/forearm region. Most of the studies extensive summary (Table 4-5) includes that examined repetition as a risk factor for information on health outcomes, covariates, and epicondylitis had several concurrent or exposure measures. All tables are presented at interacting physical work load factors. We the end of this chapter. Not all the articles attempted to select those studies in which summarized in the tables are referenced in this repetition was either the single risk factor or the narrative, but they have been reviewed and dominant risk factor based on our review evaluated and are included for information. of the study and our knowledge of the There are 19 studies referenced in Tables 4-1 occupation. This method eliminated those through 4-4, 18 cross-sectional studies and one

4-2 studies in which a combination of high levels of based on self-reported symptoms lasting more repetition and high levels of force exist, or those than 1 month since starting career, or pain for studies which selected their exposure groups more than 30 days. Exposure was based on the based on highly repetitive, forceful work. authors’ experiences as occupational health physicians and involved crude assessment of Studies Reporting on the Association exposure level and exposure repetitiveness. of Repetition and Epicondylitis Analysis dealt with exposure as “duration of Seven studies reported results on the exposure as a sewing machine operator”. association between repetition and adverse Statistical modeling controlled for age, having elbow health outcomes including epicondylitis. children, not doing leisure exercise, smoking, The epidemiologic studies that address and socioeconomic status. For this study, the repetitive work and epicondylitis compare exposure classification scheme does not allow working groups by classifying them into separation of the effects of repetition from those categories based on some estimation of of force, although repetition may be a more repetitive work, such as percent of time typing obvious exposure. [Burt et al. 1990], number of items per hour [Ohlsson et al. 1989], or number of hand Baron et al. [1991] explored epicondylitis manipulations per hour [Baron et al. 1991]. among grocery store workers, comparing the Those studies which may have measured prevalence among grocery store cashiers to repetitive work but have exposure to higher that among non-cashiers and identified work levels of force will be discussed in the “Force” risk factors while controlling for covariates. section. Detailed ergonomic assessment of grocery checking and cashiering was completed using Studies Meeting the Four Evaluation both on-site observational techniques and Criteria videotaped analyses. The majority of cashiers None of the studies (see Table 4-1 and Figure were categorized as having “medium” levels of 4-1) reviewed for the elbow summary section repetition for the hand (defined in this study as met all four evaluation criteria outlined in the making 1250 to 2500 hand movements per Introduction Section. hour). Repetitive movements were not recorded directly for the elbow; however, the Studies Meeting at Least One of the Criteria number of hand movements serve as an The studies will be summarized in alphabetical approximation for elbow repetitions. Age, order as they appear in hobbies, second jobs, systemic disease, and Table 4-1. height were considered as covariates in the multivariate analyses. The diagnosis of Andersen and Gaardboe [1993a] used a cross- epicondylitis required standard physical sectional design to compare sewing machine examination techniques of palpation and operators with a random sample of women resisted extension and flexion of the elbow. from the general population of the same region. Elbow pain, not epicondylitis, was the MSD of Burt et al. [1990] studied 834 employees using interest in this study. A case of elbow pain was computers at a metropolitan newspaper, using a

4-3 self-administered questionnaire for case assembly workers and compared these two ascertainment. Exposure assessment was based groups to a random sample from the general on self-reported typing time and observation of population. A case of elbow pain was based on employees’ job tasks, then categorization by questionnaire responses; exposure was based job title. A separate job analysis using a on job categorization as well as questionnaire checklist and observational techniques was responses. Repetitive exposure was based on a carried out for validating questionnaire self-reported frequency of task items exposure data. Workers fulfilling the case completed per hour (work pace). Results definition for elbow/forearm pain were showed no association with work pace and compared to those who did not fulfill the case elbow symptoms, and no association between definition. Prevalence of cases was associated length of employment and elbow symptoms. with percent of time typing and typing speed. Logistic regression controlled for age, gender, Punnett et al. [1985] compared neck/shoulder metabolic disorders, and job satisfaction. MSDs based on symptom reporting alone in 162 women garment workers and 76 women Automobile assembly line workers were hospital workers such as nurses, laboratory compared to a randomly selected group from technicians, and laundry workers. There was a the general population in the study by Byström low participation rate among the hospital et al. [1995]. A case of epicondylitis required workers. Eighty-six percent of the garment symptoms and physical examination. “Job title” workers were sewing machine operators and was used as a surrogate for exposure in the finishers (sewing and trimming by hand). The analysis. No assessment of repetition or sewing machine operators were described as repetitive work was completed specifically for using highly repetitive, low force wrist and the elbow. finger motions, while the finishers had shoulder and elbow motions as well. The exposed McCormack et al. [1990] had a randomly garment workers likely had more repetitive selected population of 2,261 textile workers jobs than most of the hospital workers. from over 8,000 eligible workers. Workers were analyzed by job category, after Strength of Association—Repetition observation of jobs. Epicondylitis case and Elbow MSDs ascertainment was by clinical exam. Of the 37 No studies met the four criteria to discuss cases of epicondylitis identified, 13 were strength of association. categorized as mild, 22 were moderate, and 2 were severe. Eleven examiners may have Strength of Association—Studies Not introduced an interexaminer reliability problem. Meeting the Four Criteria Age, gender, race, and years of employment For the other studies not fulfilling all the criteria, were analyzed as confounders. the odds ratio (OR) reported in the

Baron et al. [1991] study for epicondylitis Ohlsson et al. [1989] studied electrical overall was 2.3, but this was not statistically equipment and automobile assemblers, former significant.

4-4 Anderson and Gaardboe [1993a] used years addressed repetition as a physical factor alone; employed as a sewing machine operator as a all the studies were cross-sectional, so a surrogate for exposure and found no significant temporal relationship cannot be established. association with epicondylitis. However, some cross-sectional studies allow None of the other studies that looked at us to infer causality by use of restrictive case epicondylitis among working groups carried out definitions. Studies by the National Institute for independent exposure assessment of workers Occupational Safety and Health (NIOSH) or representative workers that focused on the investigators [Burt et al. 1990; Baron et al. elbow. 1991] excluded from analysis those workers who reported symptoms experienced prior to Burt et al. [1990] found a statistically significant their present job and those with acute injury to OR of 2.8 for elbow/forearm symptoms in the elbow not related to the job. newspaper employees who reported typing 80%–100% of their working day compared to Consistency in Association for those typing 0%–20%. (Typing hours has been Repetition and Epicondylitis used as a surrogate of both repetition and The studies were not consistent in showing an duration of exposure.) association between repetitive work and epicondylitis. In terms of strength of Likewise, Punnett et al. [1985] found a association, there were no studies that had significant prevalence rate ratio (PRR=2.4) of statistically significant ORs greater than 3.0, persistent elbow symptoms among garment four studies had ORs between 1.0 and 3.0, that workers performing repetitive, forceful work were statistically significant; and two studies compared to hospital workers. Analysis by job had nonsignificant ORs less than 1.0. title showed that underpressers, whose jobs consisted of ironing by hand, had a PRR of 6.0. Coherence of Evidence for Repetition Among stitchers (sewing machine operators), The evidence for epicondylitis in the the significant PRR for the task of setting linings biomechanical and sports literature does not was 7.7. When standardized to the age address repetition alone, but has consistent distribution of the hospital workers, the rate evidence with a combination of forceful ratio did not change. exertion, awkward or extreme postures, and repetitive movements. Please refer to the McCormack et al. [1990] and Ohlsson et al. discussion under Coherence of Evidence for [1989] based exposure on job title and found Force. no association between repetitive work and epicondylitis, with non-significant ORs between 0.5 and 2.8.

Temporal Relationship—Repetition and Epicondylitis There were no prospective studies which

4-5 Exposure-Response Relationship for categories based on an estimated amount of Repetition resistance or force of exertion and a In Baron et al.’s [1991] study, there was a combination of estimated rate of repetition dose-response relationship for the elbow for (e.g., Viikari-Juntura et al. [1991b]; Kurppa et the number of hours per week working as a al. [1991]; Chiang et al. [1993]) or in terms of checker, with ORs up to around 3.0, but not overall elbow stress [Dimberg 1987; Ritz for the duration of employment (the average 1995]. length of employment was 8 years). Studies Meeting the Four Evaluation Criteria Conclusions Regarding Repetition Of the studies examining epicondylitis and There is insufficient evidence for support of an forceful exertion, three studies [Chiang et al. association between repetitive work and 1993; Luopajärvi et al. 1979; Moore and Garg elbow MSDs based on currently available 1994] fulfilled all four criteria. Most of these epidemiologic data. There were no studies that studies used combinations of risk factors in their met the four criteria. Of the 7 studies examining analysis, of which forceful exertion was one. repetitive work, no studies found ORs above 3.0, 5 studies found ORs from 1–3, and 2 Chiang et al. [1993] assessed exposure though studies found an OR less than one. observational methods, recording of tasks and biomechanical movements of representative FORCE workers. With these methods, they categorized Definition of Force for Elbow MSDs fish processing workers into three exposure groups according to the ergonomic risks to the For our review, we included studies that shoulders and upper limbs: (1) those with low examined force or forceful work or heavy loads force and low repetition (the comparison to the elbow, or described exposure as group), (2) those with high force or high strenuous work involving the forearm extensors repetition, and (3) those with both high force or flexors, which could generate loads to the and high repetition. The diagnosis of elbow/forearm region. Most of the studies that epicondylitis included standard physical examined force or forceful work as a risk factor examination techniques of palpation and for epicondylitis had several concurrent or resisted extension and flexion of the elbow. interacting physical workload factors. Examination-defined cases were about one-half the number of cases defined by symptom alone. Studies Reporting on the Association The analysis was stratified by gender, and those of Force and Epicondylitis with metabolic diseases associated with MSDs Thirteen studies reported results on the were excluded. There was no significant association between force and adverse elbow difference in age between the comparison health outcomes, including epicondylitis. The groups. Multivariate analysis was not carried epidemiologic studies that addressed forceful out for the elbow in this study. work and epicondylitis compared working groups by classifying them into broad

4-6 Luopajärvi et al. [1979] determined MSDs estimated as percent of maximal strength by differences between female assembly line comparing the reported weight of the pertinent workers and shop assistants in a department object with estimated average maximal strength store (cashiers were excluded from the of the worker for different types of pinches and comparison group). Exposure assessment grasps, then categorized into five levels. involved on-site observation, video analysis and interviews. The assembly work was found to be These values were derived from population- repetitive, with up to 25,000 cycles per based data stratified according to age, gender, workday involving hand and finger motions. and hand dominance. Repetition was recorded Specific cycles were not recorded for elbow as cycle-time and exertions per minute. The motions; however, motions involving the hands exposure assessment in this study gave more and fingers involve tendons and muscles from weight to the factor of “force” than to repetition the flexors and extensors that have their origin or posture (the force variable could increase to at the elbow. Static muscle loading of the a higher categorization level if the job was forearm muscles, deviations of the wrist, and repetitive, involved jerky motions, or extreme lifting were also found. The diagnosis of postures). Work histories, demographics, and epicondylitis included standard physical pre-existing morbidity data were not collected examination techniques of palpation and on each participant. The diagnosis of resisted extension and flexion of the elbow. epicondylitis extracted from the medical Subjects with previous trauma, arthritis, and records included standard physical examination other pathologies associated with MSDs were techniques of palpation and resisted extension excluded. All participants were female. and flexion of the elbow. Analyses were based Covariates considered in the analysis included on “full-time equivalents” for jobs, not individual age, social background, hobbies, and the workers. This analysis did not control for amount of housework performed. Duration of potential confounders; there was a slight employment was not an issue because the preponderance of morbidity of all MSDs factory had only been open a short time. among females.

Moore and Garg [1994] carried out a medical Studies Meeting at Least One Criteria records review using an epicondylitis case The Andersen and Gaardboe study [1993a], definition based on symptoms and physical which did not carry out ergonomic assessment examination and a semi-quantitative ergonomic pertaining to the elbow, found a non-significant assessment of 32 jobs at a meatpacking plant. association between repetitive, forceful work The authors used their “Strain Index” to and symptoms or physical findings consistent categorize jobs as “hazardous” or “safe” based with epicondylitis. In the Andersen and on a number of factors: observation, video Gaardboe study [1993a], the exposed group analysis, and judgements based on force, consisted of sewing machine operators. repetition, posture, and grasp. Force was

Baron et al.’s [1991] measure of force was based on estimated assessment of exertion by

4-7 experienced ergonomists through observation focused on the stress to the hand/wrist, elbow, of tasks and video analysis, as well as weight of and shoulder areas. scanned items. Average forces for the grocery checkers were categorized as “low” and peak One study by Kurppa et al. [1991] was forces “medium” on a three-tiered scale (“low, prospective. Here, workers in meat processing medium, and high”). were categorized into strenuous and nonstrenuous jobs based on repetitive and Byström et al.’s [1995] study of automobile forceful work. The strenuous tasks for the assembly workers is reviewed in the Repetition meatcutters consisted of cutting approximately section. 1,200 kg of veal or 3,000 kg of pork per day; the nonstrenuous tasks consisted primarily of Dimberg’s studies [1987] fulfilled three of the office work. Workers had to have a physician criteria but did not mention if examiners were visit and diagnosis in order to be considered a blinded to exposure status. In the 1987 study, case—a restrictive definition requiring exposure was assessed by observational significant enough symptoms to seek out methods, jobs were categorized according to medical care. the amount of elbow stress in a particular job, but no individual measurements were made. Numerical results from the logistic regression Twenty-five percent of cases were diagnosed by physicians outside the plant, so examination model were not given in the paper, although employee category (blue collar versus white techniques may not have been the same as collar), gender, and degree of elbow stress those for the other 75%. The nonstrenuous were said not to be significant predictors of group was similar to the strenuous group with having any one of the three types of regards to age, gender, and duration of epicondylitis. The author classified epicondylitis employment, except for the small number of into three types: leisure-related, no known male sausage makers and male cause, and work-related groups based on meatpackers—these were excluded from history. When the author specifically looked at calculation of individual IRs. “work-related” epicondylitis (criteria for such designation was not given) with respect to Punnett et al.’s [1985] study of garment elbow stress, he found a significant trend with workers is reviewed in the Repetition section. increasing levels of elbow stress. Ritz [1995] did not mention the participation The exposure assessment approach was rate in their study of welders and pipefitters but different for the 1989 study by Dimberg et al. fulfilled the other three criteria. Workers In the 1987 study by Dimberg, the exposure studied were likely to be a representative classification scheme was focused principally sample, however, since all male employees on the elbow and identified jobs with heavy who were taking their elbow-straining work. In the 1989 study, the author focused on multiple health outcomes in the upper extremity and used an exposure classification scheme that was more broadly annual examinations during a three month

4-8 period were enrolled in the study. The multiple examinations and may have influenced results. logistic model analysis considered age and a The investigators stated the comparison group variety of confounding factors. Among these was selected similar to the study group in public gas and water work employees, the gender, age, and duration of employment. welders and pipefitters who installed and repaired pipes were considered to have high In conclusion, for the studies with less than our exposure. four criteria, four are supportive [Kurppa et al. 1991; Ritz 1995; Dimberg 1987; and Roto and Roto and Kivi [1984] based their exposure on Kivi 1984], two are non-supportive [Dimberg job title alone, but fulfilled the other three et al. 1989; Byström et al. 1995], and one is criteria. They compared meatcutters who had not very informative [Andersen and Gaardboe forceful, repetitive work to construction 1993a]. The results from the positive studies workers who had more varied tasks. The are unlikely to be due to confounding or authors stratified the analysis by age and found selection bias. Overall, these studies provide the majority of cases in the older age groups. limited support for the association of forceful They also found that the meatcutters with repetitive work and epicondylitis. epicondylitis had been exposed, on the average, five years longer than the other Strength of Association—Force and meatcutters. All the meatcutters had more than Epicondylitis 15 years in their current occupation, which the Chiang et al. [1993] did not find an association authors attributed to support of the work- between hand-intensive work (categorized relatedness of the condition, although increasing based on forceful exertion and repetition) and age may have been a confounder or effect epicondylitis when analyzing all workers at six modifier. fish processing plants. However, in examining the highest level of exposure (we calculated the Viikari-Juntura et al. [1991b] studied subjects odd ratios for men and women separately, at the same meat processing plant as Kurppa et which was not done in the article), we found a al. [1991] using 3 cross-sectional examinations significant difference between males in the covering a period of 31 months. The same highest exposed group (Group III) and males in exposure assessment scheme used in the the lowest exposed group (Group I) (OR= Kurppa et al. [1991] study mentioned above 6.75) and a non-significant OR of 1.44 for was used comparing workers in strenuous and women. Exposure in Group III was based on a nonstrenuous work. This study compared the combination of high-force exertion and high prevalence of all cases of epicondylitis; cases repetition; analysis of working techniques by due to injury or known non-occupational gender was not performed, so the reason for causes were not excluded. The diagnosis of the difference in the groups by gender is not epicondylitis included standard physical known. The Chiang et al. [1993] study examination techniques of palpation and provides limited support for the association resisted extension and flexion of the elbow; the authors stated that palpation pressure increased between high levels of forceful repetitive elbow on the second of the three cross-sectional work and epicondylitis.

4-9 Luopajärvi et al. [1979] found a non-significant difference overall in the prevalence of Strength of Association—Studies Not Meeting the Four Criteria: Force and epicondylitis and pronator teres syndrome (3 Elbow MSDs versus 11 cases, OR 3.35 [95% confidence Baron et al. [1991] found an OR of 2.3 for the interval (CI) 0.86–19.1]); for lateral combination of factors, but this was not epicondylitis only, an OR of 2.73 (95% CI statistically significant. The authors mention that 0.66–15.94). There were five cases of medial ergonomic analysis of the non-checkers epicondylitis in the assembly workers and none showed that they also performed work in the shop assistants. The increase in medial requiring repetitive motions and awkward epicondylitis (an indeterminate OR because of postures; therefore, the comparison probably “zero” cases in the shop assistants) was resulted in a lower OR than had the referent attributed to the difficult grasping movements group been truly unexposed to the ergonomic involved in the assembly line work. They found stressors. that their female assembly workers tended to have physically light work, but this work Kurppa et al. [1991] found a strong significant required highly repetitive movements of the relationship between strenuous jobs and wrists and fingers and static muscle loading of epicondylitis (IR= 6.7), while Viikari-Juntura et the forearm muscles. al. [1991b] did not (OR=0.88, nonsignificant). These results may have been influenced by Using the Strain Index, Moore and Garg allowing “cases” who had recurrence in the [1994] found a significant relationship between same elbow to be counted as new cases (12 hazardous jobs (of which force was a major out of 57 employees with epicondylitis had component) and upper extremity MSDs (of more than one episode, and were counted which epicondylitis was an important twice). There was a median of 184 days component). The results found a significant OR between the episodes. In examining this study, of 5.5 for a case of epicondylitis to occur in a it is important to see if the odds of having hazardous job. When approximating the epicondylitis would be elevated if these classification scheme for low and high force workers with recurrences were only counted used by Silverstein et al. [1987] and then by once. We recalculated the OR using only Kurppa et al. [1991], Viikari-Juntura et al. “persons” and not “single episodes of [1991b], and Chiang et al. [1993], the epicondylitis” in order to obtain a more association between forcefulness and the conservative estimate. We counted, only once, overall upper extremity morbidity in the study the employees with recurrence, as well as the was again statistically significant (p<0.02). four employees mentioned with simultaneous occurrence in both elbows and subtracted these The overall conclusion from the three studies from the strenuous job cases. This gave a total that met our four criteria is that there is of 44 cases of epicondylitis among the evidence for association between force strenuous group. and epicondylitis based on strength of association. Using this estimate, more restrictive than that

4-10 found in the article, gives an OR of 5.5 (2.4, The study conducted by Roto and Kivi [1984] 12.7) for epicondylitis among the workers with found an OR of 6.4 (95% CI 0.99–40.9) using strenuous jobs versus those with nonstrenuous an exposure assessment based on job title jobs. The Kurppa et al. [1991] prospective alone (meatcutters were assumed to have more study also found the IR of epicondylitis in forceful jobs than construction workers). Only nonstrenuous jobs to be similar to Allender's one referent had epicondylitis. [1974] population background prevalence rate (1%) for epicondylitis. In the paper by Viikari-Juntura et al. [1991b], the cases of epicondylitis not listed as insidious Ritz [1995] found a significant OR for 10 years all involved forceful, repetitive tasks (although of high exposure to elbow straining work: 1.7 some of these tasks were not related to work). for currently held jobs and 2.2 for formerly held Prevalences of “epicondylar pain” and “sick jobs. The significant OR for moderate exposure leave due to epicondylar pain” were in the current job was 1.4 for 10 years of significantly different between the two groups exposure. This study provides support for the (OR 1.9 and 2.1). There was no significant association of forceful work with epicondylitis. difference in the prevalence of epicondylitis (combined work and non-work related) We calculated odd ratios from data in between workers in strenuous versus Dimberg’s [1987] study and found an OR for nonstrenuous jobs (OR=0.88). In 95 women moderate stress versus none or light elbow sausage makers, there were four cases with stress of 2.9, and for heavy versus none or light insidious onset, while among 160 women stress of 7.4. Heavy stress in the elbows was referents there were two cases, one with assigned to job titles like blaster, driller, or insidious onset, the other related to an grinder. The major limitation of this analysis of “exceptional task of cutting cheese.” The the work-related cases is that it did not resulting OR was 6.9 (95% CI 0.74–171). This consider age, a likely confounder. Overall, this study also found that rates of “epicondylar study provides support for the association pain” and “sick leave due to epicondylar pain” between forceful work and epicondylitis, differed significantly between the two groups particularly in older workers. (OR 1.9 and 2.1, respectively). Rates of medically diagnosed cases of epicondylitis were The 1989 Dimberg et al. study was not not statistically different between the two supportive of an association between lateral groups, but the results for epicondylar pain epicondylitis and forceful repetitive work, but (causing sick leave in the two groups), and the was positive for “mental stress at work” at the fact that the majority of cases in both groups onset of symptoms for lateral epicondylitis were due to events involving strenuous, (p<0.001). As a result of the specific elbow repetitive tasks, give some support to forceful, exposure assessment, we believe that with repetitive work as regards to stressful or a cause. forceful elbow exertions that the 1987 study is Byström et al. [1995] noted that the low more informative. frequency could not be attributed to selected

4-11 subjects being absent, as all persons on leave Viikari-Juntura et al.’s [1991b] study did not participated in the investigation. The authors exclude workers with elbow symptoms or also stated that “exposure to repetitiveness and physical findings that were due to acute injury force in automobile assembly line work may be not related to the job, which may account for less than in other investigated work situations.” the contrasting result. In fact, in that study, four Because the authors did not give quantitative or workers with acute non-work-related qualitative information on the forcefulness or epicondylitis in the nonstrenuous group were repetitiveness of jobs included in the study noted in the journal article. Another group, it is difficult to know whether these jobs consideration for inconsistency is due to were appropriate to use to study epicondylitis. grouping of studies, which may all fulfill good epidemiologic criteria, may all examine the Temporal Relationship: Force and same risk factor, but may compare groups that Epicondylitis do not have similar contrasting levels of See temporal relationship above in Repetition exposure. For example, the Chiang et al. and Epicondylitis. [1993] study found statistically significant results in men when comparing high force/high Consistency of Association repetition jobs to low force/low repetition jobs. The studies that met the four criteria were fairly Baron et al. [1991], on the other hand, consistent in their strength of association compared checkers in low force, medium between force and epicondylitis, with most repetition jobs to noncheckers in low force, low ORs between 2.5 and 7.0. Focusing on those repetition jobs. studies that compared workers exposed to force that was documented to be at a high level, Two factors explain the difficulty in determining to those exposed to a low level, all studies the reasons for the apparent inconsistencies [Chiang et al. 1993; Kurppa et al. 1991; among the studies on forceful and repetitive Moore and Garg 1994] were consistent. work. First, very few of the exposure assessments were quantitative—this is due to Of those 10 studies that examined force but did existing limitations in directly measuring not fulfill the four criteria, two studies had a exposure in detail in most field studies. As a significant OR greater than 3.0, three studies result, there is likely to be frequent non- had significant ORs between 1.0 and 3.0, one differential misclassification of exposure. had a nonsignificant OR between 1.0 and 3.0, Second, most of the studies completed have and two had an OR less than 1.0. Two had been cross-sectional, and therefore subject to statistically significant findings but did not report survivor bias. ORs. Most of these studies examined workers in repetitive, forceful job tasks and did not As an example, Chiang et al. [1993] found that separate out epicondylitis was significantly associated with increasing repetitiveness and the independent effect of repetition through any analytic method. forcefulness among fish processors employed less than 12 months. For those working for 12

4-12 to 60 months, a similar trend was found, but a workers in the automobile industry. In both of reverse trend was found in those workers these studies, those cases of epicondylitis listed employed for over 60 months. The authors in the comparison groups were due to highly stated that because most of the workers were repetitive, forceful activities. The lack of a semi-skilled, they were likely to leave their job significant difference in the prevalence of the if they felt frequent muscle pain because of it. disorder between the two groups may be They went further to say that the selection because the referent, “low” exposure groups mechanism may explain the lack of significant had a higher incidence of non-work-related associations between the disorders and the lateral epicondylitis. duration of employment. There was no indication that the authors pursued this Coherence of Evidence hypothesis by trying to identify former workers The epidemiologic results of finding the majority who may have left. Turnover rate was not of cases occurring in highly repetitive, forceful discussed. This example highlights two work [Moore and Garg 1994; Chiang et al. important factors concerning the cross- 1993; Kurppa et al. 1991; Kopf et al.1988] sectional studies examining work-related are consistent with the evidence from epicondylitis: there is some evidence that older biomechanical and physiologic findings, as well workers may be at higher risk of epicondylitis as from sports medicine literature and older [Dimberg 1987; Ritz 1995], and there is also a medical clinical case series. In cases of lateral “survivor” effect, which results in the loss to the epicondylitis occurring in workplaces as well as study of affected workers. These two factors in sports, the forearm extensors are repetitively make the interpretation of duration of disease contracted and produce a force that is relationships complex and may affect the transmitted via the muscles to their origin on the estimate of the risk of disease. lateral epicondyle. These repetitive contractions There were studies that used more accurate produce chronic overload of the bone-tendon exposure assessment or had comparison junction, which in turn leads to changes at this groups with marked differences in levels of junction. The most common hypothesis is that exposure to forceful and repetitive work that microruptures occur at the attachment of the were positive, such as the Kurppa et al. [1991] muscle to bone (usually at the origin of the study of meatcutters, sausage makers, and extensor carpi radialis brevis muscle), which packers, Moore and Garg's [1994] study of causes inflammation. Peƒina et al. [1991] did pork processors; Dimberg's [1987] study of not agree with the microrupture theory; they blasters, drillers, grinders, and others in an theorized that overuse leads to avascularization engineering industry; Ritz’s [1995] study of of the affected muscle origin, which leads to pipefitters and welders in a public utility; and overstimulation of the free nerve endings and Roto and Kivi’s [1984] study of meatcutters. results in aseptic inflammation. Further There were studies with these characteristics repetition of the offending movements causes that were negative, such as the Viikari-Juntura angiofibroblastic hyperplasia of the origin. et al. [1991b] study of meatcutters, sausage Nirschl [1975] stated that the degree of makers, and packers; and the study by angiofibroblastic hyperplasia is correlated to the Dimberg et al. [1989] of blue- and white-collar duration and severity of symptoms. On

4-13 histologic analysis of severe cases of forearm extensor muscle contraction and epicondylitis, one can see the characteristic repetitive supination/ pronation of the forearm. invasion of fibroblasts and vascular tissue, the In the second group of 26, the 23 (88%) who typical picture of angiofibroblastic hyperplasia. had spontaneous onset worked in jobs with constant gripping or repetitive movements. Prior to many of the epidemiologic studies, there were numerous reports in the medical Many case studies of professional athletes have literature of clinical case series that suggest a documented that forceful, repeated dorsiflexion, relationship between epicondylitis and pronation, and supination movements with the repetitive, forceful work. For example, as early elbow extended can cause epicondylitis. as 1936 Cyriax reported that with regard to [Ollivierre et al. 1995; Priest et al. 1977; King patients with lateral epicondylitis, “those et al. 1969]. Most cases have occurred in patients who remember no special overexertion baseball pitchers and tennis players. will be found to be working at screwing, lifting, Occupations involving movements described hammering, ironing, etc., or to be violinists, above have also been found to have increases surgeons, masseurs, etc.” Cyriax had in rates of elbow MSDs. This literature has also designated a “Chronic Occupational” variety of referred to increased occurrence in occupations tennis elbow, in which he stated that “often no requiring force, awkward postures, and history of an injury is obtainable, but the repetitive use of the elbow and forearm patient's occupation at once provides the clue.” [Lapidus and Guidotti 1970; Mintz and Fraga He cited “work which entails repeated 1973; Berkeley 1985]. These reports, though pronation and supination movements with mainly case series, have lead to further studies elbow almost fully extended” to be responsible that examined the links between exposure and for epicondylitis [Cyriax 1936]. Feldman et al. epicondylitis. [1987] reported that occupations with work tasks requiring repeated pronation and An example of an early occupational study is internal/external rotation of the forearm are at one by Mintz and Fraga [1973], who found high risk of pronator teres syndrome that foundry workers (with an average of 14 (compression of the median nerve as it courses years of employment) who used tongs requiring through the pronator teres muscle in the twisting and bending of the elbows/forearms for forearm). A number of case series have eight hours per day had decreased elbow reported similar findings [Hartz et al. 1981; flexion and extension and Morris and Peters 1976]. pain on physical examination, as well as severe radiographically documented osteoarthritis Sinclair [1965] reported 2 case series of localized to the elbows. In the studies that are patients with tennis elbow (lateral epicondylitis), reviewed in Tables 4-1 44 patients treated between 1959-1961 and 38 patients treated between 1961-1963. In the through 4-4, the occupations with the highest first group of 267, the 130 (48%) whose onset rates of epicondylitis, such as drillers, packers, occurred spontaneously had occupations that meatcutters, and pipefitters, are consistent with included gripping tools with consequent the force-repetition model of the causation of

4-14 epicondylitis. The development of epicondylitis Relationship for Force in these workers is consistent with proposed The Baron et al. [1991] study is mentioned biological mechanisms and is plausible. above in the Repetition Section as showing a dose-response relationship for number of hours The lack of elbow MSDs and work factors in of work per week. Chiang et al. [1993] found some of the studies with occupations like that among men the prevalence of epicondylitis sewing workers [McCormack et al. 1990] or increased with increasing force and repetition in automobile assembly line workers [Byström et fish processors. In several studies, only al. 1995], most likely reflects the interplay of dichotomous divisions were made, so two factors. The movement of affected workers conclusions concerning an exposure-response out of high exposure jobs limits the ability of relationship cannot be drawn. However, we cross-sectional studies to accurately determine can see significantly contrasting rates of elbow associations between work factors and MSDs between high- and low-exposure epicondylitis. Our ability to accurately identify groups. Moore and Garg [1994] found a higher working conditions with an elevated risk for risk in workers with high-strain jobs compared epicondylitis may require an exposure to those with low-strain jobs. Kurppa et al. assessment of each job to a degree that has [1991] found higher risk in workers with been beyond the limits of current strenuous jobs compared to those with epidemiological methods. As a result, nonstrenuous jobs, and that female sausage misclassification of exposure may be common. makers had an increase in epicondylar Overall, the majority of the epidemiologic tenderness with increasing duration of studies are supportive of the hypothesis of an employment. While Dimberg [1987] found no increase risk of epicondylitis for occupations difference in epicondylitis between blue- and that involve forceful and repetitive work, white-collar workers, he found that workers frequent extension, flexion, supination, and with elbow pain severe enough to require a pronation of the hand and the forearm. The physician consult were significantly more often surveillance data are also supportive of this in those jobs identified independently as having hypothesis [Roto and Kivi 1984; Washington high elbow stress. Dimberg also found a State Department of Labor and Industry 1996]. statistically significant correlation coefficient for The highest relative risks for epicondylitis in lateral epicondylitis and time spent in the Finland were with mechanics, butchers, food present job. Luopajärvi et al. [1979] found a industry workers, higher rate of epicondylitis and pronator teres and packers; the highest industries in syndromes in a Washington State for 1987-1995 [Silverstein et high-exposure group of assembly line packers al. In Press] were construction workers, meat compared to the referent group of shop dealers, and foundry workers—all occupations assistants. Overall, these studies provide with repetitive, forceful work involving the arms considerable evidence for a and hands and requiring pronation and supination. difference in level of risk for epicondylitis when there are marked differences in the level of Evidence of a Dose-Response exposure to forceful and repetitive tasks.

4-15 Ritz [1995] reported a positive dose-response Strength of Association—Posture relationship between duration of exposure to and Epicondylitis gas and waterworks jobs regarded as Studies Meeting the Four Evaluation moderately and highly stressful to the elbow Criteria and epicondylitis. Roto and Kivi [1984] The Moore and Garg [1994] study (also reported that all workers with epicondylitis in discussed above) recorded wrist posture using their meat-packing facility worked for more a classification similar to Armstrong et al. than 15 years in the strenuous job category and [1982] and Stetson et al. [1991]. Pinch grasp had been exposed an average of 5 years longer was also noted to be present or absent. In this than non-diseased workers. Kopf et al. [1988] study, posture was not found to be significantly reported that in their study of brick layers, with associated with “hazardous” jobs. This may be increasing levels of job demands (defined as due to the heavier weighting given the force either heavy physical work, awkward working rating system than the posture or repetition postures, repetitive movements, or restriction in scale. For example, if a job required extreme standing position), the OR increased from 1.8 posture, the authors increased the force rating to 3.4. These studies, with less clear contrasts instead of the posture rating. If a combination of in exposure, provide support for the exposure- extreme posture and high-speed movement was response relationship between epicondylitis and required, then the force rating was raised by forceful, repetitive work. two levels, but not the posture rating. Data that would allow analysis of the incidence of POSTURE epicondylitis and the exposure to extreme Definition of Postures for posture were not presented. Elbow MSDs We chose to include those studies that Luopajärvi et al.’s [1979] assessment was addressed posture or examined workers in focused on the extreme work position of the those activities or occupations that require hands but not the elbow; it included extension, repeated pronation and supination, flexion/ flexion and deviation of the wrists. Although extension of the wrist, either singly or in there was a non-significant association between combination with extension and flexion of the assembly line work and the presence of either elbow. epicondylitis or pronator teres syndrome in shop assistants (11 cases versus 3), there were 5 cases of medial epicondylitis and 2 cases of Studies Reporting on the Association pronator teres syndrome in the assembly of Posture and Epicondylitis workers and none in the shop assistants. The The six studies in Table 4-3 addressed posture greater prevalence of medial epicondylitis in variables. Of these, only the studies by Moore and Garg [1994] and Luopajärvi et al. [1979] assembly workers was attributed to the difficult fulfilled all four criteria. The details of these grasping movements involved in the assembly studies are discussed in the Repetition and line work. The authors stated that the overall Force sections. prevalence may have been “connected with the constant overstrain of flexors in work.”

4-16 Studies Not Meeting the Four wrist deviation, thus producing more symptoms. Evaluation Criteria These conclusions should be considered to be The Dimberg [1987] study stated that over- hypothesis generating and not definitive. exertion of the extensor muscles of the wrist due to gripping and twisting movements prior to Temporal Relationship the onset of symptoms was verified in 28 of the There are no prospective studies that address 40 (70%) of the cases, of which 14 were posture and epicondylitis. The one prospective considered to be caused by work. In the study study concerning epicondylitis did not address by Dimberg et al. [1989], the guidelines for posture. classification include repeated rotation of the forearms and wrists in Group 1, large and Consistency in Association frequent rotations in extreme positions in Group 2, but fail to include work involving frequent There are too few occupational epidemiologic rotations in the highest exposed group, Group studies that address posture and epicondylitis to 3. The difference in exposure classification meaningfully discuss consistency of association. scheme may explain why there was no relationship between prevalence of epicondylitis Coherence of Evidence and increasing work strain. Please refer to the “Repetition Section and Coherence of Evidence” for a discussion of the Hughes and Silverstein [1997] found a strong, sports literature, and the combination of factors, statistically significant association (OR 37) including extreme postures that have been between elbow/forearm disorders and “the documented concerning epicondylitis. number of years of forearm twisting” in their study of aluminum workers. However, this Exposure-Response Relationship study had an overall low participation rate There is little evidence on which to base a (55%), which limits the interpretation of its discussion exposure response relationship in the result. epidemiologic studies. Once again, the reader is referred to the biomechanical sports literature. The other study that may be interpreted as related to a posture variable is the one by EPICONDYLITIS AND THE ROLE OF Hoekstra et al. [1994]. This study evaluated CONFOUNDERS video display terminal users at two work sites The model for epicondylitis clearly implies that differing only in whether adjustable office both occupational and non-occupational equipment was present. By self-reported activities can cause the disorder. Several symptoms and exposure studies [Ritz 1995; Andersen and Gaardboe 1993a; Dimberg 1987] directly observations, the Hoekstra et al. [1994] study address the issue of work-related versus non- found that having a “non-optimally adjusted” work-related exposures by assessing both. chair was associated with elbow MSDs. This Two of the most important potential improper chair adjustment was thought to confounders or effect modifiers are age and increase shoulder and elbow flexion, as well as duration of employment. In Dimberg's [1987]

4-17 and Ritz’s [1995] studies, older workers had recognize this limitation of the epidemiologic high rates of epicondylitis. Nevertheless, in both studies, there is value in assessing where we are studies the increase in the risk for epicondylitis in regards to the epidemiologic evidence of in the high-exposure group does not seem causal inference. related primarily to age, independent of intensity and duration of exposure. There is epidemiologic evidence for the Furthermore, the incidence of elbow MSDs relationship between forceful work and unlike most MSDs, has been found to decrease epicondylitis. Those studies that base their after exposure assessment on quantitative or retirement age, after peaking during the fourth semiquantitative data have shown a solid and fifth decades. relationship. We conclude that there is insufficient evidence for the association of Many of the studies controlled for several repetitive work and epicondylitis. For extreme possible confounders in their analyses. In posture in the workplace, the epidemiologic general, for epicondylitis, psychosocial factors evidence thus far is also insufficient, and we or gender do not appear to be important turn to the sports medicine literature to assist us confounders in occupational studies. in evaluating the risk of the single factors of repetition and posture. The strongest evidence CONCLUSIONS by far when examining the relationship between The epidemiologic studies reviewed in this work factors and epicondylitis is the section focused principally on the risk of combination of factors, especially at higher epicondylitis in workers performing repetitive levels of exposure. This is consistent with the job tasks requiring forceful movements. These evidence that is found in the biomechanical and forceful movements included, but were not sports literature. limited to, repeated dorsiflexion, flexion, pronation, and supination, sometimes with the Most of the relevant occupational studies were arm extended. Clinical case series of cross-sectional; the current estimates of the occupationally-related epicondylitis and studies level of exposure were used to estimate past of epicondylitis among athletes had suggested and current exposure. Despite the cross- that repeated forceful dorsiflexion, flexion, sectional nature of the studies, it is likely, in our pronation, and supination, especially with the opinion, that the exposures predated the onset arm extended, increased the risk of of disorders in most cases. epicondylitis. In general, the epidemiologic studies have When we examine all of the studies, a majority of studies are positive. The association between not quantitatively measured the fraction of forceful and repetitive work involving forceful hand motions most likely to contribute dorsiflexion, flexion, supination, and pronation to epicondylitis; rather, they have used as a of the hand is definitely biologically plausible. surrogate qualitative estimation the presence or These motions can cause the contraction of the absence of these types of hand movements muscle-tendon units that attach in the area of [Viikari-Juntura et al. 1991b]. Although we the medial and lateral epicondyles of the elbow.

4-18 The evidence for a qualitative exposure- that cause contractions of the muscles whose response relationship overall was considerable tendons insert in the area of the lateral and for the combination of exposures, with studies medial epicondyles of the elbow. While the examining differences in levels of exposure for studies do not identify the number or intensity of the elbow, and corresponding evidence for forceful contractions needed to increase the greater risk in the highly exposed group. In risk of epicondylitis, the levels are likely to be contrast, we found one study with clear substantial. Future studies should focus on the differences in exposure and no evidence of an types of forceful and repetitive hand motions increase in risk [Viikari-Juntura et al. 1991b]. such as forceful dorsiflexion, pronation, and supination that result in forceful contractions of In summary, the combination of the biological the muscle tendon units that insert in the area of plausibility, the studies with more quantitative the lateral and medial epicondyles. Common evaluation of exposure factors finding strong non-occupational activities, such as sport associations, and the considerable evidence for activities, which cause epicondylitis should be the occurrence with combinations of factors at considered. Older workers may be at some higher levels of exposure provide evidence for increased risk. Finally, even though the the association between repetitive, forceful epidemiologic literature shows that many work and epicondylitis. There are several affected workers continue to work with definite important qualifications to this conclusion. symptoms and physical findings of epicondylitis, Forceful and repetitive work is most likely a survivor bias should be addressed. surrogate for repetitive, forceful hand motions

4-19 Table 4-1. Epidemiologic criteria used to examine studies of elbow MSDs associated with repetition

Risk Investigator indicator (OR, Participation Physical blinded to case Basis for assessing elbow Study (first author and PRR, IR or p- rate $70% examination and/or exposure exposure to repetition year) value)*,† status

Met at least one criterion:

Andersen 1993a 1.7 Yes No Yes Job titles or self-reports

Baron 1991 2.3 No Yes Yes Observation or measurements Burt 1990 2.8† Yes No Yes Job titles or self-reports

Byström 1995 0.74 Yes Yes No Job titles or self-reports

McCormack 1990 0.5–1.2 Yes Yes NR‡ Job titles or self-reports

Met none of the criteria:

Ohlsson 1989 1.5–2.8 NR No NR Job titles or self-reports

Punnett 1985 2.4† No No NR Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported.

4-20

Table 4-2. Epidemiologic criteria used to examine studies of elbow MSDs associated with force

Risk indicator Investigator (OR, PRR, IR or Participation Physical blinded to case Basis for assessing Study (first author and p-value)*,† rate $70% examination and/or exposure elbow exposure to force year) status Met all four criteria:

Chiang 1993 6.75† (males) Yes Yes Yes Observation or 1.44 (females) measurements

Luopajärvi 1979 2.7 Yes Yes Yes Observation or measurements Moore 1994 5.5† Yes Yes Yes Observation or measurements

Met at least one criterion:

Andersen 1993a 1.7 Yes No Yes Job titles or self-reports

Baron 1991 2.3 No Yes Yes Observation or measurements

Byström 1995 0.74 Yes Yes No Job titles or self-reports

Dimberg 1987 NR†,‡ Yes Yes NR Observation or measurements

Dimberg 1989 NR Yes Yes NR Observation or measurements Kurppa 1991 6.7† Yes Yes NR Observation or measurements Punnett 1985 2.4† Yes No NR Job titles or self-reports

Ritz 1995 1.4–1.7† NR Yes Yes Observation or measurements Roto 1984 6.4† Yes Yes Yes Job titles or self-reports

Viikari-Juntura 1991b 0.88 Yes Yes NR Observation or measurements

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

4-22

Table 4-3. Epidemiologic criteria used to examine studies of elbow MSDs associated with posture

Physical Investigator Risk examination blinded to indicator Participation or medical case and/or Basis for assessing Study (first author and (OR, PRR, IR rate $70% records exposure elbow exposure to year) or p- status posture value)*,†

Met all four criteria:

Luopajärvi 1979 2.7 Yes Yes Yes Observation or measurements

Moore 1994 NR‡ Yes Yes Yes Observation or measurements

Met at least one criterion:

Dimberg 1987 NR† Yes Yes NR Observation or measurements

Dimberg 1989 NR Yes Yes NR Observation or measurements

Hoekstra 1994 4.0† Yes No Yes Job titles or self-reports

Hughes 1997 37.0† No Yes NR Observation or measurements

*Some risk indicators are based on a combination of risk indicators—not on posture alone (e.g., posture plus repetition, force, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

4-24

Table 4-4. Epidemiologic criteria used to examine studies of elbow MSDs associated with vibration

Risk Physical Investigator indicator examination blinded to case (OR, PRR, IR Participatio or medical and/or Basis of assessing elbow Study (first author and or n rate $70% records exposure exposure to vibration year) p-value)*,† status

Met at least one criterion: Bovenzi 1991 4.9† NR‡ Yes Yes Observation or measurements

*Some risk indicators are based on a combination of risk indicators—not on vibration alone (e.g., vibration plus repetition, force, or posture). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). † Indicates statistical significance. ‡ Not reported.

4-26

Table 4–5. Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Andersen Cross- 424 female sewing Outcome: Questionnaire: 4.5% 2.6% 1.7 0.9-3.3 Participation rate: 78.2%. and Gaardbo sectional machine operators, continuous pain lasting > 1 e compared to month since starting career; Examiners blinded to 1993a 781 females from the pain for > 30 days. control/subject status. general population of the region and an Exposure: Job categorization Adjusted for age, number of internal referent based on “authors’ children, exercising, smoking, group of 89 females experiences” as occupational socioeconomic status. from the garment health physicians and industry. involved crude assessment of exposure level and exposure repetitiveness. Jobs involving high repetitiveness (several times/min) and low or high force, and jobs with medium repetitiveness (many times/hr) combined with high force were classified as high exposed jobs; jobs with medium repetitiveness and low force and jobs with more variation and high force were classified as medium exposed. Job titles such as teachers, self-employed, trained nurses, and the academic professions were “low exposed.” Exposure also measured as years as sewing machine operator.

(Continued)

4-27

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Baron et al. Cross- Grocery checkers using Outcome: Self-administered 8% among Õ 2.3 0.5-11 Participation rate: 85% 1991 sectional; laser scanners (n=124, questionnaire and physical checkers checkers; 55% non-checkers in case- 119 females, 5 males) exam. Case defined as the field study. Following telephone referent compared to other presence of pain, numbness, survey 91% checkers and 85% grocery store workers tingling, aching, stiffness or non-checkers. (n=157, 56 females, 101 burning in the elbow region males); excluded 18 as previous non-occupational Examiners blinded to worker’s workers in meat, fish, injury; symptoms must have job and health status. and deli departments, begun after employment at workers under 18, and the supermarket of Age, hobbies, second jobs, pregnant workers. employment and in the systemic disease and height current job, and last >1 week were considered as covariates or occurred once a month in the multivariate analyses. within the past year. Total repetitions/hr ranged from Physical Exam: Tenderness 1,432 to 1,782 for right hand at the lateral/medial and 882 to 1,260 for left hand. epicondyle and pain with palpation and resisted motion. Average forces were low and peak forces medium. Exposure: Based on job category, estimates of No statistical significance repetitiveness, average and associated between duration of peak forces based on employment as a checker and observed and videotaped elbow MSDs. postures, weight of scanned items, and subjective Multiple awkward postures of assessment of exertion. all upper extremities recorded but not analyzed in models. The majority of cashiers were categorized as having Statistically significant increase “medium” levels of repetition in elbow MSD with increase in for the hand (defined in this hr/week “checking.” study as making 1250 to 2500 hand movements/hr).

(Continued)

4-28

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi et al. Cross- Vibration-exposed Outcome: Epicondylitis 29.3 6.4% For vibration Participation rate: Not reported. 1991 sectional forestry operators using syndrome: Pain at the exposed chain-saws (n=65) and epicondyle either during rest group Analysis controlled for age and maintenance workers or motion, local tenderness at >7.5 m/s 2: ponderal index. (n=31, control group). the lateral or medial OR=4.9 1.27-56 epicondyle; pain during (adjusted) Controls found to have several resisted flexion/extension of risk factors for MSDs at work- the fingers and wrist with the OR=5.99 static arm and hand overload, elbow flexed, palpated local (unadjusted) overhead work, stressful tenderness at the postures, non-vibrating hand lateral/medial epicondyle. tool use.

Exposure: Direct observation Controls actually had a greater of awkward postures, proportion of the time in work manual forces and cycles shorter than 30 sec than repetitiveness evaluated via forestry workers. checklist. Vibration measured from two chain saws.

(Continued)

4-29

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Burt et al. Cross- Newspaper employees Outcome: Self administered Male: 11% Õ 80% to 100% Participation rate: 81%. 1990 sectional (n=836, females=55%). questionnaire. Case defined Female: 14% time typing Workers fulfilling case as the presence of pain, compared to Analysis controlled for age, definitions compared to numbness, tingling, aching, 0% to 19%: gender, years on the job. those who did not fulfill stiffness, or burning in the OR=2.8 case definition. elbow region as previous 1.4-5.7 Psychosocial factors dealing non-occupational injury. Reporters with job control and job Symptoms began after compared to satisfaction were addressed in starting the job, last > 1 week others: questionnaire. or occurred once a month OR=2.5 within the past year; reported 1.5-4.0 Job analysis found significant as “moderate” (3) or greater correlation (0.56) between on a 5-point scale. reported average typing time/day and observed 8 hr Exposure: Based on period of typing (p < 0.0001). observation of job tasks, then categorized by job title. A Reporters were characterized separate job analysis using a by high, periodic demands checklist and observational (deadlines), although they had techniques was carried out high control and high job for validating questionnaire satisfaction. exposure data. Number of workers in some non- typing jobs not reported.

Case definition based on symptoms alone.

(Continued)

4-30

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Byström Cross- Automobile assembly Outcome: Epicondylitis was Tender Tender PRR for Participation rate: 96%. et al. 1995 sectional line workers (n=199) defined as tenderness to lateral lateral tender lateral Comparison group is from the compared to a randomly palpation of the lateral or epicondyle: epicondyle: epicondyle: MUSIC study (Hagberg and selected group from the medial epicondyle and pain at 4.3% 12.4% 0.74 0.04-1.7 Hogstedt, 1991). general population the same epicondyle or in the (n=186). The forearm extensors or flexors Epicondy- Epicondy- Examiners were blinded to automobile assembly on resisted wrist extension litis: 0 cases litis: 1% questionnaire responses but line workers were or flexion. not exposure status. randomly selected from a primary group of Exposure: No evaluation of Analysis stratified by gender 700 assembly line repetition, force, posture, or and age <40 years. workers. These vibration occurred in this Psychosocial variables and original 700 workers study to evaluate risk factors other potential confounders or had been randomly for epicondylitis. “Assembly effect modifiers were selected from the line worker” vs. “Population addressed by Fransson-Hall 2,334 assembly referent” was used. Hand et al. [1995]. workers of a Swedish grip strength was evaluated. automobile factory. Forearm muscular load and Pain-pressure threshold (PTT) wrist angle were evaluated was evaluated. PTT was not for a subgroup in this related to age. It was higher population but were not used among women with short in this analysis [Hägg et al employment compared to those 1996]. who had been employed for a long time.

No correlation was found between low MCV and subjective or objective signs.

(Continued)

4-31

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Chiang et al. Cross- 207 fish processing Outcome: Prevalence of Group II: Group I: Crude ORs Participation rate: Authors 1993 sectional workers, 67 males and lateral or medial epicondylitis 15% 10% calculated reported: “In order to prvent 140 females, divided (local tenderness, pain in Male: 10%; Male: 6%; from data selective bias all employees in in 3 groups: (I) low resisted extension or flexion Female: 17% Female: 14% presented: the fatories were observed force, low repetition of the wrist and fingers, Group II vs. initially.” (comparison group, decreased hand grip strength Group III: Group I, n=61); (II) high force compared to the opposite 21% males: Workers examined in random or high repetition hand). Male: 33%; OR=1.7 0.3-9.2 sequence to prevent observer (n=118); (III) high Female: 18% bias, examiners blinded to case force and high Exposure: Assessed by Group II vs. status. repetition (n=28). observation and recording Physician Group I, of tasks and biomechanical observed females: Analysis stratified by gender. movements of three workers, epicondy- OR=1.2 0.4-3.4 No significant age difference in each representing one of litis, all exposure groups. 3 study groups. Highly cases: Group III vs. repetitive jobs with cycle time 14.5 % Group I, Logistic regression not <30 sec or >50% of males: performed for epicondylitis cycle-time performing the OR=6.75 1.6-32.7 because of lack of significant same fundamental cycles. trend with increasing exposure. Hand force from EMG Group III vs. recordings of forearm Group I, Workers with hypertension, flexor muscles. females: diabetes, history of traumatic Classification of workers into OR=1.44 0.3-5.6 injuries to upper limbs, arthritis, 3 groups according to or collagen diseases excluded the ergonomic risks of the from study group. shoulders and upper limbs: Group I: low rep. and low Physician observed cases had force; Group II: high repetition about ½ the prevalence of or high force; Group III: high symptoms of elbow pain (9.8 repetition and high force. vs. 18.0; 5.3 vs. 19.5; 35.7 vs. 17.9).

No dose-response for elbow pain or physician observed epicondylitis.

(Continued)

4-32

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Dimberg 1987 Cross- A questionnaire was Outcome: Only workers Lateral Õ Epicondylitis, Participation rate: 98.9%. sectional distributed to every fifth reporting elbow problems humeral blue vs. Physician blinded to exposure person in the automobile were examined by the epicondylitis white collar status: not reported. workers: company’s personnel physician. Physical exam: among all 0.7 0.3-1.2 Results age stratified. file selected by random case defined as physical subjects: Distribution of Physician-consulted elbow pain numbers. Final sample findings of lateral elbow pain 7.4% significantly greater in jobs with consisted of and pain with palpation over epicondylitis increased elbow stress. cases by 546 workers, 494 males lateral epicondyle and pain Blue collar type of work Work considered to be the and 52 females. increase with dorsiflexion of workers: stress: cause in 35%. Authors found (25 were excluded due wrist with resistance. 5.3% that work-related group had Leisure work defined by high stress to military service, related (categorized by low, moderate, pregnancy, or study Exposure: Observation of White collar and high) compared to leisure- epicondylitis: related epicondylitis and away). the work site then workers: low work epicondylitis of no-known- categorization of jobs “with 11% stress: 85%; cause. respect to elbow stress” by medium work Authors reported that a Physical Work Stress Blue collar: stress: 15%; proportion of workers who Group composed of a under age 40 high work consulted a physician for their stress: 0% elbow problems was physician, physiotherapist, years: 4.6% significantly greater with and safety engineer. Table 2 No-known- increasing elbow stress (p < in the article lists types of Blue collar: cause group: 0.05). epicondylitis: jobs with respect to over age 40 low work Multiple regression analyses subjects’s elbow stress. years: 8.9% included gender, employee stress: 75%; category, age, and degree of medium work stress as independent White collar: stress: 25%; variables—only age significantly under age 40 high work related to prevalence. years: 6.1% stress: 0% Overexertion of the extensor Work-related muscles of the wrist due to gripping and twisting White collar: epicondylitis: movements prior to onset was over age 40 low work verified in 28 (70%) of those stress: 14%; with epicondylitis. years: medium work 13.9% Tennis players among stress: 36%; “sufferers”: 15% total high work population: 12%. All racquet stress: 50% sports: 20% among sufferers, 15% among total population.

(Continued)

4-33

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Dimberg et al. Cross- 2,814 automotive Outcome: Questionnaire Blue collar White collar Univariate Participation rate: 96%. 1989 sectional workers, both blue- results of elbow trouble Results: Not stated whether examiner and white-collar (pain, ache, discomfort) p<0.001: blinded to exposure status. workers: 2,423 males, preventing normal work in higher age; Multivariate analysis performed, 382 females. last 12 months. longer time in present job; although the confounders controlled for were not stated Physical exam performed on ponderal index, more by authors, nor were ORs 615 of 641 symptomatic symptoms; presented. Vibrating tools, workers. Epicondylitis: more mental ponderal index, and mental tenderness at the stress at the stress at work listed as lateral/medial epicondyle onset of significant. and pain with resistance. symptoms. Guidelines for classification of p<0.05: jobs as listed in the article do Exposure: Observation of salaried staff not seem to reflect increasing vs. others; elbow stress. Group 1 includes jobs, then classification into 3 “repeated rotation of the Physical Work Stress Groups heavy forearms and wrists occurs by physician, weight; less racquet sporadically”; Group 2 includes physiotherapist, and safety sports, more less specifically “large and engineer. Guidelines for symptoms. frequent rotations in extreme classification with respect to positions”; Group 3 does not the strain on the subject’s p<0.01: include any reference to vibrating repeated rotation or extreme neck and upper extremities hand tools, position of the forearms or listed for light, moderately more wrists. The classification used heavy, and heavy work symptoms; seems unlikely to pick up included in article. time in increased elbow stress that present job, would reflect higher strain and more risk of epicondylitis. symptoms. Increased ponderal index p>0.05: correlated with elbow gender; strain symptoms in multivariate group; full analysis. time; hrs/week; Mental stress at work with the piece-work; onset of symptoms correlated fixed pay; with right-sided lateral smoking, epicondylitis. house- Mental stress variables not owner. uniformly collected, so this may impact interpretation.

(Continued)

4-34

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Fishbein et al. Cross- 2212 musicians Outcome: Outcome based on 10% right Õ Severe Participation rate: 55%. Low 1988 sectional performing on a regular self-reported responses from elbow: 6 % medical response rate due to the fact (mailed basis with one or more survey. Self-reported elbow severe problem and that many orchestras were not survey) of the International pain, with severity defined its affect on in season at the time of the Conference of in terms of the effect of 8% left performance, survey. Symphony and Opera the problem on the musician’s elbow: 4% females vs. Musicians (ICSOM). performance. severe males: Statistical weighting performed; Total population of the OR=2.04 1.6-2.6 "severe" pain was defined as membership was 4,025 Exposure: Questionnaire pain that affects performance. musicians in 48 ICSOM responses to orchestral orchestras. One instrument, age they began Health habits, such as extent of orchestra did not playing, age they joined the exercise, use of cigarettes, participate. orchestra, number of weeks alcohol, beta blockers, and each year spent playing other drugs. professionally. Average age beginning playing instrument is 10 years. Average age joining a professional orchestra is 23 years. Average age: male musicians–43 years, female musicians–40 years.

Severe problems were more likely in ages under 35 than over 45 years. Authors speculated that musicians with severe problems leave the orchestra.

Low participation rate limits interpretation.

(Continued)

4-35

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Hales et al. Cross- 518 telecommunication Outcome: Pain, aching, 7% Õ Fear of being Participation rate: 93%. 1994 sectional workers (416 females stiffness, burning, replaced by and 117 males). numbness, or tingling >1 computers: ORs for psychosocial represent Workers fulfilling week or >12 times a year; OR=2.9 1.4-6.1 risk at scores one standard outcome definition occurring after employment deviation (SD) above the mean compared to those not on current job within the last Lack of compared to risk at scores one fulfilling outcome year and positive physical decision- SD below mean. May be a definition. examination (PE): Moderate making problem with non-normal to worst pain experienced opportunities: distribution. with medial or lateral OR=2.8 1.4-5.7 epicondyle palpation. Analysis controlled for age, Surges in gender, individual factors, and Exposure: Assessed by workload: number of keystrokes/day. questionnaire. Questions OR=2.4 1.2-5.0 addressed number of Physician examiners blinded to overtime hr, co-worker use Race (non- case and exposure status. of same workstation, task white) rotation, hr spent at the OR=2.4 1.2-5.0 Although keystrokes/day was (VDT) workstation, hr spent not significant–workers only typing, number and types of typed average of 8 words/min work breaks, length of time over 8-hr period. sitting, frequency of arising from a chair, number of 97% of workers “used” VDTs $ keystrokes estimated for 6 hr/day–not enough variance each directory assistance to adequately evaluate hr operator. typing.

Number of hr on hobbies and recreation not significant.

Over 70 variables analyzed in models–may have multiple comparison problem.

(Continued)

4-36

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Hoekstra Cross- 108 of 114 teleservice Outcome: Self administered Center A 19% Õ Participation rate: 95%. et al. 1994 sectional representatives working questionnaire. Case defined at 2 government as the presence of pain, "Non- Analysis controlled for gender. administration centers: numbness, tingling, aching, optimally" A and B. stiffness, or burning in the adjusted Interactions evaluated. elbow region as previous Center B 21% chair: 4.0 1.2-13.1 non-occupational injury; Variables considered in logistic symptoms began after model included location, age, starting the job, last > 1 week seniority, hr spent typing at or occurred once a month VDT, hr on the phone, 3 chair within the past year; reported variables: (1) Perceived as “moderate” (3) or greater adequacy of chair adjustment, on a 5-point scale. VDT screen, (2) Perceived adequacy of keyboard Exposure: Measurement and adjustment, VDT screen, evaluation of work station; (3) Perceived adequacy of desk observation of postures to adjustment, job control, provide descriptive workload variability. differences between the two locations. Linear regression also performed on psychosocial variables in separate models for job dissatisfaction and exhaustion.

Center B generally had nonadjustable chairs and work stations. Authors noted elevated arms, hunched shoulders and other "undesirable" postures.

Did not include non-work- related variables in analyses.

(Continued)

4-37

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Hughes and Cross- 104 male aluminum Outcome: Symptoms 11.6% with Õ Model based Participation rate: Carbon Silverstein sectional smelter workers: occurring in the positive on MSD setters: 65%; crane operators: 1997 62 carbon setters, elbow/forearm > symptoms defined by 56%; carbon plant: 33%. 36 crane operators, once/month or lasting and physical symptoms 9 carbon plant workers. longer than one week in exam and physical Examiners blinded to exposure There were 14 workers the previous year, no exam and health status: not stated. who were not from acute or traumatic onset; 24% had selected jobs and were occurrence since working symptoms in Age: Analysis controlled for age, excluded. at the plant, no systemic the OR=0.96 0.9-1.2 smoking status, sports, and/or disease. elbow/forear hobbies. m in the Low decision Physical examination: Active, previous latitude: Psychosocial data collected passive, and week OR=3.5 0.6-19 individually; physical factors resisted motions, pinch based on estimates of each job. and grip strength, 128 Years of Hz vibration sensitivity, two- forearm Job risk factors entered into the point discrimination. twist: OR=37 model for hand/wrist included: 3.0-470 (1) the number of years Psychosocial scales from Model based handling > 2.7 kg/hand, questionnaire based on on MSD (2) push/pull, (3) lift/carry, Theorell and Karasek Job defined by (4) pinching, (5) wrist Stress Questionnaire, and symptoms flexion/extension, (6) ulnar on Work Apgar deviation, and (7) forearm Questionnaire. Age: twisting. OR=0.96 Exposure: For carbon 0.9-1.2 Health interview included setters and crane Years of information about metabolic operators (non-repetitive ulnar diseases, acute traumatic jobs) a modified job- deviation: injuries, smoking, hobbies. surveillance checklist OR=0.005 0.0-16 method was used. Job task Low participation rate limits analysis used a formula Years interpretation. based on the relative forearm frequency of occurrence twist: OR=4 0.18- 4 of postures during (a) task(s).

(Continued)

4-38

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kopf et al. Cross- Bricklayers (n=163) Outcome: Questionnaire Not reported Not reported Painful left Not Participation rate: bricklayers: 1988 sectional compared to other based, self-reported elbow, reported 65%, manual workers: 69%. manual workers (n=144) symptoms. Self-reported bricklayers employed by state pain in the elbow. vs. other Controlled for confounders: agencies in Hamburg, manual age, job satisfaction, job Germany. Exposure: Based on job workers: security, vibration, moistness, categories, bricklayer vs. OR=2.8 Scheuerman’s disease. other manual laborers. Physical stress of bricklayers Karasek’s model of job latitude described as lifting and and job demands were included carrying bricks weighing 5 to in the questionnaire. 24 kg up to 100 times/hr with the left hand and handling the Physically demanding previous bricklayer’s trowel with the tasks, medical disposition for right hand. MSD, being a member of a trade union included in analysis.

64% attributable risk proportion of elbow pain is explained by being a bricklayer.

For increasing levels of job demands (heavy physical work, awkward working positions, repetitive movements, and restriction in standing position), OR increased from 1.8 to 3.4.

(Continued)

4-39

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kurppa et al. Cohort; 31 Sausage makers (107 Outcome: Tenderness to Sausage Workers in IR of males in Participation rate: 93% of 1991 month females) compared to palpation of the epicondyle makers Non- strenuous strenuous workers retained during study; 90% of follow-up nonstrenuous jobs (197 and epicondylar pain (females): strenuous jobs vs. nonstrenuous workers. females). provoked by resisted 11.1 jobs: 1.1 nonstrenuous extension or flexion of the cases/100 cases /100 jobs: 5.7 Examiners blinded to exposure Meatcutters (102 males) wrist and fingers with the person- person- or past episodes: not reported. Diagnoses made by different compared to elbow extended. Incidence years years IR of females physicians at different nonstrenuous jobs based on visits to doctor in strenuous locations. Plant physicians (n=141). during 31 month visit. Workers in jobs vs. agreed to the diagnostic criteria Meatcutters non- nonstrenuous and made 75% of diagnoses. 25% of physicians were not Packers (118 females) Disease considered "new" (males): strenuous jobs: 8.1 involved in agreement of compared to episode if new sick leave 6.4 jobs: 0.9 diagnostic criteria. 13% of nonstrenuous jobs (197 with same diagnosis cases/100 cases/100 IR of total epicondylitis diagnosed by females). occurred at same anatomic person- person- number of consulting specialists at the nearby medical center, 12% site within 60 days after end years years cases of elsewhere, usually at municipal of former sick leave. epicondylitis health centers. Workers in in strenuous Exposure: Data obtained Nonstrenu- jobs vs. No adjustment for confounders, but referent group selected from “previous published Packers ous jobs: 1.1 nonstrenuous similar to strenuous group with literature” and walkthrough. (males): cases/100 jobs: 6.7 3.3-13.9 regards to age, gender, and 7.0 person- duration of employment, except “Cutting of veal (appx. 1,200 cases/100 years for male sausage makers and male packers who were kg/day) or pork (appx. 3,000 person- younger than the rest of the kg/day) (meatcutters); years study population–these were spraying the sausages and excluded from calculations of hanging them on bars incidence rates. (sausage makers); peeling “New" episode of epicondylitis sausages, inserting them into may be recurrence of same slicing machine, setting the disease. 12 employees slices into packages, setting reafflicted with epicondylitis with median of 184 days packages on a conveyor belt, between episodes. collecting finished packages into bags; room temperature There were 68 diagnoses of 8E to 10E (packers); epicondylitis among 57 nonstrenuous tasks included individuals. primarily office work.”

(Continued)

4-40

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Luopajärvi et Cross- Female assembly line Outcome: Epicondylitis 5.9% 2.3% 2.7 0.66- Participation rate: 84%. al. 1979 sectional workers (n=152) diagnosed by interview and 15.9 Workers excluded from compared to female physical exam. participation for previous shop assistants in a trauma, arthritis and other department store Symptoms include muscle pathologies. (n=133). Cashiers pain during effort, local excluded from swelling, and local ache at Examiner blinded to case comparison group. rest. Signs include status: yes, according to the tenderness at the ateral or Waris et al. 1979, epidemiologic medial epicondyle on screening procedure, which palpation, pain during was used in study. resisted extension/flexion of the wrist and fingers with the No association between age elbow extended. and MSDs or length of Physiotherapist examined employment and MSDs. Gender workers, diagnoses were not an issue because study from pre-determined criteria population was all female. (Waris 1979). In problem cases orthopedic and Factory opened only short time physiatric teams handled so no association between cases. duration of employment and MSDs possible. Exposure: Exposure to repetitive work, awkward Social background, hobbies, hand/arm postures, and amount of housework not static work assessed by significant. observation, video analysis and interviews. Video recordings showed repetitive motins of the hands and fingers up to 25,000 cycles/day, static muscle loading of the forearm muscles, and deviations of the wrist, lifting.

(Continued)

4-41

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments McCormack Cross- Randomly selected Outcome: Based on Boarding Non-office Boarding vs. Participation rate: 91%. et al. 1990 sectional population of 2,261 physician administered workers: workers: non-office: textile workers from physical exams. 1.0% 1.9% OR=0.5 0.09-2.1 Physician or nurse examiners not blinded to case or exposure 8,539 eligible workers; Reproducible tenderness status (personal 4 groups compared with with direct pressure on the Sewing Sewing vs. communication). 468 non-office workers lateral epicondyle. Severity workers: non-office: graded as mild, moderate, 2.1% OR=1.1 0.4-2.9 Age, gender, race, and years Manufacturing workers: and severe. of employment analyzed. Packaging/ Packaging vs. A. Packaging/folding Exposure: Assessment by folding non-office: Prevalence higher in workers with < 3 years of employment. workers (41 males, observation of jobs. workers: OR=1.1 0.4-3.2 238 females). Exposure to repetitive finger, 2.2% Questionnaire asked types of wrist and elbow motions jobs, length of time on job, B. Sewing workers assumed from job title; no Knitting: Knitting vs. production rate, nature and type (28 males, 534 females). objective measurements 1.4% non-office: of upper extremity complaint, performed. OR=1.2 0.5-3.4 and general health history. C. Non-office workers 11 physician examiners; (204 males, 264 interexaminer reliability potential females). problem acknowledged by authors. D. Boarding workers (19 males, 277 females). Epicondylitis significantly associated with years of employment, age, race. Job category not related to epicondylitis, however no measurement of force, repetition, posture analysis, etc. Of 37 cases of epicondylitis identified: 13 were categorized as mild, 22 were moderate, and 2 were severe.

(Continued)

4-42

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Moore and Cross- Workers employed in Outcome: OSHA logs Workers in Workers in Odds of Participation rate: Cases Garg 1994 sectional 32 jobs at a pork verified by medical records “hazardous “safe jobs”: epicondylitis identified from medical records. processing plant data for 20 months. jobs”: 23% 3% in workers in Jobs analyzed from (n=230). Epicondylitis: localized elbow “hazardous observational methods. pain that increased with jobs” Investigators blinded to Workers in jobs tension of muscle-tendon unit compared to exposure, case outcome classified as and direct palpation. A case workers in status, and personal identifiers “hazardous” compared required that a physical “safe jobs”: on medical records. to those in “safe” jobs. examination specific to OR=5.5 1.5-62 Repetitiveness and “type of epicondylitis was performed. (based on grasp” were not significant personal factors between hazardous- Exposure: Observation and communi- and safe-job categories. video analysis, semi- cation) quantitative methods using No pattern of morbidity accord- motion and time methods ing to date of clinic visits. (MTM), force estimated as % Strength demands significantly maximal strength (5 levels), greater for hazardous job wrist posture (3 levels), type categories compared to safe. of grasp (2 levels), high IR based on full-time equivalents speed work (yes or no), and not individual workers, may localized mechanical have influenced overall results. compression (yes or no), vibration (yes or no), and Workers had a maximum of cold (yes or no). Observed 32 months of exposure at videotaped representative plant–duration of employment worker in each job. analysis limited. Repetition as cycle-time and Duration of exposure not exertions/min measures. collected on study sample. Jobs classified as Average maximal strength "hazardous" or "safe" based derived from population-based on data, experience of data stratified for age, gender, authors, and judgements. and hand dominance. Using estimates of Silverstein’s Work histories, demographic, classification, association pre-existing morbidity data between forcefulness, and not collected on each overall observed morbidity was participant. statistically significant; repetition was not. 31 of 32 jobs were in high repetitive category–no variance to find difference.

(Continued)

4-43

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ohlsson Cross- Electrical equipment and Outcome: Questionnaire: Elbow pain in Elbow pain Participation rate: Not reported. et al. 1989 sectional automobile assemblers Any elbow pain, elbow pain last 12 in last 12 (n=148), former female affecting work ability, months: 21% months: Work pace assessed by assembly workers who and elbow pain in the last 17% 1.5 0.6-3.4 questionnaire, the number of quit within 4 years seven days and the last Elbow pain in items completed/hr. (n=76) compared to 12 months. last 7 days: Elbow pain randomly sampled 14% in last 7 No association between length females from general Exposure: No exposure days: 11% 1.9 0.7-5.3 of employment and elbow population (n=60). measurements; based on job Work inability symptoms. categorization. in last 12 Work months: 10% inability in No statistical significance Work pace divided into last 12 associated with work pace 4 classes: months: 3% 2.8 0.8-10.7 (data not present). (1) Slow <100 items/hr; (2) Medium 100 to 199 Logistic models evaluated for items/hr; (3) Fast 200 to 700 interaction and controlled for items/hr; (4) Very Fast >700 age. items/hr. Study group consisted of females only.

(Continued)

4-44

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Punnett et al. Cross- 162 female garment Outcome: Self-administered Garment Hospital Elbow Participation rate: 97% 1985 sectional workers, 85% were questionnaire concerning workers: employees: Symptoms in (garment workers), 40% employed as sewing symptoms 6.5% 2.8% Garment (hospital workers). machine operators and workers vs. sewing and trimming by Cases defined as the Hospital Analysis stratified for number hand. presences of persistent employees: of years employed, decade of elbow pain, numbness or OR= 2.4 1.2-4.2 age, native language. Comparison: 76 of 190 tingling (lasted for most full or part-time workers days for one month or more Health outcome based on on day shift in a hospital within the past year); were Persistent symptoms alone for elbow who worked as nurses not associated with previous elbow pain in MSDs. or aids; lab technicians injury; and, began after finishers vs. or therapists; food first employment in garment hospital Age and length of employment service workers. manufacturing or hospital employees: not a predictor of risk of elbow employment. Key questions OR=5.6 MSDs. Employees typing based on the arthritis >4 hr/day excluded supplement questionnaire of Persistent Prevalence of pain not from comparison group. National Health and Nutrition elbow pain in associated with years of Examination Survey underpresser employment in garment (NHANES). vs. hospital workers. employees: Exposure: Self-administered OR=5.0 Non-English speakers questionnaire; # of years in significantly less likely to report the industry, job category, pain (RR 0.6 ; p<0.05). previous work history. Native English speakers significantly older than non- native English speakers (p<0.03).

(Continued)

4-45

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Ritz 1995 Cross- 290 males from the Outcome: Physician 41 10 years of Participation rate: Not reported. sectional public gas and water diagnosed; required local employees: high works of Hamburg, tenderness to palpation at the 14% had exposure to Examiner blinded to exposure Germany examined epicondyle and pain during epicondylitis elbow status. during routine medical resisted movement of the straining Logistic regression model check-up at the wrist and fingers (extension 11% fulfilled work for controlled for age, age- company occupational or flexion of the wrist or Waris’s currently held squared, and an indicator term health center. fingers with an extended criteria for job: OR=1.7 1.0-2.7 for “history of cervical spine Employees, excluded if elbow) AND elbow pain epicondylitis symptoms” (yes, no). on sick leave, came for during the lifting of a chair. (Waris, High The following variables tested medical treatment, pre- Epicondylitis was 1979) exposure to for confounding: having ever employment checkups, categorized as severe elbow played tennis, squash, other or to file a worker’s (Grade II and Grade III) if both straining racquet sports, rowing, compensation claim. functional tests were positive work for bowling, the duration of having and as moderate (Grade I) if formerly held played these sports, injuries only symptom was a severe job: involving the elbow joint, ponderal index, handedness, tenderness to palpation or a OR= 2.16 1.1-4.3 and former surgical treatment moderate pain in the for epicondylitis. resistance test. Clinical signs 10 years of of epicondylitis > Grade 0 at high The variable “time in years one or more of the four exposure to since retiring from a job with anatomical sites was elbow high or moderate exposure” considered sufficient for the straining was retained in the model for workers formerly employed in diagnosis. work for high exposure jobs when currently held duration of exposure was Exposure: All current and job using tricotomized. former job titles evaluated by diagnostic members of the team criteria for Mean length of employment according to possible bio- epicondylitis was not significantly different mechanical strain to the [Waris et al. between cases and non-cases. elbow and grouped into 1979]: Increasing duration of current categories of high, moderate, OR=1.89 1.2-3.1 exposure increased the risk of and non work-related being diagnosed with exposure. Exposure epicondylitis. categorization was based on company job descriptions, interviews with employees, and workplace observations.

Exposure duration was defined for all subjects as the

(Continued)

4-46

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Roto and Kivi Cross- Meatcutters, (n=90) Outcome: Defined by Meatcutters: Construc- 6.4 0.99-40.9 Participation rate: 100% for 1984 sectional compared to physical exam: local 8.9% tion meat cutters, 94% for construction workers tenderness, pain during workers: p= 0.05 construction workers. (n=72) not exposed to resisted extension/flexion of 1.4% repetitive movements. the wrist and fingers, and Authors state that examiners decreased hand grip power were blinded to occupation of in comparison to other hand. subjects because part of larger group of meat processing Exposure: Based on job title workers examined, but it is (meatcutter vs. construction unclear whether construction worker). foremen (referents) were examined separately.

Serologic testing for rheumatoid arthritis was done to control for potential confounding (none detected).

7 additional meatcutters had local tenderness in epicondylar region.

All with epicondylitis had > 15 years of employment.

Authors stated that on average, meatcutters with epicondylitis had been exposed five years longer than other meatcutters, supporting the association with meatcutting.

(Continued)

4-47

Table 4–5 (Continued). Epidemiologic studies evaluating elbow musculoskeletal disorders

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Viikari- Cross- All permanent workers Outcome: Elbow trouble Epicondy- Epicondy- The Odds Participation rate: 94%. Juntura sectional exposed to repetitive (pain, ache, discomfort) litis: 0.8% litis: 0.8% Ratio of 1991b and manually stressful preventing normal work in epicondylitis No adjustment for confounders tasks in a meatpacking last 12 months and physical Lateral: Lateral: in strenuous in analysis. Authors stated that plant (102 meatcutters, exam: tenderness at the 0.6% 0.6% jobs vs. non- the comparison group was 150 packers, and lateral/medial epicondyle and Medial: Medial: strenuous selected similar to the study 125 sausage makers) pain with resistance. 0.2% 0.3% jobs: 0.88 0.27-2.8 group to sex, age, and duration were compared to of employment. 332 workers in Exposure: Based on Elbow Pain nonstrenuous jobs observation: (without the Examiners blinded to case and (supervisors, physical exposure status. maintenance men, Meatcutters: High force/high exam): accountants, and office repetition. Male: 1.8 1.1-2.8 Male packers and male sausage workers). Female: 1.6 1.2-2.3 makers younger and length of Sausage makers: High employment shorter than other repetition/low force with high groups. force tasks. Palpation pressure increased on Packers: High repetition/low 2nd of cross-sectional force with high force jobs. examinations–may have influenced results. Nonstrenuous jobs, mainly office jobs. For female sausage makers, elbow pain for preceding 12 “Cutting of veal (appx. 1,200 months increased with age and kg/day) or pork (appx. 3,000 duration of employment. No kg/day) (meatcutters); such associations in other spraying the sausages and groups. hanging them on bars (sausage makers); peeling Age and current occupational sausages, inserting them into correlated (r=0.52) for female slicing machine, setting the sausage makers. slices into packages, setting packages on a conveyor belt, Cases were not excluded due collecting finished packages to direct trauma. into bags; room temperature 8E to 10E (packers); nonstrenuous tasks included primarily office work.”

4-48 CHAPTER 5 Hand/Wrist Musculoskeletal Disorders (Carpal Tunnel Syndrome, Hand/Wrist Tendinitis, and Hand-Arm Vibration Syndrome): Evidence for Work-Relatedness

Musculoskeletal disorders (MSDs) of the hand/wrist region have been separated into three components for the purpose of this review: (a) Carpal Tunnel Syndrome (CTS), (b) Hand/Wrist Tendinitis, and (c) Hand-Arm Vibration Syndrome (HAVS). Each of these are described with regard to the evidence for causality between workplace risk factors and development of MSDs.

5-1 CHAPTER 5a Carpal Tunnel Syndrome

SUMMARY Over 30 epidemiologic studies have examined physical workplace factors and their relationship to carpal tunnel syndrome (CTS). Several studies fulfill the four epidemiologic criteria that were used in this review, and appropriately address important methodologic issues. The studies generally involved populations exposed to a combination of work factors, but a few assessed single work factors such as repetitive motions of the hand. We examined each of these studies, whether the findings were positive, negative, or equivocal, to evaluate the strength of work-relatedness using causal inference.

There is evidence of a positive association between highly repetitive work alone or in combination with other factors and CTS based on currently available epidemiologic data. There is also evidence of a positive association between forceful work and CTS. There is insufficient evidence of an association between CTS and extreme postures. Individual variability in work methods among workers in similar jobs and the influence of differing anthropometry on posture are among the difficulties noted in measuring postural characteristics of jobs in field studies. Findings from laboratory-based studies of extreme postural factors support a positive association with CTS. There is evidence of a positive association between work involving hand/wrist vibration and CTS.

There is strong evidence of a positive association between exposure to a combination of risk factors (e.g., force and repetition, force and posture) and CTS. Based on the epidemiologic studies reviewed above, especially those with quantitative evaluation of the risk factors, the evidence is clear that exposure to a combination of the job factors studied (repetition, force, posture, etc.) increases the risk for CTS. This is consistent with the evidence that is found in the biomechanical, physiological, and psychosocial literature. Epidemiologic surveillance data, both nationally and internationally, have also consistently indicated that the highest rates of CTS occur in occupations and job tasks with high work demands for intensive manual exertion–for example, in meatpackers, poultry processors, and automobile assembly workers.

INTRODUCTION [Washington State Department of Labor and In 1988, CTS had an estimated population Industry 1996]. These data suggest that about prevalence of 53 cases per 10,000 current 5 to 10 workers per 10,000 workers will miss workers [Tanaka et al. (in press)]. Twenty work each year due to work-related CTS. percent of these individuals reported absence from work because of CTS. In 1994, the In recent years, the literature relating Bureau of Labor Statistics (BLS) reported that occupational factors to the development the rate of CTS cases that result in “days away of CTS has been extensively reviewed from work” was 4.8 cases per 10,000 by numerous authors [Moore 1992; Stock workers. The agency also reported that the 1991; Gerr et al. 1991; Hagberg et al. 1992; median number of days away from work for Armstrong et al. 1993; Kuorinka and CTS was 30, which is even greater than the Forcier 1995; Viikari-Juntura 1995]. Most median reported for back pain cases [BLS of these reviews reach a similar 1995]. In 1993, the incidence rate (IR) for conclusion—work factors are one of CTS workers’ compensation cases was 31.7 the important causes of CTS. One review cases per 10,000 workers; only a minority of [Moore 1992] found the evidence these cases involved time off of work

5a-1 more equivocal, but stated that the of labeling workers as having “CTS” or epidemiologic studies revealed a fairly “median nerve mononeuropathy” based on consistent pattern of observations regarding the abnormal sensory nerve conduction alone spectrum and relative frequency of CTS (without symptoms). The reason for this view is [among other musculoskeletal disorders illustrated in a recent prospective study by (MSDs)] among jobs believed to be hazardous. Werner et al. [1997]. On follow-up six to The epidemiologic studies which form the basis eighteen months after initial evaluation, they for these reviews are outlined in Tables 5a–1 to found that asymptomatic active workers with 5a–4 of this chapter. abnormal sensory median nerve function (by Nerve Conduction Studies [NCS]) were no Thirty studies of occupational CTS are listed on more likely to develop symptoms consistent Tables 5a–5. Twenty-one are cross-sectional with CTS than those with normal nerve studies, six are case-control, and three involve function. Studies which have used nerve a longitudinal phase; all have been published conduction tests for epidemiologic field studies since 1979. We included one surveillance study have employed a variety of evaluation methods [Franklin et al. 1991] because it has been and techniques [Nathan et al. 1988, 1994b; included in many of the earlier reviews. The few Bernard et al. 1993; Osorio et al. 1994]. earlier studies of CTS identified were clinical Normal values for nerve conduction studies case series, or did not identify work place risk have also varied from laboratory to laboratory. factors and were not included in the tables NCS results have been found to vary with related to CTS. electrode placement, temperature, as well as age, height, finger circumference and wrist ratio OUTCOME AND EXPOSURE [Stetson 1993], suggesting that “normal” values MEASURES may need to be corrected for those factors. In four of 30 studies listed in Tables 5a–1 to 5a–4, CTS was assessed based on symptoms Several epidemiologic studies have used a alone; in another nine studies, the case surveillance case definition of CTS based definition was based on a combination of on symptoms in the median nerve distribution symptoms and physical findings. and abnormal physical examination findings Electrophysiological tests of nerve function using Phalen’s test and Tinel’s sign, and have were completed in 14 studies. Electrodiagnostic not included NCS. Two recent studies testing (nerve conduction studies) has been [Bernard et al. 1993; Atterbury et al. 1996] considered by some to be a requirement for a looked at CTS diagnosis based on valid case definition of CTS, as is similarly used questionnaire and physical examination findings for a clinical diagnosis in individuals with CTS. and its association with the “gold standard” of A few studies which have looked at the nerve conduction diagnosed median relationship of occupational factors to CTS mononeuropathy. Both studies found have used a health outcome based on statistically significant evidence to support the electrodiagnostic testing alone [Nathan et al. use of an epidemiologic CTS case definition 1988; Schottland et al. 1991; Radecki 1995.] based on symptoms and physical examination However, some authors [Nilsson 1995; (not requiring NCS) for Werner et al. 1997] have discouraged the use epidemiologic surveillance studies. Nathan

5a-2 [1992a] also found a strong relationship of Repetition and CTS between symptoms and prolonged sensory Nineteen studies reported on the results of the median nerve conduction. (It is important to association between repetition and CTS. note here that a case definition used for Several studies in Table 5a-1 quantitatively epidemiologic purposes usually differs from one measured [Moore 1992; Chiang et al. 1990, used for medical diagnosis and therapeutic 1993; Silverstein et al. 1987] or observed intervention.) [Stetson et al. 1993; Nathan et al. 1988, 1992a; Barnhart et al. 1991; Osorio et al. Researchers have relied on a variety of 1994] and categorized repetitive hand and wrist methods to assess exposure to suspected movements in terms of: a) the frequency or occupational risk factors for CTS. These duration of tasks pertaining to the hand/wrist, b) methods include direct measurement, the ratio of work-time to recovery time, c) the observation, self-reports, and categorization by percentage of the workday spent on repetitive job titles. Most investigators agree that use of activities, or d) the quantity of work performed observational or direct measurement methods in a given time. The rest of the studies generally increases the quality (both the precision and used job titles or questionnaires to characterize accuracy) of ergonomic exposure assessments, exposure. but these methods also tend to be costly and time consuming. In general, misclassification Studies Meeting the Four Evaluation Criteria errors tend to dilute the observed associations Five epidemiologic studies of the hand/wrist between disease and physical workload area addressing repetitiveness and CTS [Viikari-Juntura 1995]. [Chiang et al. 1990, 1993; Moore and Garg 1994; Osorio et al. 1994; Silverstein et al. REPETITION 1987] met the four criteria. Chiang et al. [1990] Definition of Repetition for CTS studied 207 workers from 2 frozen food For our review, we identified studies that processing plants. Investigators observed job examined repetition or repetitive work for the tasks and divided them into low or high hand and wrist for CTS as cyclical or repetitive repetitiveness categories of wrist movement work activities that involved either 1) repetitive based on cycle time, as previously described by hand/finger or wrist movements such as hand Silverstein et al. [1987]. Jobs were also gripping or wrist extension/flexion, ulnar/radial classified according to whether or not workers’ deviation, and supination or pronation. Most of hands were exposed to cold work conditions. the studies that examined repetition or repetitive The resulting exposure groups were: work as a risk factor for CTS had several Group 1–Not Cold, Low Repetitiveness concurrent or interacting physical workload (mainly office staff and technicians); factors. Therefore, repetitive work should be Group 2–Cold Exposure or High considered in this context, with repetition as Repetitiveness; and Group 3–Cold Exposure only one exposure factor, accompanied by and High Repetitiveness. CTS diagnosis was others such as force, extreme posture, and, less based on abnormal clinical examination and commonly, vibration. nerve conduction studies. Prevalence of CTS Studies Reporting on the Association

5a-3 was 3% in Group 1, 15% in Group 2, and 37% females. The OR for repet-itiveness was 1.5 in Group 3. Statistical modeling that also (95% CI 0.8–2.8), con-trolling for oral included gender, age, length of employment, contraceptive use and force. and cold resulted in an odds ratio (OR) of 1.87 (p=0.02) for CTS among those with highly Moore and Garg [1994] evaluated 32 jobs in a repetitive jobs. The OR for CTS among those pork processing plant and then reviewed past exposed to cold conditions and high OSHA illness and injury logs and plant medical repetitiveness was 3.32 (p=0.03). The authors records for CTS cases in these job categories. cautioned that cold exposure may have at least A CTS case required the recording of partially acted as a proxy for forceful suggestive symptoms (numbness and tingling) hand/wrist exertion in this study group. combined with electrodiagnostic confirmation (as reported by the attending Chiang et al. [1993] studied 207 workers from electromyographers) of a case. Incidence ratios 8 fish processing factories in Taiwan. Jobs were (IRs) were calculated using the full-time divided into 3 groups based on levels of equivalent number of hours worked reported repetitiveness and force. The comparison group on the logs. The exact number of workers was (low force/low repetitiveness) was comprised not reported. Exposure assessment included of managers, office staff, and skilled craftsmen videotape analysis of job tasks for (group 1). The fish-processing workers were repetitiveness and awkward postures. The divided into high repetitiveness or high force force measure was an estimate of the percent (group 2), and high force and high maximum voluntary contraction (%MVC) repetitiveness (group 3). Repetition of upper based on weight of tools, and parts and limb movements (not specifically the wrist) was population strength data adjusted for extreme defined based on observed cycle time posture or speed. Jobs were then categorized [Silverstein et al. 1987]. CTS was defined on as hazardous or safe (for all upper extremity the basis of symptoms and positive physical MSDs, not for CTS), based on exposure data examination findings, ruling out systemic and the judgment of the investigators. The diseases and injury. CTS prevalence for the hazardous jobs had a relative risk (RR) for overall study group was 14.5%. CTS CTS of 2.8 (95% CI 0.2–36.7) compared to prevalence increased from group 1, to group 2, the safe jobs. Due to the lack of data from and to group 3 (8.2%, 15.3%, and 28.6%, individual workers, the study was unable to respectively), a control for common confounders. Potential for statistically significant trend (p<0.01). survivor effect (79% of the workforce was laid Repetitiveness alone was not a significant off the year prior to the study), a limited latency predictor of CTS (OR 1.1). Statistical modeling period (8–32 months), and the potential for showed that women in this study group had a incomplete case ascertainment (underreporting higher prevalence of CTS than men (OR 2.6, is known to be a problem with OSHA illness 95% confidence interval [CI] 1.3–5.2). and injury logs) limit confidence in this estimate. Because the proportion of women varied by This study did not specifically address the exposure group (48%, 75%, and 79% from relationship between repetitiveness and CTS. group 1 to 3), further analyses were limited to No significant association was identified

5a-4 between repetitiveness and the grouped “upper force/low repetitiveness, low force/high extremity musculoskeletal disorders,” but there repetitiveness, and high force/high was very little variability in repetitiveness (31 of repetitiveness. Fourteen cases (2.1% the 32 jobs had a cycle time less than 30 prevalence) of CTS were diagnosed based on seconds). standardized physical examinations and structured interviews. Osorio et al. [1994] studied 56 supermarket workers. Exposure to repetitive and forceful The OR for CTS in highly repetitive jobs wrist motions was rated as high, moderate, or compared to low repetitive jobs, irrespective of low, following observation of job tasks. The force, was 5.5 (p<0.05) in a statistical model CTS case definition was based on symptoms that also included age, gender, years on the job, and nerve conduction studies. CTS-like and plant. The OR for CTS in jobs with symptoms occurred more often (OR 8.3, 95% combined exposures to high force and high CI 2.6–26.4) among workers in the high repetition was 15.5 (p<0.05), compared to exposure group compared to the low exposed jobs with low force and low repetition. Age, group. The odds of meeting the symptom and gender, plant, years on the job, hormonal NCS-based CTS case definition among the status, prior health history, and recreational high exposure group were 6.7 (95% CI activities were analyzed and determined not to 0.8–52.9), compared to the low exposure confound the associations identified. group. Studies Meeting at Least One Criterion Silverstein et al. [1987] studied 652 workers in Fourteen additional studies met at least one of 39 jobs from 7 different plants (electronics, the criteria. appliance, apparel, and bearing manufacturing; metal casting, and an iron foundry). Barnhart et al. [1991] studied ski manufacturing Investigators divided jobs into high or low workers categorized as having repetitive or repetitiveness categories, based on analysis of nonrepetitive jobs based on observational videotaped job tasks of 3 representative exposure methods for hand/wrist exposure. The participation rate for this study was below workers in each job. High repetitiveness was 70%. Three different case definitions were used defined as cycle time less than 30 seconds or at for CTS based on symptoms, physical exam least 50% of the work cycle spent performing findings, and NCS using the mean median-ulnar the same fundamental movements. Jobs were difference in each group. Each case definition also divided into high or low force categories used the NCS results. The authors reported a based on EMGs of representative workers’ significant prevalence ratio (PR) of 2.3 for the forearm flexor muscles while they performed mean median-ulnar sensory latency nerve their usual tasks. EMG measurements were difference among those in repetitive jobs averaged within each work group to compared to those in non-repetitive jobs. characterize the force requirements of the job. However, the difference was found in the ulnar High force was defined as a mean adjusted rather than in the median nerve. The median force >6 kg. Jobs were then classified into 4 nerve latencies were not statistically different groups: low force/low repetitiveness, high between the two groups.

5a-5 Baron et al. [1991] studied CTS in 124 occupational factors: grocery store checkers and 157 other grocery 1) uninterrupted shoulder rotation with elevated store workers who were not checkers. The arm (OR 1.8, 95% CI 1.2–2.8) and 2) CTS case definition required symptoms that protection from repeated finger tapping (OR met pre-determined criteria on a standardized 0.4, 95% CI 0.2–0.7). The authors note that questionnaire and physical examinations. The the latter observation presented “difficulties of OR for CTS among checkers was 3.7 (95% interpretation.” Limitations of this study concern CI 0.7–16.7), in a model that included age, the lack of exposure assessment for repetition, hobbies, second jobs, systemic disease, and and the questionable reliability for reported limb obesity. Participation rates at the work sites movements as an accurate measure of were higher among the exposed group repetition. (checkers: 85% participation, non-checkers: 55% participation). After telephone interviews Feldman et al. [1987] studied electronic in which 85% of the non-checkers completed workers at a large manufacturing firm using a questionnaires, investigators reported that the questionnaire survey and biomechanical job proportion of non-checkers meeting the case analysis. Four work areas with 84 workers definition did not increase. were identified as “high risk” with highly repetitive and forceful tasks. Workers in these Cannon et al. [1981] in a case-control study of high risk areas had physical examinations and aircraft engine workers did not find a significant NCS. Sixty-two workers from the high risk area had repeat NCS one year later. association with the performance of repetitive Comparing these high risk workers to the motion tasks (OR 2.1, 95%CI 0.9–5.3), but others, one can calculate ORs for symptoms of found a significant association with self- numbness and tingling [OR 2.26 (p<0.05)] and reported use of vibrating hand tools, history of a positive Phalen’s sign [2.7 (p<0.05)]. gynecologic surgery, and an inverse relationship Longitudinal NCS of workers in the high risk with years on the job. One must assume from area showed significant worsening in the the article that “repetitive motion tasks” were median motor latency and sensory conduction defined by job title. The diagnosis of CTS was velocity in the left hand, and motor changes based on medical and workers’ compensation over a year’s period, which the authors records. attributed to work exposure. A limitation of this study concerns inadequate exposure In English et al.’s [1995] case-control study of information about the extent of worker upper limb disorders diagnosed in orthopedic exposure to repetitive and forceful work. clinics, the case series included 171 cases of CTS and 996 controls. Exposure was based on McCormack et al. [1990] studied 1,579 textile self-reports; repetitiveness was defined as a production workers and compared them to 468 motion occurring more than once per minute. other nonoffice workers, a comparison group The logistic regression model of CTS found that included machine maintenance workers, transportation workers, cleaners, and significant associations with height (negative), sweepers. The textile production workers were weight (positive), presentation at the clinic as a divided into four broad job categories based on result of an accident (negative), and two similarity of upper extremity exertions. No

5a-6 formal exposure assessment was conducted. both a general population comparison group Health assessment included a questionnaire and and a low exposed checker group. The screening physical examination followed by a limitations of this study are: 1) the use of an diagnostic physical examination. CTS was overly sensitive health outcome measure, for diagnosed using predetermined clinical criteria. example, 32% of the surveyed population The severity of cases was also reported as reported numbness; and 2) the use of self- mild, moderate, or severe. The overall prevalence for CTS was 1.1%, with 0.7% in reported exposure. boarding, 1.2% in sewing, 0.9% in knitting, 0.5% in packaging/folding, and 1.3% in the Nathan et al. [1988] studied median nerve comparison group. None of the differences conduction of 471 randomly selected workers were statistically significant. A statistical model from four industries (steel mill, meat/food that also included age, gender, race, and years packaging, electronics, and plastics of employment showed that CTS occurred manufacturing). Median nerve sensory latency more often among women in this study values were adjusted for age for statistical (p<0.05). Interpretation of these data, analyses. Thirty-nine percent of the study especially with a low prevalence disorder like subjects had impaired sensory nerve CTS, is difficult since gender varied with job conduction, or “slowing” of the median nerve. (94% of boarding workers were female, The five exposure groups were defined as compared to 56% in the comparison group), follows: Group 1 is very low force, low and the comparison group (machine maintenance workers, transportation workers, repetition (VLF/LR); Group 2 is low force, cleaners and sweepers) may have also been very high repetition (LF/VHR); Group 3 is exposed to upper extremity exertions. moderate force, moderate repetition (MF/MR); Interactions among potential confounders were Group 4 is high force/moderate repetition not addressed, but they are suspected because (HF/MR); and Group 5 is very high force/high of significant associations between race and repetition (VHF/HR). There was no significant three MSDs. difference between Group 1 and Group 2, the groups that had the greatest differences in Morgenstern et al. [1991] mailed repetition. The authors reported a significantly questionnaires to 1,345 union grocery checkers higher number of subjects with median nerve and a general population group. Exposure was slowing in Group 5 (VHF/HR) compared to based on self-reported time working as a Group 1 (VLF/LR), but not in other groups, checker. Symptoms of CTS were significantly using a statistical method described as a associated with age and the use of diuretics, “pairwise unplanned simultaneous test and nonsignificantly associated with average procedure” [Sokal and Rohlf 1981]. The hours worked per week, and years worked as authors also reported that when individual a checker. A positive CTS outcome was based hands were the basis of calculations rather than on the presence of all four symptoms: pain in subjects, Group 3 had a significantly higher the hands or wrist, nocturnal pain, tingling in the prevalence of median nerve slowing. hands or fingers, or numbness. The estimated Calculations of the data using PRs and chi- attributable fraction of CTS symptoms to squares [Kleinbaum et al. 1982] resulted in working as a checker was about 60%, using significantly higher prevalences of median nerve

5a-7 slowing in each of Groups 3, 4, and 5 drop in prevalence of median nerve slowing in (moderate to high repetition, with moderate to Group 5 between 1984 and 1989 might be very high force) compared to Group 1 explained by the higher drop-out rate among (VLR/LF). PRs are 1.9 (95% CI 1.3–2.7), 1.7 cases in Group 5 compared to Group 1 (PR (95% CI 1.1–2.5), and 2.0 (95% CI 1.1–3.4) 2.9, 95% CI 1.3–6.6). This was not addressed for Groups 3, 4, and 5, respectively. A by the authors. conservative (Bonferroni) adjustment of the significance level to 0.0125 for multiple Punnett et al. [1985] compared the symptoms comparisons [Kleinbaum et al. 1982] would and physical findings of CTS in 162 women result in Group 5 no longer being statistically garment workers and 76 women hospital significantly different from Group 1 (p=0.019), workers such as nurses, laboratory technicians, but Group 4 (p=0.009), and Group 3 and laundry workers. Eighty-six percent of the (p=0.000) remain statistically significantly garment workers were sewing machine higher than Group 1 in prevalence of median operators and finishers (sewing and trimming by nerve slowing. hand). The sewing machine operators were described as using highly repetitive, low force In 1992, Nathan et al. [1992a] reported on a wrist and finger motions, whereas finishing follow-up evaluation in the same study group. work also involved shoulder and elbow Sixty-seven percent of the original study motions. The exposed garment workers subjects were included. Hands (630), rather probably had more repetitive jobs than most of than subjects, were the basis of analysis in this the hospital workers. CTS symptoms occurred study. Novice workers (those employed less more often among the garment workers (OR than 2 years in 1984) were less likely to return 2.7, 95% CI 1.2–7.6) compared to the hospital than non-novice workers (56% compared to workers. There was a low participation rate 69%, p=.004). Maximum latency differences in (40%) among the hospital workers. median nerve sensory conduction were determined as in the Nathan et al. [1988] study. Schottland et al. [1991] carried out a The authors state that there was no significant comparison of NCS findings in poultry workers difference in the prevalence of median nerve and job applicants as referents. No exposure slowing between any of the exposure assessment was performed, and applicants categories in Nathan et al. [1988] using the were not excluded if they had prior same statistical method described in the Nathan employment in the plant. Results indicated that et al. 1988 study. However, calculations using the right median nerve sensory latency was common statistical methods result in the significantly longer in 66 female poultry workers following PRs for slowing: Group 3–1.5 (95% compared to 41 female job applicants. In these CI 1.0–2.2), Group 4–1.4 (95% CI 0.9–2.1), two groups of women there were less and Group 5–1.0 (95% CI 0.5–2.2), pronounced differences in the left median compared to Group 1. Group 5 had the same sensory latency. The latencies in the 27 male prevalence of slowing (18%) as Group 1 in poultry workers did not differ significantly from 1989. In 1984 the prevalence of slowing was the 44 male job applicants, although the power 29% in Group 5, and 15% in Group 1. The calculations presented in the paper noted

5a-8 limited power to detect differences among male symptomatic industrial workers, the mean participants. The OR for percentage of female exposure for the symptomatic industrial poultry workers who exceeded the criteria workers was nonsignificantly slightly greater for value for the right median sensory latency is all exposure factors except for repetitiveness. 2.86 (95% CI 1.1–7.9). The major limitations The median sensory amplitudes were of this study are the absence of detailed significantly smaller information on exposure and the inclusion of (p<0.01) and latencies longer (p<0.05 ) for former poultry workers into the applicant industrial workers with exposure to high grip group, as well as the inadequate sample size, forces compared to those without. Mean and the personal characteristics of these sensory amplitudes were significantly smaller workers. This study found a significant (p<0.05) and motor and sensory latencies were association between highly repetitive, highly significantly longer (p<0.01) in the industrial forceful work and abnormal NSC consistent asymptomatic workers compared to the control with CTS. It does not allow analysis of group. These findings for the motor latencies repetition alone. are similar to Feldman et al. [1987]. Since most of the industrial workers were exposed to Stetson et al. [1993] used measurements of repetitive work, it is not clear whether this sensory nerve conduction velocity of the study population allowed a comparison median nerve as indicators of nerve impairment between repetitive and non-repetitive work. or CTS; clinical examination results were not Overall this study suggests that repetitive work reported in this article. Three groups were combined with other risk factors is associated studied: a reference group of 105 workers with slowing of median nerve conduction. without occupational exposure to highly forceful or repetitive hand exertions, 103 industrial The Wieslander et al. [1989] case-control workers with hand/wrist symptoms, and 137 study used self-reported information collected asymptomatic industrial workers. Exposure was via telephone interview about the duration of assessed with a checklist by trained workers. exposure (number of years and hours per Factors considered included repetitiveness week) to several work attributes including (Silverstein criteria), force defined by the repetitive work. Definitions for these work weight of an object that is carried or held, attributes were not provided. Three categories localized mechanical stress, and posture. of duration of exposure were defined for each Exposure assessments were available on 80% attribute (<1 year, of the industrial workers. Most of the industrial 1–20 years, and >20 years), but the asymmetry workers were on repetitive jobs (76%), a of the categories was not explained. A minority carried more than ten pounds some of significant OR for reporting repetitive the time (32%), and gripped more than six movements of the wrist comparing CTS pounds at least some of the time (44%). The patients to hospital referents (OR 4.6) and analysis controlled for several confounders general population referents (OR 9.6) was including age, gender, finger circumference, reported, but only among those employed height, weight, and a square-shaped wrist. In greater than 20 years. Those employed from the comparison of the asymptomatic to 1–20 years compared to the referent

5a-9 population had elevated ORs for repetitive to cold may be at least partially explained by movements of the wrist (1.5 for CTS patients forceful motions among workers who were also compared to hospital referents, and 2.3 exposed to cold. Force was not evaluated in compared to population referents), but these this study. were not significant. Jobs with increasing numbers of work risk factors gave increasing Silverstein et al. [1987] reported an OR of 5.5 ORs (from 1.7 to 7.1) among CTS cases when (p<0.05) for repetition as a single predictor of compared to referents; these were statistically CTS. Among workers exposed to high significant when there were two or more risk repetition and high force, the OR was 15.5 factors. Given the limited quality of the (p<0.05). exposure data and findings (repetition is a significant risk factor only after 20 years of Chiang et al. [1993] reported a significant trend exposure), this is only suggestive of a of increasing prevalence of CTS with increasing relationship between repetition alone and CTS. exposure to repetition and/or force (8.2%, 15.3%, and 28.6%, p<0.05). Repetition (of the Studies Not Meeting Any of the Criteria whole upper limb, not the wrist) alone did not Liss et al. [1995] conducted a mail survey significantly predict CTS (OR 1.1). concerning CTS among 2,124 Ontario dental hygienists compared to 305 dental assistants In summary, three studies that met all four who do not scale teeth. Both groups had a low criteria reported ORs for CTS associated with response rate (50%). The age adjusted OR repetition. The statistically significant ORs for was 5.2 (95% CI 0.9–32) for being told by a CTS attributed to repetition alone ranged from physician that you had CTS and 3.7 (95% CI 1.9 to 5.5. The statistically significant ORs for 1.1–1.9) using a questionnaire-based definition CTS attributed to repetition in combination with of CTS. The major limitations of this study are force or cold ranged from 3.3 to 15.5. Gender, the low participation rate, the lack of a detailed age, and other potential confounders were exposure assessment for repetitiveness, and addressed and are unlikely to account for the self-reported health outcome. associations reported.

Strength of Association—Repetition Five other studies observed job tasks, then and CTS grouped them into categories according to Three of the five studies that met all four criteria estimated levels of repetitiveness combined evaluated the effect of repetitiveness alone on with other risk factors [Feldman et al. 1987; CTS: Chiang et al. [1990], Silverstein et al. Moore and Garg 1994; Nathan et al. 1988, [1987], and Chiang et al. [1993]. 1992a; and Osorio et al. 1994]. CTS case definitions reported here required more than Chiang et al. [1990] reported an OR of 1.9 symptom-defined criteria. Moore and Garg (p<0.05) for CTS among those with highly [1994] reviewed medical records; Nathan et al. repetitive jobs. The OR for CTS among those [1988] and Osorio et al. [1994] performed exposed to high repetitiveness and cold was nerve conduction studies. 3.32 (p<0.05). The additional effect attributed

5a-10 Feldman et al. [1987] reported an OR of 2.7 comparison groups was 8.3 (95% CI 2.6– (p<0.05) for a positive Phalen’s test among 26.4). workers in high exposure jobs, compared to low exposure jobs. To summarize, three of the five studies reviewed resulted in statistically significant Moore and Garg [1994] reported an OR of positive findings for CTS associated with 2.8 (0.2, 36.7) for CTS among workers in combined exposures. Feldman et al. [1987] “hazardous” jobs compared to workers in reported an elevated OR for CTS with high “nonhazardous” jobs. combined exposure. Nathan et al.’s [1988] data resulted in elevated PRs for CTS among Nathan et al.’s [1988] data result in PRs the three highest combined exposure groups. for four groups with varying levels of Nathan et al.’s [1992a] data resulted in an repetitiveness and force from very low (VL) to elevated PR for CTS among one of the high very high (VH), compared to a very low force, combined exposure groups. There was low repetition group (VLF/LR): evidence of survivor bias in the highest LF/VHR versus VLF/LR: 1.0 (95% CI exposure group. 0.5–2.0) MF/MR versus VLF/LR: 1.9 (95% CI The following studies used job title or job 1.3–2.7) category to represent exposure to HF/MR versus VLF/LR: 1.7 (95% CI 1.1– repetitiveness combined with other exposures 2.5) and defined CTS based on physical VHF/HR versus VLF/LR: 2.0 (95% CI examination [Baron et al. 1991, McCormack et 1.1–3.4). al. 1990, Punnett et al. 1985] or nerve conduction studies [Schottland et al. 1991]. Nathan et al. [1992a] data, a 5-year follow-up of the 1988 study, result in PRs for the Baron et al. [1991] reported an OR of 3.7 following groups: (95% CI 0.7–16.7) for CTS, defined by LF/VHR versus VLF/LR: 1.0 (95% CI 0.6– symptoms and physical examination, among 1.9) grocery checkers compared to other grocery MF/MR versus VLF/LR: 1.5 (95% CI 1.0– workers. 2.2) HF/MR versus VLF/LR: 1.4 (95% CI 0.9– McCormack et al. [1990] reported the 2.1) following ORs for CTS among workers in each VHF/HR versus VLF/LR: 1.0 (95% CI 0.5– of four broad job categories that were 2.2). considered exposed, compared to a comparison group of maintenance workers and Osorio et al. [1994] reported an OR of 6.7 cleaners that was considered to have low (95% CI 0.8–52.9) for CTS among workers in exposure: high exposure jobs, compared to workers in Boarding versus Low: 0.5 (95% CI 0.1–2.9) low exposure jobs. Using a symptom-based Sewing versus Low: 0.9 (95% CI 0.3–2.9) case definition, the OR for the same

5a-11 Packaging versus Low: 0.4 (95% CI 0.0–2.4) combined with other job risk factors and CTS. Knitting versus Low: 0.6 (95% CI 0.1–3.1) Temporal Relationship: Repetition Punnett et al. [1985] reported an OR of 2.7 and CTS (95% CI 1.2–7.6) for CTS among garment The question of which occurs first, exposure or workers versus hospital workers. disease, can be addressed most directly in prospective studies. However, study limitations Schottland et al. [1991] reported an OR of such as survivor bias can cloud the 2.86 (95% CI 1.1–7.9) for prolonged right interpretation of findings. In our analysis of median sensory latency among female poultry Nathan et al.’s [1992a] data, 2 of 3 groups that workers, compared to female applicants for the were exposed to forceful hand/wrist exertions same jobs. No significant differences were were more likely to have median nerve slowing identified among males. when nerve conduction testing was repeated 5 years later. The highest exposure group had the In summary, two of the four studies reviewed same prevalence of slowing as the lowest above reported significantly elevated ORs for exposure group in 1989, whereas they had a CTS or median sensory nerve conduction higher prevalence rate in 1984. As discussed slowing. above, this apparent decrease in prevalence over 5 years can probably be explained by a Wieslander et al. [1989] reported an OR for higher drop-out rate among cases in the highest CTS (surgical cases, confirmed by NCS) of exposure group, compared to the lowest 2.7 (95% CI 1.3–5.4) among those with self- exposure group. These interpretations of the reported exposure to repetitive wrist movement data differ from those of the authors. Further >20 years, compared to hospital referents, and study is needed to clarify these issues. 4.5 (95% CI 2.0–10.4), compared to However, to our knowledge, there is no population referents. Significant OR s for CTS evidence demonstrating that those with CTS among those with combined job risk factors would be more likely to be hired in jobs that ranged from 3.3 to 7.1. involve high exposure to repetitive hand/wrist exertions and combined job risk factors, The remaining two studies relied on self- compared to those without CTS. In fact, reported symptoms and self-reported employment practices tend to exclude new exposures from mail [Morgenstern et al. 1991] workers with CTS from jobs that require or telephone surveys [Liss et al. 1995]. Data repetitive and intensive hand/wrist exertion. quality and response rates limit interpretation of findings. Feldman et al. [1987] reported longer median motor (but not sensory) latencies among In conclusion, among the studies that measured workers with combined exposure to hand/wrist repetition alone, there is evidence that repetition exertion, compared to nerve conduction is positively associated with CTS. The majority findings in the same group one year earlier. of studies provide evidence of a stronger positive association between repetition Cross-sectional studies provide evidence that

5a-12 exposure occurred before CTS, by using case then it may trigger a fibrosis which damages the definitions that exclude pre-existing cases, and function of the nerve. The interplay between by excluding recently hired workers from the acute increases in pressure and chronic changes study. The studies that provide evidence that to the nerve could partially explain why there is repetitive and combined job exposures are not a stronger correlation between symptoms of associated with CTS followed these practices, CTS and slowing of the median nerve. Both therefore the associations identified cannot be symptoms and slowing of the median nerve are explained by disease occurring before likely to have both acute and chronic exposure. components in many cases of CTS.

Consistency in Association for The work determinants of pressure in the Repetition and CTS carpal tunnel are wrist posture and load on the One study [English et al. 1995] reported a tendons in the carpal tunnel. For example, the statistically significant negative association normal resting pressure in the carpal tunnel with between repetitive work and CTS. The specific the wrist in a neutral posture is about 5 exposure was self-reported repeated finger millimeters of mercury (mmHg), and typing with tapping; the investigators stated that they had the wrist in 45E of extension can result in an difficulty interpreting this finding. All of the other acute pressure of 60 mmHg. Substantial load statistically significant findings pointed to a on the fingertip with the wrist in a neutral positive association between repetitive work posture can increase the pressure to 50 mmHg. and CTS. The non-significant estimates of RR A parabolic relationship between wrist posture were also mostly greater than one. and pressure in the carpal tunnel has been found. In laboratory studies of normal subjects, Coherence of Evidence for Repetition elevated carpal tunnel pressures quickly return One of the most plausible ways that repetitive to normal once the repetitive activity stops; hand activities may be associated with CTS is patients with CTS take a long time for the thorough causing a substantial increase in the pressure to return to their baseline values. One pressure in the carpal tunnel. This in turn can of the supporting observations for this model is initiate a process which results in either that at surgery for CTS, edema and vascular reversible or irreversible damage to the median sclerosis (fibrosis due to ischemia) are common nerve [Rempel 1995]. The increase in pressure, [Rempel 1995]. if it is of sufficient duration and intensity, may reduce the flow of blood in the epineural This model of the etiology of work-related CTS venules. If prolonged, this reduction in flow is consistent with two observations from the may affect flow in the capillary circulation, epidemiological literature. First, it illustrates resulting in greater vascular permeability and why both work and nonwork factors such as endoneural and synovial edema. Because of the obesity may be important because anything that structure of the median nerve and the carpal increases pressure in the carpal tunnel may tunnel, this increase in fluid and resulting contribute to CTS. Second, it explains why increase in pressure may persist for a long repetitiveness independent of wrist posture and period of time. If the edema becomes chronic, load on the flexor tendons may not be a major

5a-13 risk factor for CTS. [Nathan et al. 1988; Stetson et al. 1993; Barnhart et al. 1991]. Exposure-Response Relationship for Repetition There is evidence of a positive association Evidence of an exposure-response relationship between highly repetitive work alone and CTS. is provided by studies that show a correlation There is strong evidence of a positive between the level or duration of exposure and association between highly repetitive work in either the number of cases, the illness severity, combination with other job factors and CTS, or the time to onset of the illness. Silverstein et based on currently available epidemiologic al. [1987] showed an increasing prevalence of data. CTS signs and symptoms among industrial workers exposed to increasing levels of FORCE AND CTS repetition and forceful exertion. This Definition of force for CTS relationship was not seen when repetition alone The studies reviewed in this section determined was assessed. Similar findings on an exposure- hand/wrist force exposure by a variety of response relationship were reported by Chiang methods. Some investigators [Armstrong and et al. [1993], Osorio et al. [1994], Wieslander Chaffin 1979; Chiang et al. 1993; Silverstein et et al. [1989], and by Stock [1991] in her al. 1987] measured force by EMGs of reanalysis of the Nathan et al. [1988] data. representative workers’ forearm flexor muscles while they performed their usual tasks. EMG Morgenstern et al. [1991] and Baron et al. measurements were averaged within each work [1991] reported increased prevalence of CTS group to characterize the force requirements of with increasing length of time working as a the job; jobs were then divided into low or high grocery cashier. categories if the average force was above or below a cutoff point. Moore and Garg [1994] Conclusions Regarding Repetition estimated force as %MVC, based on weight of Based on the epidemiologic studies noted tools and parts and population strength data, above, especially those with quantitative adjusted for extreme posture or speed. Jobs evaluation of repetitive work, the strength of were then predicted to be either hazardous or association for CTS and repetition has been safe (for any upper extremity musculoskeletal shown to range from an OR of 2 to 15. The disorder), based on exposure data and higher ORs are found when contrasting highly judgment. Stetson et al. [1993] estimated repetitive jobs to low repetitive jobs, and when manipulation forces based on weights of tools repetition occurred in combination with high and parts and systematically recorded levels of forceful exertion. Those studies with observations of one or more workers on each certain epidemiologic limitations have also been job. Jobs were then ranked according to grip fairly consistent in showing a relationship force cutoffs. Nathan et al. [1988, 1992a] and between repetition and CTS. The evidence Osorio et al. [1994] estimated relative levels of from those studies which defined CTS based force (e.g., low, moderate, high) after on symptoms, physical findings, and NCS is observation of job tasks. McCormack et al. limited, due to the variety of methods used [1990] grouped jobs into broad job categories

5a-14 based on similarity of observed job tasks; one force/low repetitiveness) was managers, office job group (boarding) required forceful staff, and skilled craftsmen. The fish-processing hand/wrist exertions. Baron et al. [1991] and workers were divided into high force or high Punnett et al. [1985] used job title as a repetitiveness (group 2), and high force and surrogate for exposure to forceful hand/wrist high repetitiveness (group 3). Hand force exertions. requirements of jobs were estimated by electromyographs of forearm flexor muscles of Much of the epidemiologic data on CTS and a representative worker from each group force overlaps with those studies discussed in performing usual job tasks. High force was the above section on repetition. Repetitive defined as an average hand force of >3 kg work is frequently performed in combination repetition of the upper limb (not specifically the with external forces, and much of the wrist) was defined based on observed cycle epidemiologic literature has combined these time [Silverstein et al. 1987]. CTS was defined two factors when determining association with on the basis of symptoms and positive physical CTS. examination findings, ruling out systemic diseases and injury. CTS prevalence for the Studies Reporting on the Association overall study group was 14.5%. CTS of Force and CTS prevalence increased from group 1 to group 3 Eleven studies reported results on the (8.2%, 15.3%, and 28.6%), a statistically association between force and CTS. The significant trend p<0.01). Statistical modeling epidemiologic studies that addressed forceful showed that women in this study group had a work and CTS tended to compare working higher prevalence of CTS than men (OR 2.6, groups by classifying them into broad 95% CI 1.3–5.2). Force also significantly categories based on estimates of the predicted CTS (OR 1.8, 95% CI 1.1–2.9), but forcefulness of hand/wrist exertions in not repetitiveness. Because the proportion of combination with estimated repetitiveness. In women varied by exposure group (48%, 75%, most studies the exposure classification was an and 79% from groups 1 to 3), the possibility of ordinal rating (e.g., low, moderate, or high); in an interaction between gender and job some studies job categories or titles were used exposure exists, but this was not statistically as surrogates for exposure to force exertions. examined. In an analysis limited to females, the 2 significant predictors of CTS were oral Studies Meeting the Four Evaluation Criteria contraceptive use (OR 2.0, 95% CI 1.2–5.4), Four studies that evaluated the relationship and force (OR 1.6, 95% CI 1.1–3.0). Concern between forceful hand/wrist exertion and CTS over interpretation of these findings is raised met all four criteria: Chiang et al. [1993], because oral contraceptive use varies with age, Moore and Garg [1994], Osorio et al. [1994], and age may vary with job exposures. Silverstein et al. [1987]. Chiang et al. [1993] studied 207 workers from 8 fish-processing factories in Taiwan. Jobs were divided into 3 These potential interactions were not examined, groups based on levels of force and and women’s ages by job group were not repetitiveness. The comparison group (low reported.

5a-15 jobs studied allowed for up to a 32-month Moore and Garg [1994] evaluated 32 jobs in a latency period. The possibility of differential pork processing plant and then reviewed past case ascertainment between exposed and OSHA 200 logs and plant medical records for unexposed jobs exists, both because of CTS cases in these job categories. IRs were different observation periods, as well as the calculated using the full-time equivalent (FTE) likelihood that turnover may have been greater number of hours worked as reported on the in the exposed jobs. It is also unclear whether logs. The exact number of workers was not employees worked full-time or part-time hours. reported. Exposure assessment included videotape analysis of job tasks for repetitiveness and awkward postures. The Osorio et al. [1994] studied 56 supermarket force measure was an estimate of the %MVC, workers. Exposure to repetitive and forceful based on weight of tools and parts and wrist motions was rated as high, moderate, or population strength data, adjusted for extreme low, following observation of job tasks (97% posture or speed. Jobs were then predicted to initial concordance with 2 independent be either hazardous or safe (for all Upper observers). The CTS case definition was based Extremity MSDs), based on exposure data and on symptoms and nerve conduction studies. judgment. CTS was determined by reviewing CTS-like symptoms occurred more often (OR OSHA 200 logs and plant medical records. 8.3, 95% CI 2.6–26.4) among workers in the The proportion of CTS in the overall study high exposure group compared to the low group during the 20 months of case exposed group. The odds of meeting the ascertainment was 17.5 per 100 FTEs. If the symptom and NCS-based CTS case definition occurrence of CTS did not vary over this among the high exposure group were 6.7 (95% period, the proportion of CTS in a 12-month CI 0.8–52.9), compared to the low exposure period would be 10.5 per 100 FTEs. The group. hazardous jobs had a RR for CTS of 2.8 (0.2, 36.7) compared to the safe jobs. Potential for Silverstein et al. [1987] measured force by survivor effect (79% of the workforce was laid electromyographs of representative workers’ off the year before the study), limited latency forearm flexor muscles while they performed period (8-32 months), and the potential for their usual tasks. EMG measurements were incomplete case ascertainment (underreporting averaged within each work group to is common on OSHA 200 logs, and logs were characterize the force requirements of the job; not reviewed for the first 12 months of the jobs were then divided into high or low study) limit confidence in this estimate. One of categories if the mean adjusted force was the more hazardous jobs, the Ham Loaders, above or below required extreme wrist, shoulder and elbow 4 kg. Jobs were then classified into 4 groups posture and was rated 4 on a 5-point scale for that also accounted for repetitiveness: low force, yet there was no observed morbidity. force/low repetitiveness, high force/low Since this job did not start until 1989, the repetitiveness, low force/high repetitiveness, period of observation for musculoskeletal and high force/high repetitiveness. Fourteen disorders for this job was only 8 months. Other cases (2.1% prevalence) of CTS were

5a-16 diagnosed based on standardized physical sweepers. The textile production workers were examinations and structured interviews. divided into four broad job categories based on similarity of upper extremity exertions. The The OR for CTS in high force jobs compared Boarding group required the most physical to low force jobs, irrespective of repetitiveness, exertion. No formal exposure assessment was was 2.9 (p>0.05). The plant- adjusted OR for conducted. Health assessment included a CTS in jobs with combined exposures to high questionnaire and screening physical force and high repetition was 14.3 (p<0.05), examination followed by a diagnostic physical compared to jobs with low force and low examination. CTS was diagnosed using repetition. Age, gender, plant, years on the job, predetermined clinical criteria. The severity of hormonal status, prior health history, and cases was also reported as mild, moderate or recreational activities were analyzed and severe. The overall prevalence for CTS was determined not to confound the associations 1.1%, with 0.7% in Boarding, 1.2% in Sewing, identified. The OR for CTS in jobs with 0.9% in Knitting, 0.5% in Packaging/Folding, combined exposure from the multiple logistic and 1.3% in the comparison group. None of analysis was 15.5 (95% CI 1.7–142.) the differences were statistically significant. A statistical model that also included age, gender, Studies Meeting at Least One Criterion race, and years of employment showed that Baron et al. [1991] studied CTS in 124 CTS occurred more often among women in this grocery store checkers and 157 other grocery study (p<0.05). Interpretation of these data, store workers who were not checkers. The especially with a low prevalence disorder like CTS case definition required symptoms that carpal tunnel syndrome, is difficult since gender met pre-determined criteria on a standardized varied with job (e.g., 94% of Boarding workers questionnaire. Physical examinations were also were female, compared to 56% in the performed, but participation rates at the work comparison group), and the comparison group sites were higher among the exposed group may have also been exposed to upper extremity (checkers: 85% participation, non-checkers: exertions (machine maintenance workers, 55% participation). Telephone interviews to transportation workers, cleaners and non-checkers resulted in questionnaire sweepers). Interactions among potential completion by 85% of the non-checkers. confounders were not addressed, but they are Based on a questionnaire case definition, the suspected because of significant associations OR for CTS among checkers was 3.7 (95% between race and three musculoskeletal CI 0.7–16.7), in a model that included age, disorders. hobbies, second jobs, systemic disease, and obesity. Nathan et al. [1988] studied median nerve conduction of 471 randomly selected workers McCormack et al. [1990] studied 1,579 textile from four industries (steel mill, meat/food production workers compared to 468 other packaging, electronics, and plastics nonoffice workers, a comparison group that manufacturing). Jobs were grouped into 5 included machine maintenance workers, relative levels of force (from very light to very transportation workers, cleaners, and high) after observation of job tasks. Jobs were

5a-17 also rated for repetitiveness (5 levels). Thirty- significantly higher than Group 1 in prevalence nine percent of the study subjects had impaired of median nerve slowing. sensory conduction, or “slowing” of the median nerve. The 5 exposure groups were defined as In 1992 Nathan et al. [1992a] reported on a follows: Group 1 is very low force, low follow-up evaluation in the same study group. repetition (VLF/LR); Group 2 is low force, Sixty-seven per cent of the original study very high repetition (LF/VHR); Group 3 is subjects were included. Hands (630), rather moderate force, moderate repetition (MF/MR); than subjects, were the basis of analysis in this Group 4 is high force/moderate repetition study. Novice workers (those employed less (HF/MR); and Group 5 is very high force/high than 2 years in 1984) were less likely to return repetition (VHF/HR). The most logical than non-novice workers (56% compared to comparisons to evaluate the effect of force 69%, p=0.004). Probable CTS was defined on would be Groups 3, 4, and 5 (moderate, high, the basis of symptoms reported during a and very high force) compared to Group 1 (low structured interview and a positive Phalen’s or force). Group 2 jobs are not a good Tinel’s test. Maximum latency differences in comparison because they are very highly median nerve sensory conduction were repetitive, which may confound the determined as in the 1984 study. The authors comparisons. The authors reported a state that there was no significant difference in significantly higher number of subjects with the prevalence of slowing between any of the median nerve slowing in Group 5 (VHF/HR) exposure categories in 1989. However, compared to Group 1 (VLF/LR), but not in calculations using common statistical methods other groups, using an uncommon statistical show significantly higher prevalences of slowing method (pairwise unplanned simultaneous test in Group 4 (PR 1.4, 95% CI 0.9–2.1) procedure [Sokal and Rohlf 1981]). The compared to Group 1. Group 3's prevalence of authors also reported that when individual slowing was 26% compared to Group 1's hands were the basis of calculations rather than 18%, but this difference was not statistically subjects, Group 3 had a significantly higher significant (p=0.07). Group 5 had the same prevalence of median nerve slowing. prevalence of slowing (18%) as Group 1 in Calculations of the more familiar PRs and chi- 1989; the prevalence of slowing in Group 5 squares [Kleinbaum et al. 1982], using the was 29% in 1984. The drop in prevalence of published data, result in higher prevalences of slowing in Group 5 between 1984 and 1989 median nerve slowing in each of Groups 3, 4, might be explained by the higher drop-out rate and 5, compared to Group 1 (PRs: 1.9, 95% among cases in Group 5 compared to Group 1 CI 1.3–2.7; 1.7, 95% CI 1.1–2.5; and 2.0, (PR 2.9, 95% CI 1.3–6.6). This was not 95% CI 1.1–3.4, respectively). A conservative addressed by the authors. adjustment (Bonferroni) of the significance level to 0.0125 for multiple comparisons [Kleinbaum Punnett et al. [1985] compared the symptoms et al. 1982] would result in Group 5 no longer and physical findings of CTS in 162 women being statistically significantly different from garment workers and 76 women hospital Group 1 (p=0.019), but Group 4 (p=0.009) workers such as nurses, laboratory technicians, and Group 3 (p=0.000) remain statistically and laundry workers. Eighty-six percent of the

5a-18 garment workers were sewing machine ulnar nerve function showed lower amplitudes operators and finishers (sewing and trimming by and longer latencies (p<0.05) among the hand). The sewing machine operators were asymptomatic automotive workers; differences described as using highly repetitive, low force were greater between the symptomatic wrist and finger motions, whereas finishing automotive workers and the white collar work also involved shoulder and elbow workers. The symptomatic automotive workers motions. The exposed garment workers likely had lower amplitudes and longer latencies for 5 had more repetitive jobs than most of the of 6 median sensory measures (p<0.05), hospital workers. CTS symptoms occurred compared to the asymptomatic automotive more often among the garment workers (OR workers; there were no significant differences in 2.7, 95% CI 1.2–7.6) compared to the hospital ulnar nerve function between these two groups. workers. There was a low participation rate Asymptomatic automotive workers had (40%) among the hospital workers. “healthier” median nerves than automotive workers with CTS symptoms, but there were Stetson et al. [1993] conducted nerve no differences between these 2 groups in ulnar conduction studies on 105 administrative and nerve function, suggesting that the case professional workers, and 240 automotive definition was specific for CTS. workers. Hand/wrist forces were estimated based on weights of tools and parts and Of the studies that addressed CTS, almost all systematically recorded observations of one or examined occupations and jobs in which force more workers on each job. Jobs were then was combined with another exposure factor ranked according to grip force cutoffs: <6 lb, (such as repetition or awkward postures). >6 lb, >10 lb. Median nerve measures differed Chiang et al. [1993] estimated exposure to among the groups: index finger sensory hand/wrist force independent of repetitiveness amplitudes were lower and distal sensory and found statistically significant RRs for CTS latencies were longer among automotive ranging from 1.6 to 1.8. Estimates of RR that workers in jobs requiring grip force >6 lb and were not statistically significant ranged from 0.4 >10 lb, compared to those requiring less than 6 to 6.7 [McCormack et al. 1990; Osorio et al. lb (p<0.05 for all). At the wrist, median sensory 1994]. Relative risk estimates for CTS among amplitudes were also lower and distal median workers exposed to a combination of forceful sensory latencies were also longer among the and repetitive hand/wrist exertions ranged from >6 lb, and the >10 lb exposure groups (p<0.05 1.0 to 15.5 [Nathan et al. 1988, 1992a; for 3 of 4 differences). Age, height, and finger Silverstein et al. 1987]. circumference were included in statistical models. The automotive workers were then Study limitations may impact the interpretation divided into two groups, symptomatic (n=103) of findings. One limitation to consider is gender and asymptomatic (n=137), based on whether effect. Of the studies listed above reporting or not they met standard interview criteria for statistically significant associations between CTS symptoms. When comparisons were forceful hand/wrist exertions and CTS, gender made to the administrative and professional effect was controlled for in the analyses. Other workers, 15 of 16 measures of median and potential limitations such as selection factors

5a-19 impact the interpretation of the studies limitations, such as survivor bias, can cloud the reviewed. Survivor bias can be a concern. If findings of even prospective studies. In our re- workers with CTS are more likely to leave jobs analysis of Nathan et al.’s [1992a] data, 2 of 3 that require forceful and repetitive hand/wrist groups exposed to forceful hand/wrist exertions exertions than jobs without those demands, were more likely to have median nerve slowing then the workers in the highest risk jobs may be when nerve conduction testing was repeated 5 “survivors” (those who did not get CTS). Our years later. The highest exposure group had the analysis of Nathan’s [1992a] data from a same prevalence of slowing as the lowest follow-up of industrial workers shows that exposure group in 1989, whereas there had cases (with median nerve slowing) were more been a higher prevalence rate in 1984. As likely to drop out of the most highly exposed discussed above, this apparent decrease in group than the unexposed group, which might prevalence over 5 years can likely be explained explain why the RR for high exposure by survivor bias. Our interpretations of the data decreased from 2.0 to 1.0 over a 5-year differ from those of the author. Further study is period. Survivor bias results in an needed to clarify these issues. To our underestimate of the RR. knowledge, there is no evidence that workers with pre-existing CTS are more likely to seek Refined or exact measures of exposure to or to be employed in jobs with high force forceful hand/wrist exertions are not always requirements. We believe that employment used in epidemiologic studies (e.g., sometimes practices would, if they had any influence, tend exposure is based on job category and not to exclude new hires with CTS from jobs with actual forceful measurements); this can result in high force requirements for the hand/wrist. some study subjects being assigned to the wrong exposure category. When this occurs, Case definitions in most of the cross-sectional the usual effect is again to underestimate the RR studies excluded cases that occurred before between exposure groups. working on the current job. This limits CTS cases studied to those that occurred following Stetson et al. [1993] did not report RR current exposure. Several of the studies estimates for exposure variables, but they reviewed also required a minimum time period reported that median sensory amplitudes were of working on the job before counting CTS significantly smaller and distal sensory latencies cases. This increases the likelihood that were significantly longer in groups with forceful exposure to forceful hand/wrist exertion hand exertions (p<0.05). Age, height, and occurred for a sufficient length of time to finger circumference were included in statistical develop CTS. models. There is evidence that CTS is also attributable to nonwork causes (hobbies, sports, other medical conditions, and hormonal status in Temporality, Force and CTS women, etc.). One issue which deals with temporality is whether those with Temporal issues can usually best be addressed nonwork-related CTS would be more likely to using longitudinal studies. However, study be hired into jobs requiring more forceful

5a-20 hand/wrist exertions than those without CTS. that increasing levels of force alone resulted in Again, it seems unlikely that those with increased risk for CTS. The only evidence for pre-existing CTS would be preferentially hired an increasing risk for CTS that can be into jobs requiring highly forceful hand/wrist attributed to increasing levels of force alone is exertions. from a comparison across 2 studies that used the same methods. Chiang et al. [1993] and Consistency of Association for Force Silverstein et al. [1987] used the same methods and CTS to measure hand/wrist force requirements and repetitiveness of jobs. Chiang et al. [1993] Most of the statistically significant estimates of used a lower cutoff point (3 kg compared to 4 RR for CTS among workers with exposure to kg) in Silverstein et al.’s [1987] study for forceful hand/wrist exertions were positive. No classifying jobs as “high force”; these studies found statistically significant negative investigators used identical definitions of associations between forceful hand/wrist repetitiveness. Therefore, a comparison of the exertions and CTS. One study reported ORs RR estimates between the 2 studies provides that were less than one among the groups that some information about the level of risk were described as exposed to repetitive hand associated with different levels of force. Chiang movements; chance and study limitations et al. [1993] reported an OR of 2.6 (95% CI cannot be ruled out as possible explanations for 1.0–7.3) for the high force and repetitive this finding. The other nonsignificant estimates (HF/HR) (>3 kg) group (limited to females to of RR were, with one exception, greater than avoid confounding) compared to the low force one. and repetitive (LF/LR) group; whereas Silverstein et al. [1987] reported an OR of Statistical significance can be a function of 15.5 (95% CI 1.7–142) for the HF/HR group power (the ability of a study to detect an (in a statistical model that included gender, age, association when one does exist). In general, years on the job, plant and exposure level) larger studies are necessary in order to have compared to the LF/LR group. This sufficient power to detect associations with rare comparison provides limited evidence of an diseases. CTS is a less frequently observed increased RR for CTS with increasing level of disorder than tendinitis, for example, and so hand/wrist force. larger studies are required to detect associations with confidence. There is more evidence of a dose-response relationship for CTS with increasing levels of Coherence of Evidence, Force and force and repetition combined. Chiang et al. CTS [1993] reported a statistically significant trend of increasing prevalence of CTS with increasing Please refer to the Repetition and CTS Section. exposure level (8.2% [LF/LR], 15.3% [HF or HR], and 28.6% [HF/HR], p<0.01). Silverstein et al. [1987] suggested a multiplicative effect Exposure-Response Relationship, when exposure to high force and high Force and CTS repetitiveness were combined (15.5), compared to high force (1.8) or high None of the studies reviewed demonstrated repetitiveness (2.7) alone.

5a-21 Of the remaining nine studies, seven are Studies Meeting the Four Evaluation Criteria consistent with the combined effect of force and Two studies fulfilled the four criteria for posture repetition [Stetson et al. 1993; Moore and and CTS: Moore and Garg [1994], Silverstein Garg 1994; Osorio et al. 1994; Armstrong and et al. [1987]. The overall study designs are Chaffin 1979; Nathan et al. 1988; Punnett et al. mentioned above; the following section will 1985; Baron et al. 1991], one is not cover the posture assessment. [McCormack et al. 1990]; and one is equivocal [Nathan et al. 1992a]. For the exposure assessment of the posture variables in the Silverstein et al. [1987] study, In conclusion, there is evidence that force three representative workers from each alone is associated with CTS. There is strong selected job performing the jobs for at least evidence that a combination of forceful three cycles were videotaped using two hand/wrist exertion and repetitiveness are cameras. The authors then extrapolated the associated with CTS. posture data to non-observed workers.

POSTURE AND CTS Moore and Garg [1994] used a wrist Definition of Extreme Postures For classification system similar to that used by CTS Stetson et al. [1993], classifying the wrist angle We selected those studies which addressed estimated from videotape as neutral, non- posture of the hand/wrist area including those neutral or extreme if the flexion/extension angle addressing pinch grip, ulnar deviation, wrist was 0° to 25°, 25° to 45° and greater than 45°, flexion/extension. Posture is a difficult variable respectively; or if ulnar deviation was less than to examine in ergonomic epidemiologic studies. 10°, 10° to 20°, and greater than 20°, It is hypothesized that extreme or awkward respectively. postures increase the required force necessary to complete a task. Posture may increase or Strength of Association: Posture and decrease forceful effort; its impact on MSDs CTS may not be accurately reflected in measurement Silverstein found no significant association of posture alone. Reasons that the variable between percentages of cycle time observed in “extreme posture” has not been measured or extreme wrist postures or pinch grip and CTS. analyzed in many epidemiologic studies are: 1) “CTS jobs” had slightly more ulnar deviation because of the extreme variability of postures and pinching but these differences were not used in different jobs as well as the extreme statistically significant. The authors noted that variability of postures between workers among all the postural variables recorded, the performing the same job tasks, variability between individuals with similar or 2) because several studies have taken into identical jobs was probably the greatest for account the effects of posture when determining wrist postural variables. This individual variation other measured variables such as force within jobs was not taken into account in the [Silverstein et al. 1987; Moore and Garg analysis, creating a potential for 1994]; and 3) stature often has a major impact misclassification of individuals by using the on postures assumed by individual workers variable “job category” in the analysis. The during job activities. effect of exposure misclassification is usually to decrease differences between exposure groups

5a-22 and decrease the magnitude of association. a pinched grasp. Given the period of recall for self-reported exposure (0–5 years), and no Moore and Garg’s [1994] classification of jobs independent observation or attributes of did not separate the posture variables from exposure, these results must be interpreted with other work factors, and used posture along caution (meaning that within the limitations of with other variables to classify jobs into the data and conclusions, when considered with “hazardous” and “safe” categories. The RR of other studies that have more stringent methods, CTS occurring in hazardous jobs was 2.8 but the RRs seem consistent and supportive and do not statistically significant (p=0.44). not offer alternate conclusions).

Studies Not Meeting All Four Evaluation Armstrong and Chaffin’s [1979] pilot study of Criteria female sewing machine operators with deKrom et al. [1990] compared certain symptoms and/or signs for CTS compared to exposure factors between 28 CTS cases from controls found that pinch force exertion a community sample and 128 CTS cases from (exposure measurements estimated from EMG, a hospital (a total of 156 CTS cases) to 473 film analysis) was significantly associated (OR community “non-cases” (n=473). The authors 2.0). Pinch force was a combination of relied on self-reported information about factors—posture and forceful exertion. The duration of exposure (hours per week) to CTS authors reported that CTS-diagnosed subjects risk factors (flexed wrist, extended wrist, used deviated wrist postures more frequently extended and flexed wrists combined; pinch than nondiseased, particularly during forceful grasp and typing), with respondents recalling exertions. What is unable to be answered due exposure from the present to 5 years prior from to the study design, was whether the deviated the questionnaire date. Four groups of duration postures were necessitated due to symptoms were used in the analyses (0; 1–7; 8–19, and signs of CTS, or the deviated postures 20–40 hours/week). In this study, the selection caused or exacerbated the symptoms and signs. process of cases was not consistent. Initially, a random population sample was used, then Stetson et al. [1993] found that “gripping hospital outpatients were used to supplement greater than 6 pounds” per hand was a the number of CTS cases when numbers were significant risk factor for median distal sensory found to be insufficient. This may be a problem dysfunction (an indicator of CTS) when the when estimating the etiologic role of workload, study population was divided into exposed and as cases seeking medical care may cause a non-exposed groups. “Gripping greater than 6 referral bias. However, the authors stated that pounds” is a variable which combines two they came up with the same relationship work-related variables, posture and forceful between flexed and extended wrist using only exertion. As seen with other studies referenced CTS cases from the population-based data. above, the single work-related variable was not The risk of CTS was found to increase with the found to be associated with median nerve reported duration of activities with flexed wrist dysfunction, but the combination of variables (RRs from 1.5 to 8.7, with increasing hours) or was significant. Looking specifically at wrist activities with extended wrist (RR from 1.4 to deviation in the Stetson et al. [1993] study, the 5.4 with increasing hours) over the past 5 midpalm to wrist sensory amplitude was smaller years, but not for working with a flexed or in the group not exposed to wrist deviation extended wrist in combination, or working with

5a-23 (p=0.04) compared to those exposed to wrist significantly associated with CTS when deviation (contrary to what was expected). comparing jobs where grip strength was three Also, no significant differences were found in times greater than in the low risk jobs. In those the mean measurements between nonexposed studies which used self-reports for categorizing and exposed groups for use of pinch grip. posture, the associations were also positive.

Tanaka et al. [1995] analysis of the Temporal Relationship Occupational Health Supplement of the NHIS There were no longitudinal studies which population survey depended on self-reported examined the relationship between extreme CTS, self-reported exposure factors, and posture and CTS. Two cross-sectional studies occupation of the respondent for analysis. Self- that met the evaluation criteria addressed the reported bending and twisting of the hand and association between posture and CTS. wrist (OR 5.9) was found to be the strongest Silverstein et al. [1987] did not find a significant variable associated with “medically-called relationship between CTS and extreme CTS” among recent workers, followed by race, posture, but exposure assessment was limited gender, vibration and age (repetition and force to representative workers; inter-individual were not included in the logistic models). variability limited the ability to identify actual Limitations of self-reported health outcome and relationships between postures and CTS. In the exposure do not allow the conclusions of this Stetson et al. [1993] study, the authors study to stand alone; however, when examined mentioned the limitations of interpretation of with the other studies, it suggests a relationship their posture results due to misclassification of between posture and CTS. workers. They extrapolated exposure data to non-observed workers, so individual variability The two other studies which examined posture in work methods and differing anthropometry and its relationship to CTS did not focus on the are not accounted for. These limitations all hand and wrist. English et al. [1995] found a influence outcome, and the conclusions must be relationship between self-reported rotation of interpreted with caution, and considered along the shoulder and elevated arm and CTS, an OR with biomechanical and laboratory studies. of 1.8. Liss et al. [1995] found an OR of 3.7 for self-reported CTS comparing risk factors Coherence of Evidence from dental hygienists to dental assistants, with Flexed wrist postures may reduce the area of self-reported percent of time the trunk was in a the carpal tunnel thus potentially increasing the rotated position relative to the lower body as pressure in the tunnel with a concomitant one of the factors. increase in the risk of CTS [Skie et al. 1990; Armstrong et al. 1991]. Marras and Given these limitations of categorizing posture, Shoenmarklin [1993] found that the variables of three studies [Stetson et al. 1993; Loslever and wrist flexion, extension, angular velocity, and Ranaivosoa 1993; Armstrong and Chaffin wrist flexion, extension, angular acceleration 1979] using different methods to measure discriminated between jobs with a high versus a posture and estimate force, found that the low risk of having an upper extremity combination of significant force and posture reportable injury (an OSHA recordable was significantly related to CTS. Marras and disorder due to repetitive trauma). The authors Shoenmarklin [1993] also found posture to be suggested that this result was due to high

5a-24 accelerations requiring high forces in tendons. [Chatterjee 1992; Silverstein et al. 1987]. Szabo and Chidgey [1989] showed that Chatterjee et al. [1982] performed independent repetitive flexion and extension of the wrist exposure assessment of the vibrating tools, and created elevated pressures in the carpal tunnel found the rock drillers to be exposed to compared to normal subjects, and that these vibration between the frequencies of 31.5 and pressures took longer to dissipate than in 62 Hertz. normal subjects. Observed repetitive passive flexion and extension appeared to “pump up” Silverstein et al. [1987] is discussed above. the carpal tunnel pressure; active motion of the Silverstein [1987] had no quantitative measures wrist and fingers also had an effect over and of vibration, but observed exposure from above that of the passive motions tested. videotapes and found all jobs with vibration Laboratory studies demonstrate that carpal exposure to be highly repetitive and mostly canal pressure is increased from less than forceful jobs. 5mmHg to more than 30 mmHg during wrist flexion and extension [Gelberman et al. 1981]. Studies Not Meeting the Evaluation Criteria There are seven studies on Table 5a–4 that Exposure-Response Relationship, meet at least one of the four criteria. CTS and Posture Few studies address exposure-response In addition, there are 2 clinical case studies of relationship between CTS and extreme vibration and CTS [Rothfleish and Sherman posture. deKrom et al. [1990] reported an 1978; Lukas 1970] that were not controlled for increased risk of CTS with workers reporting confounders and not referenced in Table 5a–4. increasing weekly hours of exposure to wrist Rothfleisch and Sherman [1978] found an flexion or extension (but not a combination of excess of power hand tool users among CTS flexion/extension). Laboratory studies also patients. Lucas [1970] examined workers using support a dose-response relationship of vibrating hand tools including stone cutters, increased carpal tunnel pressure due to tunnelers, coal miners, forest workers and increasing wrist deviation from neutral [Weiss grinders (all with a mean of 14 years exposure et al. 1995] and pinch force [Rempel 1995]. to vibration) and found CTS in 21%. He found that the prevalence of CTS in some groups was In conclusion, there is insufficient evidence in as high as 33% (neither study had a referent the current epidemiologic literature to group.) demonstrate that awkward postures alone are associated with CTS. Cannon et al. [1981] found that the self- reported use of vibrating tools, in combination VIBRATION AND CTS with reported forceful and repetitive hand Definition of Vibration for CTS motions, was associated with a greater We selected studies that addressed manual incidence of CTS than was repetitive motion work involving vibrating power tools and CTS alone. specifically. Bovenzi’s study in 1994 compared stone Studies Meeting the Four Evaluation Criteria workers (145 quarry drillers and 425 stone Two studies examining the association between carvers) exposed to hand-transmitted vibration vibration and CTS fulfilled the four criteria to 258 polishers and machine operators who

5a-25 performed manual activity only not exposed to a strong potential for confounding by hand or hand-transmitted vibration. CTS was assessed wrist posture and forceful exertion. by a physician, and exposure was assessed through direct observation to vibrating tools and Temporal Relationship by interview. Vibration was also measured in a There were no longitudinal studies which sample of tools. examined the relationship between vibration and CTS. Strength of Association: Vibration and CTS Consistency in Association Chatterjee et al. [1982] found a significant All studies on Table 5a–4 examining vibration difference between rock drillers with symptoms and CTS found a significantly positive and signs of CTS and the controls using the relationship between CTS and vibration following NCS measurements: median motor exposure. Most studies had ORs greater than latency, median sensory latency, median 3.0, so that results were less likely to be due to sensory amplitude, and median sensory confounding. duration, all at the p<0.05 level. Based on nerve conduction measurements, they also Coherence of Evidence and Vibration found an OR of 10.9 for rock drillers having The mechanism by which vibration contributes abnormal NCS amplitudes in the median and to CTS and tendinitis development is not well ulnar nerves compared to controls. Bovenzi et understood, probably because vibration al. [1991] found an OR of 21.3 for CTS based exposure is usually accompanied by exposure on symptoms and physical exam comparing to forceful and repetitive movements. Muscles vibration-exposed forestry operators using exposed to vibration exhibit a tonic vibration chain-saws to maintenance workers performing reflex that leads to increasing involuntary manual tasks. Bovenzi’s study in 1994 found an muscle contraction. Vibration has also been OR of 0.43 for CTS defined by signs and shown to produce short-term tactility symptoms, controlling for several confounders. impairments which can lead to an increase in In the Silverstein et al. [1987] study the crude the amount of force exerted during manipulative OR for high force/high repetition jobs with tasks. Vibration can also lead to mechanical vibration compared to high force/high repetition abrasion of tendon sheaths. Neurological and without vibration was 1.9, but not statistically circulatory disturbances probably occur significant. This suggested that there may have been confounding (the OR was not statistically independently by unrelated mechanisms. significant) between high force/high repetition Vibration may directly injure the peripheral and vibration. Nilsson et al. [1990] found that nerves, nerve endings, and mechanoreceptors, platers operating tools such as grinders and producing symptoms of numbness, tingling, chipping hammers had a CTS prevalence of pain, and loss of sensitivity. It has been found in 14% compared to 1.7% among office workers. rats that vibration has caused epineural edema Nathan et al. [1988] found a PR of 2.0 (95% in the sciatic nerve [Lundborg et al. 1987]. CI 1.3–3.4) for slowing of nerve conduction Vibration may also have direct effects on the velocity when grinders were compared to digital arteries. The innermost layer of cells in administrative and clerical workers. Cannon et the blood vessel walls appears especially al. [1981] found an OR of 7.0 for CTS with the susceptible to mechanical injury by vibration. If use of vibrating hand tools, although there was damaged, these vessels may become less

5a-26 sensitive to the actions of certain vasodilators epidemiologic studies of CTS that address that require an intact endothelium. The NIOSH work factors also take into account potential Criteria Document on exposure to hand-arm confounders. vibration NIOSH [1989] quoted Taylor [1982] as follows: “ It is not known whether vibration Almost all of the studies reviewed controlled for directly injures the peripheral nerves thereby the effects of age in their analysis [Chiang et al. causing numbness and subsequent sensory loss, 1990, 1993; Stetson et al. 1993; Silverstein et or whether the para-anaesthesia of the hands is al. 1987; Wieslander et al. 1989; Baron et al. secondary to the vascular constriction of the 1991; Tanaka et al. 1995, In Press; blood vessels causing ischemia . . . in the nerve McCormack et al. 1990]. Likewise, most organs.” studies included gender in their analysis, either by stratifying [Schottland et al. 1991; Chiang et Exposure-Response Relationship, al. 1993], by selection of single gender study CTS and Vibration groups [Morganstern et al. 1991; Punnett et al. In the studies examined, only dichotomous 1985] or by including the variable in the logistic categorizations were made, so conclusions regression model [Silverstein et al. 1987; concerning an exposure-response relationship Stetson et al. 1991; Baron et al. 1991]. cannot be drawn. However, we can see Through selection of the study population and significantly contrasting rates of CTS between exclusion of those with metabolic diseases, high and low exposure groups. Wieslander et most studies were able to eliminate the effects al. [1989] found that based on exposure from these conditions. Other studies did control information obtained from telephone interviews, for systemic disease [Chiang et al. 1993; Baron CTS surgery was significantly associated with et al. 1991]. Anthropometric factors have also vibration exposure. Exposure for 1–20 years been addressed in several studies [Stetson et al. gave an OR of 2.7, more than 20 years gave an 1993; Nathan et al. 1997; 1992b; Werner et OR of 4.8. al. 1997]. As more is learned about confounding, more variables tend to be Conclusion addressed in more recent studies (smoking, In conclusion, there is evidence supporting caffeine, alcohol, hobbies). In those older studies which may not have controlled for an association between exposure to vibration multiple confounders, it is unlikely that they are and CTS. highly correlated with exposure, especially those with ORs above 3.0. When examining CONFOUNDING AND CTS those studies that have good exposure It is clear that CTS has several non- assessment, widely contrasting levels of occupational causes. When examining the exposure, and that control for multiple relationship of occupational factors to CTS, it is confounders, the evidence supports a positive important to take into account the effects of association between occupational factors and these individual factors; that is, to control for CTS. their confounding or modifying effects. Studies that fail to control for the influence of individual CONCLUSIONS factors may either mask or amplify the effects There are over 30 epidemiologic studies which of work-related factors. Most of the have examined workplace factors and their

5a-27 relationship to CTS. These studies generally jobs with exposure to vibration and CTS. compared workers in jobs with higher levels of There is strong evidence for a relationship exposure to workers with lower levels of between exposure to a combination of risk exposure, following observation or factors (e.g., force and repetition, force and measurement of job characteristics. Using posture) and CTS. Ten studies allowed a epidemiologic criteria to examine these studies, comparison of the effect of individual versus and taking into account issues of confounding, combined work risk factors [Chiang et al. bias, and strengths and limitations of the studies, 1990, 1993; Moore and Garg 1994; Nathan et we conclude the following: al. 1988, 1992a; Silverstein et al. 1987; Schottland et al. 1991; McCormack et al. There is evidence for a positive association 1990; Stetson et al. 1993; Tanaka et al. [In between highly repetitive work and CTS. Press]. Nine of these studies demonstrated Studies that based exposure assessment on higher estimates of RR when exposure was to a quantitative or semiquantitative data tended to combination of risk factors, compared to the show a stronger relationship for CTS and effect of individual risk factors. Based on the repetition. The higher estimates of RR were epidemiologic studies reviewed above, found when contrasting highly repetitive jobs to especially those with quantitative evaluation of low repetitive jobs, and when repetition is in the risk factors, the evidence is clear that combination with high levels of forceful exposure to a combination of job factors exertion. There is evidence for a positive studied (repetition, force, posture, etc.) association between force and CTS based on increases the risk for CTS. This is consistent currently available epidemiologic data. There is with the evidence that is found in the insufficient evidence for a positive biomechanical, physiologic, and psychosocial association between posture and CTS. There is literature. evidence for a positive association between

5a-28 Table 5a-1. Epidemiologic criteria used to examine studies of carpal tunnel syndrome (CTS) associated with repetition

Physical Investigator examination, blinded to Risk indicator and/or nerve case and/or (OR, PRR, IR or Participatio conduction exposure Basis for assessing Study (first author and p-value)*,† n rate $70% studies status hand exposure to repetition year)

Met all four criteria:

Chiang 1990 1.87† Yes Yes Yes Observation or measurements

Chiang 1993 1.1 Yes Yes Yes Observation or measurements

Moore 1994 2.8 Yes Yes Yes Observation or measurements

Osorio 1994 6.7 Yes Yes Yes Observation or measurements

Silverstein 1987 5.5† Yes Yes Yes Observation or measurements

Met at least one criterion:

Barnhart 1991 1.9–4.0† No Yes Yes Observation or measurements

Baron 1991 3.7 No Yes Yes Observation or measurements

Cannon 1981 2.1 NR‡ Yes NR Job titles or self-reports

English 1995 0.4 Yes Yes Yes Job titles or self-reports

Feldman 1987 2.26† Yes No NR Observation or measurements

McCormack 1990 0.5 Yes Yes NR Job titles or self-reports

Morgenstern 1991 1.88 Yes No No Job titles or self-reports

Nathan 1988 1.0 NR Yes NR Observation or measurements

Nathan 1992a 1.0 No Yes NR Observation or measurements

Punnett 1985 2.7† No Yes NR Job titles or self-reports

Schottland 1991 2.86†, NR Yes NR Job titles or self-reports 1.87

Stetson 1993 NR Yes Yes NR Observation or measurements

Weislander 1989 2.7† Yes Yes No Job titles or self-reports

Met none of the criteria:

Liss 1995 5.2 No No No Job titles or self-reports 3.7†

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. 5a-29

Table 5a-2. Epidemiologic criteria used to examine studies of carpal tunnel syndrome (CTS) associated with force

Physical Investigato examination, r blinded to Risk indicator and/or nerve case and/or (OR, PRR, IR, Participatio conduction exposure Basis for assessing Study (first author and year) or p-value)*,† n rate $70% studies status hand exposure to force

Met all four criteria:

Chiang 1993 1.8† Yes Yes Yes Observation or measurements

Moore 1994 2.8 Yes Yes Yes Observation or measurements

Osorio 1994 6.7 Yes Yes Yes Observation or measurements

Silverstein 1987 15.5† Yes Yes Yes Observation or measurements

Met at least one criterion:

Armstrong 1979 2.0† NR‡ No No Observation or measurements

Baron 1991 3.7 No Yes Yes Observation or measurements

McCormack 1990 0.4-0.9 Yes Yes NR Job titles or self-reports

Nathan 1988 1.7-2.0† NR Yes NR Observation or measurements

Nathan 1992a 1.0, 1.4†, 1.6 No Yes NR Observation or measurements

Punnett 1985 2.7† No Yes NR Job titles or self-reports

Stetson 1993 NR† Yes Yes NR Observation or measurements

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

5a-31

Table 5a-3. Epidemiologic criteria used to examine studies of carpal tunnel syndrome (CTS) associated with posture

Physical Investigator examination, blinded to Risk indicator and/or nerve case and/or (OR, PRR, IR, Participato conduction exposure Basis for assessing Study (first author and or p-value)*,† n rate studies status hand exposure to posture year) $70%

Met all four criteria:

Moore 1994 2.8 Yes Yes Yes Observation or measurements

Silverstein 1987 NR‡ Yes Yes Yes Observation or measurements

Met at least one criterion:

Armstrong 1979 2.0† NR No No Observation or measurements

deKrom 1990 5.4† Yes Yes NR Job titles or self-reports

English 1995 1.8† Yes Yes Yes Job titles or self-reports

Stetson 1993 NR† Yes Yes NR Observation or measurements

Tanaka 1995 5.9† Yes No No Job titles or self-reports

Met none of the criteria:

Liss 1995 3.7† No No No Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on posture alone (i.e., posture plus repetition, force, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

5a-33

Table 5a-4. Epidemiologic criteria used to examine studies of carpal tunnel syndrome (CTS) associated with vibration

Physical Investigato examination, r blinded to Risk indicator and/or nerve case and/or (OR, PRR, IR, Participation conduction exposure Basis for assessing hand Study (first author and or p-value)*,† rate $70% studies status exposure to vibration year)

Met all four criteria:

Chatterjee 1992 10.9† Yes Yes Yes Observation or measurements

Silverstein 1987 5.3† Yes Yes Yes Observation or measurements

Met at least one criterion:

Bovenzi 1991 21.3† NR‡ Yes Yes Observation or measurements

Bovenzi 1994 3.4† Yes Yes No Observation or measurements

Cannon 1981 7.0† NR Yes NR Job titles or self-reports

Färkkilä 1988 NR† NR Yes NR Job titles or self-reports

Koskimies 1990 NR† NR Yes No Observation or measurements

Tanaka In Press 1.8† Yes No No Job titles or self-reports

Weislander 1989 3.3† Yes Yes No Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on vibration alone (i.e., vibration plus repetition, posture, or force). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

5a-35

Table 5a–5. Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Armstrong Case- 18 female sewing Outcome: CTS defined as Õ Õ For pinch force Participation rate: Not reported. and Chaffin control machine operators with history of symptoms, surgical exertion: 2.0 1.6-2.5 1979 CTS histories compared decompression of the median All cases of CTS diagnosed prior to to 18 female sewing nerve, positive Phalen’s test, or For hand study in working sewing machine machine operators thenar atrophy. force: 1.05 1.0-1.2 operators, may cause referral bias without CTS histories. in estimating role of workload. Exposure: Hand/wrist postures and estimation of Subjects excluded if history of forearm flexor force in various fractures, metabolic or soft tissue wrist and hand postures disease. assessed by film analysis and EMG. No association found between hand size or shape and CTS.

CTS diagnosed subjects used deviated wrist more frequently than non-diseased, particularly during forceful exertions.

(Continued)

5a-37

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Barnhart Cross- Ski manufacturing Outcome: CTS determined by: Case 1: Participation rate: 70% (repetitive et al. 1991 sectional workers: 106 with (1) Case 1: Electro-diagnosis 34% 19% 1.9 1.0-3.6 jobs), 64% (non-repetitive jobs). repetitive jobs compared of median-ulnar difference to 67 with non-repetitive (latency on response time); Case 2: Examiner blinded to subject’s job jobs. (2) Case 2: Either Tinel's or 15.4% 3.1% 3.95 1.0-15.8 status but clothing may have biased Phalen's test and electro- observations. diagnosis; (3) Case 3: Ever Case 3: having symptoms of hand pain, 32.5% 18.2% 1.6 0.8-3.2 Controlled for age and gender. tingling, numbness, or nocturnal hand pain and Tinel's Found for both right and left hand or Phalen's test and electro- of those with repetitive jobs; mean diagnosis. difference between distal sensory latencies of median and ulnar Exposure: Jobs classified as nerves were primarily due to a repetitive and non-repetitive. shorter mean sensory latency of Repetitive jobs entailed the ulnar nerve. repeated or sustained flexion, extension, or ulnar deviation of There was no difference in median the wrist by 45E, radial nerve distal sensory latencies deviation by 30E, or pinch grip between groups. (determined by observation). Hormonal status, systemic disease included in questionnaire.

Diabetes significantly more frequent in those with CTS than without (p=0.01).

(Continued)

5a-38

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Baron et al. Cross- 119 female grocery Outcome: CTS case defined 11% 4% 3.7 0.7-16.7 Participation rate: 85% checkers; 1991 sectional checkers vs. 56 other as having moderate to severe 55% non-checkers in field study. female grocery store symptoms of pain, stiffness, Following telephone survey 91% employees (comparison numbness, tingling. Symptoms checkers and 85% non-checkers. group). begun after employment in the current job; lasted > one week Adjusted for duration of work. or occurred > once a month during the past year; no history Total repetitions/hr ranged from of acute injury to part of body 1,432 to 1,782 for right hand and in question and a positive 882 to 1,260 for left hand. physical exam of either Phalen's or Tinel's test. Multiple awkward postures of all upper extremities recorded but not Exposure: Based on job analyzed in models. category, estimates of repetitive, average, and peak Examiners blinded to worker’s job forces based on observed and and health status. videotaped postures, weight of scanned items, and subjective Controlled for duration of work, assessment of exertion. hobbies.

Exposure level in checkers: Average forces: Low Peak force: Medium Repetition: Medium

Exposure level in referents: Average force: Medium Peak force: Medium to low Repetition: Medium.

(Continued)

5a-39

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi Cross- 65 vibration-exposed Outcome: CTS cases defined 38.4% 3.2% 21.3 (adjusted) p=0.002 Participation rate: Not reported. et al. 1991 sectional forestry operators using as having symptoms of pain, chain-saws compared to numbness, or tingling in the Examiners blinded to case status. referents composed of median nerve distribution, and 31 maintenance workers physical exam findings of Controlled for age and ponderal (electricians, mechanics, Tinel's or Phalen's test, index (height and weight variable). and painters). diminished sensitivity to touch Metabolic disease also considered. or pain in 3½ fingers on radial side, weakness in pinching or Controls also found to have several gripping. risk factors for MSDs at work—static arm and hand Exposure: Direct observation overload, overhead work, stressful of awkward postures, manual postures, non-vibrating hand-tool forces, and repetitiveness use. evaluated via checklist. The focus of the study was to Controls had a greater proportion of compare vibration-exposed time in work cycles shorter than workers to controls doing 30 sec than forestry workers. manual work. Vibration measured from two chain- Chain saw operators worked saws. Vibration exposure for outdoors and were exposed to each worker assessed in lower temperatures than terms of 4-hr energy- maintenance workers. equivalent frequency-weighted acceleration according to ISO 5349.

(Continued)

5a-40

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi and Cross- Case group: Stone Outcome: CTS assessed by 8.8% 2.3% 3.4 1.4-8.3 Participation rate: 100%. “All the the Italian sectional workers employed in physician assessment. CTS active stone workers participated in Group 1994 9 districts in Northern defined as symptoms, the study, so self-selection was and Central Italy; (1) parathesias, numbness, or not a source of bias.” 145 quarry drillers and pain in median nerve 425 stone carvers distribution; (2) nocturnal Physician administered exposed to vibration. exacerbation of symptoms and questionnaires containing work positive Tinel's or Phalen's test. history and examinations, so Referent group: unlikely to be blinded to case Polishers and machine Exposure: Direct observation status. operators (n=258) who of vibrating tools assessed by performed manual interview. Vibration measured Adjusted for age, smoking, alcohol activity but were not in a sample of tools. consumption, and upper limb exposed to hand- injuries. transmitted vibration. Leisure activities and systemic All stone workers diseases included in questionnaire. employed in 6 districts participated in the survey Univariate analysis showed no (n=578, 69.8%), association between systemic whereas, in the three diseases and vibration so were not other districts they were criteria for exclusion. selected on basis of random sampling of the Dose-response for CTS and lifetime quarries and mills in the vibration exposure not significant. geographic areas (n=250, 30.2%).

(Continued)

5a-41

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Cannon et al. Case- Aircraft engine workers Outcome: CTS cases identified Õ Õ For vibrating Participation rate: Participation rate 1981 control at 4 plants: 30 CTS through worker’s hand tool use: unable to be calculated from data cases identified through compensation claims and 7.0 3.0-17 presented. 30 cases identified worker’s compensation medical department records through record review of 20,000 claims and medical during a 2-year period. For repetitive workers. department records motion tasks: during a 2-year period Exposure: Based on job 2.1 0.9-5.3 Cases and controls on gender. compared to 90 controls category, years on the job, from the same plant, identified through record History of Controlled for gynecologic surgery, 16 workers receiving review and interviews. gynecologic race, diabetic history, years on the compensation benefits Exposure to vibrating tools, surgery: job, use of low-frequency vibrating for treatment of CTS, and repetitive motion. 3.7 1.7-8.1 tools. 14 cases who had not received compensation Buffing, grinding, and hand Years on the Information obtained through self- benefits. tools were measured with an job: administered questionnaires and accelerometer and found to be 0.9 0.8-1.0 personal interviews on cases and Three controls randomly in the range of 10 to 60 Hz. controls on age, sex, race, weight, chosen from the same occupation, years employed, plant for each CTS case. worker compensation status, history of metabolic disease, hormonal status of females, history of gynecologic surgery.

Number of years employed significantly different among cases (5.5 years) and controls (11.7 years). Range of years employed among cases included 0.1 year to 28 years.

(Continued)

5a-42

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Chatterjee Case- 16 rock drillers compared Outcome: CTS was 44% 7% Abnormal Participation rate: 93%. et al. 1982 control with 15 controls. determined by symptoms from amplitudes of questionnaire and interview by digital-action Examiners blinded to case status. medical investigator, clinical potentials from exams carried out blindly, and fingers Groups standardized for age and nerve conduction studies. For supplied by the gender. Table 5-7, CTS based solely on median and NCS results; Table 5-9 based ulnar nerves; Exclusionary criteria: History of on symptoms and NCS. the OR in constitutional white finger, vibration secondary causes of Raynaud’s Exposure: To vibration carried exposed vs. phenomenon, > one laceration or out by measurement of controls: fracture in the hands or digits, vibration spectra of the rock OR=10.89 1.02-524 severe or complicated injury drills and observation of jobs. involving nerve or blood vessels or Exposed group were those significant surgical operation, miners who regularly used history of exposure to vibration rock-drills in the fluorspar from tools other than rock drills. mines or other miners using similar rock-drills. Exposure Significant differences found varied from 18 months to between controls and vibration 25 years (mean 10 years). group for symptoms of numbness The rock drillers were exposed and tingling: median motor latency; to vibration level in excess of median sensory latency; median the damage level criterion sensory amplitude; median sensory between the frequencies of duration. All at the p< 0.05 level. 31.5 and 62 Hz. Skin temperature controlled for in NCVs.

(Continued)

5a-43

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Chiang et al. Cross- 207 active workers from Outcome: CTS defined as Group 1: Participation rate: Not specifically 1990 sectional 2 frozen food plants symptoms of numbness, pain, 4% clinical mentioned, however, paper states divided into 3 groups: tingling in the fingers plus 2% that “in order to prevent selective (1) low-cold, low- innervated by the median sub-clinical bias, all of the employees in the repetition (comparison nerve, onset since work in factories were observed initially.” group, mainly office staff current job, no relationship to Group 2: Group 2 vs. and technicians, n=49), systemic disease or injury and 40.5% Group 1: Examiners blinded to exposure (2) low-cold, high- physical exam of Tinel's test or clinical plus OR=8.28 1.18-58.3 status and medical history. repetition (non-frozen Phalen's sign. Nerve 8.1% food packers, n=37), conduction testing was sub-clinical Controlled for age, sex, and length (3) high-cold, high- performed on motor and of employment. Interaction terms repetition (frozen food sensory nerves of both upper Group 3: Group 3 vs. tested. packers, n=121). limbs. If subject had abnormal 37.2% Group 1: results and symptoms and clinical plus OR=11.66 2.92-46.6 Excluded subjects with diabetes, physical exam findings, was 22.3% thyroid function disorders, history considered CTS. If no sub-clinical of forearm fracture, unspecified symptoms, considered as polyneuropathy, rheumatoid subclinical CTS. Logistic arthritis. Regression Exposure: Job analyses Model: Workers in cold groups wore conducted by industrial Cold: gloves and exerted higher forces hygienist, to cold and repetition OR=1.85 than workers in non-cold groups. assessed by observation. (p<0.22) Force was not evaluated in this study. Confounding is possible Highly repetitive jobs had cycle Repetitiveness: according to authors. times <30 sec. >50% of cycle OR=1.87 time cold exposure was (p<0.018) CTS was independent of age and defined as whether the job length of employment. Authors required hands to be locally Cold x considered this to be due to healthy exposed to cold. The mean Repetitive- worker effect. skin temperature of their hands ness: was in the range of 26 to OR=1.77 OR for group 1 vs. group 2 is 8.3 28EC, even with wearing (p<0.03) (1.2-58.3) when adjusted for sex gloves. but 2.2 (0.2-21.1) when adjusted for sex, age, and length of employment suggesting survival bias.

(Continued)

5a-44

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Chiang et al. Cross- 207 fish processing Outcome: CTS defined as Group 2 Group 1 Participation rate: Paper stated that 1993 sectional workers divided in having symptoms of (Male): (Male): 2 vs. 1 (male): all of the workers who entered the 3 groups: (1) low-force, numbness, pain, or tingling in 6.9% 3.1% OR= 2.2 0.2-22.0 fish-processing industry before June 1990 and were employed low-repetition the fingers innervated by the there full-time were part of the (comparison group, median nerve, onset after job Group 2 Group 1 2 vs. 1 cohort. n=61); (2) high-force or began, and no evidence of (Female): (Female): (female): high-repetition (n=118); systemic disease or injury and 18.0% 13.8% OR=1.3 0.5-3.5 Workers examined in random sequence to prevent observer bias; (3) high-force and high- physical exam findings of examiners blinded to case status. repetition (n=28). positive Tinel's sign or Phalen's Group 3 test. (Male): 3 vs. 1 (male): Analysis controlled for age, 0.0% Õ Õ stratified by gender. Exposure: Assessed by Contraceptive use (females): observation and recording of Group 3 significant (OR=2.0, 95% CI 1.2 to tasks and biomechanical (Female): 3 vs. 1 5.4); tubal ligation not significant. movements of 3 workers, each 36.4% (female): Workers with hypertension, representing 1 of 3 study OR=2.6 1.0-7.3 diabetes, history of traumatic groups. Highly repetitive jobs injuries to upper limbs, arthritis, with cycle time <30 sec or collagen diseases excluded from study group. >50% of cycle time performing the same fundamental cycles. Repetition: No significant age difference in Hand force from EMG OR=1.1 0.7-1.8 exposure groups. recordings of forearm flexor Physician-observed cases about muscles. Classification of Force: ½ the prevalence of symptoms of workers into 3 groups OR=1.8 1.1-2.9 elbow pain (9.8 vs. 18.0; 15.3 vs. according to the ergonomic 19.5; 35.7 vs. 17.9). risks of the shoulders and Repetition and Dose-response for symptoms both upper limbs: Group 1: low- force: in the hand and in the wrist repetition and low-force; Group OR=1.1 0.7-1.8 (p<0.03) and physician-observed 2: high-repetition and high- CTS (p<0.015). force; Group 3: high-repetition Male vs. Age, gender, repetitiveness, or high-force. female: forceful movement of upper limbs OR=2.6 1.3-5.2 and interaction of repetitiveness and forceful movement calculated in logistic regression. Significant trend for duration of employment in <12 months but not 12 to 60 months or >60 months.

(Continued)

5a-45

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments deKrom et al. Nested 28 CTS cases from a Outcome: Tingling pain and 5.6% Õ For work: Participation rate: 70% response rate obtained for both hospital and 1990 case community sample and numbness in median prevalence 20 to 40 hr/wk community samples. control 128 CTS cases from a distribution, frequency in the with flexed Controlled for age, weight, slimming hospital (total n=156) $2/week, awakened at night general wrist: OR=8.7 3.1-24.1 courses, gender, and checked for compared to community and nerve conduction studies. population interactions. non-cases (n=473). Motor latency < 4.5 months, (28 cases For work: Cases seeking medical care may different median to ulnar DSL < from 20 to 40 hr/wk cause referral bias in estimating etiologic role of work-load. Participants blinded to 4.0 months, controlled for 501 subject with extended However, authors came up with aim of study—told it was temperature. community wrist: same relationship between flexed about “general health.” sample) OR= 5.4 1.1-27.4 and extended wrist using only CTS CTS diagnosed by clinical cases from population-based data. history and neurophysiological The associations from this study are based on very small sample tests. sizes. >64% of cases reported 0 hr/wk to each of the exposures. Exposure: Awkward In random sample, age, and sex hand/finger postures and pinch stratified, included twice as many grasps assessed by females as males. No significant relationship between questionnaire: Self-reported pinch grasp or typing. information about duration of Dose-response found for duration exposure (hr/wk) to flexed of activities with flexed or extended wrist, extended wrist, wrist statistically significant; dose- response relationship for both extended and flexed wrist present but not statistically combined, pinched grasp. significant. Typing hr categorized as Typing hr not significant but very 0, 1 to 7, 8 to 19, 20 to small numbers (<5 in comparison 40 hr/wk of exposure 0 to 5 groups); may have been unable to detect a difference. years ago, responses Females with hysterectomy without truncated at 40 hr/wk. oophorectomy significantly increased risk, PRR=2.0 (1 to 3.6), compared to females not operated on; increase may be detection bias. Wrist fractures, thyroid disease, rheumatism, and diabetes not significant for CTS. Varicosis significant risk for males 12.0 (3.6-40.1). Oral contraceptives not significantly associated with CTS.

(Continued)

5a-46

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments English et al. Case- Cases: CTS patients Outcome: CTS based on Õ Õ Rotating Participation rate: 96%. 1995 control (n=171) ages 16 to 65 agreed criteria diagnosed by shoulder with years from orthopedic orthopedic surgeons using elevated arm Due to design of study (cases clinics. Controls: common diagnostic criteria (not and CTS: selected by diagnoses), blinding of (n=996) 558 males and specified). OR=1.8 1.2-2.8 examiners not an issue. 438 females attending the same clinics Exposure: Based on self- Repeated Adjusted for height, weight, and diagnosed with reported risk factors at work: finger tapping gender. conditions other than questions addressed: and CTS: diseases of the upper awkward postures, grip types, OR=0.4 0.2-0.7 Significant negative association limb, cervical, or thoracic wrist motions, lifting, shoulder with height and presentation at the spine; ages 16 to 65 postures, static postures, etc. clinic as a result of an accident and years. and job category. CTS.

A significantly positive association with height.

Included “frequency of movements” in regression analysis.

(Continued)

5a-47

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Färkkilä Cross- 79 chain saw users Outcome: CTS based on nerve 26% Õ Significant Õ Participation rate: 100% of et al. 1988 sectional randomly selected from conduction studies, motor and correlation professional forestry workers. 186 forestry workers sensory conduction velocity, between with >500 hr of distal and proximal latencies, numbness in Significant correlation between CTS sawing/year. Tinel’s and Phalen's tests and the hands and HAVs found. subjective symptoms. (r=0.38, p<0.05) and Randomly selected from EMG out of Exposure: Chain saw vibration CTS and 186. not measured. Duration of muscle fatigue chain saw use determined by (r=0.47, Alcohol consumption did not interview. p<0.05) and correlate with numbness in the CTS. hands or arms (r=0.14, p=NS) or sensory disturbances.

Only motor nerve recordings were analyzed for this study.

(Continued)

5a-48

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Feldman Cross- 586 electronics workers Outcome: Based on Wrist Wrist Numbness and Participation rate: 84%. et al. 1987 sectional at a manufacturing firm questionnaire survey and in tingling and tingling and tingling in Examiners blinded to case and for with 700 employees. some an abbreviated numbness: numbness: fingers: exposure status: Not stated. neurologic examination that symptom involved tests of hand 18% 8.7% OR=2.26 1.4-4.46 Analysis not controlled for survey sensation, finger grip, and confounders. strength of thenar muscles. High-risk vs. Questionnaire obtained data on Pro- Tinel’s and Phalen's done. low-risk jobs: past medical history, exposure to spective “Standard nerve conduction” of p<0.005 neurotoxins, cigarettes, hobbies, for nerve left and right median nerves. and symptoms. con- For nerve conduction testing, the Exposure: Two subjects duction temperature of limbs was monitored randomly selected for and controlled for. studies biomechanical analyses from each of four high-risk areas, More females were in high-risk determined from questionnaire areas and jobs than males. and walk-through observations There were no workers >60 years of tasks involving repetitive old in high-risk group. There were flexion, extension, pinching, 34 workers >60 years in and deviated wrist postures. comparison groups. Videotaping and Rheumatoid arthritis more prominent electromyography done. in low-risk group (8.2%) than high- risk (2.4%) group. Highly repetitive job task defined as <30 sec cycle or Nerve conduction in high-risk >50% of cycle performing the workers performed year 1 and fundamental cycle. year 2. Right sensory amplitude abnormal (<8µV) in 22% of Wrist posture characterized in workers at year 1 and 35.5% at terms of flexion and extension: year 2. Left sensory amplitude >45 flexed, 15 to 45 flexion, abnormal in 16.7% and 29% at neutral, 15 to 45 extension, and year 2. >45 extension and deviation. Most apparent changes (increases) Hand posture characterized by seen in bilateral sensory velocities 6 types of grip. and motor latencies (abnormal >4.5). Right motor latency abnormal No quantitative measures of in 8% at year 1 and 11% in year 2. vibration were obtained. Left motor latency abnormal in 2% in year 1 and 23% at year 2. Authors offered parameters for staging CTS in high-risk subjects (0 to 4 stages).

(Continued)

5a-49

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Franklin Retro- Workers in Washington Outcome: Assessed using 25.7 claims/ 1.74 14.8 (oyster 11.2- 19.5 Participation rate: This is a records et al. 1991 spective State (n=1.3 million full- workers’ compensation claims 1,000 FTEs claims/ and crab review so it does not apply. cohort: time workers in 1988). for CTS using ICD codes 354.0 (oyster and 1,000 FTEs packers) from and 354.1. Incident claim was crab (industry Among claimants, the female-to- 1984 to Worker’s compensation the first appearance of a paid packers) wide rate) male ratio was 1.2:1. 1988 data for Washington bill for claimant with a State, using physician diagnosis. Algorithm Mean age of claimants was 37.4. compensable (time loss) was developed to identify 23.9 claims/ and non-compensable unique claimants which 1,000 FTEs 13.8 (meat and 11.6- 16.4 Diagnosis and data entry errors claims for January 1984 removed multiple claims. (meat and poultry comprised 25% of CTS surgery to December 1988. poultry workers) claims—cases were not coded as Exposure: Not measured. workers) CTS. Workers in the same industrial classification assumed to 82% of claims were true cases of share similar workplace CTS. exposures.

(Continued)

5a-50

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Koskimies Cross- 217 forestry workers Outcome: 125 randomly Active Alcohol Participation rate: Not reported. et al. 1990 sectional who used chain saw selected for EMG of sensory vibration: consumption >500 hr during previous and motor nerves both hands. 5% white and CTS cases Examiners may not have been 3 years. finger r=0.15 p=NS blinded to exposure status because CTS diagnosis based on Vibration of design of study. symptoms, exclusion of other CTS: 20% exposure time conditions, results of Phalen’s and motor NCV No comparison group because and Tinel’s test, and findings in in median study was part of longitudinal study sensory and motor nerve EMG. nerve of right of workers followed since 1972. hand: r=-0.27 p=0.01 but not left Exposure: Number of years of hand: r=-0.12 p=NS Most of 25 CTS workers had mild vibration exposure (only symptoms at work despite severe workers who had 500 hr Exposure time reduction of sensory NCS of during previous 3 years were with both median nerve. included. motor NCV in ulnar nerve of Males with primary Raynaud’s right hand r=-0.26 and left p=0.05 disease, rheumatoid arthritis, hand diabetes, or positive urine glucose r=-0.39. p<0.001 slide test results excluded from study. Distal latencies in median 12 (48%) of those with CTS had nerve and bilateral diagnosis. The authors exposure in right hand stated that the left hand is the r=0.17; p=0.05 dominant working hand in sawing, left hand the right hand acting more to direct r=0.21. p=0.05 the saw during the operation. Numbness and sensory NCS of median nerve; right hand r=0.679; p<0.001 left hand r=0.53. p<0.01

(Continued)

5a-51

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Liss et al. Cross- 1,066 of 2,142 dental Outcome: Mailed survey, Responder Responder Participation rate: 50% response 1995 sectional hygienists from Ontario 2 CTS case definitions: told that told that rate from both groups. Canada Dental (1) based on positive response they had they had Hygienists Association to "told by a physician that you CTS: CTS: 0.9% OR=5.2 0.9-32 Study population >99% female. compared to referent had CTS", (2) if during last 12 7% group, 154 of 305 dental months, for >7 days Question- OR were age adjusted. assistants. experienced numbness and Question- naire tingling, pain, or burning in naire based based Confounders considered included distribution of median nerve, CTS: 11% CTS: 3.0% OR=3.7 1.1-11.9 typing, hobbies, and taking night pain or numbness in estrogens. hands, and no previous wrist/hand injury.

Exposure: Based on mailed survey: Length of practice, days/wk worked, patients/day, patients with heavy calculus, percent of time trunk in rotated position relative to lower body, instruments used, hr of typing/wk, type of practice.

(Continued)

5a-52

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Loslever and Cross- 17 selected jobs with Outcome: Occupational Mean High force with Participation rate: Cases selected. Ranaivosoa sectional frequent and repeated physician from each factory prevalence high flexion 1993 absences of workers involved in the study completed rate among and CTS: Occupational doctor supplied due questionnaire concerning each jobs (jobs r=0.62 information on gender, age, years to CTS investigated at job and the number of CTS chosen at on the job, hand orientation, has or the request of cases. The prevalence of CTS workplaces High force and has not contracted CTS. occupational doctors and was then calculated from ratio where CTS high extension managers. of CTS cases and total number had been and CTS: Subjects spent 60 to 80% of their Biomechanical data of employees that worked at reported): r=0.29 time in extension ranging from 13 to recorded on a number of that place. 35% (range 30E. workers from each job, 8 to 66%); ranging from 1 to Exposure: Videotaping of prevalence Vibratory tools more often used in 4 workers. Involving movements, use of vibrating of CTS in tasks with high prevalence of CTS 961 workers. tools, and two measurement both hands: (27%) than in ones with low techniques used: (1) Flexion- 20% prevalence of CTS (13%). extension measurements: Subjects recorded at several 92% of population were female. points during the day for 15 min. An angle meter used to Non-standard data analysis measure flexion-extension approaches, no statistical testing. angles of the wrist: Rated high flexion, low flexion, low Examiners not blinded. extension, and high extension using fuzzy cutting functions. Authors believe higher rate of CTS Each modality characterized by in both hands (20%) vs. dominant its arithmetic mean and its hand (100%) argue for non- relative duration. (2) Force: occupational factors being more Electromyography used; values important. under 2 daN considered as low forces. Calculated time spent over 2 daN, maximal force, number of peak exertions, and the arithmetic mean of the n values during a period.

(Continued)

5a-53

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Marras and Cross- 40 volunteers at a highly Outcome: CTS was High-risk Low-risk Model for Participation rate: Not reported. Shoenmarklin sectional repetitive, hand- determined from evaluation of job: 8 job: 0 predicting high 1993 intensive industrial jobs OSHA illness and injury logs incidents/ incidents vs. low job risk Examiners blinded: not stated. based upon in 8 different plants. Half and medical records. The 200,000 hr motion Confounders controlled for: Age, the workers were independent variable was exposure component: gender, handedness, job employed in jobs that had exposure to jobs in which CTS satisfaction. OSHA recordable had occurred previously. A Position repetitive trauma low-risk job was defined as Radial/ulnar All the jobs required gloves except incidents, half the having a zero incidence rate; a ROM: OR=1.52 1.1-2.1 two-one “low-risk” and one “high- workers were in jobs high-risk job was defined as Flexion/exten- risk.” sion ROM: with no history of having an incidence rate of OR=1.3 1.0-1.7 Significant difference between recordable repetitive eight or more recordable Pronation/ groups with regards to age, years trauma incidents. Two repetitive trauma. supination with the company, and trunk depth. subjects from 10 ROM: repetitive, hand- Exposure: Included number of OR=1.2 0.9-1.6 No significant difference in job intensive jobs were wrist motions/8-hr shift, weight satisfaction, number of wrist Velocity movements, age, weight, stature, randomly chosen to of loads, handgrip types and Radialulnar vel: hand dimensions. participate. forces, work heights, and OR= 2.4 1.3-4.3 motion descriptions. Wrist Flexion/ Turnover rate: High-risk jobs: 33%; motion monitors measured in extension vel: low-risk jobs: 0.5%. the radial/ulnar, OR=3.8 1.5-9.6 flexion/extension, and Pronation/ Grip forces were three times as pronation/supination planes: supination vel: great in the high-risk jobs than in OR=1.9 1.2-3.2 the low-risk jobs. wrist angles, angular velocity, angular acceleration. Acceleration Variance between subjects within Radial/ulnar jobs accounted for a substantial accel: percentage of total variance in OR=2.7 1.5-4.9 wrist motion. Flexion/ extension accel: OR=6.1 1.7-22 Pronation/ supination accel: OR=2.96 1.4-6.4

(Continued)

5a-54

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

McCormack Cross- Textile workers: Outcome: Assessed by Prevalences 1.3% Participation rate: 91%. et al. 1990 sectional 4 broad job categories questionnaire and screening of CTS (non- involving intensive physical examination initially by office) Physician or nurse examiners not upper extremity use. nurse. CTS diagnosed on Boarding: Boarding vs. blinded to case or exposure status Workers randomly clinical grounds of symptoms 0.7% non-office (personal communication). chosen: Sewing and positive Tinel's sign and OR=0.5 0.05-2.9 workers (n=562); Phalen's test. Physician Prevalence higher in workers with boarding workers reassessed physical findings Sewing: Sewing vs. <3 years of employment. Race and (n=296); packaging by “standardized methods.” 1.2% non-office age not related to outcome. workers (n=369); and OR=0.9 0.3-2.9 Females found to have significantly knitting workers Exposure: Assessment by more CTS than males. (n=352) compared to observation of jobs. Exposure Packaging: Packaging vs. other non-office to repetitive finger, wrist and 0.5% non-office Job category not found to be workers (n=468). elbow motions assumed from OR=0.4 0.04-2.4 significant, however no job title; no objective measurement of force, repetition, measurements performed. Knitting: Knitting vs. posture analysis, etc. 0.9% non-office OR=0.6 0.1-3.1 Questionnaire asked types of jobs, length of time on job, production rate, nature and type of upper extremity complaint, and general health history.

11 physician examiners; interexaminer reliability potential problem acknowledged.

(Continued)

5a-55

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Moore and Cross- 32 jobs in which 230 Outcome: CTS identified from 13.7% 4.9% 2.8 0.2-36.7 Participation rate: Study based on Garg 1994 sectional workers were employed. OSHA logs and medical records. This study was more an records. A case required Investigators blinded to exposure, evaluation of jobs than of electrophysiologic testing, case outcome status, and personal individuals. confirmed as abnormal by identifiers on medical records. electromyographer and Repetitiveness, “type of grasp” presence of suggestive were not significant factors symptoms. between hazardous and safe job categories. Exposure: Observation and No pattern of morbidity according to videotape analysis of jobs. date of clinic visits. Force, wrist posture, grasp Strength demands significantly type, high-speed work, increased for hazardous job localized mechanical stress, categories compared to safe job vibration, cold, and work time categories. assessed via observation of IR based on full-time equivalents videotape. Jobs classified as and not individual workers, may hazardous or safe based on have influenced overall results. data and judgement. Workers had a maximum of 32- months of exposure at plant–so duration of employment analysis limited. Average maximal strength derived from population-based data stratified for age, gender, and hand dominance. Using estimates of Silverstein’s classification, association between forcefulness and overall observed morbidity was statistically significant; repetition was not. No control for confounders. No information on work history, number of unaffected workers, or exposure duration.

(Continued)

5a-56

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Morgenstern Cross- 1,058 female grocery Outcome: Defined CTS as self- For a Participation rate: 82%. et al. 1991 sectional cashiers from a single reported hand/wrist pain, difference of union. nocturnal pain, tingling in the 25 hr/wk: 1.88 0.9-3.8 Controlled for age. hands or fingers, and 12% 5.4% Comparison group was numbness. Information collected on age, sex, those who reported no pregnancy status, work history as symptoms. Exposure: Duration, use of a checker, specific job-related laser scanner determined from tasks, use of selected drugs, Cashiers were also survey (no measurements). history of wrist injury. compared to results from a general population In logistic regression, “Use of study from Rochester, diuretics” significantly associated Minnesota (Stevens et al. with CTS, OR=2.66 (1.00-7.04); 1988). thought to be related to fluid retention by authors.

Laser scanning found not to be significant factor.

(Continued)

5a-57

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Nathan et al. Cross- 471 industrial workers Outcome: Case defined as Prevalence Prevalence Participation rate: Not reported. 1988 sectional from 27 occupations in NCS-determined impaired of abnormal of 4 industries. Jobs sensory conduction (sensory nerve abnormal Analysis controlled for age and grouped into 5 classes latency). Sensory latencies conduction nerve gender. assessed antidromically for sensory conduction based on resistance and eight consecutive 1-cm latency: sensory No description of symptom status repetition rate. segments of the nerve. A latency: for defining CTS. maximum latency difference of 0.4 ms or greater used to Group II: Group I: Group II vs. I: Method of categorization of jobs define impaired sensory 27% 28% PR=1.0 0.5-2.0 and occupations not described. conduction. Case definition did not deal with symptoms. Group III: Group I vs. III: Classification system is based on 47% PR=1.9 1.3-2.7 only repetition and not resistance Exposure: Jobs grouped into as listed. 27 occupations with similarities Group IV: Group I vs. IV: of characteristics as to type of 38% PR=1.7 1.3-2.7 Initially excluded cases of CTS in grip, wrist position, study population, yet was handedness pattern, Group V: Group I vs. V: supposedly identifying prevalences resistance, frequency, and 61% PR=2.0 1.1-3.4 of CTS in exposure groups. duration of grasp and presence of vibratory and For nerve conduction analysis, ballistic components. The wrongly assumed that each hand’s 27 occupations then grouped nerve conduction study results in into 5 classes. Resistance an individual were independent. (Res.) rated from very light to The 2 hands in a single individual very heavy; repetition rate are not independent of each other. rated from low to high. Group I: very light resistance and low repetition Group II: light resistance and very high repetition Group III: moderate resistance and moderately high repetition Group IV: heavy resistance and moderate repetition Group V: very heavy resistance and high repetition.

(Continued)

5a-58

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Nathan 1992a Long- 315 workers using both Outcome: Case defined as Group II: Group 1: Groups II vs. Participation rate: Overall: 67%; itudinal hands (each hand NCS-determined impaired 19% 18% Group I: Group 3 participation rate was analyzed separately) sensory conduction (sensory PR=1.1 0.6-1.9 59%. from four industries. latency). Sensory latencies Examiners blinded: Not reported. These represented 67% assessed antidromically for Group III: Group III vs. of original group of eight consecutive 1-cm 26% Group I: Analyzed using gender, hand workers from 1988 segments of the nerve. A PR=1.5 1.0-2.2 dominance, occupational hand use, published study maximum latency difference of duration of employment, and randomly selected from 0.4 ms or greater used to Group IV: Group IV vs. industry. four industries (67% of define impaired sensory 24% Group I: 76% of participants employed in original subjects) conduction. PR=1.4 0.9-2.1 same occupational hand-use class as in 1988. A lower percentage of Group I: Very light Probable CTS: Presence of Group V: Group V vs novice workers returned (56%) resistance and low any two primary symptoms 18% Group I: than non-novice workers (69%) for repetition (numbness, tingling, nocturnal PR=1.0 0.5-2.2 follow-up study. awakening) or one primary Analysis of “hands” instead of Group II: Light symptom and 2 secondary individual would cancel contribution resistance and very high symptoms (pain, tightness, of exposure effect if there was repetition clumsiness). unilateral slowing. Group III: Moderate Exposure: For this article, Data in table two for 1984 subjects resistance and previous exposure is not the same data as presented moderately high classification was used from in previous article; numbers have repetition 1988 Nathan article. Jobs had shifted to other groups. The been grouped into 27 significant difference seen Group IV: Heavy occupations with similarities of between nerve slowing between resistance and moderate characteristics as to type of Class 1 and Class 5 in 1988 paper repetition grip, wrist position, is no longer significantly different. handedness pattern, Authors note that “130 hands Group V: Very heavy resistance, frequency, and experienced a decrease in resistance. duration of grasp and occupational use.” No parameters presence of vibratory and given for decrease and assumption ballistic components. The 27 is made that both hands in an occupations then grouped into individual had similar decrease in 5 classes. Resistance rated use. from very light to very heavy; With one-third of cohort missing repetition rate rated from low to from 1984 study, there is no way to high. determine if homogeneity in symptoms prevalence in 1984 and 1989 reflects absence of progression or drop-out.

(Continued)

5a-59

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Nathan 1994b Long- 101 Japanese furniture Outcome: Case defined as 8 cm. 8 cm. Participation rate: For Japanese itudinal factory workers. There NCS-determined impaired Sensory Sensory Workers: 100% were 27 managers, 35 sensory conduction (sensory latency: latency: Americans: Overall: 67%; Group 3 clerical workers, 21 assembly-line or food latency). Sensory latencies 0.30 0.31 participation rate was 59%. service workers and 18 assessed antidromically for machine operators. Their eight consecutive 1 cm. 14 cm. 14 cm. Examiners blinded: Not reported. NCS results were segments of the nerve. A Sensory Sensory compared to 315 maximum latency difference of latency: latency: Analyzed using gender, hand workers using both 0.4 ms or greater used to 0.36 0.45 dominance, occupational hand use, hands (each hand define impaired sensory duration of employment, and analyzed separately) from four industries. conduction. industry. (These represented 67% Probable CTS: Presence of Probable Probable of original group of any two primary symptoms CTS: 2.5% CTS: Analysis of “hands” instead of workers from 1988 (numbness, tingling, nocturnal 2.0% individual would cancel contribution published study awakening or one primary of exposure effect if there was randomly selected from Definite Definite unilateral slowing. four industries (67% of symptom and 2 secondary original subjects) and are symptoms (pain, tightness, CTS: CTS: the subject of a separate clumsiness). 2.0 8.3 Conducted step-wise regression table entry in this analysis for Probable CTS and document. Exposure: Exposure was not reported that repetitions and addressed except is assumed duration of employment were Group I: Very light to be self-reported by protective. Cigarettes and Age resistance and low were also retained in the model. repetition. questionnaire for the Japanese workers. The jobs were Group II: Light grouped into 5 classes. resistance and very high Resistance rated from very repetition. light to very heavy; repetition Group III: Moderate rate rated from low to high resistance and repetition. moderately high repetition. Group IV: Heavy resistance and moderate repetition. Group V: Very heavy resistance.

(Continued)

5a-60

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Osorio et al. Cross- 56 supermarket workers. Outcome: CTS assessed via Symptoms: 0% for 8.3 2.6-26.4 Participation rate: 81%. 1994 sectional Comparison was medical history, physical exam, 63% in low- (for CTS- between high and low median nerve conduction high- exposure symptoms high Adjusted for age, gender, alcohol exposure groups. studies, and vibratory exposure; group vs. low consumption, and high-risk medical thresholds. 10% in exposure history. moderate- groups) A. CTS-like syndrome: exposure Interview and testing procedures Probable diagnosis: (1) Pain group performed by personnel blinded to tingling numbness in median case status. nerve distribution and Positive (2) symptoms last >1 wk or $ NCS: 33% 0% for 6.7 (for 0.8-52.9 Skin surface temperature not 12 times in last year, no acute in high- low- abnormal NCS, controlled. trauma or systemic disease, exposure; exposure high vs. low onset or exacerbation since 7% in group exposure Dose response for presumptive working on current job. moderate- groups) (symptoms of) exposure to exposure forceful, repetitive wrist motion: B. Median neuropathy: group CTS-prevalence 63% high Sensory median nerve exposure; 10% medium exposure; conduction velocity 44 m/sec 0% low exposure. or less. Dose response for prevalence of Exposure: Observation of jobs abnormal median nerve velocity: by ergonomist and industrial 33% high; 7% medium; 0% low. hygienist. Analysis based on categorization by job title after Linear regression showed observation. Jobs divided into significant relationship between 3 categories based on the years worked and worsening of likelihood of exposure to nerve conduction (decreased forceful and repetitive wrist nerve conduction velocity and motions (low, moderate, high), decreased nerve conduction years worked at this store, amplitude) adjusted for total years worked as checker, confounders (above), however total years using laser small sample size. scanners.

(Continued)

5a-61

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Punnett et al. Cross- 162 female garment Outcome: CTS assessed by 18% 6% 2.7 1.2-7.6 Participation rate: 97% (garment 1985 sectional workers; 85% were symptom questionnaire and workers), 40% (hospital workers). employed as sewing physical exam. Cases defined Controlled for age, hormonal status, machine operators who as the presence of persistent and native language. sewed and trimmed by pain (lasted for most days for hand. one month or more within the Pain in the wrist and hand past year); were not significantly correlated (p<0.01; Comparison: 76 of 190 associated with previous r=0.41). full- or part-time workers injury; and, began after first on day shift in a hospital employment in garment Age distribution not significantly different metabolic disease. who worked as nurses manufacturing or hospital or aids; lab technicians employment. Key questions Symptoms of CTS showed trend by or therapists, or food based on the arthritis age (p<0.01). service workers. supplement questionnaire of the National Health and Nutrition Prevalence of pain not associated Employees typing >4 Examination Survey (NHANES). with years of employment in hr/day excluded from Median nerve symptoms (pain, garment workers. comparison group. 162 numbness, or tingling) if Length of employment not predictor female garment workers present at night or early in the of risk. compared to 73 female morning or met 2 of 3 criteria: hospital workers. (1) accompanied by weakness Change in hormonal status in pinching or gripping; significantly associated with CTS (2) alleviated by absence from symptoms but negatively associated with employment in work for >1 wk; garment shop. (3) aggravated by housework or other non-occupational Logistic model found garment work tasks. and age significant for symptoms of CTS. Exposure: Observation of job Neither metabolic disease nor tasks. Information on work change in hormonal status history obtained by statistically significant risk. questionnaire. Job title used as a proxy for exposure in analyses.

(Continued)

5a-62

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Schottland Cross- Poultry workers (27 Outcome: Defined as 41% 20% 2.86 1.1-7.9 Participation rate: Not reported. et al. 1991 sectional males, 66 females) prolonged motor or sensory exceeding exceeding compared to job median latencies. No 2.2 ms for 2.2 ms for Not mentioned whether examiners applicants (44 males, symptoms or physical exam sensory median blinded to case status or exposure. 41 females). included in case definition. latency sensory value of latency Controlled for age and gender. Exposure: Based on current median value employment status at plant. No nerve on (right- Referents not excluded if prior measurements made. NCS (right- hand, employment at poultry plant; Repetitive tasks (15 to 50 hand, females, 15 referents had previous complex operations/min not females, corrected employment in poultry plant; this rare), requiring firm grip, with corrected for age) would result in poor selection of wrists in flexion or extension, for age) controls, would tend to bias results with internal deviations. towards the null. 24% 15% 1.87 0.6-9.8 Right-hand of female applicants exceeding exceeding who never worked in a poultry 2.2 ms for 2.2 ms for plant had significantly longer median median median palmar latency (MPS) on nerve nerve nerve conduction than referents sensory sensory (p<0.04). latency latency value on value on Symptoms of CTS not inquired. NCS (left- NCS (left- Right hand of male workers had hand, hand, longer MPS on nerve conduction females, females, but not significant (p<0.07). corrected corrected for age) for age) From Table 5-2 in paper it shows there is inadequate sample size for detecting differences in female’s left-hand and male’s left- and right- hand MPS. Concluded there is an elevated risk of CTS, roughly equal to risk from aging for the right hands of female workers, less risk for male both hands and female left hands.

(Continued)

5a-63

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Silverstein Cross- 652 industrial workers in Outcome: CTS determined by 1.0 0.6 Group 2 vs. Participation rate: 90% response et al. 1987 sectional 4 groups: (1) low-force, medical examination and (Group 2) Group 1: rate obtained. low-repetition interviews. OR=1.8 0.2-21 (comparison group, Controlled for age, gender, plant, n=93 males, 64 females); Symptoms of pain, numbness 2.1 Group 3 vs. years on the job. No interactions (2) high-force, low- or tingling in median nerve (Group 3) Group 1: found. repetition (n=139 males, distribution. OR=2.7 0.3-28 Jobs evaluated by investigators 56 females); (3) low- blinded to worker health status. force, high-repetition Nocturnal exacerbation; 5.6 Group 4 vs. (n=43 males, 100 symptoms >20 times or >1 wk (Group 4) Group 1: Examiner blinded to medical history females); (4) high-force, in previous year; no history of OR=15.5 1.7-142 and exposure. high-repetition (n=83 acute trauma; no history of males, 74 females). rheumatoid arthritis; onset of Random sample of 12 to 20 active symptoms since current job; In separate workers/job with 1 year’s seniority, positive modified Phalen’s test logistic models: stratified by age and gender. (45 to 60 sec) or Tinel’s sign; rule out cervical root thoracic (1) Repetitive- Interview data included prior health outlet, pronator teres ness: OR=5.5 and injuries, chronic diseases, reproductive status of females, syndrome. (p<0.05) recreational activities, prior job activities. Exposure: To (1) forceful, (2) Force: (2) repetitive, and (3) awkward OR=2.9 (non- No association found with wrist hand movements assessed by significant) posture, type of grasp, or use of EMG and video analysis of vibrating tool. jobs. Three workers in each selected job videotaped for (at Positive associated with age but least) 3 cycles. High-force job: not statistically significance. A mean adjusted force >6 kg (mean adjusted force = No differences in health history or [(variance/mean force)+ mean recreational activities. force]); low-force job: A mean No association with gender, or adjusted force <6 kg. industrial plant. High repetition = work cycles Negatively associated with years <30 sec or work cycles on the job but not statistically constituting >50% of the work associated. cycle. Repetitiveness found to be stronger risk factor than force. No association with hormonal status.

(Continued)

5a-64

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Stetson et al. Cross- Comparison of 137 Outcome: Symptoms Õ Õ Participation rate: 71% seen, 16% 1993 sectional asymptomatic industrial consistent with CTS defined as refused, others unavailable workers, 103 industrial numbness, tingling, or burning because of layoffs, transfers, or workers with hand/wrist localized to median nerve sick leave. symptoms, and 105 anatomic area, not caused by control subjects acute injury, and occurred >20 Industrial population randomly randomly selected not times in previous year. Nerve selected. exposed to highly conduction studies conducted forceful or repetitive on the dominant hand; median Controlled for age, height, skin hand exertions. sensory and motor, ulnar temperature, and dominant index sensory, distal amplitudes and finger circumference. latencies were measured. Temperature monitored. Comparing the means of the nerve conduction measures, the following Exposure: Observation and were statistically significantly worker interviews using different between: (1) the ergonomic checklist. One or asymptomatic hand group and the more workers on each job controls: median sensory amplitude were evaluated based on and distal latency, and median to repetitiveness, forcefulness, ulnar comparison measures; (2) the mechanical stress, pinch grip, symptomatic hand group and and wrist deviation, then data controls: median sensory distal extrapolated to other workers latency, and median to ulnar performing jobs. A 3-point comparison measures. ordinal scale used to estimate exposure (none, some, Median sensory amplitudes were frequent or persistent). smaller and distal latencies longer in symptomatic compared to asymptomatic hand group.

Forceful hand and upper extremity exertions were significantly different between exposed and non-exposed groups. Repetition not significantly different, but little statistical power to detect difference.

(Continued)

5a-65

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Tanaka et al. Cross- Data from the Outcome: Outcomes included Prevalence Logistic model Participation rate: 91.5%. In Press sectional Occupational Health those “Recent Workers” who of self- for medically interview Supplement of 1988 worked anytime during the past reported called CTS Multiple logistic regression used to survey National Health Interview 12 months (excluding armed CTS among among recent examine age, gender, race, Survey conducted by the forces). Self-reported carpal recent workers exposure to vibration, and National Center for tunnel syndrome= “yes” to workers: bending/twisting of the hand/wrists Health Statistics. question: During the past 1.47% Bend/twist: to odds of reporting CTS. Households are selected 12 months, have you had a OR=5.9 3.4-10.2 Interactions were checked for. by multistage probability condition affecting the wrist Prevalence sampling strategy. One and hand called carpal tunnel of medically White race: Self-reported CTS prevalence adult, 18 years or older, syndrome? Medically called called CTS OR=4.2 1.9-15.6 among recent workers higher in was randomly selected CTS = a response of “carpal among whites compared to non-whites, for interview. 44,233 tunnel syndrome” to the recent Female gender: highest in white females. interviews completed. question: “What did the medical workers: OR=2.4 1.6-3.8 person call your hand 0.53% When vibration was not in the discomfort?” Vibration: model the bend/twist OR=5.99. OR=1.85 1.2-2.8 When bend/twist is not in the Exposure: By questionnaire: model, vibration OR=3.00. Did the most recent job require BMI $25: you to bend or twist your OR=2.1 1.4-3.1 Major limitation is CTS is based on hands or wrists many times an self-reports without medical hr? Did you work with hand- Cigarette use: validation. held or hand-operated tools or OR=1.6 1-2.5 machinery. No temporal relationship could be Age $40: found between reported CTS and OR=1.3 0.2-1.9 the reported occupation/industry or exposure to bending/twisting of the Annual income hand/wrist. $$20,000: OR=1.5 1-2.4

Education >12: OR=1.2 0.8-1.8

(Continued)

5a-66

Table 5a–5 (Continued). Epidemiologic studies evaluating work-related carpal tunnel syndrome (CTS)

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Weislander Case- 34 male CTS patients, Outcome: CTS diagnosed Õ Õ Cases Participation rate: 93%. et al. 1989 control each matched to 2 other clinically by a hand surgeon, compared to all hospital referents confirmed by electro-diagnostic referents Referents matched for gender and (drawn from among studies. (hospital- and age (±3 years.), hospital referents population- for year of operation. other based): surgical cases, one Exposure: To vibrating tools, Vibrating tool Hospital referents and population referent had been repetitive wrist movements, use: OR=3.3 1.6-6.8 referents statistically different operated on for gall and loads on the wrist comparing: use of vibrating tool, bladder surgery and the assessed via telephone Use of hand- repetitive movements of wrist, other for varicose veins) interview using a standardized held vibrating workload on wrist, obesity. and 2 population questionnaire. The degree of tools 1-20 referents (from a general exposure was evaluated both years: OR=2.7 1.1-6.7 Hospital-based population may not reflect industrial workplace. population register and with regard to the total number Loads on the telephone directory) of work years and the average wrist: Interviewers not blinded to case (total comparison number of exposed hr a wk. OR=1.8 1.0-3.5 status. group=143 males). Repetitive movements classified independently by Cases Elevated OR for repetitive physician interviewer and compared to movements of the wrist only occupational hygienist. population statistically significant for the Exposure to repetitive wrist referents category ‘>20 years.’ alone: movements was considered to Vibrating tool Odds ratios (OR) for any of the exist if they agreed. use: three diseases (thyroid disease, OR=6.1 2.4-15 diabetes, rheumatoid arthritis) found to be statistically significant Repetitive among cases with CTS compared wrist to 143 referents; OR=2.8 (1.0-7.6). movement for >20 years: ORs tended to increase with OR=4.6 1.8-11.9 increasing number of risk factors present. One factor, OR=1.7 (0.6- Repetitive 4.4); two factors, OR=3.3 (1.2-9.1); wrist >two factors, OR=7.1 (2.2-22.7). movement: OR=2.7 1.3-5.4 Obesity is >10% above reference weight. Obesity: OR=3.4 1.2-9.8

5a-67 CHAPTER 5b Hand/Wrist Tendinitis

SUMMARY Eight epidemiologic studies have examined physical workplace factors and their relationship to hand/wrist tendinitis. Several studies fulfill the four epidemiologic criteria that were used in this review, and appropriately address important methodologic issues. The studies generally involved populations exposed to a combination of work factors; one study assessed single work factors such as repetitive motions of the hand. We examined each of these studies, whether the findings were positive, negative, or equivocal, to evaluate the strength of work-relatedness, using causal inference.

There is evidence of an association between any single factor (repetition, force, and posture) and hand/wrist tendinitis, based on currently available epidemiologic data. There is strong evidence that job tasks that require a combination of risk factors (e.g., highly repetitious, forceful hand/wrist exertions) increase risk for hand/wrist tendinitis.

INTRODUCTION example, one study used very strict criteria Since the hand/wrist area may be affected by [Byström et al. 1995]. The case definition more than one musculoskeletal disorder, only required observation of swelling along the those studies that specifically address tendon at the time of the physical examination. hand/wrist tendinitis are considered here. The only cases of tendinitis diagnosed were Studies with outcomes described as hand/wrist deQuervain’s disease; no other cases of disorders or symptoms in general, or those in tenosynovitis or peritendinitis were diagnosed which hand/wrist tendinitis was combined with among 199 automobile assembly line workers. epicondylitis, e.g., were excluded from this In contrast, the studies with the highest section because it was not possible to evaluate prevalence rates either did not clearly state evidence for work-related hand/wrist tendinitis what diagnostic criteria were used to determine from the data. The seven studies referenced in the case definition, or the case definition Table 5b-1 provided data specifically considered recurrences of tendinitis new cases. addressing hand/wrist tendinitis. In each of Whether case definitions were inclusive or these studies the outcome was determined using exclusive would not affect the relative risk (RR) physical examination criteria, although the case as long as they were applied non-differentially definitions varied among studies. Prevalence or between groups designated as exposed or incidence rates of hand/wrist tendinitis reported unexposed. in these exposed groups ranged from 4% to 56%, and in unexposed groups from 0% to Although several studies reported odds ratios, 14%. Such wide ranges of prevalence rates published data were reanalyzed and the results probably reflect the variability in diagnostic presented here and in criteria as much as they do the range of Tables 5b1-3 as prevalence ratios (PRs). This workplace exposures in these studies. For was done because odds ratios may overestimate RR when prevalence rates are

5b-1 high, and to make estimates of RR comparable though the risk estimates mostly refer to across studies. In studies that presented odds combined exposures. ratios in the original articles, the recalculation of data as PRs resulted in lower estimates of the REPETITION RR. In the one prospective cohort study [Kurppa et al. 1991] incidence rates and risk Definition of Repetition for ratios are presented. Hand/Wrist Tendinitis Armstrong et al. [1987a] analyzed videotaped Except for a study reported by Armstrong et al. job tasks of a sample of workers, then divided [1987a], risk estimates were not reported job tasks according to level of repetitiveness: separately for single risk factors. Only the high repetition (cycle time <30 sec, or $50% of Armstrong et al. study used a formal the cycle spent performing the same quantitative exposure assessment as the basis fundamental motions) or low repetition. for determining exposure groups. Other studies Kuorinka and Koskinen [1979] created a grouped jobs with similar risk factors together “workload index” based on the number of and compared them to jobs without those risk pieces handled per hour multiplied by the factors. Typically, the selection of jobs for the number of hours worked, for a dose-response exposed and unexposed groups was based on analysis within the exposed group. Comparison general knowledge of the jobs, previously groups in the other studies were job categories; published literature, or questionnaire data. selection of the groups to be compared was Repetition, force, and extreme postures were based on observations, questionnaire data, or considered in combination to determine which surveillance data. workers were exposed or unexposed. Formal exposure assessment (such as videotape Studies Reporting on the Association analysis for cycle time, repetition, extreme of Repetition and Hand/Wrist postures, and estimates of force), was usually Tendinitis conducted on a sample of jobs and used as Seven studies addressed repetition: Amano et rationale in the grouping of jobs into exposed al. [1988]; Armstrong et al. [1987a]; Byström and unexposed categories, rather than to create et al. [1995]; Luopajärvi et al. [1979]; Roto quantitative measures of risk factors. In some and Kivi [1984]; Kuorinka and Koskinen cases (e.g., Luopajärvi et al. [1979]), [1979]; and McCormack et al. [1990]. investigators noted the difficulty in examining risk factors separately because of job rotation. Studies Meeting the Four Evaluation For the purpose of this review, we have Criteria grouped study findings according to the risk Two of the seven studies that addressed factors present in the exposed job categories, repetition met all four of the evaluation criteria: based on the information in published articles. Armstrong et al. [1987a], and Luopajärvi et al. In Tables 5b1–3, studies are listed under single [1979]. Armstrong et al. studied 652 industrial risk factors if there was evidence that the workers at seven manufacturing plants exposed and unexposed groups differed in that (electronics, sewing, appliance, bearing risk factor, fabrication, bearing assembly, and investment

5b-2 casting). Exposure assessment of jobs included interviews and physical examinations conducted videotape analysis and electromyography by a physiotherapist (active and passive (EMG) of a sample of workers. Data from this motions, grip-strength testing, observation, and assessment were then used to categorize jobs palpation). Diagnoses of tenosynovitis and according to level of repetitiveness and force. peritendinitis were later determined by medical Health assessment of workers focused on specialists using these findings and deQuervain’s disease, trigger finger, tendinitis, predetermined criteria. The PR for tendinitis and tenosynovitis. The hand/wrist tendinitis among the assembly line packers compared to case definition required abnormal physical the shop assistants was 4.13 (95% CI examination findings (increased pain with 2.63–6.49). Age, hobbies and housework resisted but not passive motion or tendon were addressed and no associations with locking with a palpable nodule, or a positive musculoskeletal disorders were identified. Finkelstein’s test) in addition to meeting symptom criteria on standardized interviews. Studies Meeting at Least One Criteria The PR for the high repetition/low force group Amano et al. [1988] reported the prevalence of (n=143) compared to the low repetition/low cervicobrachial disorders, including force group (n=157) was 5.5 (95% confidence tenosynovitis, among 102 assembly line interval [CI] 0.7–46.3). The PR for the high workers in an athletic shoe factory and 102 repetition/high force group (n=157) compared age- and gender-matched non-assembly line to the low repetition/low force group (n=157) workers (clerks, nurses, telephone operators, was 17.0 (95% CI 2.3–126.2). The effect of cooks, and key punchers). Exposure age, gender, years on the job, and plant were assessment was based on videotape analysis of analyzed. A higher prevalence of tendinitis was the tasks of 29 workers on one assembly line. noted among women but was not significantly Assembly line workers produced about 3,400 associated with personal factors, whereas shoes a day. All but one task had cycle times significant differences in posture were observed less than 30 seconds. No formal exposure between males and females. assessment of the comparison group was reported. Diagnoses were determined by Luopajärvi et al. [1979] compared the physical examination, including palpation for prevalence of hand/wrist tendinitis among 152 tenderness. The PRs for tenosynovitis of the female assembly line packers in a food right and left index finger flexors among the production factory to 133 female shop shoe factory workers were 3.67 (95% CI assistants in a department store. Exposure to 1.85–7.27) and 6.17 (95% CI 2.72–13.97) repetitive work, awkward hand/arm postures, respectively, compared to the non-factory and static work was assessed by observation workers. Tenosynovitis of the other digits was and videotape analysis of factory workers. No not diagnosed in the comparison group. Shoe formal exposure assessment was conducted on assembly workers held shoe lasts longer in the the department store workers; their job tasks left hand and had greater frequency of were described as variable. Cashiers were symptoms in the left hand. Comparison subjects excluded, presumably because their work was were matched to shoe factory workers on repetitive. The health assessment consisted of gender and age (within five years).

5b-3 cycle was multiplied by the number of pieces Byström et al. [1995] studied forearm and hand handled per hour and the number of hours disorders among 199 automobile assembly line worked to create a workload index. Cycle workers and compared them to 186 randomly times ranged from 2 to 26 seconds; the number selected subjects from the general Swedish of pieces handled per hour ranged from 150 to population. For both groups, exposure was 605. No formal exposure assessment was assessed using rating scales on nurse- conducted on the shop assistants. Health administered questionnaires that addressed assessment involved interview and physical daily duration of hand and finger movements, examination by a physiotherapist following a wrist position, grip, and hand tool use standard protocol. Diagnoses of tenosynovitis [Fransson-Hall et al. 1995]. Videotape analysis and peritendinitis were later determined from and electromyograms were conducted on a these findings using predetermined criteria subgroup [Hägg et al. 1996]. A diagnosis of (localized tenderness and pain during tenosynovitis or peritendinitis required the movement, low-grip force, swelling of wrist observation of swelling and pain during active tendons [Waris et al. 1979]). In equivocal movement on physical examination. A diagnosis cases, orthopedic and physiatric teams of deQuervain’s disease required a positive determined case status. The PR for muscle- Finkelstein’s test. No cases of tenosynovitis or tendon syndrome among the scissor makers peritendinitis, other than deQuervain’s disease, was 1.38 (95% CI 0.76–2.51) compared to were found in this study, probably because of the department store assistants. Whether or not strict clinical criteria used for the case definition. cashiers were excluded from the comparison The PR for deQuervain’s disease among the group in this study, as they were in the automobile assembly line workers was 2.49 Luopajärvi et al. [1979] study is unclear. The (95% CI 1.00–6.23) compared to the general study group was 99% female. No relationship population group. Psychosocial variables and was found between age- or body-mass index other potential confounders or effect modifiers and muscle-tendon syndrome. The number of were addressed by Fransson-Hall et al. [1995]. symptoms increased with the number of parts A higher prevalence of deQuervain’s disease handled per year. Analyses of subgroups of was noted among men than women. scissor makers showed non-significant increased prevalence of muscle-tendon Kuorinka and Koskinen [1979] studied syndrome in short versus long cycle tasks and occupational rheumatic diseases and upper limb in manipulation versus inspection tasks. The strain among 93 scissor makers and compared authors noted a lack of contrast in exposures them to the same group of department store between the subgroups. A non-significant trend assistants (n=143) that Luopajärvi used as a of increasing prevalence of diagnosed muscle- comparison group. Temporary workers and tendon syndrome with increasing number of those with recent trauma were excluded from pieces handled per year was noted in a nested the scissor case-control analysis (n=36). makers group. Exposure assessment included videotape analysis of scissor maker tasks. The McCormack et al. [1990] studied tendinitis and time spent in deviated wrist postures per work related disorders of the upper extremity among

5b-4 1,579 textile production workers compared to tendinitis (p=0.01), but job category was a 468 non-production textile workers, a stronger predictor (p=0.001). reference group that included machine maintenance workers, transportation workers, Roto and Kivi [1984] studied the prevalence of cleaners, and sweepers. The textile production tenosynovitis among 92 male meatcutters workers were reported as being exposed to compared to 72 male construction foremen. No repetitive finger, wrist and elbow motions based formal exposure assessment was conducted. on knowledge of jobs; no formal exposure Meatcutters’ work entailed repetitive physical assessment was conducted. Health assessment exertion of upper extremities and shoulders. included a questionnaire and screening physical Construction foremen’s work did not involve examination followed by a diagnostic physical repetitive movements of the upper extremities. examination. The diagnosis of tendinitis Health assessment was by questionnaire and required positive physical findings suggestive of physical examination. Tenosynovitis was inflammation. The textile production workers defined as swelling, local pain, and finger were divided into four broad job categories: weakness during movement. The prevalence of boarding (n=296), which was noted to require tenosynovitis among the meatcutters was 4.5%. forceful work as well as the repetitive hand- The PR for tenosynovitis as defined by physical intensive work of the other categories; sewing examination could not be calculated because (n=562); packaging (n=369); and knitting there were no cases among the comparison (n=352). The PR for tendinitis among all textile group. The PR of tendinitis-like symptoms production workers was 1.75 (95% CI reported on the questionnaire among the 0.9–3.39), compared to the reference group meatcutters was 3.09 (1.43, 6.67) compared non-production textile workers. The PRs and to the construction foremen. Serologic testing 95% CIs comparing tendinitis among each for rheumatoid arthritis was done to control for broad category of textile production workers to potential confounding, none was detected. the reference group are as follows: Authors noted that tenosynovitis occurred in boarding—3.0 (1.4, 6.4); sewing—2.1 (1.0, younger age groups. 4.3); packaging—1.5 (0.7, 3.5); and knitting—0.4 (0.1, 1.4). The authors noted that Strength of Association—Repetition the knitting work was more automated than the and Hand/Wrist Tendinitis other textile production job categories. Race The PRs for repetitive work and hand/wrist and age were not related to outcome, but the tendinitis in the studies reviewed above ranged prevalence of tendinitis was higher in workers from 1.4 to 6.2: with less than three years of employment. Female gender was a significant predictor of

5b-5 Repetition and Hand/Wrist Tendinitis

PR and 95% CI Authors Exposed/Unexposed Groups 5.5 (0.7–46.3) Armstrong et al. [1987a]* HI REP & LO FORCE/LO REP & LO FORCE 17.0 (2.3–126.2) HI REP & HI FORCE/LO REP & LO FORCE 3.7 (1.9–7.3) to Amano et al. [1988] Shoe assemblers/clerks, nurses, 6.2 (2.7–14.0) operators, cooks, keypunchers 2.5 (1.0–6.23) Byström et al. [1995] Auto assemblers/general population 1.4 (0.8–2.5) Kuorinka and Koskinen [1979] Scissor makers/department store assistants 1.8 (0.9–3.4) McCormack et al. [1990] Textile production/ maintenance workers, etc. 3.1 (1.4–6.7) Roto and Kivi [1984] Meatcutters/construction foremen 4.1 (2.6–6.5) Luopajärvi et al. [1979]* Food packers/department store assistants excluding cashiers *Study met all four criteria. cleaners and sweepers, whose exposure to In evaluating these RR estimates, study repetition was not measured. If there were limitations should be considered in addition to some exposure to repetitive work in the statistical significance. Statistical significance comparison group, then this would tend to addresses the likelihood that the results are not decrease the RR for hand/wrist tendinitis among due to chance alone, whereas study limitations the textile workers. Another concern with this can bias the RR estimates in either direction. All study is the possibility that the knitting workers of the PRs were greater than one, and four of may not have been exposed to very repetitive the seven were statistically significant. The work due to greater automation in the knitting range (1.4–6.2) might reflect the level of process. The effect of this potential contrast in repetitiveness between the exposed misclassification of exposure would also be to and comparison groups. For example, in decrease the RR. McCormack et al. [1990], the comparison group consisted of machine maintenance Note that Kuorinka and Koskinen and workers, transportation workers, and Luopajärvi et al. both used the same

comparison group, but the number of subjects

5b-6 in the department store assistant group was 143 Temporal Relationship—Repetition for Kuorinka and Koskinen, and 133 for and Hand/Wrist Tendinitis Luopajärvi (who excluded cashiers from the All of the studies reviewed for this section were comparison group). If Kuorinka and Koskinen cross-sectional, so proving that exposure to did not exclude cashiers, this might tend to repetitive work occurred before hand/wrist decrease the RR. tendinitis is not possible. However, information in several of the studies suggests the likelihood The highest RR (6.2) reported for repetitive that exposure to repetitive work occurred work was by Amano et al. [1988]. In this study before the diagnosis of tendinitis. For example, it is unclear whether the examiner was blinded recently employed workers were excluded by to whether the subjects were shoe assemblers Kuorinka and Koskinen [1979]. In Luopajärvi or in the comparison group of non-assembly et al.’s [1979] study group, the minimum length line workers that included clerks, nurses, of employment was telephone operators, cooks, and key punchers. 3 years. In the McCormack et al. [1990] study, Because the occupational groups were the minimum average length of employment in examined on separate dates blinding seems the job categories was more than 7 years. unlikely. The lack of a clear case definition Byström et al. [1995] noted that subjects were leaves open the possibility of examiner bias, selected for clinical examination 5 months after which might lead to an increased RR. completion of questionnaires on exposure. Roto Alternatively, if there were a significant number and Kivi’s [1984] subjects had all worked in of key punchers in the comparison group, who the food industry for more than one year. may have been exposed to repetitive work, that Armstrong and Chaffin [1979] required a would tend to decrease the contrast in minimum length of employment of one year. exposure and might lead to a decrease in the Case definitions generally required that RR. symptoms began after starting the current job or employment at the plant. This also suggests In summary, the potential for underestimation of that exposure occurred before disease. the RR has been noted in studies where the RR is at the low end of the range, and the potential Consistency in Association for for overestimation of the RR has been noted at Repetition and Hand/Wrist Tendinitis the high end of the range. Considering these All of the studies reviewed showed positive RR concerns and statistical significance, the RR for estimates for hand/wrist tendinitis among hand/wrist tendinitis attributable to occupational groups exposed to repetitive repetitiveness is probably more likely to be in work, ranging from 1.4 to 6.2. Four of the the middle range of the estimates, based on the seven studies resulted in statistically significant studies reviewed. The statistically significant PRs. Considering only statistically significant estimates of RR in this middle group range from estimates from studies not noted to have serious 2.5 to 4.1. limitations (which might bias the RR), the range narrows to 2.5–4.1.

5b-7 Coherence of Evidence for Repetition Schoenmarklin [1991] reported that velocity and Hand/Wrist Tendinitis and acceleration significantly predicted upper DeQuervain’s disease and other extremity musculoskeletal disorders (including tenosynovitis of the hand, wrist, and forearm tendinitis) among industrial workers performing have been associated for decades with hand-intensive job tasks. repetitive and forceful hand activities as one of the possible causal factors [Amadio 1995]. Dose-Response Relationship For DeQuervain’s disease is the entrapment of the Repetition and Hand/Wrist Tendinitis tendons of the extensor pollicis brevis and Kuorinka and Koskinen [1979] reported that abductor pollicis longus. Other similar within the group of scissor makers, increased conditions are trigger thumb and triggering of prevalence of muscle-tendon syndrome the middle and ring fingers, characterized by occurred in short versus long cycle tasks and in pain with motion of the affected tendon. manipulation versus inspection tasks. These Despite the fact that the tendon and its sheath increases were not statistically significant. The may be swollen and tender, the histopathology authors noted a lack of contrast in exposures shows peritendinous fibrosis without between the subgroups. A non-significant trend inflammation, and fibrocartilaginous metaplasia of increasing prevalence of diagnosed muscle- of the tendon sheath tissue. The role of tendon syndrome with increasing number of inflammation early in the process is not clear pieces handled per year was also noted in a [Hart et al. 1995]. As in carpal tunnel nested case-control analysis (n=36) in the same syndrome or epicondylitis, acute classical study. inflammation does not seem a critical pathophysiological component of the clinical The Armstrong et al. [1987a] data resulted in a condition, at least once it becomes chronic. PR of 17.0 (2.3, 126.2) for jobs that were Despite the observations that too much forceful highly repetitious and required highly forceful and repetitive activity contributes to carpal exertions. This suggests a synergistic effect tunnel syndrome and epicondylitis, the response when both risk factors are present because the of the tendons and the muscles to repetitive estimate is greater than the sum of the RR activity is likely that of a U-shaped curve. Too estimate for force or repetition alone. little and too much activity may be harmful, but intermediate levels of activity are probably Conclusions on Repetition and beneficial. The studies of tendon and muscle Hand/Wrist Tendinitis physiology suggest that a certain amount of There is strong evidence for a positive activity maintains the normal state of these association between highly repetitive work, in tissues and leads to adaptive changes. These combination with other job risk factors, and tissues have the ability to repair significant hand/wrist tendinitis based on currently amounts of damage from some overuse; the available epidemiologic data. All seven of the poorly understood issue is when overuse studies reviewed reported positive RR exceeds the ability of the tissue to repair the damage or triggers a more harmful type of estimates. Four of these estimates were statistically significant. Potential confounders damage [Hart et al. 1995]. Marras and

5b-8 (factors associated with both exposure and according to level of repetitiveness and force. outcome that may distort interpretation of Health assessment of workers focused on findings) considered in the studies of hand/wrist deQuervain’s disease, trigger finger, tendinitis, tendinitis included gender, age, other medical and tenosynovitis. The hand/wrist tendinitis conditions, and outside activities. There is no case definition required abnormal physical evidence that the associations reported here examination findings (increased pain with between repetitive work and hand/wrist resisted but not passive motion or tendon tendinitis are distorted by gender, age, or other locking with a palpable nodule, or a positive factors. Finkelstein’s test) in addition to meeting symptom criteria on standardized interviews. FORCE The PR for the high force/low repetition group (n=195) compared to the low force/low Definition of Force for Hand/Wrist repetition group (n=157) was 4.8 (95% CI Tendinitis 0.6–39.7). The PR for the high repetition/high Armstrong et al. [1987a] based high and low force group (n=157) compared to the low force categories on electromyographs of repetition/low force group (n=157) was 17.0 forearm flexor muscles of representative (95% CI 2.3–126.2). The effect of age, workers. Comparison groups in the other gender, years on the job and plant were studies were job categories; selection of the analyzed. A higher prevalence of tendinitis was groups to be compared was based on noted among women, but was not significantly observations, questionnaire data, or associated with personal factors, whereas surveillance data. significant differences in posture were observed between males and females. Studies Reporting on the Association of Force and Hand/Wrist Tendinitis Studies Meeting at Least One Criteria Five studies addressed force: Armstrong et al. Byström et al. [1995] studied forearm and hand [1987a]; Byström et al. [1995]; Kurppa et al. disorders among 199 automobile assembly line [1991]; McCormack et al. [1990]; and Roto workers and compared them to 186 randomly and Kivi [1984]. selected subjects from the general Swedish population. For both groups, exposure was Studies Meeting the Four Criteria assessed using rating scales on nurse- One of the studies that addressed force met all administered questionnaires that addressed four of the evaluation criteria: Armstrong et al. daily duration of hand and finger movements, [1987a]. Armstrong et al. studied 652 industrial wrist position, grip, and hand-tool use workers at seven manufacturing plants [Fransson-Hall et al. 1995]. Videotape analysis (electronics, sewing, appliance, bearing and electromyograms were conducted on a fabrication, bearing assembly, and investment subgroup [Hägg et al. 1996]. A diagnosis of molding). Exposure assessment of jobs tenosynovitis or peritendinitis required the included videotape analysis and EMG of a observation of swelling and pain during active sample of workers. Data from this assessment movement on physical examination. A diagnosis were then used to categorize jobs of deQuervain’s disease required a positive

5b-9 Finkelstein’s test. No cases of tenosynovitis or among the meatcutters (100% males) peritendinitis, other than deQuervain’s disease, compared to the male comparison group was were found in this study, probably because of 14.0 (5.7, 34.4); the RR for tendinitis among strict clinical criteria used for the case definition. the meatpackers (79% female) compared to The PR for deQuervain’s disease among the the female comparison group was 38.5 (11.7, automobile assembly line workers was 2.49 56.1); and the risk ratio for tendinitis among the (95% CI 1.00–6.23) compared to the general sausage makers (86% female) was 25.6 (19.2, population group. Psychosocial variables and 77.5). A limitation of the study is the fact that other potential confounders or effect modifiers the subjects were not actively evaluated for were addressed by Fransson-Hall et al. [1995]. musculoskeletal disorders. Investigators relied A higher prevalence of deQuervain’s disease on workers to seek medical care. This could was noted among men than women. result in a difference in case ascertainment between the exposed and unexposed groups Kurppa et al. [1991] conducted a prospective because workers in non-strenuous jobs may cohort study of tenosynovitis or peritendinitis not have sought medical care for (and epicondylitis) in a meat processing factory musculoskeletal disorders since they might still in Finland. Three hundred seventy-seven be able to perform their jobs, whereas workers meatcutters, meatpackers, and sausage makers with MSDs in strenuous jobs might not be able were compared to 338 office workers, to perform their jobs, and would be more likely maintenance workers, and supervisors. to seek medical care. If subjects sought medical Exposure assessment was based on previously care, investigators were very likely to capture published literature and knowledge of jobs at the information, even if medical care was the plant. Job categories were selected based provided outside the plant, plant nurses on whether or not strenuous manual work was received and reimbursed the bills, and recorded required. The cohort was followed for 31 the diagnosis and sick leave. However, when months. Health assessment consisted of diagnoses were made by physicians outside the physical examinations by plant physicians who plant, diagnostic criteria were unknown; this were on-site daily, using predetermined criteria occurred in 25% of the cases. Exposed and for diagnosing tenosynovitis or peritendinitis comparison groups were similar in age and (swelling or crepitation and tenderness to gender mix, although gender varied with job. palpation along the tendon and pain at the tendon sheath, in the peritendinous area, or at McCormack et al. [1990] studied tendinitis and the muscle-tendon junction during active related disorders of the upper extremity among movement) and deQuervain’s disease (positive 1,579 textile production workers compared to Finkelstein’s test). Incidence density rates (if a 468 referents that included machine recurrence of tendinitis occurred after 60 days, maintenance workers, transportation workers, it was considered a new case) for tendinitis cleaners, and sweepers. The textile production were compared between each of the strenuous workers job categories and either the male or female comparison group of combined non-strenuous were reported, based on knowledge of the jobs job categories (office workers, maintenance to be exposed to repetitive finger, wrist and workers and supervisors). The RR for tendinitis elbow motions; no formal exposure assessment

5b-10 was conducted. Health assessment included a Strength of Association—Force and questionnaire and screening physical Hand/Wrist Tendinitis examination followed by a diagnostic physical Estimates of RR for hand/wrist tendinitis among examination. The diagnosis of tendinitis those in jobs requiring forceful exertion range required positive physical findings suggestive of from 2.5 to 38.5: inflammation. The textile production workers were divided into four broad job categories. The very large risk ratios reported by Kurppa Boarding (n=296), was the only category noted et al. [1991] could be biased upward because to require forceful work. The PR for tendinitis of the difference in case ascertainment between among the boarding workers was 3.0 (95% CI the exposed and unexposed groups. 1.4–6.4), compared to the reference group. Investigators did not actively evaluate subjects Race and age were not related to outcome, but for MSDs, but relied on workers to seek the prevalence of tendinitis was higher in medical care. As the authors noted, workers in workers with less than three years of non-strenuous jobs may not have sought employment. Female gender was a significant medical care for MSDs since they might still be predictor of tendinitis (p=0.01), but job able to perform their jobs, while workers in category was a stronger predictor (p=0.001). strenuous jobs may not have been able to perform their jobs and would be more likely to Roto and Kivi [1984] studied the prevalence of seek medical care. This potential for differential tenosynovitis among 92 male meatcutters case ascertainment between the exposed and compared to 72 male construction foremen. No unexposed groups undermines the credibility of formal exposure assessment was conducted. the magnitude of the risk estimate. Meatcutters’ work entailed repetitive physical exertion of upper extremities and shoulders. Statistically significant estimates of RR for Construction foremen’s work did not involve hand/wrist tendinitis among workers who repetitive movements of the upper extremities. perform strenuous tasks from the remaining Health assessment was by questionnaire and studies range from 2.5 to 3.1. physical examination. Tenosynovitis was defined as swelling, local pain, and finger weakness during movement. The prevalence of tenosynovitis among the meatcutters was 4.5%. The PR for tenosynovitis as defined by physical examination could not be calculated because there were no cases among the comparison group. The PR of tendinitis-like symptoms reported on the questionnaire among the meatcutters was 3.09 (1.43, 6.67) compared to the construction foremen. Serologic testing for rheumatoid arthritis was done to control for potential confounding, none was detected. Authors noted that tenosynovitis occurred in younger age groups.

5b-11 Force and Hand/Wrist Tendinitis

PR and 95% CI Authors Exposed/Unexposed Groups 4.8 (0.6–39.7) Armstrong et al. [1987a] HI FORCE & LO REP/LO FORCE & LO REP 17.0 (2.1–26.2) HI FORCE & HI REP/ LO FORCE & LO REP 2.5 (1.0–6.23) Byström et al. [1995] Auto assemblers/general population 14.0 (5.7–34.4) to Kurppa et al. [1991] Meat processors/office 38.5 (11.7–56.1) workers, maintenance workers, supervisors 3.0 (1.4–6.4) McCormack et al. [1990] Textile boarding workers/ maintenance workers, etc. 3.1 (1.4–6.7) Roto and Kivi [1984] Meatcutters/construction foremen * Study met all four criteria.

Temporal Relationship—Force and in the job categories studied was more than 7 Hand/Wrist Tendinitis years. Roto and Kivi’s The Kurppa et al. [1991] study determined exposure status of 83% of the cohort on [1984] subjects had all worked in the food October 2, 1982, and followed their health industry for more than one year. Armstrong et status until April 30, 1985. The remaining al. [1987a] required a minimum of 1 year of subjects entered the study when they became employment to be included in the study. permanent employees, and were also followed until April 30, 1985. Consistency of Association—Force and Hand/Wrist Tendinitis Although the remaining studies that addressed All of the studies reviewed reported positive force were cross-sectional, the following RR estimates for hand/wrist tendinitis among information increases the likelihood that occupational groups exposed to forceful exposure to forceful work occurred before the exertions, ranging from 1.8 to 38.5. Four of the occurrence of tendinitis; Byström et al. [1995] five studies reported statistically significant noted that subjects were selected for clinical findings. If only statistically significant estimates examination from studies in which limitations were not noted 5 months after completion of questionnaires on are considered, RR estimates for force and exposure. McCormack et al. [1990] reported hand/wrist tendinitis range from 2.5 to 3.1. that the minimum average length of employment

5b-12 Coherence of Evidence—Force and Tendinitis Hand/Wrist Tendinitis Kuorinka and Koskinen [1979] determined the See Repetition Section. time spent in deviated wrist postures per work cycle as part of their “workload index” that was used in a dose-response analysis Evidence of a Dose-Response Relationship—Force and Hand/Wrist within the exposed group. Comparison groups Tendinitis in the other studies were job categories; Armstrong et al. [1987a] demonstrated a dose- selection of the groups to be compared was response relationship between jobs requiring based on observations, questionnaire data, or forceful exertions and hand/wrist tendinitis. The surveillance data. estimate of RR for hand/wrist tendinitis among workers with jobs that were classified as HIGH Studies Reporting on the Association FORCE & LOW REPETITION was 4.8 (0.6, of Posture and Hand/Wrist Tendinitis 39.7), while the estimate for HIGH FORCE & Four studies addressed posture: Amano et al. HIGH REPETITION jobs was 17.0 (2.3, [1988]; Byström et al. [1995]; Luopajärvi et al. 126.2), compared to the comparison group of [1979]; and Kuorinka and Koskinen [1979]. LOW FORCE & LOW REPETITION jobs. Studies Meeting the Four Criteria Conclusions on Force and Luopajärvi et al. [1979] met all four evaluation Hand/Wrist Tendinitis criteria. Luopajärvi et al. [1979] compared the There is strong evidence for an association prevalence of hand/wrist tendinitis among 152 between work that requires forceful exertions, female assembly line packers in a food in combination with other job risk factors, and production factory to 133 female shop hand/wrist tendinitis based on currently assistants in a department store. Exposure to available epidemiologic data. All five of the repetitive work, awkward hand/arm postures, studies reviewed reported data that resulted in and static work was assessed by observation positive RR estimates. Four of the five and videotape analysis of factory workers. No estimates were statistically significant. formal exposure assessment was conducted on Eliminating one estimate of RR from a study the department store workers; their job tasks with noted limitations that might bias the were described as variable. Cashiers were estimate upward does not change this excluded, presumably because their work was conclusion. Potential confounders such as age repetitive. The health assessment consisted of and gender were examined in these studies (see interviews and physical examinations conducted discussion of potential confounders on page by a physiotherapist (active and passive 5b-16) and there was no evidence that motions, grip-strength testing, observation, and reported associations were distorted by palpation); and diagnoses of tenosynovitis and confounders. peritendinitis were later determined by medical specialists using these findings and POSTURE predetermined criteria. The PR for tendinitis among the assembly line packers compared to Definition of Posture for Hand/Wrist the shop assistants was 4.13 (95% CI

5b-13 2.63–6.49). Age, hobbies, and housework assessed using rating scales on nurse- were addressed, and no associations with administered questionnaires that addressed musculoskeletal disorders were identified. daily duration of hand and

Studies Meeting at Least One Criteria finger movements, wrist position, grip, and Amano et al. [1988] reported the prevalence of hand-tool use [Fransson-Hall et al. 1995]. cervicobrachial disorders, including Videotape analysis and electromyograms were tenosynovitis, among 102 assembly line conducted on a subgroup [Hägg et al. 1996]. A workers in an athletic shoe factory and 102 diagnosis of tenosynovitis or peritendinitis age- and gender-matched non-assembly line required the observation of swelling and pain workers (clerks, nurses, telephone operators, during active movement on physical cooks, and key punchers). Exposure examination. A diagnosis of deQuervain’s assessment was based on videotape analysis of disease required a positive Finkelstein’s test. the tasks of 29 workers on one assembly line. No cases of tenosynovitis or peritendinitis, Characteristic basic postures were summarized other than deQuervain’s disease, were found in by the investigators as: holding a shoe or a tool, this study, probably because of strict clinical extending or bending the arms, and keeping the criteria used for the case definition. The PR for arms in a certain position. Assembly line deQuervain’s disease among the automobile workers produced about 3,400 shoes a day. assembly line workers was 2.49 (95% CI All but one task had cycle times less than 30 1.00–6.23) compared to the general population seconds. No formal exposure assessment of the group. Psychosocial variables and other comparison group was reported. Diagnoses potential confounders or effect modifiers were were determined by physical examination, addressed by Fransson-Hall et al. [1995]. A including palpation for tenderness. The PRs for higher prevalence of deQuervain’s disease was tenosynovitis of the right and left index finger noted among men than women. flexors among the shoe factory workers were 3.67 (95% CI 1.85–7.27) and 6.17 (95% CI Kuorinka and Koskinen [1979] studied 2.72–13.97) respectively, compared to the occupational rheumatic diseases and upper limb non-factory workers. Tenosynovitis of the other strain among 93 scissor makers and compared digits was not diagnosed in the comparison them to the same group of department store group. Shoe assembly workers held shoe lasts assistants (n=143) that Luopajärvi used as a longer in the left hand and had greater comparison group. Temporary workers and frequency of symptoms in the left hand. those with recent trauma were excluded from Comparison subjects were matched to shoe the scissor makers group. Exposure assessment factory workers on gender and age (within five included videotape analysis of scissor maker years). tasks. The time spent in deviated wrist postures per work cycle was multiplied by the number of Byström et al. [1995] studied forearm and hand pieces handled per hour and the number of disorders among 199 automobile assembly line hours worked to create a workload index. workers and compared them to 186 randomly Cycle times ranged from 2 to 26 seconds; the selected subjects from the general Swedish number of pieces handled per hour ranged from population. For both groups, exposure was 150 to 605. No formal exposure assessment

5b-14 was conducted on the shop assistants. Health significant estimates. As noted in the Repetition assessment involved interview and physical Section, the possibility of examiner bias might examination by a exist in the study reported by Amano et al. [1988], potentially biasing the RR estimate upward. The middle of the range of statistically physiotherapist following a standard protocol. significant estimates for RR for hand/wrist Diagnoses of tenosynovitis and peritendinitis tendinitis is 2.5 to 4.1. were later determined from these findings using predetermined criteria (localized tenderness Temporal Relationship and pain during movement, low-grip force, Although all of the studies reviewed in this swelling of wrist tendons [Waris et al. 1979]). section were cross-sectional, at least two of the In equivocal cases, orthopedic and physiatric studies addressed temporality by reporting a teams determined case status. The PR for minimum length of employment (Luopajärvi et muscle-tendon syndrome among the scissor al. [1979]—5 years) or by evaluating exposure makers as 1.38 (95% CI 0.76–2.51) before health outcomes [Byström et al. 1995], compared to the department store assistants. as discussed in the previous sections on Whether or not cashiers were excluded from Repetition and Force. the comparison group in this study, as they were in the Luopajärvi et al. [1979] study is Consistency unclear. The study group was 99% female. No All of the studies reviewed showed positive RR relationship was found between age or body estimates for hand/wrist tendinitis among mass index and muscle-tendon syndrome. The occupational groups exposed to extreme number of symptoms increased with the number postures, ranging from 1.4 to 6.2. Three of the of parts handled per year. Analyses of four studies reviewed resulted in statistically subgroups of scissor makers showed non- significant PRs. Considering only statistically significant increased prevalence of muscle- significant estimates from studies not noted to tendon syndrome in short versus long cycle have design limitations that might bias the RR, tasks and in manipulation versus inspection narrows the range to 2.5 to 4.1. tasks. The authors noted a lack of contrast in exposures between the subgroups. A non- Coherence of Evidence significant trend of increasing prevalence of See Repetition Section. diagnosed muscle-tendon syndrome with increasing number of pieces handled per year Dose-Response was noted in a nested case-control analysis See Repetition Section. (n=36).

Strength of Association—Extreme Posture and Hand/Wrist Tendinitis The PRs for extreme postures and hand/wrist tendinitis ranged from 1.4 to 6.2. All of the PRs were greater than one and three of the four studies reported statistically

5b-15 Posture and Hand/Wrist Tendinitis

PR and 95% CI Authors Exposed/Unexposed Groups 4.1 (2.6–6.5) Luopajärvi et al. [1979] Food packers/department store assistants 3.7 (1.9–7.3) to Amano et al. [1988] Shoe assemblers/clerks, nurses, 6.2 (2.7–14.0) operators, cooks, keypunchers 2.5 (1.0–6.23) Byström et al. [1995] Auto assemblers/general population 1.4 (0.8–2.5) Kuorinka and Koskinen [1979] Scissor makers/department. store assistants

There is strong evidence for a positive tendinitis among women but found no significant association between work that requires associations with other medical factors or extreme postures, in combination with other job activities outside of work. However, significant risk factors, and hand/wrist tendinitis, based on differences in posture were observed between currently available epidemiologic data. All of males and females. Differences in postures may the studies reviewed reported be due to differences in height between men data that resulted in positive RR estimates. and women whose workstations have uniform Three of the four estimates from these studies dimensions. In McCormack et al.’s [1990] were statistically significant. Taking into account study of textile workers, three of the four the effect of potential confounders (See exposed groups were largely female Repetition Section) such as gender, age, and (89%–95%), limiting the ability to separate the study limitations does not alter this conclusion. effect of gender from job effect. However, in an analysis that included gender and job as risk Potential Confounders factors, they reported that gender was a significant predictor of tendinitis (p=0.01), but Gender not as significant a predictor as job category The association between gender and tendinitis (p=0.001). The other studies reviewed did not is not uniform. Byström et al. [1995] reported a have both male and female subjects. higher prevalence of deQuervain’s tendinitis in men than in women, and proposed the Age explanation that men in their study group used Several investigators noted that tendinitis hand tools more often than women. Ulnar appears to be more prevalent in younger age deviation and static muscle loading were groups. Byström et al. [1995] reported that likewise more often reported among men. most of the cases of deQuervain’s tendinitis Armstrong et al. [1987a] reported a higher occurred in the <40-yr age group. prevalence of McCormack et al. [1990] reported that age

5b-16 was not a significant predictor of tendinitis, but There is no evidence in the studies reviewed years on the job was inversely here that the associations reported between associated—prevalence was higher if less than work factors and hand/wrist tendinitis are 3 years on the job. Armstrong et al. [1987] explained by gender, age, or other factors. noted that “a significant interaction between sex, age, and years on the job suggested that CONCLUSIONS the risk of hand/wrist tendinitis might actually Eight epidemiologic studies have examined decrease with an increased number of years on physical workplace factors and their the job, but the effect was too small to merit relationship to hand/wrist tendinitis. Several further discussion.” Roto and Kivi [1984] noted studies fulfill the four epidemiologic criteria that that “The few cases of tenosynovitis occurred in were used in this review, and appropriately younger workers.” Kuorinka and Koskinen address important methodologic issues. The [1979] and Luopajärvi et al. [1979] found no studies generally involved populations exposed significant association between age and to a combination of work factors; one study tendinitis. assessed single work factors such as repetitive motions of the hand. We examined each of Other Potential Confounders these studies, whether the findings were McCormack et al. [1990] reported that race positive, negative, or equivocal, to evaluate the was not associated with tendinitis. Armstrong et strength of work-relatedness, using causal al. [1987a] found no significant associations inference. with personal factors—birth control pills, hysterectomy, oophorectomy, recreational There is evidence of an association between activities. No subjects with seropositive any single factor (repetition, force, and posture) rheumatic diseases were included in the and hand/wrist tendinitis, based on currently Kuorinka and Koskinen [1979] study. They available epidemiologic data. There is strong reported that their earlier unpublished evidence that job tasks that require a questionnaire found no correlations between combination of risk factors (e.g., highly illness and extra work, work outside the repetitious, forceful hand/wrist exertions) factory, work at home, or hobbies. Luopajärvi increase risk for hand/wrist tendinitis. et al. [1979] excluded subjects with previous trauma, arthritis, and other pathologies.

5b-17 Table 5b-1. Epidemiologic criteria used to examine studies of hand/wrist tendinitis associated with repetition

Investigator Risk blinded to indicators Participation Physical case and/or Basis for assessing Study (first author and (OR, PRR, IR rate $70% examination exposure hand/wrist exposure to year) or p-value)*,† status repetition

Met all four criteria:

Armstrong 1987a 5.5, Yes Yes Yes Observation or measurements 17.0† Luopajärvi 1979 4.1† Yes Yes Yes Observation or measurements

Met at least one criterion:

Amano 1988 3.7–6.2† NR‡ Yes NR Job titles or self-reports

Byström 1995 2.5† Yes Yes No Job titles or self-reports§

Kuorinka 1979 1.4 Yes Yes NR Observation or measurements

McCormack 1990 1.8 Yes Yes NR Job titles or self-reports

Roto 1984 3.1† Yes Yes Yes Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on repetition alone (i.e., repetition plus force, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §EMG and video analysis of subgroup reported in Hägg et al. [1996].

5b-18

Table 5b-2. Epidemiologic criteria used to examine studies of hand/wrist tendinitis MSDs associated with force

Investigator Risk blinded to indicator (OR, Participation Physical case and/or Basis for assessing Study (first author and PRR, IR or p- rate $70% examination exposure hand/wrist exposure to year) value)*,† status force

Met all four criteria:

Armstrong 1987a 17.0†, Yes Yes Yes Observation or 4.8 measurements

Met at least one criterion:

Byström 1995 2.5† Yes Yes No Job titles or self-reports§

Kurppa 1991 14.0–38.5† Yes Yes NR‡ Observation or measurements

McCormack 1990 3.0† Yes Yes NR Job titles or self-reports

Roto 1984 3.1† Yes Yes Yes Job titles or self-reports

*Some risk indicators are based on a combination of risk factors—not on force alone (i.e., force plus repetition, posture, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §EMG and video analysis of subgroup reported in Hägg et al. [1996].

5b-20

Table 5b-3. Epidemiologic criteria used to examine studies of hand/wrist tendinitis MSDs associated with posture

Investigator Risk blinded to case indicator (OR, Participation Physical and/or Basis for assessing Study (first author and PRR, IR or p- rate $70% examination exposure hand/wrist exposure to year) value)*,† status posture

Met all four criteria:

Luopajärvi 1979 4.1† Yes Yes Yes Observation or measurements

Met at least one criterion:

Amano 1988 3.7–6.2† NR‡ Yes NR Job titles or self-reports

Byström 1995 2.5† Yes Yes No Job titles or self-reports§

Kuorinka 1979 1.4 Yes Yes NR Observation or measurements

*Some risk indicators are based on a combination of risk factors—not on posture alone (i.e., posture plus force, repetition, or vibration). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §EMG and video analysis of subgroup reported in Hägg et al. [1996].

5b-22

Table 5b–4. Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence

Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Amano Cross- 102 assembly line Outcome: Examination Tenosynovitis, Tenosynovitis Participation rate: Not reported. et al. 1988 sectional workers in an by a physician: palpation right index finger right index athletic shoe factory for tenosynovitis and flexors: 32.35% finger flexors: Unclear whether examiner was blinded to job category (occupational groups compared to 102 tenderness. 8.82% PRR=3.67 1.85-7.27 examined on separate dates). No clear age and gender case definition provided. Potential for matched non- Exposure: One line of Tenosynovitis, Tenosynovitis examiner bias exists. assembly line 29 shoe assembly left index finger left index workers (clerks, workers was selected flexors: 36.27% finger flexors: Comparison group was matched in nurses, telephone for job analysis. 5.88% PRR=6.17 2.72-13.97 gender and age (within 5 years). operators, cooks, Videotapes were Tenosynovitis of other digits was not and key punchers). evaluated diagnosed in the comparison group. for movements of the upper extremities and Neurological exam and clinical tests of shoulders and cycle pinch strength, tapping, pressure, and and holding times. vibration sensibility were also done. No significant differences between groups in finger-pinch strength. Shoe No formal exposure workers failed the tapping test more assessment of often, had lower pressure-sensibility in comparison group. 1 of 10 fingers tested, and had lower vibration-sensibility in 2 of 10 fingers. One of 3 neurological maneuvers (Morley’s test) was more often positive in shoe workers. Exposure to toluene is noted and is a potential confounder for neurological findings. Assembly line workers produced about 3,400 shoes a day. All but one task had cycle times <30 sec. Assembly workers held shoe lasts longer in the left hand and had greater frequency of symptoms in left hand vs. non-assembly workers, who were assumed to use right hand (dominant hand) more frequently.

(Continued)

5b-24

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Armstrong Cross- 652 industrial Outcome: Positive 3.1% (Group 2) 0.6% PRR=4.8 0.6-39.7 Participation rate: 90% of workers 1987a sectional workers divided into findings on interview originally selected for inclusion actually 4 groups: (1) low and physical exam were 3.5% (Group 3) PRR=5.5 0.7-46.3 participated. force, low repetition required for case (comparison group, definition. 10.8% (Group 4) PRR=17.0 2.3-126.2 The effect of age, gender, years on the n=157), (2) high Tendinitis/teno-synovitis: job, and plant were analyzed. Higher force, low repetition localized pain or prevalence of tendinitis among women, (n=195), (3) low swelling lasting > a but not significantly associated with force and high week, and increased personal factors. Significant repetition (n=143), pain with resisted but differences in posture were observed and (4) high force not passive motion. between males and females. and high repetition Trigger finger: locking in (n=157). extension or flexion and Examiners were blinded to exposure a palpable nodule at status of study participants. base of finger.

DeQuervain’s: positive Finkelstein test with localized pain score of >=4 (range 1 to 8).

Exposure: To force and repetition assessed by EMG and video analysis of jobs performed by a sample of workers.

(Continued)

5b-25

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Byström Cross- 199 automobile Outcome: Tenosynovitis 8.04% 3.23% PRR=2.49 1.00-6.23 Participation rate: 96%. Study group et al. 1995 sectional assembly line or peri-tendinitis were (deQuervain’s randomly selected from assembly workers, compared diagnosed based on tendinitis) division of a plant. Comparison group is to 186 general physical examination from the MUSIC study [Hagberg and population. observations: swelling Hogstedt 1991]. and pain at the tendon sheath, peritendinous Examiners blinded to exposure status: area or muscle-tendon no, everyone examined by the authors junction during active was in the exposed group. movement of the tendon. deQuervain’s tendinitis: Results are reported separately for Positive Finkelstein’s males and females, and for age <40 test. years. Psychosocial variables and other potential confounders or effect Exposure: Daily modifiers were addressed by duration of hand and Fransson-Hall et al. [1995]. finger movements, manual handling, wrist Higher prevalence of deQuervain’s position, grip type, and tendinitis in males than in hand-tool use were females—possibly related to greater rated by workers on use of hand tools, ulnar deviation, 6-point scales in and/or static muscle loading. questionnaires [Fransson-Hall et al. No cases of tenosynovitis or 1995]. Forearm peritendinitis were found in this study, muscular-load and wrist probably because of strict clinical angle were evaluated criteria (required observation of by EMG and videotape swelling). analysis for a subgroup [Hägg et al. 1996].

(Continued)

5b-26

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Kuorinka Cross- 93 scissor makers Outcome: Tenosynovitis 18.3% 13.5% PRR=1.38 0.76-2.51 Participation rate: 81%. and sectional compared to 143 and peritendinitis Koskinen shop assistants. diagnosed by interview Examiner was not blinded to case 1979 and physical exam. status, but diagnosis was made Phase One: physical Physiotherapist separately, using predetermined criteria examination and examined workers, [Waris et al. 1979]. interview. diagnoses were from predetermined criteria Study group was 99% female. No Phase Two: work [Waris 1979] (localized relationship found between age or analysis. 10-month tenderness and pain body mass index and “muscle-tendon interval between during movement and syndrome.” phases. low grip-force and swelling of wrist The number of symptoms increased Comparison group tendons). In problem with the number of parts handled/year. was from another cases orthopedic and Workers were paid by piece rate. study that used the physiatric teams same method determined case status. Within the group of scissor makers, [Luopajärvi et al. non-significant increased prevalences 1979]. Exposure: Work history, of muscle-tendon syndrome in short hr, and production rates vs. long cycle tasks and in manipulation for the previous year vs. inspection tasks was reported. were taken from The authors noted a lack of contrast in company records. A exposures between the subgroups. A workload index was non-significant trend of increasing based on videotape prevalence of diagnosed muscle- analysis of scissor tendon syndrome with increasing maker workstations: number of pieces handled/year was time spent in deviated noted in a nested case-control analysis wrist-posture (n=36). (>20E)/work cycle; multiplied by number pieces handled multiplied by hr worked. No exposure assessment of shop assistants.

(Continued)

5b-27

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Kurppa Cohort: 377 meatcutters, Outcome: Defined as 12.5/100 person 0.9/100 14 (meat- 5.7-34.4 Participation rate: >70%. Job transfers et al. 1991 31-month meatpackers and physician-diagnosed years person years cutters) and employee termination followed up follow-up sausage makers tenosynovitis or (meatcutters) (males) with questionnaire. Questionnaire compared to peritendinitis of the hand response rate over 70%. 388 office workers, or forearm. Criteria 25.3/100 person 0.7/100 38.5 (meat- 11.7-56.1 maintenance were swelling or years (meat- person years packers) Exposed and comparison groups were workers, and crepitation and packers) (females) similar in age and gender mix, although supervisors. tenderness to palpation gender varied with job. along the tendon and 16.8/100 person 25.6 19.2-77.5 pain at the tendon years (sausage (sausage If same diagnosis occurred at same sheath, in the makers) makers) site in worker after 60 days, it was peritendinous area, or at considered new episode. Therefore, the muscle-tendon separate episodes may be junction during active recurrences, and thus influence movement of the tendon. results. Median interval of 233 days deQuervain’s tendinitis: between episodes. positive Finkelstein’s test (if not positive, Packers worked in temperatures 8E to included in tendinitis 10EC; sausage makers worked in group). 25% of temperatures 8E to 20EC. diagnoses made by physicians outside plant, Examiners were not blinded to criteria unknown. occupation of subjects.

Exposure: Job Plant selected because of high number categories selected of reports of musculoskeletal based on whether or disorders. All permanent workers in not strenuous manual meat cutting, sausage making and work was required. packing departments were included, Exposure data obtained after 3 months of work. from previous published Case ascertainment: Workers in non- literature and plant walk- strenuous jobs may not have sought throughs. medical care for MSDs since they might still be able to perform their jobs.

(Continued)

5b-28

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments Luopajärvi Cross- 152 female Outcome: Tenosynovitis 55.9% 13.5% PRR= 4.13 2.63-6.49 Participation rate: 84%. Workers et al. 1979 sectional assembly line and peritendinitis excluded from participation for packers in a food diagnosed by interview previous trauma, arthritis and other production factory and physical exam. pathologies. were compared to Physiotherapist 133 female shop performed active and Examiner blinded to case status: Not assistants in a passive motions, grip stated in article. department store. strength tests, Cashiers were observation and No association between age and MSDs excluded from palpation. Medical or length of employment and MSDs. comparison group. specialists used these Factory opened only short time. findings later to Hobbies and housework were not diagnose disorders significantly associated with outcome. using predetermined criteria [Waris 1979]. Unable to examine effect of job- specific risk factors because of job rotation. Exposure: Exposure to repetitive work, awkward hand/arm postures, and static work assessed by observation and video analysis of factory workers. No formal exposure assessment of shop assistants.

(Continued)

5b-29

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments McCormack Cross- Textile workers: 4 Outcome: Assessed by Boarding: 6.4% Other non- PRR=3.0 1.4-6.4 Participation rate: 90.5% for screening; et al. 1990 sectional broad job categories questionnaire and office: 2.1% 93.6% of those screened went on to involving intensive screening physical Sewing: 4.4% PRR=2.1 1.0-4.3 complete physical examination. upper extremity exam, followed by use—sewing diagnostic physical Packaging: 3.3% PRR=1.5 0.7-3.5 Stratified random sampling within (n=562), boarding examination. occupational groups. (n=296), packaging Knitting: 0.9% PRR=0.4 0.1-1.4 (n=369), and knitting Tendinitis: Positive Not mentioned whether examiners (n=352); compared physical findings Overall exposed blinded to exposure status (job to other non-office suggestive of group: 3.75% PRR=1.75 0.9-3.39 category). workers (n=468), inflammation. including machine Prevalence higher in workers with maintenance Severity reported as <3 years of employment. Race and workers, mild, moderate or age not related to outcome. Female transportation severe. gender was a significant predictor of workers, and tendinitis (p=0.01), but job category cleaners and Exposure: To repetitive was a stronger predictor (p=0.001). sweepers. finger, wrist and elbow motions based on 10/12 physician examiners recorded knowledge of jobs; no diagnoses within 12% of the mean for formal exposure the group. assessment performed. 47.9% of workers who had either positive screening physical exams or reported symptoms on questionnaire were diagnosed with tendinitis or tendinitis-related syndromes.

(Continued)

5b-30

Table 5b–4 (Continued). Epidemiologic studies evaluating work-related hand/wrist tendinitis

MSD prevalence Study Outcome and Exposed Referent RR, OR, Study design Study population exposure workers group or PRR 95% CI Comments

Roto and Cross- 90 meatcutters Outcome: Tenosynovitis 4.5% 0.0% Indetermin- Õ Participation rate: 100% for Kivi 1984 sectional compared to defined as swelling, ate meatcutters, 94% for comparison reference group of local pain and finger Symptom Symptom group. 72 construction weakness during prevalence rate: prevalence foremen who had movement (determined 30.0% rate: 10.0% PRR=3.09 1.43-6.67 Authors state that examiners were not been exposed to by questionnaire and blinded to occupation of subjects repetitive physical exam). because part of larger group of meat movements of the processing workers examined, but it is upper extremities in Exposure: Based on job unclear whether construction foremen their work. All title. Study groups were (referents) were examined separately. participants were selected based on male. general knowledge of Serologic testing for rheumatoid job tasks: meatcutters’ arthritis was done to control for work entailed physical potential confounding (none detected). exertion of upper extremities and Relatively strict diagnostic criteria used shoulders. Construction to avoid false positive cases. Authors foremen’s work did not note that tenosynovitis occurred in involve repetitive younger age groups. movements of the upper extremities. No formal Although the only diagnosed cases of exposure assessment. tenosynovitis occurred in the meatcutters (none in the referents), the authors were reluctant to infer association with meatcutting because of the relatively low prevalence rate (4.5%).

5b-31 CHAPTER 5c Hand-Arm Vibration Syndrome

SUMMARY In general, the studies listed in Table 5c–1 show strong evidence of a positive association between high level exposure to hand-arm vibration (HAV) and vascular symptoms of hand-arm vibration syndrome (HAVS). These studies are of workers with high levels of exposures such as forestry workers, stone drillers, stone cutters or carvers, shipyard workers, or platers. These workers were typically exposed to HAV acceleration levels of 5 to 36 m/s². These studies typically were cross sectional studies which examined the relationship between workers with high levels of exposures to HAV with a non-exposed control group. There is substantial evidence that as intensity and duration of exposure to vibrating tools increase, the risk of developing HAVS increases. There also is evidence that an increase in symptom severity is associated with increased exposure. As intensity and duration of exposure are increased, the time from exposure onset and beginning of symptoms is shortened.

As described above, the relationship between vibration exposure and HAVS was evaluated favorably with regard to other epidemiological causality criteria, including consistency and coherency of available information and evidence describing the temporal sequence of exposure and outcome.

INTRODUCTION In addition to the four criteria we used to The 20 epidemiologic studies discussed in this evaluate the studies, we determined whether review were selected according to criteria that studies demonstrated statistically significant appear in the introduction of this document. In associations between exposure attributes and our review, we evaluated the studies according health outcomes. We also examined whether to criteria that enabled us to assess the the observed associations were likely to be research. These criteria, including adequate caused or substantially influenced by major participation rate, definition of health outcome study flaws, including confounding and selection by both symptoms and medical exam criteria, bias. Some of these limitations are shown in the blinding of investigators to exposure/outcome descriptions of individual studies (Table 5c–2). status, and independent/objective measure of exposure, also are described in detail in the We then reviewed and summarized the studies Introduction. with regard to standard criteria used by epidemiologists to evaluate the causal In reviewing the studies, we gave greatest relationship between a health outcome and an qualitative weight to those which fulfilled all four exposure of interest. These criteria included of the above criteria. Table 5c-1 (all tables are strength of association, temporal relationship, presented at the end of the chapter) consistency of association, coherence of characterizes the 20 reviewed Hand-Arm association, and exposure-response Vibration studies according to the four relationships. evaluation criteria. Full summary descriptions of all the studies appear at the end of the chapter. In this review, results of each of the studies

5c-1 examined, whether negative, positive, or sectional studies are limited in their ability to equivocal, contributed to the pool of data used ascertain temporal relationships between to make our decision regarding the strength of exposure and outcome. Because results are the causal relationship between HAVS and obtained at only one point in time, the cross- workplace risk factors. Greater or lesser sectional study design also is subject to confidence in the findings reflected the underassessment of the health outcome evaluation criteria described above. (particularly in groups with longer durations of employment and higher participant attrition). Definition of HAV for HAVS Hand-Arm Vibration is defined as the transfer The studies included in this review varied in of vibration from a tool to a worker’s hand and design and quality of information. Sixteen were arm. The amount of HAV is characterized by cross-sectional in design, and three were the acceleration level of the tool when grasped prospective cohort in design. One study was by the worker and in use. The vibration is both cross-sectional and prospective, including typically measured on the handle of tool while 10 cross-sectional follow-ups over time and a in use to determine the acceleration levels cohort group [Koskimies et al. 1992]. Thirteen transferred to the worker. of the 20 studies reported assessing case status using physical exams, while other studies used EVIDENCE FOR THE WORK- only a questionnaire to determine outcomes. Of RELATEDNESS OF HAVS the studies in which the subjects underwent a The hazardous effects of occupational exposure physical exam, five performed a cold to HAV have been discussed in hundreds of provocation test [Bovenzi et al. 1988; Bovenzi studies dating to the work of Loriga in 1911. et al. 1995; Brubaker et al. 1983; Brubaker et The composite of vibration-induced signs and al. 1987; McKenna et al. 1993], three symptoms referred to as hand-arm vibration performed a nail compression test [Mirbod et syndrome includes episodic numbness; tingling al. 1992b; Nagata et al. 1993; Saito 1987], and blanching of the fingers, with pain in one performed a nerve conduction test response to cold exposure; and reduction in [Virokannas 1995], one performed grip strength and finger dexterity. These signs sensorineural physician testing [Bovenzi and and symptoms are known to increase in Betta 1994], one performed a neurological severity as exposure to vibration increases in exam [Shinev et al. 1992], one performed an intensity and duration. Allan test [Nilsson et al. 1989] and one used physician judgement based on workers’ A review of pertinent epidemiologic studies of complaints and history [Koskimies et al. 1992]. HAVS has been previously presented [NIOSH 1989]; therefore, Table 5c–2 includes only Twelve of the 20 studies conducted an those studies completed after 1989. Except for exposure assessment of the tools subjects were a few longitudinal studies of chain sawyers in using; an additional study used exposure the United Kingdom, Finland, and Japan, the assessment information the authors had literature comprises largely cross-sectional collected in a previous investigation. The studies carried out within an industry. Cross- remaining studies estimated exposures by self-

5c-2 report or job title. Koskimies et al. [1992] examined vibration syndrome in a group of forestry workers The one study that met all four criteria and the employed by the National Board of Forestry in four studies which met the three criteria are Finland. All those employed in one parish discussed in the following section. Detailed participated in a series of 10 cross-sectional descriptions for all 20 investigations can be studies from 1972 to 1990. Results also were found at the end of the chapter. reported for a cohort of 57 individuals who remained in the study from 1972 to 1986. HAVS symptoms were assessed by Comments Related to Specific questionnaire and physical exam criteria. Studies of HAVS Exposure to chain saw vibration was The Bovenzi et al. [1995] cross-sectional determined by measurement of front handle investigation of forestry workers compared acceleration. Cross-sectional analysis results vibration white finger (VWF) in this group with showed a monotonic decrease in prevalence of shipyard worker referents. VWF was VWF from 40% in 1972 to 5% in 1990. In the diagnosed by symptom report and cold cohort of 57, VWF increased from 30% in provocation test; vibration exposures were 1972 to 35% in 1975. VWF decreased estimated by questionnaire report on frequency monotonically to approximately 6% in 1986. of chain saw work and types of saws used, Over the same time period, modifications of along with direct measurement of vibration chain saws used by the workers resulted in a produced by 27 antivibration and 3 non- decrease in saw vibration acceleration from 14 antivibration saws. Daily exposure to saw m/s2 to vibration was estimated by linking the two 2 m/s2. The authors attributed the reduction in assessments. The prevalence rates for VWF VWF to saw changes, although exposures and were 23.4% in forestry workers and 2.6% in outcomes were never linked for individual shipyard referents [Odds ratio (OR) 11.8, 95% workers. Strengths of the study included Confidence Interval (CI) 4.5–31.1]. For observation of similar results from the series of workers using only antivibration saws, the OR cross-sectional analyses and full participation was 6.2 (95% CI 2.3–17.1); for those using on the part of the 57 subjects. Limitations non-antivibration saws, the OR was 32.3 (95% included failure to assess chain saw exposure CI 11.2–93). A dose-response was observed measures at the individual level. The study for VWF and lifetime vibration dose (OR 34.3, demonstrated the potential for symptom 95% CI 11.9–99, for the highest category). improvement after exposure reduction. Although participation rates were not stated for referents, the participation appeared to be In the Nilsson et al. [1989] cross-sectional 100% for forestry workers. Authors included study of male pulp mill machine manufacturing 10 retired workers to lessen the problems with employees, VWF was examined in a group of selection out of the workforce. Results 89 platers and 61 office workers. VWF was demonstrated that antivibration saw use was ascertained by physical exam and interview. associated with a lower prevalence of VWF. For platers, vibration exposure was assessed by measuring acceleration intensity on a sample

5c-3 of tools and linking results to subjective ratings by participant attrition. of exposure time. Current and past exposures were estimated for both platers and office The Bovenzi et al. [1988] cross-sectional workers (some office workers had experienced investigation examined VWF in vibration- exposures in the past). Prevalence for platers exposed stone drillers and stone with current exposure was 42%, in comparison cutters/chippers and a reference group of to 2.3% for office workers with no exposures quarry and mill workers. VWF was assessed (OR 85, 95% CI 15–486). When those by questionnaire and physical exam. Exposure exposed to vibration (platers plus office was assessed by measuring acceleration workers with previous vibration exposure) intensity on a sample of tools and linking it with were compared to unexposed office workers, self-reported exposure time. VWF prevalence prevalences were 40.0% and 2.3% rates were 35.5% in exposed and 8.3% in respectively (OR 56, 95% CI 12–269). A unexposed groups (OR 6.06, 95% CI dose-response was observed for VWF and 2.0–19.6; OR 4.26, 95% CI 1.8–10.4). A years of exposure. The relationships between significant association was observed between outcome and exposure, after adjustment for vibration acceleration level and severity of age, were strong. Representativeness of the VWF symptoms (0% and 18.4% in the lowest referent group of office workers could not be and highest categories, respectively). determined. Strength of Association Bovenzi [1994] examined HAVS cross- sectionally in 570 quarry drillers and stone One of the studies examined met all four of the carvers, along with a referent group of polishers evaluation criteria [Bovenzi et al. 1995]. Five and machine operators who were not exposed investigations met three of the criteria [Bovenzi to hand-transmitted vibration. HAVS was et al. 1988, 1994; Kivekäs et al. 1994; assessed by physician interview, and Koskimies et al. 1992; and Nilsson et al. sensorineural symptoms were staged and 1989]. The criterion that was not met (or not graded. Exposure to vibrating tools was reported) by four of the studies was blinding of assessed by interview and linked to vibration the physician with regard to worker job status. measurements obtained from assessment of a However, most studies used objective sample of tools. Prevalences of HAVS were measures for determining case status: cold 30.2% in the exposed and 4.3% in the provocation [Bovenzi et al. 1988, 1995], unexposed groups (OR 9.33, 95% CI sensorineural physician grading [Bovenzi and 4.9–17.8). Symptoms of VWF increased with the Italian Study Group 1994], and the Allan lifetime vibration dose (OR 10.2, 95% CI test [Nilsson et al. 1989]. Use of objective 4.8–21.6, for the highest category). Study measures lessens the likelihood that case status strengths included detailed exposure was influenced by knowledge of participants’ assessment and modeling of relationships, exposures. 100% participation, and a very stable work population. Because of the work population In the Bovenzi et al. [1988] cross-sectional stability, results were unlikely to be influenced investigation, vibration-exposed stone drillers

5c-4 and stone cutters/chippers showed a 6.06-fold from 1972 to 1986 showed a decrease in (95% CI 2.0–19.6) increase in risk of VWF in VWF from 30% to 6%. The reductions were comparison to unexposed quarry and mill attributed to modifications in chain saws during workers. Similar results were observed in the same time period. another study of stone workers conducted by Bovenzi in 1994. Quarry drillers and stone The remaining, less rigorous, studies showed carvers exposed to vibration showed an OR for varying relationships between HAVS and VWF of 9.33 (95% CI 4.9–17.8) when exposure. The majority of the studies compared to a reference group of polishers and demonstrated moderate to strong positive machine operators. A dose-response relationship was observed for VWF and associations. Most compared exposed to lifetime vibration dose, with an OR of 10.2 unexposed groups with little or no detailed (95% CI 4.8–21.6) for the highest exposure analysis by exposure level. Two investigations category. A study of forestry workers [Bovenzi examined HAVS in exposed groups and found et al. 1995] demonstrated an OR of 11.8 (95% an increase in risk by years of employment, CI 4.5–31.1) for VWF when comparing with ORs of 8.4 and 8.9 (95% CI 2.9–28.9) when comparing the highest and lowest forestry workers with exposure to chain saw vibration to an unexposed group of shipyard categories [Mirbod et al. 1992b; Kivekäs et al. workers. A lower risk of VWF (OR 6.2, 95% 1994]. Another study that examined HAVS CI 2.3–17.1) was observed for those using only prevalence in power tool users found no antivibration saws. A dose-response between association with duration of employment (with a VWF and vibration exposure also was observed participation rate of only 38%) [Musson et al. in this investigation, with an OR of 34.3 (95% 1989]. For other investigations, exposed and CI 11.9–99) for the highest exposure category. unexposed groups were defined by job titles. Nilsson et al. [1989] observed very strong ORs for these studies ranged from 3.2 to 40.6 relationships between VWF and exposure to (relative risk [RRs] from 3.2 to 16) [Brubaker vibration in machine manufacturing platers. In et al. 1983; Dimberg and Oden 1991; Letz et comparison to office workers with no al. 1992; McKenna et al. 1993; Mirbod et al. exposure, platers had an OR of 85 (95% CI 1992a; Mirbod et al. 1994; Nagata et al. 15–486). Kivekäs et al. [1994] found a 1993]. Three studies demonstrated varying significantly increased OR in the cumulative HAVS rates for exposed groups, but included incidence of HAVs in a 7-year cohort study no referents [Shinev et al. 1992; Starck et al. (OR 6.5, 95% CI 2.4–17.5). Koskimies et al. 1990; Virokannas and Tolonen 1995]. [1992] examined a dynamic cohort of forestry workers at 10 intervals from 1972 to 1990 Two investigations produced conflicting during which time saws were being modified in evidence related to the effects of chain saw weight, vibration frequency, and vibration modifications on HAVS in forestry workers. acceleration. Over the 18-year period, a The Brubaker et al. [1987] study, observed a monotonic decrease in VWF was observed in 28% increase in prevalence of VWF in a the 10 cross-sectional examinations, with an cohort of tree fellers over a 5-year period and overall eight-fold reduction in prevalence. A claimed that saw modifications were ineffective. subset of workers followed Saito [1987] found no new HAVS symptom

5c-5 development over 6 years in a cohort of chain saws [Pyykkö and Starak 1986; Koskimies et sawyers in reducing symptoms. al. 1992].

Comparing construction workers to office Consistency workers, one study demonstrated an OR of 0.5 The literature consistently shows that workers (95% CI 0.1–11.8) for HAVS. This study met exposed to HAV develop HAVS at a none of our four criteria [Miyashita et al. 1992]. substantially higher rate than workers not exposed to vibration. Although there is In general, the studies in Table 5c-1 show considerable variation in the occurrence of strong evidence of a positive association HAVS among different groups using similar between exposure to HAV and vascular types of vibrating tools, the lack of consistency symptoms of HAVS. probably is explained by methodological differences between studies (i.e., some Temporality researchers did not account for exposure variation over time in the summary estimate of The temporal relationship between HAV exposure) or by differences in work methods, exposure and symptoms of HAVS is well work processes, and work organization established by studies which have determined [Gemne et al. 1993]. Important also is the the latency between exposure and symptom difference in the intensity and duration of onset. Of 52 studies reviewed by NIOSH in exposure. 1989, 44 included some information about the latency period for the development of vascular HAVS symptoms following initial exposure. Coherence of Evidence Latency ranged from 0.7 to 17 years, with a The mechanisms by which HAV produces mean of 6.3 years. Unfortunately, because most neurological, vascular, and musculoskeletal of these studies were cross-sectional (i.e., damage are supported by some experimental latency was determined retrospectively) and evidence [Armstrong et al. 1987b; Lundborg et because HAVS develop slowly, the possibility al. 1990; Necking et al. 1992]. Neurological of recall bias is strong [Gemne et al. 1993]. and circulatory disturbances probably occur However, longitudinal studies provide support independently and by unrelated mechanisms. for the temporal nature of the association. Vibration may directly injure the peripheral Kivekäs et al. [1994], in a 7-year follow-up of nerves, nerve endings, and mechanoreceptors, Finnish lumberjacks, found a cumulative producing symptoms of numbness, tingling, incidence rate (IR) of 14.7%, compared to a pain, and loss of sensitivity. Vibration also may cumulative IR of only 2.3% among referents. have direct effects on the digital arteries. The The cumulative IR of lumberjacks who had innermost layer of cells in the blood vessel walls more than 25 years of exposure at the end of appears especially susceptible to mechanical the follow-up period was 30.6%. Other studies injury by vibration. If these vessels are of Finnish forestry workers also showed a damaged, they may become less sensitive to the marked decrease in HAVS prevalence actions of following the introduction of improved chain

5c-6 certain vasodilators that require an intact relationship between the dose and the latency endothelium. Experiments involving of symptom onset. lumberjacks exposed to chain saw vibration support this hypothesis [Gemne et al. 1993]. Support for the first relationship is provided by There also is evidence that the walls of the a few longitudinal studies of workers exposed digital blood vessels are thickened in persons to HAV. In general, all show strong evidence with HAVS [Takeuchi et al. 1986]. During that decreasing the acceleration level of a hand- cold exposure, vessels with these changes will held vibrating tool has a positive relationship become abnormally narrow and may close with prevalence of HAVS. In a study of Finnish entirely [Gemne 1982]. Symptoms of numbness forestry workers using chain saws, Koskimies and tingling which characterize HAVS may be et al. [1992] found that the prevalence of secondary to vascular constriction of the blood HAVS symptoms declined from a peak of 40% vessels, resulting in ischemia in the nerve-end to 5% after the introduction of light-weight, organs. low-vibration chain saws with reduced acceleration from 14 to 2 m/s2. Likewise, a Other evidence concerning the coherence of study of similar workers in Japan found that the information regarding the association between prevalence of vascular symptoms among chain vibration exposure and HAVS relates to saw operators who began their jobs before the background prevalence of similar disorders in introduction of various engineering and the general population. One estimate placed the administrative controls peaked at 63%. prevalence of Raynaud's phenomenon at 4.6% (Vibration acceleration levels for chain saws for females and 2.5% for males in the general used during this period ranged from 111 to 304 population [Iwata and Makimo 1987]. Only 7 m/s2.) In contrast, the peak prevalence for of the studies examined in this review found chain saw operators who began working after prevalence rates less than 20% among workers the introduction of antivibration chain saws exposed to HAV. In the 1989 NIOSH review, (acceleration level: 10-33 m/s2) and exposure only 9 of 52 cross-sectional studies reported a duration limits (2 hrs/day) was only 2% prevalence rate of less than 20% among [Futatsuka and Uneno 1985, 1986]. workers exposed to HAV. This provides strong evidence that individuals working in NIOSH authors ranked 23 cross-sectional vibration-exposed occupations are at much studies that measured HAV acceleration levels higher risk of these disorders than those in the and estimated a prevalence rate for vascular general population. symptoms [NIOSH 1989]. To test whether a linear relationship existed between the HAV level and the prevalence of vascular symptoms, Exposure-Response Relationship a correlation coefficient was calculated. The Exposure-response relationships involving correlation analysis found a statistically HAV have been postulated, including: (1) a significant linear relationship between HAV relationship between the prevalence of HAVS acceleration level and prevalence of vascular and vibration acceleration (and cumulative symptoms (R 0.67, p<0.01), indicating that exposure time), (2) a relationship between the prevalence of vascular symptoms tends dose and symptom severity, and/or (3) a to increase as the HAV acceleration level

5c-7 increases. However, the absorption of vibration latency of symptom onset is provided by British energy by the hand is influenced by the studies conducted in the 1970s among various vibration intensity, as well as by frequency, occupational groups, including chain sawyers, transmission direction, grip and feed forces, grinders, chiselers and swagers [Gemne et al. hand-arm postures, and anthropometric factors 1993]. Exposure to 10-25 m/s2 chainsaw [Gemne et al. 1993]. vibration correlated with a latency of about 3 years. Pedestal grinders using machines with Several studies reviewed for the current zirconium wheels were exposed to vibration document found relationships between levels of 50 m/s2 and demonstrated a mean prevalence of HAVS and duration of vibration- latency of 1.8 years, whereas grinders who exposed work [Bovenzi 1994; Bovenzi et al. used softer wheels with accelerations of 10-20 1995; Letz et al. 1992; Nilsson et al. 1989]. m/s2 had a mean latency of 14 years. Exposure One cross-sectional study with a very poor to 70 m/s2 vibration during swaging correlated response rate found no association with with a mean latency of about 7 months, duration of exposure [Musson et al. 1989]. although some swagers developed symptoms in as few as 6 weeks. Justification for a relationship between dose and HAVS prevalence and symptom severity is Confounding and HAVS provided by Bovenzi et al. [1988] and Mirbod Age and metabolic disease are the primary et al. [1992b]. In a study of stone-cutters using potential confounders for HAVS. rock drills and chisel hammers, Bovenzi found that HAVS prevalence increased linearly with It is important that epidemiologic studies the total number of working hours, from about examine non-occupational factors, and control 18% for persons with 6,000 hrs of exposure, to for them. Most of the studies were able to more than 50% among persons with more than address “age” by stratification in their analyses, 26,000 hrs of exposure. Likewise, in a study of or through use of multiple logistic regression. 447 workers using chain saws, Mirbod et al. [Bovenzi and Betta 1994; Bovenzi et al. 1995; [1992b] found that the prevalence of HAVS Brubaker et al. 1983, 1987; Kivekäs et al. increased from 2.5% among workers with less 1994; Letz et al. 1992; McKenna et al. 1993; than 14 years of exposure to 11.7% among Mirbod et al. [1994]. Several authors workers with 20–24 years exposure, to 20.9% controlled for metabolic disease [Bovenzi and among workers exposed 30 years or more. Betta 1994; Bovenzi et al. 1995; Letz et al. Both studies found a statistically significant 1992; McKenna et al. 1993]. This is important correlation between the severity of symptoms because of the effects that some disorders have (graded according to the Taylor-Pelmear scale) on peripheral circulation which may have and a dose measure based on total exposure symptoms similar to HAVS. time.

Support for a relationship between dose and Nonoccupational Raynaud’s phenomena - a

5c-8 rare disorder which mimics HAVS has been HAVS increases. Most of the studies showed a known to occur in individuals with metabolic positive association between high level disorders, peripheral neuropathy, alcohol- exposure to HAV and vascular symptoms of related illness, as well as other conditions. HAVS. For many of the studies there is a strong association between HAVS and In reviewing the methods and results of these exposure to vibrating tools in the workplace. studies, taking into account substantially The temporal relationships and consistency elevated ORs and evidence of dose-response between exposure and symptoms of HAVS are relationships, it appears that potential well established in these studies. The confounders do not account for the consistent mechanisms by which HAV produces relationships seen. neurological, vascular, and musculoskeletal damage are supported by some experimental Review of the 20 studies, leads us to the evidence. Many of the studies showed an conclusion that there is substantial evidence that exposure-response relationship between dose as intensity and duration of exposure to of HAV and the HAVS prevalence and vibrating tools increase, the risk of developing symptom severity.

5c-9 Table 5c-1. Epidemiologic criteria used to examine studies of hand/wrist and hand/arm MSDs associated with vibration

Risk Investigator indicator Physical blinded to (OR, PRR, IR examination case and/or Basis for assessing Study (first author and or p- Participation or cold exposure hand/wrist or hand/arm year) value)*,† rate $70% provocation status exposure to vibration

Met all four criteria:

Bovenzi 1995 6.2–32.3† Yes Yes Yes Observation or measurements

Met at least one criterion:

Bovenzi 1988 6.06† NR‡ Yes NR Observation or measurements

Bovenzi 1994 9.33† Yes Yes No Observation or measurements

Brubaker 1983 NR Yes Yes NR Job titles or self-reports

Brubaker 1987 NR No Yes NR Observation or measurements

Dimberg 1991 NR† Yes No NR Job titles or self-reports

Kivekäs 1994 3.4–6.5† Yes Yes Yes Job titles or self-reports

Koskimies 1992 NR Yes Yes NR Observation or measurements

Letz 1992 5.0–40.6† Yes No No Job titles or self-reports— previous study results used

McKenna 1993 24.0† NR Yes No Job titles or self-reports

Mirbod 1992a, 1994 3.77† NR No NR Observation or measurements

Mirbod 1992b NR NR Yes No Observation or measurements

Musson 1989 NR No No NR Observation or measurements

Nagata 1993 7.1† NR Yes No Job titles or self-reports

Nilsson 1989 14–85† Yes Yes NR Observation or measurements

Saito 1987 NR No Yes NR Job titles or self-reports

Shinev 1992 NR NR Yes NR Observation or measurements

Starck 1990 NR NR No No Observation or measurements

Virokannas 1995 NR† NR Yes NR Observation or measurements

Met none of the criteria:

Miyashita 1992 0.5 NR No No Job titles or self-reports

*Some risk indicators are based on a combination of risk factors–not on vibration alone (i.e., vibration plus force, posture, or repetition). Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If combined with NR, a significant association was reported without a numerical value. ‡Not reported.

5c-10

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi et Cross- Vibration-exposed stone Outcome: Assessed by 35.5% 8.3% 6.06 2.01-19.6 Participation rate: Participation al. 1988 sectional drillers (n=32) and stone physical examination and rate cannot be determined from cutters/chippers (n=44); questionnaire. VWF data in the study. quarry and mill workers symptoms staged using the not exposed to vibration Taylor-Pelmear scale. Significant association (control group, n=60). between vibration acceleration Exposure: Vibration level and severity of VWF exposure assessed by symptoms. measuring the acceleration intensity on a sample of tools, Mean latency period to together with subjective symptom onset =12.3 yr. ratings of exposure time. Frequency-weighted acceleration levels ranged from 19.7 to 36.4 m/s 2 (rock drills and chipping hammers) and from 2.4 to 4.1 m/s 2 (grinders and hand cutters).

(Continued)

5c-12

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi Cross- Case group: Stone Outcome: HAVs assessed 30.2% 4.3% 9.33 4.9-17.8 Participation rate: 100% “All and the sectional workers employed in nine by physician-administered the active stone workers Italian districts in Northern and interview; sensineural participated in the study, so Group Central Italy: 145 quarry symptoms staged according self-selection was not a source of bias.” 1994 drillers and 425 stone to Brammer [1992]. Graded carvers exposed to according to the Stockholm Physician administered the vibration. Referent group: scale [Gemne 1987]. questionnaires containing work polishers and machine history and examinations, so unlikely to be blinded to case operators (n=258) who Exposure: To vibrating tools status. performed manual activity assessed by interview. only not exposed to Vibration measured in a Adjusted for age, smoking, hand-transmitted sample of tools used. alcohol consumption, and vibration. upper limb injuries. Leisure activities, systemic diseases included in questionnaire. Univariate analysis showed no association between systemic diseases and vibration so was not criteria for exclusion. Univariate analysis showed no association between systemic diseases and vibration so was not criteria for exclusion. Dose–response for CTS and lifetime vibration exposure not significant. Frequency-weighted acceleration levels = 15 m/s 2 (stone drills), 21.8 m/s 2 (stone hammers), 2.84 m/s 2 (rotary grinding tools). Percent of workers affected with HAVs increased in proportion to the square root of the exposure duration.

(Continued)

5c-13

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Bovenzi et Cross- 222 active forestry Outcome: (1) History of All Forestry Shipyard (adjusted Participation rate: 95% vibrating al. 1995 sectional workers and 10 retired episodes of cold provoked workers: workers: OR’s) tool users, not reported for forestry workers with well–demarcated blanching in 23.4% 2.6% 11.8 4.5-31.1 control. >400 hr of sawing one or more fingers and (2) occurrence after Workers using Analysis controlled for age, compared with 195 smoking, drinking habits. randomly chosen employment and exposure to only AV chain shipyard workers never hand vibration and vibration saws: 13.4% 6.2 2.3-17.1 Physicians blinded to case exposed to hand white finger (VWF) attacks in status–since cold provocation vibration. Controls last 2 years and (3) abnormal Workers using test was used, it was not an excluded for digital arterial response to chain saws 32.3 11.2-93 issue. cold provocation. Clinically, without cardiovascular and Smoking, alcohol, metabolic, metabolic disease. VWF graded using Stockholm vibration VWF scale. isolation operators of cardiovascular, neurologic, systems: non-AV and previous musculoskeletal Exposure: Vibration 51.7% AV saws vs. injuries, use of medicines measured on front and rear Operators of included in questionnaire and of 27 antivibration (AV) chain antivibration accounted for in logistic saws used in the forest; for saws only: regression model. previous exposure OR=4.0 assessment, 3 non-AV chain Cold provocation testing saws were measured. performed on both forestry Lifetime workers and controls. Vibration measurements vibration were made in the field during dose Exposure–response cross-cutting operations by in 9m relationship found between skilled workers according to (m2S-4 hd) VWF and vibration exposure: ISO 7505. <19: the expected prevalence of Forestry workers gave OR=4.1 1.1-16.4 VWF increased almost linearly 19-20: to either the 8-hr detailed list of chain saws energy–equivalent used. OR=4.7 1.3-16.1 20-21: frequency–weighted Workplace questionnaires OR=9.4 3.1-28.4 acceleration or the number of validated by direct interviews >21: years of exposure (with with employers and OR=34.3 11.9-99 equivalent acceleration employees, employment unchanged). records, and amount of fuel used by chain saws Daily exposure to saw vibration assessed in terms of 8-hr energy–equivalent frequency–weighted acceleration.

(Continued)

5c-14

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Brubaker et Cross- 146 tree fellers in 7 Outcome: VWF symptoms With With Õ Õ Participation rate: 100%. al. 1983 sectional camps employed for $1 staged using Taylor-Pelmear symptoms: symptoms: year compared to 142 scale. 51% 5% Smoking, no significant workers not exposed to differences. vibration matched for Ischemic water bath testing Stage 3: 22% Stage 3: 2% location. for VWF completed on all Age was significantly different subjects. Excluding between cases and controls. other vibration Exposure was based on exposure and Height and weight not questionnaire data. medical 2% significantly different. history: 54% Stage 3: 1% Mean latency period between Stage 3: 25% work and symptoms 8.6 years. Records of duration of exposure.

(Continued)

5c-15

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Brubaker et Cohort: Fellers at Canadian Outcome: Defined as HAVs Raynaud’s Raynaud’s Õ Õ Participation rate: 53%. al. 1987 5-year lumber camps (n=71) symptoms, assessed by symptoms: symptoms: follow-up who had questionnaire and digit 53% 51% Original group (1979 to 1980) of been interviewed systolic blood pressure. (1984 to 1985) (1979 to included 146 fellers. exposed and tested in 1979 to 1980) group. 1980 then again in VWF symptoms staged using Tingling, 16 fellers excluded because of 1984 to 1985. Taylor-Pelmear scale. numbness: Tingling potential confounders. 56% numbness: Ischemic water bath testing (1984 to 1985) 65% Author concluded antivibration for VWF completed on all (1979 to saws not effective at subjects. 1980) preventing HAVs. Exposure: Vibration 15% of fellers reported new measurements recorded from symptoms of VWF over 5-year a representative sample of period. chain saws used in the logging camp. 28% increase in prevalence of VWF in workers using antivibration chain-saws. Correlation between objective test and symptoms poor: 54% reporting symptoms with positive findings on objective tests.

(Continued)

5c-16

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Dimberg Cross- 2,814 Aircraft engine Outcome: Exposure to 23% Õ Multivariate Õ Participation rate: 96% and Oden sectional workers. vibrating hand-tools (polishers/ analysis questionnaire. 1991 assessed by questionnaire. grinders) showed 68 Sheet metal workers. White fingers as a spasm in increased Vibrating tool use significantly 26 Polishers/grinders. blood vessels occurring in 19% (sheet symptoms correlated with HAVs symptom one or more fingers in metal with prevalence. 20 Cleaners. connection with cooling workers) increasing 40 Forklift-truck drivers. leading to reversible pallor age, work Analysis was stratified by 46 Engine testers. followed by redness. 15% with vibrating gender, age, and employment 146 Fitters. (cleaners) hand tools category. Exposure: Vibration and weight 49 Storemen assessed by questionnaire: loss 38 Electric welders. working with vibrating tools, time in present job, leisure No control group used. activities.

(Continued)

5c-17

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Kivekäs et Cohort 213 lumberjacks and Outcome: HAVs assessed Prevalence Prevalence Participation rate: 76% among al. 1994 with 140 referents. by questionnaire, clinical (HAVs) (HAVs): exposed workers, 78% among 7-year examination, and control. follow-up radiographs. 1978: 5.0% For 1978: (1978 to 1978: 16.9% OR= 3.4 1.7-6.9 Follow-up group included 76% 1985) Exposure: Not measured. of lumberjacks and 78% of Exposure history determined 1985: 5.7% referents from original group. via questionnaire. For 1985: 1985: 24.9% Cumulative OR= 4.4 2.3-8.1 Adjusted for age. incidence Cumulative HAVs (7 X-ray films read by radiologists incidence, years): blinded to case status HAVs (7 2.3% years): After adjusting for age, no 14.7% OR=6.5 2.4-17.5 difference in lumberjacks with <15-years exposure and referents, but risk increased with increasing duration of exposure. For those exposed RR=8.9 (2.9-28.9). No X-ray differences in prevalence of detectable translucencies or osteoarthritic changes in wrists or hands.

(Continued)

5c-18

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Koskimies Cohort Finnish forest workers Outcome: HAVs assessed Prevalence of Prevalence Õ Õ Participation rate: 100% of et al. 1992 (18-year (n=118-124). by questionnaire and HAVs among of HAVs those who had a yearly follow-up) physical examination. forestry among physical exam. workers in forestry Exposure: Vibration 1990: 5% workers in Decrease in prevalence acceleration of the front 1972: 40% attributed to reduction in weight handle of chain saws of saws, increase in vibration analyzed. frequency, and reduction in vibration acceleration (from 14 to 2 m/s 2).

(Continued)

5c-19

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Letz et al. Cross- Shipyard workers with Outcome: HAVs assessed Vascular Vascular Part-time Participation rate: 79%. 1992 sectional full-time vibration by self-administered symptoms symptoms: vibration- exposure (n=103); part- questionnaire; graded among part- 5.7% exposed Participants randomly selected time vibration exposure according to the Stockholm time vibration- workers vs. within departments. (n=115), and no vibration scale. exposed controls: exposure (n=53, workers: 33% OR=8.23 2.3-35.4 Significant exposure–response comparison group). Vibration measurements from relationship found after 51 pneumatic tools made in 3 Vascular Full-time adjustment for smoking, not studies. Extreme variability symptoms vibration- age or race. precluded direct comparison among full- exposed of tools. Number of hours per time vibration- workers vs. Average latency to symptom week and years using tools exposed controls: onset <5 years. asked. workers: OR=40.6 11-177 70.9%; Alcohol consumption, past medical conditions considered Sensorineural Sensori- Part-time in analysis. symptoms neural vibration- among part- symptoms: exposed Exposure–response time vibration- 17% workers vs. relationship found regarding exposed controls: self-reported cumulative workers: OR= 5.0 2.1-12.1 exposure to vibratory tools, 50.4% sensorineural stages, and corresponding vascular Sensorineural Full-time classifications but no further symptoms vibration- increases in workers with > among full- exposed 17,000 hr of exposure. time vibration- workers vs. exposed controls: Median latency for appearance workers: OR=24.7 9.5-67 of symptoms of white finger 83.5% was 8,400 hr of vibratory tool/use and 8,200 hr for numbness. Participants not blinded to purpose of questionnaire may have been over-reporting.

(Continued)

5c-20

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments McKenna Cross- 46 pairs of riveters and Outcome: Defined as cold- 35% 2% 24 3.1-510 Participation rate: Not reported. et al. 1993 sectional matched control subjects induced digital vasospasm. (machine operators who Matched on age and smoking had never used vibrating Exposure: To specific tools habits. tools). assessed via questionnaire. Only males studied.

Excluded those with injury to neck, trunk, upper limbs.

44% of riveters had <2.5 years of vibration exposure.

Did not of blind examiners because they tested the most symptomatic finger.

No differences in resting finger systolic pressure, vibration perception, or finger temperature between cases and controls.

17% of riveters reported symptoms of VWF.

(Continued)

5c-21

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Mirbod et Cross- Forestry workers Outcome: HAVs assessed 9.6% overall Õ Õ Õ Participation rate: Not reported. al. 1992b sectional (n=447) by interview and physical examination. Symptoms 20.9% among HAVs symptom severity No control group used. graded using the workers with positively correlated with Stockholm scale. 30 or more exposure duration. years Frequency-weighted experience Chain saw vibration levels vibration-acceleration ranged from 2.7 to 5.1 m/s 2. measurements made on the 2.5% among Low prevalence attributed to hands of chain workers <14 recent improvements in saw operators during years working conditions. different job processes. 11.7% 20 to 24 years

(Continued)

5c-22

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Mirbod et Cross- (A) 164 male dental Outcome: HAVs assessed (See first Males: H vs. 2.1-6.8 Participation restricted to al. 1994; sectional technicians, (B) 54 male by questionnaire, interviews, column for job 2.7% unexposed workers age 30 to 59 years. Mirbod et orthopedists, (C) 256 field visits, or annual health categories) Females: Males: 3.77 Subjects stratified by age in al. 1992a male aircraft technicians, examinations. 3.4% analysis. (D) 79 male laborers, A: 4.8% (E) 27 male grinders, Exposure: To vibrating B: 3.7% Hand-transmitted vibration (F) 46 female sewing- tools assessed by C: 2.3% levels in groups A to G ranged machine operators, questionnaire and D: 2.5% from 1.1 to 2.5 m/s 2. Hand- (G) 23 male tea- interviews. Hand- E: 3.7% transmitted vibration levels in harvesting-machine transmitted vibration F: 4.3% group H ranged from 2.7 to 5.1 operators, (H) 272 male measured among a sample G: 0.0% m/s 2. chain-saw operators; of workers using H: 9.6% compared with 1,027 representative tools in actual males and 1,301 females work activities. not exposed to vibration.

(Continued)

5c-23

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Miyashita Cross- 355 Male construction Outcome: HAVs assessed 1.1% 2.3% 0.5 0.1-11.8 Participation rate: Not reported. et al. 1992 sectional workers (machine by self-administered operators) compared questionnaire. Participation restricted to male with 44 male office workers age 30 to 49. workers. Exposure: Status assumed (A) 184 power shovel from job title (no objective Vibration due to construction- operators. measurements performed). machinery operation. (B) 127 bulldozer operators. (C) 44 forklift operators.

(Continued)

5c-24

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Musson et Cross- Impact power-tool users Outcome: HAVs based on 17% Õ Õ Õ Participation rate: 38% al. 1989 sectional in The Netherlands symptoms, assessed via questionnaire. (n=169). postal questionnaire. Adjusted for age. No control group used. Exposure: Vibration intensity measured using five Exposure duration not related representative tools. to HAV symptoms. Duration of vibration exposure assessed via questionnaire.

(Continued)

5c-25

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Nagata et Cross- 179 chain-saw workers Outcome: HAVs assessed >20-years 2.9% 7.1 for >20- 2.5-19.9 Participation rate: Not reported. al. 1993 sectional and 205 local inhabitants by dermatological tests and exposure: years who had never used physical examination. 16% vibration Adjusted for age. vibrating tools (control exposure group). Exposure: Vibration not < 20-years Examiners not blinded to measured directly; exposure exposure: exposure status. duration expressed as years 2.4% since commencement of occupation.

(Continued)

5c-26

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments Nilsson Cross- Platers (n=89) and office Outcome: Assessed by Platers with Office Participation rate: 79% among et al.1989 sectional workers (n=61) divided physical examination current workers platers, not reported among into 4 groups according and interview. VWF exposure: with no control. to current and past symptoms staged using 42% exposure: vibration exposure. the Taylor-Pelmear scale. 2% 85 15- 486 Controlled for age.

Exposure: Vibration Platers with Office Vibration acceleration levels exposure assessed by current and workers =5.5 m/s 2 (grinders), 10.3 m/s 2 measuring the former with no (hammers), 1.5 m/s 2 (die acceleration intensity on exposure. vibration grinders). a sample of tools, exposure subjective ratings, and and former Mean latency to symptom onset objective measures of exposure. 14 5-38 = 9.8 years. exposure time. Platers and Office Odds ratio increased by 11% office workers for each year of exposure. No workers with with no correlation between the Taylor- current or vibration Pelmear stage and years of former experience. 56 12-269 exposure. exposure.

(Continued)

5c-27

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Saito 1987 Cohort: Chain sawers without Outcome: Assessed by 0% in 1983 0% in 1978 Õ Õ Participation: Follow-up of 6-year HAV symptoms in 1978 symptoms, skin temperature, cohort. follow-up (n=155) followed up in vibration threshold, nail prospect- 1983. compression, pain sense, Improvements in chain saw ive and cold provocation. design, age restrictions, and a decrease in weekly operating Exposure: Chain saw time credited for preventing operating time determined by HAV. questionnaire. Recovery rates of skin temperature after 10-min provocation test significantly better in 1982 and 1983 compared to 1978.

Vibratory sense thresholds at 5th minute after cold provocation significantly better in 1980, 1982, and 1983 compared with 1978.

Age significance correlated to recovery rates from 1978 to 1983.

(Continued)

5c-28

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Shinev et Cross- 77 male fettlers; 59 male Outcome: HAV assessed by 22.1% Õ Õ Õ Participation rate: Not reported. al. 1992 sectional molders; 85 male neurological examination. (fettlers) polishers. 6.8% Percussive vibration had Exposure: Vibration (molders) greater effect on muscle and No control group used. characteristics of chipping 25% bone pathology than constant and caulking hammers, air (polishers) high-frequency vibration. tampers, and polishing machines measured.

(Continued)

5c-29

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Starck et Cross- Forest workers (n=200), Outcome: HAV based on 40% (forest Õ Õ Õ Participation rate: Not reported. al. 1990 sectional pedestal grinders (n=12), symptoms, assessed via workers using shipyard workers questionnaire. 1st generation No demographic data about (n=171), stone workers chain saw) study participants provided. (n=16), and platers (n=5). Exposure: Vibration measurements taken on a 16% (forest Poor correlation between No control group used. sample of tools during normal workers using vibration exposure and HAV operation at the workplace. 2nd when tools were highly generation impulsive. chain saw)

<7% (forest workers using 3rd generation chain saw)

100% (for pedestal grinders with zirconium wheels)

5% (shipyard workers)

75% (stone workers using pneumatic hammers)

50% (stone workers using chisel heads)

40% (platers)

(Continued)

5c-30

Table 5c–2. Epidemiologic studies evaluating work-related hand-arm vibration syndrome

MSD prevalence Study Exposed Referent RR, OR, Study design Study population Outcome and exposure workers group or PRR 95% CI Comments

Virokannas Cross- Railway workers (n=31) Outcome: “History of attack” Railway Õ Õ Õ Participation rate: Not reported. and sectional and lumberjacks (n=32) of white finger reported by workers: 45% Tolonen exposed to HAV. No subjects. VWF Total exposure to HAV had 1995 controls used. Article significant correlation with VPT in railway workers (r=0.55- evaluates the vibration VPT and electroneuro- Lumberjacks: 0.47; p=0.017) and lumberjacks perception threshold myography used as 38% VWF (r=0.77-0.59; p=0.003). (VPT) among exposed indicators of sensory nerve workers and tries to damage (outcome measure). Increase in VPT approximately determine a dose- 2 times greater in railway response relationship Exposure: To vibrating tools workers. between exposure to assessed by interview. (No HAV and the VPTs. measurements performed). 7 workers excluded—2 railway workers with Groups asked about polyneuropathy; 4 railway exposure time with self- workers with CTS; estimated annual use of 1 lumberjack with CTS. These vibrating tools and vehicles may have been related to (hr) and number of years of vibration exposure. exposure to vibration. Mean (SD) duration of exposure to Lumberjacks used chain saws vibration was 8,050 (3,500) daily >1,000 hr per year. Railway workers used hand- among railway workers and held tamping machines -500 21,250 (10, 950) hrs among hrs per year. lumberjacks. Found peak value differences for hand-held tamping machines (40 to 60 Hz) and chain saws (120 to 150 Hz). Nerve-conduction measurements adjusted for skin temperature.

5c-31 CHAPTER 6 Low-Back Musculoskeletal Disorders: Evidence for Work-Relatedness

SUMMARY Over 40 recent articles provided evidence regarding the relationship between low-back disorder and the five physical workplace factors that were considered in this review. These included (1) heavy physical work, (2) lifting and forceful movements, (3) bending and twisting (awkward postures), (4) whole-body vibration (WBV), and (5) static work postures. Many of the studies addressed multiple work-related factors. All articles that addressed a particular workplace factor contributed to the information used to draw conclusions about that risk factor, regardless of whether results were positive or negative.

The review provided evidence for a positive relationship between back disorder and heavy physical work, although risk estimates were more moderate than for lifting/forceful movements, awkward postures, and WBV. This was perhaps due to subjective and imprecise characterization of exposures. Evidence for dose- response was equivocal for this risk factor.

There is strong evidence that low-back disorders are associated with work-related lifting and forceful movements. Of 18 epidemiologic studies that were reviewed, 13 were consistent in demonstrating positive relationships. Those using subjective measures of exposure showed a range of risk estimates from 1.2 to 5.2, and those using more objective assessments had odds ratios (ORs) ranging from 2.2 to 11. Studies using objective measures to examine specific lifting activities generally demonstrated risk estimates above three and found dose-response relationships between exposures and outcomes. For the most part, higher ORs were observed in high-exposure populations (e.g., one high-risk group averaged 226 lifts per hour with a mean load weight of 88 newtons [N]) . Most of the investigations reviewed for this document adjusted for potential covariates in analyses; nevertheless, some of the relatively high ORs that were observed were unlikely to be caused by confounding or other effects of lifestyle covariates. Several studies suggested that both lifting and awkward postures were important contributors to the risk of low-back disorder. The observed relationships are consistent with biomechanical and other laboratory evidence regarding the effects of lifting and dynamic motion on back tissues.

The review provided evidence that work-related awkward postures are associated with low-back disorders. Results were consistent in showing positive associations, with several risk estimates above three. Exposure-response relationships were demonstrated. Many of the studies adjusted for potential covariates and a few examined the simultaneous effects of other work-related physical factors. Again, it appeared that lifting and awkward postures both contribute to risk of low-back disorder.

There is strong evidence of an association between exposure to WBV and low-back disorder. Of 19 studies reviewed for this document, 15 studies were consistent in demonstrating positive associations, with risk estimates ranging from 1.2 to 5.7 for those using subjective exposure measures, and from 1.4 to 39.5 for those using objective assessment methods. Most of the studies that examined relationships in high- exposure groups using detailed quantitative exposure measures found strong positive associations and exposure-response relationships between WBV and low back disorders. These relationships were observed after adjusting for covariates.

Both experimental and epidemiologic evidence suggest that WBV may act in combination with other work-related factors, such as prolonged sitting, lifting, and awkward postures, to cause

6-1 increased risk of back disorder. It is possible that effects of WBV may depend on the source of exposure (type of vehicle).

With regard to static work postures and low-back disorder, results from the studies that were reviewed provided insufficient evidence that a relationship exists. Few investigations examined effects of static work postures, and exposure characterizations were limited.

INTRODUCTION The relationship of the disorder with Low-back pain (LBP) is common in the employment can be complex: individuals may general population: lifetime prevalence has been experience impairment or disability at work estimated at nearly 70% for industrialized because of back disorders whether the latter countries; sciatic conditions may occur in one was directly caused by job-related factors or quarter of those experiencing back problems not. The degree to which ability to work is [Andersson 1981]. Studies of workers’ impaired is often dependent on the physical compensation data have suggested that LBP demands of the job. Furthermore, when an represents a significant portion of morbidity in individual experiences a back disorder at work, working populations: data from a national it may be a new occurrence or an exacerbation insurer indicate that back claims account for of an existing condition. Again, originally it may 16% of all workers’ compensation claims and have been directly caused by work or by 33% of total claims costs [Snook 1982; nonwork-related factors. Those suffering back Webster and Snook 1994b]. Studies have pain may modify their work activities in an demonstrated that back disorder rates vary effort to prevent or lessen pain. Thus, the substantially by industry, occupation, and by relationship between work exposure and job within given industries or facilities [see disorder may be direct in some cases, but not in Bigos et al. 1986a; Riihimäki et al. 1989a; others. Schibye et al. 1995; Skovron et al. 1994]. When discussing causal factors for low-back Back disorder is multifactorial in origin and may disorders, it is important to distinguish among be associated with both occupational and the various outcome measures, such as LBP, nonwork-related factors and characteristics. impairment, and disability. LBP can be defined The latter may include age, gender, cigarette as chronic or acute pain of the lumbosacral, smoking status, physical fitness level, buttock, or upper leg region. Sciatic pain refers anthropometric measures, lumbar mobility, to pain symptoms that radiate from the back strength, medical history, and structural region down one or both legs; lumbago refers abnormalities [Garg and Moore 1992]. to an acute episode of LBP. In many cases of Psychosocial factors, both work- and LBP, specific clinical signs are absent. Low- nonwork-related, have been associated with back impairment is generally regarded as a loss back disorders. These relationships are of ability to perform physical activities. Low- discussed at length in Chapter 7 and Appendix back disability is defined as necessitating B. restricted duty or time away from the job. Although it is not clear which outcome measure

6-2 is best suited for determining the causal It is important to include subjectively defined relationship between low-back disorder and health outcomes in any consideration of work- work-related risk factors, it is important to related back disorders because they comprise consider severity when evaluating the literature. such a large subset of the total. It may be too restrictive to define cases of back disorder In addition to level of severity, outcomes may using “objective” medical criteria. Therefore, in be defined in a number of other ways, ranging contrast to chapters for musculoskeletal from subjective to objective. Information on disorders or other anatomic regions, this review symptoms can be collected by interview or of literature on the back used slightly different questionnaire self-report. Back “incidents” or evaluation criteria. For consideration of back “reports” include conditions reported to disorders, use of a subjective health outcome medical authorities or on injury/illness logs; was not necessarily considered a study these may be symptoms or signs that an limitation. Furthermore, because back individual has determined need for medical or disorders were rarely defined by medical other attention. They may be due to acute examination criteria, the evaluation criterion symptoms, chronic pain, or injury related to a related to blinding of assessors (to health or particular incident, and may be subjectively or exposure status) was also less relevant to a objectively determined. Whether an incident is discussion of this literature. reported depends on the individual’s situation and inclinations. Other back disorders can be In this review, epidemiologic studies of all diagnosed using objective criteria—for forms of back disorder were included. The example, various types of lumbar disc term “back disorder” is used to encompass all pathology. health outcomes related to the back. It should be pointed out that, in some studies, disorders There are many conditions in the low back of the low back were not distinguished from which may cause back pain, including muscular total back disorders. We assumed that a or ligamentous strain, facet joint arthritis, or significant portion of these related to the low disc pressure on the annulus fibrosis, vertebral back, and articles using such a definition were end-plate, or nerve roots. In most patients, the included in our review. anatomical cause of LBP, regardless of its relationship to work exposures, cannot be The 42 epidemiologic studies discussed below determined with any degree of clinical certainty. were selected according to criteria that appear Muscle strain is probably the most common in the introduction of this document. Most (30) type of work or nonwork back pain. While used a cross-sectional design, followed by there is sometimes a relationship between pain prospective cohort (5), case-control (4), and and findings on magnetic resonance imaging retrospective cohort (2) designs. One study (MRI) of disc abnormalities (such as a combined both cross-sectional and cohort herniated disc and clinical findings of nerve analyses. Full descriptions of the studies appear compression), unfortunately, the most common in Table 6-6. Twenty-four investigations form of back disorder is “non-specific defined the health outcome only by report of symptoms,” which often cannot be diagnosed. symptoms on questionnaires or in interviews

6-3 (for example, total back pain, LBP, and are discussions of the evidence for each work- sciatica); used symptoms plus medical related physical risk factor. examination (back pain, low-back syndrome, sciatica, back insufficiency, lumbago, herniated HEAVY PHYSICAL WORK lumbar disc, and lumbar disc pathology), Definition 2 used sick leaves and medical disability retirements, and 6 used injury/illness reports. Heavy physical work has been defined as work The last category included outcomes defined as that has high energy demands or requires some “low-back complaints, injuries caused measure of physical strength. Some specifically by lifting or mechanical energy,” and biomechanical studies interpret heavy work as “acute industrial back injury.” Clearly, the 42 jobs that impose large compressive forces on studies used outcome definitions that the spine [Marras et al. 1995]. In this review, correspond to several regions of the back and the definition for heavy physical work includes include disorders that may have been acute or these concepts, along with investigators’ chronic and subjectively or objectively perceptions of heavy physical workload, which determined. range from heavy tiring tasks, manual materials handling tasks, and heavy, dynamic, or intense In the studies included in this review, exposures work. In several studies, evaluation of this risk were assessed primarily by questionnaire or factor was subjective on the part of participant interview (n=17), followed by observation or or investigator, and in many cases, “heavy direct measurement (n=15) and by job title only physical work” appeared to include other (n=10). Study groups included general potential risk factors for back disorder, populations (Swedish, Dutch, U.S., Finnish, particularly lifting and awkward postures. and English) and occupational groups (nurses, clerical employees, school lunch preparers, Studies Reporting on the Association baggage handlers, and individuals working in Between LBP and Heavy Physical Work construction, agriculture, maritime, petroleum, Eighteen studies appeared to address the risk paper products, transportation, automobile, factor related to heavy physical work, although aircraft, steel, and machine manufacturing none of them fulfilled all four evaluation criteria industries). (Table 6-1, Figure 6-1). In fact, most (78%) had acceptable participation rates, but only This review of epidemiologic studies of low- three defined health outcomes using both back disorder examined the following potential symptoms and medical exam criteria, and only risk factors related to physical aspects of the two assessed exposure independent of self- workplace: (1) heavy physical work, (2) lifting report. and forceful movements, (3) bending and twisting (awkward postures), (4) WBV, and In nearly all of these studies, covariates were (5) static work postures. Psychosocial addressed in at least minimal fashion, such as workplace factors were also included in a restricting the study population as to number of studies; these relationships are gender and conducting age-stratified or discussed separately in Chapter 7. Following

6-4 adjusted analyses; in many, multivariate symptoms. analyses were carried out. With regard to health outcome, while only three used medical Burdorf and Zondervan [1990] carried out a exams, in addition to symptoms or injury cross-sectional study comparing 33 male reports, to arrive at case definitions, in many workers who operated cranes with age- instances standard questionnaire instruments matched workers from the same Dutch steel were used. The major study limitations, overall, plant who did not operate cranes. Symptoms of were related to relatively poor ascertainment of LBP and sciatica were assessed by exposure status. questionnaire. Exposure was assessed by job title (crane operators were noted to experience Following are descriptions of seven studies that frequent twisting, bending, stooping, static were most informative. Detailed descriptions sedentary postures, and WBV) and by for all 18 investigations can be found in Table questionnaire (exposures to sedentary postures, 6-6. WBV, heavy physical work, and frequent lifting were assessed for both current and past jobs). Bergenudd and Nilsson [1988] followed a Crane operators were significantly more likely Swedish population-based cohort established in to experience LBP (OR 3.6, 95% CI 1938. Back pain (total) presence and severity 1.2–10.6). Among crane operators alone, the were self-assessed by questionnaire, as of OR for heavy work was 4.0 (95% CI 1983; exposures (light, moderate, or heavy 0.76–21.2) after controlling for age, height, and physical work) were assessed based on weight. It was determined that this heavy work questionnaires completed by the cohort from occurred in past and not in current jobs. 1942 onward. Univariate results demonstrated Among crane operators alone, the OR for that those with moderate or heavy physical frequent lifting was 5.2 (95% CI 1.1–25.5). demands in their jobs had more back pain than The frequent lifting in crane operators was also those with light physical demands (OR 1.83, determined to be from jobs held in the past. 95% Confidence Interval [CI] 1.2-2.7). When Among workers who were not crane stratified by gender, the relationship was slightly operators, history of frequent lifting was not stronger for females (OR 2.03, 95% CI associated with LBP (OR 0.70, 95% CI 1.1–3.7) than for males (OR 1.76, 95% CI 0.14–3.5). Among crane operators, univariate 1.01–3.1). When prevalence was examined by ORs for WBV and prolonged sedentary exposure category, rates were 21.4%, 32.8%, postures were 0.66 (95% CI 0.14–3.1) and and 31.3% for males (no trend was available 0.49 (95% CI 0.11–2.2), respectively. In for females, as none worked in the highest multivariate analyses controlled for age, height, exposure category). Analyses were stratified by weight, and current crane work, most of the gender but did not account for other potential associations with specific work-related factors covariates. The longitudinal design ensured that were substantially reduced. The high exposures preceded health outcomes. prevalence of LBP in crane operators was Shortcomings included a relatively low explained only by current crane work. No response rate (67%), minimal exposure measures of dose-response were examined. assessment, limited adjustment for covariates in analyses, and self-reporting of health

6-5 Limitations included a relatively low response drivers, welders, smiths, and operators of rate for crane operators (67%)—with some several types of machines (lathes, punch suggestion that those with illness may have been presses, and milling). Outcome information was under-represented (perhaps underestimating obtained by questionnaire. Exposure data were the OR)—and self-reporting of health also obtained by questionnaire and included outcomes and exposures. The investigators information on occupational, psychosocial, and attempted to clarify the temporal relation physical workloads, including sitting, carrying, between exposure and outcome by excluding pushing, pulling, lifting, work postures, and cases of back pain with onset before the repetitive movements. Questionnaire items present job. related to carrying, pushing, pulling, and lifting were combined to produce an index of manual As part of a Finnish population-based health materials handling. The prevalence of work- survey, Heliövaara et al. [1991] conducted a related LBP was significantly higher in blue- cross-sectional analysis of chronic low-back collar employees than in white-collar workers syndrome, sciatica, and LBP. Health outcomes (RR 1.8, p<0.05). In both white and blue- were determined by interview and examination; collar workers, work-related LBP was not work-related exposure information was significantly associated with either heavy or light obtained by a self-administered questionnaire, materials handling, or bent or twisted work which included items related to lifting, carrying postures, after adjustment for age and gender. heavy objects, awkward postures, WBV, LBP was significantly associated with extreme repeated movements, and paced work. The work postures (blue-collar workers only) and total number of factors was designated the monotonous working movements (white-collar “sum index of occupational physical stress.” workers only). In these analyses, relationships Mental work stress measures were also were presented as partial correlations; thus, a included. A dose-response was observed for comparison of risk estimates was not possible. sciatica and the physical stress score (with an Limitations of the study included the cross- OR of 1.9, 95% CI 0.8–4.8 for the highest sectional design, collection of outcome and score) and for low-back syndrome and exposure data by self-report, and potential physical stress (OR 2.5, 95% CI 1.4–4.7), problems with multiple comparisons, as many after adjusting for a number of covariates. The independent variables were examined in study did not address temporal relationships, analyses. Many of the exposed group (blue- and exposure information was derived from collar workers) were engaged in machine self-reports. Strengths included a high response operation tasks with perhaps limited rate, objective measure of health outcomes, and opportunity for exposure to work with heavy multivariate adjustment for covariates. physical demands. Also, heavy physical work and lifting were combined into a single index. Johansson and Rubenowitz [1994] examined Strengths included consideration of age and low-back symptoms cross sectionally in 450 gender as covariates and inclusion of both blue- and white-collar workers employed in physical and psychosocial workplace eight Swedish metal companies. The exposed measures. group included assemblers, truck

6-6 Svensson and Andersson [1989] examined 83% for light, intermediate, and heavy LBP in a population-based cross-sectional categories). The trend was not observed in study of employed Swedish women. older age groups, nor for sciatica in any age Information on LBP and sciatica was obtained group. LBP and sciatica rates were slightly by questionnaire, as were exposure-related higher for nurse aides than for qualified nurses, items. Physical exposures included lifting, although the differences were not statistically bending, twisting, other work postures, sitting, significant. The authors suggested that aides standing, monotony, and physical activity at had higher rates of back pain because of work. Lifetime incidence rates (IRs) varied by heavier workload, including patient handling occupation, with ranges from 61%–83% in and lifting. Lack of consistency of LBP OR younger age groups and 53%–75% in older across exposure and age groups suggested that groups. A posteriori, the authors noted that, a healthy worker effect was operating and that for these women, the highest lifetime incidence injured workers might be leaving the field, a of LBP was not found in the jobs with the phenomenon that the cross-sectional study highest physical demands. The measure for design could not address. “physical activity at work” was also not significantly associated with LBP in univariate Videman et al. [1990] carried out a cross- analyses. Bending forward (RR 1.3), lifting (RR sectional study of 86 males who died in a 1.2), and standing (RR 1.3) were associated Helsinki hospital to determine degree of lumbar with lifetime incidence of LBP in univariate spinal pathology. Disc degeneration and other analyses (p<0.05). None of the measures of pathologies were assessed in the cadaver physical workplace factors were associated specimens by discography and radiography. with lifetime incidence of LBP in multivariate Subjects’ symptoms and work analyses. exposures—heavy physical work, sedentary work, driving, and mixed—were determined by A cross-sectional study of LBP in Finnish interview of family members. In comparison to nurses was conducted [Videman et al. 1984]. those with mixed work exposures, those with LBP and sciatica were ascertained by sedentary and heavy work had increased risk questionnaire; exposure information was also of symmetric disc degeneration with ORs of self-reported and included items related to both 24.6 (95% CI 1.5–409) and 2.8 (95% CI physical loading factors at work and to work 0.3–23.7), respectively). Similar relationships history. Exposures were reclassified as were seen for vertebral end-plate defects and “heavy,” “intermediate,” and “light,” based on facet joint osteoarthrosis. Risk of vertebral questionnaire responses. The derivation of this osteophytosis was highest for those in the heavy classification was not clear, but it may have work category (OR 12.1, 95% CI 1.4–107). been a combination of responses to questions For most pathologic changes, sedentary work on lifting, bending, rotation, standing, walking, appeared to have a stronger relationship than and sitting. A dose-response was observed heavy work. Back pain symptoms were between prevalence of previous LBP and consistently higher in those with any form of workload category in younger women (77%, spinal pathology, although the difference was 79%, and significant only for anular ruptures. Results of

6-7 this study were notable in that anular rupture, a overall physical workload in a group of Belgian classic pathologic condition of the disc, was not steelworkers, although LBP was related to associated with exposure. This study was heavy shoulder efforts. In a study of blue-and unusual in design in that it examined a white-collar workers, Johansson and combination of spinal pathological outcomes, Rubenowitz [1994] found higher LBP rates in symptoms, and workplace factors. However, blue-collar workers (RR 1.8, p<0.05). participation in the study was dependent on However, in more detailed analyses of obtaining information from family members; exposure, back pain was not associated with participation rates were not stated. While recall indices for heavy or light materials handling bias is often a problem in studies of the after adjustment for age and gender (with deceased, in this case, it should have been partial correlation coefficients of less than nondifferential, if present. 0.10). Burdorf and Zondervan’s 1990 study of crane operators demonstrated increased risk of Strength of Association LBP with exposure to heavy work (OR 4.0, The most informative studies were generally 95% CI 0.8–21.2) after controlling for age, those that carried out exposure assessments height, and weight. Two studies used indices of which ranked physical workload based on physical stress to create questionnaire questionnaire report. In a prospective study of responses related to lifting, carrying heavy back injury reports, Bigos et al. [1991b] found objects, awkward postures, repeated no associations with physical job characteristics movements, and others. Heliövaara et al. (although the authors stated that the study [1991] found that both low-back syndrome population had low overall exposures). This and sciatica were associated with physical study described the biomechanical methods that stress scores, with ORs of 2.5 (p<0.05) and were used to directly assess spinal loads 1.9 (not significant) for the highest scores, associated with jobs, but no results related to respectively. A study of Finnish nurses these measures were presented. Svensson and classified exposures as “heavy,” “intermediate,” Andersson [1989] appear to have examined a and “light” based on questionnaire response measure for physical activity at work and its scores [Videman et al. 1984]; prevalence of relationship to LBP in Swedish women. No LBP was slightly higher in the heavy category associations were observed. In a population- than in the light (RR 1.1, not significant) for based study, Bergenudd and Nilsson [1988] younger women only. Sciatica was also observed significantly more back pain in those examined, and no relationships were found. with heavier physical work (OR 1.8 for moderate/heavy versus light work, p<0.01). The other studies that examined heavy physical ORs were slightly higher for females (OR 2.0) work as a risk factor for back disorder than for males (OR 1.8). Leigh and Sheetz classified exposure in a simpler manner, either [1989] found that back symptoms were by job title alone or by grouping jobs based on associated with self-reporting that “job requires prior knowledge of the work or questionnaire a lot of physical effort” (OR 1.5, 95% CI responses. Burdorf et al. [1991] found that 1.0–2.2). Masset and Malchaire [1994] observed that LBP was not associated with

6-8 heavy physical work was associated with back impact injuries, with RRs of 2.2 and 4.3 (no pain in concrete workers in univariate, but not significance testing was done) in comparison to multivariate models (no risk estimate was control room operators and maintenance reported). Hildebrandt [1995] found that professionals, those with the lowest rates. A individuals in jobs described as “heavy non- study of hospital employees that matched cases sedentary” were more likely to experience with controls by department found that those on back pain than those in sedentary jobs (OR the day shift had an OR of 2.2 (p<0.005) in 1.2, p<0.05). In a cadaver study of lumbar disc comparison to those working other shifts pathology, Videman et al. [1990] found that [Ryden et al. 1989]. In the last two studies, the those with jobs involving heavy physical work authors determined a posteriori that job titles had increased risk of disc pathology in (or shifts) that were observed to have high back comparison to those with mixed work disorder rates were those requiring the heaviest exposures (e.g., an OR of 2.8, 95% CI physical effort. 0.3–23.7, for symmetric disc degeneration and an OR of 12.1, 95% CI 1.4–107, for vertebral Although in all 18 of these studies the authors osteophytosis). For most pathologic changes, stated that “heavy physical effort or work” was sedentary work had a stronger relationship than at least one of the risk factors of interest, the heavy work. actual estimates of these exposures varied from assumptions based on job title to self-reported Finally, several studies examined back disorder scores based on self-reported work activities. rates by job title or occupation alone. In no case were measured physical loads used Hildebrandt et al. [1996] observed differences as independent variables. Study populations in back symptom rates by unit and task group included individuals working in health care, in “nonsedentary” steel workers. The reference office work, manufacturing, construction, and group also had high symptom rates; general populations, all with varying degrees of comparisons between the two groups did not physical work requirements. Some studies yield significant differences. In multivariate created physical “stress” indices that included analyses, Riihimäki et al. [1989b] found no more than one risk factor. Since most estimates significant difference in sciatic pain for of physical load were subjective, they tended to carpenters and office workers (OR 1.0, 95% reflect the relative requirements of the jobs and CI 0.8–1.3). Partridge and Duthie [1968] individuals included in each study. Health found that dock workers had slightly higher outcomes also varied. LBP rates than civil servants (RR 1.2, not significant). In a similar study, Åstrand [1987] In summary, the strength of the relationship classified pulp mill jobs as heavy and the between back disorder and heavy physical referent group of clerical jobs as light; mill work in some of the studies with more workers were 2.3 times more likely to quantitatively defined exposures ranged from experience back pain than clerical staff none [Bigos et al. 1991b; Johannsson and (p=0.002). Clemmer et al. [1991] found that Rubenowitz, 1994; Masset and Malchaire floor hands, roustabouts, and derrickhands had 1994; Svensson and Andersson 1989; the highest rates for low-back strains and Videman et al. 1984] to ORs of 1.9 (not

6-9 significant) for sciatica and 2.5 (p<0.05) for Temporal Relationship low-back syndrome [Heliövaara et al. 1991], Fourteen of the 18 reviewed studies had a 1.5 (95% CI 1.0–2.2) [Leigh and Sheetz cross-sectional design that could not directly 1989], 1.8 (95% CI 1.2–2.7) [Bergenudd and address this issue. Three mentioned potential Nilsson 1988], and 4.0 (p<0.05) for LBP problems related to this study design. Åstrand [Burdorf and Zondervan 1990]. In another [1987] suggested that exposure study, which used a scoring system and focused misclassification occurred in her study of paper on a subject group of nurses, the RR was 1.1 mill workers (some individuals were transferred (not significant) for the high-exposure category to clerical jobs—the unexposed group—after [Videman et al. 1984]. experiencing a back injury in the mill). In the Videman et al. 1984 study of nurses, lack of Dichotomous estimates of physical workload consistency of LBP OR by age and exposure yielded ORs of 1.2 [Hildebrandt 1995], 2.8- group suggested that injured workers were 12.1 [Videman et al. 1990], and no association leaving the field. A study of cadavers carried (results were observed in univariate but not out by Videman et al. [1990] seemed to have multivarate analyses, with no risk estimates potential for problems with temporal reported) [Burdorf et al. 1991]. Exposures relationships, as exposure information for past based on job title alone yielded estimates from periods depended on recall of study none [Hildebrandt et al. 1996], nonsignificant participants’ activities by family members. ORs of 1.0 and 1.2 [Partridge and Duthie 1968; Riihimäki et al. 1989b], to significant Two cross-sectional studies attempted to clarify ORs of 2.2–4.3 [Åstrand 1987; Clemmer et al. temporal relationships by excluding from 1991; Ryden et al. 1989]. Half of the studies analysis the cases with disorder onset prior to had positive point estimates for this risk factor current job [Burdorf et al. 1991; Burdorf and but were low to moderate in magnitude. In five Zondervan 1990]. Both showed results studies that found no association between back suggesting a positive relationship between disorder and heavy physical work, no details exposure and back disorders. Three studies were given. Two of the highest significant ORs had cohort designs in which temporal were based on exposed groups in the oil and relationships between outcome and exposure steel industries [Burdorf and Zondervan 1990; could be determined [Bergenudd and Nilsson Clemmer et al. 1991]. For these, true exposure 1988; Bigos et al. 1991b; Clemmer et al. to heavy physical work was probably more 1991]: in one, no association was observed, in likely than for some of the other study another, a modest increase in risk was seen. In populations. For many of the investigations, the third, exposure (assessed a posteriori by exposure estimates were subjectively assessed. job title) was significantly associated with back In many cases, study groups had potentially low injuries. A case-control study conducted using exposures or exposure to heavy physical work hospital personnel records appeared free from in combination with other risk factors. recall bias and showed a significant association between low-back injury and working the day shift (assessed a posteriori as having the heaviest workload) [Ryden et al. 1989].

6-10 Although the majority of studies were limited by materials handling are due to a combination of their cross-sectional designs, results were the weight lifted, and the person’s method of similar for these and other studies with designs handling the load. The internal reaction forces that could assess temporal relationships. needed to equilibrate the body segment weights and external forces such as weight of the load For most studies, the data are compatible with being lifted are supplied by muscle contraction, a temporal relationship in which exposure ligaments, and body joints. Injury to the preceded disorder. supporting tissues can occur when the forces from the load, body position, and movements Consistency in Association of the trunk create compressive, shear, or Half of the 18 studies examined demonstrated rotational forces that exceed the capacities of no significant association between exposure the discs and supporting tissues needed to and outcome. All of those which showed counteract the load moments. Rowe [1985] significant associations (n=9) were positive in hypothesized that disc and facet degeneration direction, (one OR of 1.2, two ORs between and ligament strain are responsible for the 1.5 and 2, and six ORs between 2.2 and 12.1). potentially high rates of LBP disability in those whose jobs demand heavy physical activity. Study groups included males working in industrial environments, office workers, health The Videman et al. [1990] cross-sectional care employees—female, for the most study of cadavers addressed two aspects of the part—and population-based groups that causal chain linking exposure to heavy physical included both genders and many occupations. work and back disorder. First, the study That some consistency in results was noted demonstrated an association between among these diverse groups, particularly after subjective health outcome measures and more adjustment for covariates, suggests that the objective measures: back pain symptoms observed associations have validity and can be (assessed from family members) were generalized across working populations. consistently higher in those with signs of spinal pathology. Second, the study demonstrated an Coherence of Evidence association between objective measures of Information derived from a large number of disorder and heavy work exposures: individuals laboratory and field studies using a wide variety whose jobs included heavy work exposures of approaches provides a plausible explanation showed increased risk of symmetric disc for associations between LBP and physically degeneration, vertebral osteophytosis, and facet demanding jobs [Waters et al. 1993]. Research joint osteoarthrosis. Significant relationships conducted in the 1950s demonstrated that disc were also found for back pain and disability. degeneration occurs earlier in life among We agree with the conclusion of Videman et al. workers who perform heavy physical work [1990] that states that “back injury and than among those who perform lighter work. Similar findings are reported in more recent investigations [Videman et al. 1990]. The stresses induced at the low back during manual

6-11 sedentary or heavy (but not mixed) work Only a few studies examined exposure in contributed to the development of pathologic sufficient detail to assess exposure-response findings in the spine. The severity of back pain relationships with low-back disorders. Results was related to the heaviness of work. Work- were mixed. Heliövaara et al. [1991] observed related factors may be responsible for the an exposure-response between sciatica and development of pathologic changes and for physical stress score; the Videman et al. [1984] increased episodes of LBP and disability.” results demonstrated a dose-response between LBP prevalence and workload categories in Another important contribution to the younger nurses, but not in older groups, or for coherence of evidence is that the Bureau of sciatica in any age group. In Åstrand’s 1987 Labor Statistics Annual Survey of Injuries and “high exposure group” (pulp mill workers), Illnesses has demonstrated significant elevations duration of employment was associated with in overexertion injuries and disorders in back pain. Bergenudd and Nilsson [1988] and industries which are associated with heavy Johansson and Rubenowitz [1994] observed work, such as nursing and personal care and air no exposure-response relationships between transportation. Some broad population surveys back disorders and their exposure measures. such as the National Health Interview Survey On the whole, evidence of exposure-response (NHIS) from 1988 and the 1990 Ontario is equivocal, based on the paucity of Health Survey (OHS) found increased back information available. pain or long-term back problems with exposure to factors such as lifting, pulling, and physical Conclusions: Heavy Physical Work pushing [Guo et al. 1995; Liira et al. 1996]. In The reviewed epidemiologic investigations the NHIS, the two occupations with the highest provided evidence that low-back disorders are significant rates of work-related LBP were associated with heavy physical work. Despite male construction laborers (with a prevalence the fact that studies defined disorders and ratio [PR] of 2.1) and female nursing aides, assessed exposures in many ways, all studies orderlies, and attendants (PR 2.8) [Guo et al. which demonstrated significant associations 1995]. In the OHS, the number of simultaneous between exposure and outcome were positive physical exposures was directly related to risk in direction and showed low to moderate increase after adjustment for covariates. For increased risk. Exposures were assessed the highest exposure index level, the adjusted subjectively, for the most part; and in some OR was 3.18 (95% CI 1.72–5.8), which cases, classification schemes were crude. This occurred in 3% of the population [Liira et al. study limitation may have led to 1996]. It is important to point out that truly misclassification of exposure status to the extent heavy work probably occurs in only a tiny that it caused a dampening effect on risk proportion of all jobs in most industries and in estimates, where nondifferential only a minority of many high-risk industries, misclassification caused bias toward a null value which is why misclassification of exposures is for the measure of association. This may likely in population-based studies. account for the moderate ORs that were

Exposure-Response Relationships

6-12 observed. A few studies were able to examine Between LBP and Lifting and dose-response relationships between outcomes Forceful Movements and exposure; these results were equivocal. Eighteen studies examined relationships Most studies utilized cross-sectional study between back disorders and lifting or forceful designs; however, five of six studies which used movements. Only one, Punnett et al. 1991 specific methodologies to address temporality case-control study of back pain in auto showed positive associations between exposure workers, fulfilled the four evaluation criteria and outcome. Many studies addressed potential (Table 6-2, Figure 6-2). The majority (66%) effects of covariates by restriction in selection had adequate participation rates; four defined of study participants, stratification, or outcomes using both symptoms and medical multivariate adjustment in statistical analyses. exam criteria. Blinding of investigators with regard to case/exposure status was not In many studies, “heavy physical work” mentioned in most, but it could be confirmed in exposure appeared to include other work- two papers and inferred (by study methodology related physical factors (particularly lifting and ) in two others. Seven studies used an exposure awkward postures). assessment that included observation or direct measurement; an additional nine obtained LIFTING AND FORCEFUL exposure information by self-report on MOVEMENTS questionnaire or interview. Only two relied on Definition job title alone to characterize exposure. Lifting is defined as moving or bringing something from a lower level to a higher one. Thirteen investigations were cross-sectional in The concept encompasses stresses resulting design; three were case-control, and two were from work done in transferring objects from prospective. Eleven defined the health outcome one plane to another as well as the effects of by symptom report on interview or varying techniques of patient handling and questionnaire. transfer. Forceful movements include movement of objects in other ways, such as Descriptions of seven studies which provided pulling, pushing, or other efforts. Several the most information regarding the relationship studies included in this review used indices of between low-back disorder and lifting and physical workload that combined lifting/forceful forceful movements follow. Detailed movements with other work-related risk factors descriptions for all 18 investigations can be (particularly heavy physical work and awkward found in Table 6-6. postures). Some studies had definitions for lifting which include criteria for number of lifts The Punnett et al. [1991] case-control study per day or average amount of weight lifted. examined the relationship between back pain and occupational exposures in auto assembly workers. Back pain cases (n=95) were

Studies Reporting on the Association

6-13 determined by symptoms at interview and group of maintenance workers. Back pain medical examination; controls included those symptoms were assessed by questionnaire. free of back pain. For all participants (or Exposures were measured using the Ovako proxies in the same jobs), jobs were Working Posture Analysis System, which videotaped and work cycles were reviewed assessed postures for the back and lower limbs using a posture analysis system. Exposures along with lifting load. Information on included time spent in various awkward exposures in previous jobs was also collected. postures. Peak biomechanical forces were Concrete workers experienced significantly estimated for up to nine postures where a load more back symptoms than referents (OR 2.8, weighing at least 10 lb was held in the hands. In 95% CI 1.3–6.0). Univariate results showed multivariate analyses that adjusted for a number associations between back pain and both of covariates (age, gender, length of posture index and WBV in current job employment, recreational activities, and medical (correlations were presented). Lifting was not history), time in non-neutral postures (mild or found to be associated with back pain (and severe flexion and bending) was strongly exposure was found not to vary significantly associated with back disorder (OR 8.09, 95% across the six job categories examined in the CI 1.4–44). Lifting was also associated with study). In multivariate analyses adjusting for age, both posture index and WBV were back disorder (OR 2.16, 95% CI 1.0–4.7). significantly associated with back pain, with When the subset with physical medical findings ORs of 1.23 (p=0.04) (for an ordinal scale of was examined, associations were more 6) and 3.1 (p=0.01) (dichotomous), pronounced. Although few study subjects were respectively. These two measures were highly unexposed to all of the postures studied, a correlated and analyzed separately. Strengths strong increase in risk was observed with both of the study include use of a standard symptom intensity and duration of exposure. It was not questionnaire, high participation rates, an possible to determine the relative contributions objective measure of exposure, and an attempt of different awkward postures because all were to clarify the temporal relation between highly correlated. Only participants’ current exposure and outcome by excluding cases of jobs (for referents), or job when symptoms back pain with onset before the present job. started (for cases) were analyzed; the study design thus assumed a short-term relationship Chaffin and Park [1973] carried out a between exposure and outcome (although prospective study of back complaints in 411 length of time in job was also included in the employees of four electronics manufacturing models). The authors attempted to ensure that plants. The outcome included visits to the plant exposure preceded disease by identifying time medical department because of back of onset and measuring exposures in the job complaints over a one-year period. Exposure held just prior. The strong associations, after was assessed by evaluating 103 jobs with a adjustment for covariates, are notable. range of manual lifting for lifting strength rating (LSR) and load weights. The LSR is a ratio of Burdorf et al. [1991] examined back pain the maximum weight lifted on the job to the symptoms in a cross-sectional study of male lifting strength, in the same load position, for a concrete fabrication workers and a referent large/strong man. Results

6-14 showed a strong increase in back complaint 1.3–7.5). The highest risk was observed for incidence with LSR for both males and females simultaneous lifting and twisting with straight (with an approximate five-fold increase in risk knees (OR 6.1, 95% CI 1.3–27.9). Despite comparing males in the highest and lowest the fact that exposures were self-reported, LSR). A similar increase was observed for these associations were notably strong. The females, although there were no women in the potential existed for differential recall bias for highest exposure category. No dose-response cases and controls because study subjects were was observed by frequency of lifts (a relatively interviewed about work-related factors after high risk of back complaints was observed for case status was established. Interviewers may the lowest exposure category). Covariates not have been blinded to case/control status. (age, weight, and stature) were examined and found not to contribute to back complaints. The In Liles et al. [1984] prospective study of 453 prospective study design helped increase the individuals working in jobs with manual material likelihood that exposure preceded disorder. handling requirements, incidence of back Study limitations include lack of information on injuries was examined with regard to lifting. The participation rates and an outcome consisting of study group included those who lifted frequently incident reports. Time of true onset was not (at least 25 lifts per day of not less than 4.53 ascertained, and it is possible that symptom kg, with exposure of at least two hours per onset preceded or coincided with exposure day). The outcome included reported or assessment despite the longitudinal study recorded lifting injuries to the back. Lifting design. The detailed exposure assessment exposures were assessed until job change (up addressed only lifting as a risk factor; presence to a two-year period) using the Job Severity of other risk factors related to back disorders Index (JSI). The JSI is a measure of the was not identified. physical stress level associated with lifting jobs and is a function of the ratio of job demands to A case-control study of prolapsed lumbar disc the lifting capacities of the person performing was carried out using a hospital population- the job. Information on weight, frequency of based design [Kelsey et al. 1984]. Cases lifting, and task geometry is collected through (n=232) included individuals diagnosed with comprehensive task analysis. When the study prolapsed lumbar disc; an equal number of group (working in 101 jobs from 28 plants) controls matched on sex, age, and medical was classified into 10 equal categories service were selected. Exposure was assessed according to JSI, a dose-response relationship using a detailed occupational history that was with injury was observed (RR 4.5, 95% CI not described but presumably was obtained by 1.02–19.9 for total injuries, comparing interview. An association with work-related category 10 to category 1). Study limitations lifting without twisting the body was observed included no statement relating to response rate at the highest lifting level (25 lb or more) or participant selection, no adjustment for (OR 3.8, 95% CI 0.7–20.1). Twisting without confounders, and no statistical testing. The lifting was associated with disc prolapse (OR outcome definition specified that the back injury 3.0, 95% CI 0.9–10.2); a combination of both be lifting- risk factors had an OR of 3.1 (95% CI

6-15 related, which increased the likelihood that the for measurement of lifting motions. However, outcome would be related to the exposure the unit of analysis was job, and each was measured. The prospective designassured that characterized by measurement of at least one measured exposures preceded injury onset. study subject. Effects of covariates were not Other strengths included objective assessment addressed (multivariate analyses appeared to of exposure. include only biomechanical variables). The study results emphasized the multifactorial Using an unusual cross-sectional study design, etiology of back disorders, including Marras et al. [1993, 1995] examined the contributions of lifting frequency, loads, and relationship between low-back disorders and trunk motions and postures. The study design spinal loading during occupational lifting. A total did not allow for examination of temporal of 403 jobs from 48 diverse manufacturing relationships. companies were assessed for risk of low-back disorder using plant medical department injury Walsh et al. [1989] examined the relationship reports. Jobs were ranked into three categories between self-reported LBP and work-related according to risk, then assessed for position, factors in a population-based cross-sectional velocity, and acceleration of the lumbar spine study of 436 English residents. LBP was during lifting motions in manual materials ascertained by interview, as was lifetime handling using electrogoniometric techniques. occupational history (including exposures to Those in high-risk jobs averaged 226 lifts per standing, walking, sitting, driving, lifting, and hour, with an average load weight of 88.4 N. A using vibrating machinery). Exposures were combination of five factors distinguished ascertained either as of the birthday prior to between high- and low-risk jobs: lifting onset of symptoms or by lifetime occupational frequency, load moment, trunk lateral velocity, history prior to onset of symptoms. Using the trunk twisting velocity, and trunk sagittal angle. most recent job (as of the birthday prior to The highest combination of exposure measures symptoms), driving was associated with produced an OR of 10.7 (95% CI 4.9–23.6 in symptoms in males (RR 1.7, 95% CI 1.0–2.9), comparison to the lowest combined measures). as was lifting or moving weights of 25 kg or In univariate analyses, the most powerful single more (RR 2.0, 95% CI 1.3–3.1), when all variable was maximum moment (a combination exposures were considered in multivariate of both weight of the object and distance from analyses. For women, lifting (RR 2.0, 95% CI the body), which yielded a significant OR of 3.3 1.1–3.7) was associated with symptoms. When between low- and high-risk groups [Marras et lifetime exposures were considered, lifting al. 1995]. The study design was unusual in that remained significantly associated for males (RR the unit of analysis appeared to be the job 1.5, 95% CI 1.0–2.4). Both sitting (RR 1.7, rather than the individual. Neither participation 95% CI 1.1–2.6) and use of vibrating rates nor total number of participants was machinery (RR 5.7, 95% CI 1.1–29.3, based stated. No information appeared regarding the on one case) were associated with symptoms in proportions of individuals within jobs who were females. The multivariate analyses stratified on recruited sex and adjusted for age and simultaneous work exposures. While information on

6-16 symptoms and exposures was obtained the highest rate was seen for those who did it crosssectionally, the authors attempted to often, with a dose-response for three categories construct a retrospective cohort design by (10.9, 11.3, and 18.0, respectively, with a RR gathering data on lifetime work exposures and of 1.65 [95% CI 1.3–2.1]) when comparing back symptoms. While in the design lifetime lowest to highest). Liles et al. [1984] found a exposures were cumulated only prior to significant association between incidence of disorder onset, it would not be expected that back injuries related to lifting and lifting participants could recall these relationships exposures as assessed by JSI: the RR was 4.5 accurately. Temporal relationships were (95% CI 1.02–19.9) comparing the highest and unclear. lowest exposure categories. Burdorf et al. [1991] found no association between back pain Strength of Association symptoms and lifting load (the latter did not The most informative studies included those that vary across the six job categories examined in employed independent measures of exposure to the study). Huang et al. [1988] conducted assess lifting demands, as they provided the detailed ergonomic evaluations of two school best contrast among levels of exposure and lunch preparation centers with differing rates of were subject to the least misclassification. A musculoskeletal (including back) disorders. The case-control study by Punnett et al. [1991] center with higher disorder rates had greater found an OR of 2.16 (95% CI 1.0–4.7) for the lifting and other work-related demands. relationship between back pain (ascertained by Unfortunately, the study was ecologic in design symptoms and medical exam) and lifting, after and did not link exposures and outcomes to adjusting for covariates (including awkward calculate risk estimates for the study groups, postures). In their 1973 investigation, Chaffin although several areas for ergonomic and Park found a strong increase in incidence intervention were identified. of medical visits related to back problems with increased LSR (with an approximate five-fold Other studies assessed exposures by self- increase in risk comparing males in the highest report on interview or questionnaire. Johansson and lowest categories); they did not find a and Rubenowitz [1994] examined low-back similar dose-response relationship for symptoms by index of manual materials frequency of lifts. Marras et al. [1993, 1995] handling (which included lifting and other risk examined the relationship between low-back factors). In neither white- nor blue-collar injury reports and spinal loading during lifting, workers was LBP significantly associated with and found an OR of 10.7 (95% CI 4.9–23.6) the index. In Kelsey’s 1975 case-control study for simultaneous exposures to lifting frequency, of herniated lumbar discs, cases and controls load weight, two trunk velocities, and trunk had similar histories of occupational lifting (RR sagittal angle. Both lifting and postures 0.94, p=0.10). In a second case-control study contributed to the high ORs. In Magora’s of prolapsed lumbar disc, Kelsey et al. [1984] [1972, 1973] studies of LBP and occupational found that an association with work-related physical efforts, the highest LBP rate was lifting without twisting was observed only at the observed in those who lifted rarely. When LBP was ranked by level of sudden maximal effort,

6-17 highest lifting level (OR 3.8, 95% LBP rates were higher for registered nurses CI 0.7–20.1). A combination of both risk than for nursing aides, whom they stated had factors at moderate levels yielded an OR of 3.1 more lifting responsibilities (OR 1.2, p=0.04). (95% CI 1.3–7.5). The highest risk was seen After adjusting for hours worked, however, for simultaneous lifting and twisting with straight aides had the higher rate (RR 1.3, no statistical knees (OR 6.1, 95% testing done). Undeutsch et al. [1982] CI 1.3–27.9). Svensson and Andersson [1989] examined back pain in baggage handlers, a found a significant association between lifetime group characterized by frequent bending, lifting, incidence of LBP and lifting in univariate and carrying of loads. Although no exposures analyses (RR 1.2, p<0.05), but not in were estimated for this group, symptoms were multivariate analyses. Holmström et al. [1992] significantly associated with length of found an association between one-year employment after adjusting for age (p=0.035). prevalence of LBP and an index of manual materials handling (OR 1.27, 95% CI 1.2–1.4), In the studies using more quantitative exposure after adjusting for age. No association was assessments, strengths of association for the observed in multivariate analyses. Toroptsova relationships between low-back disorder and et al. [1995] found that LBP and lifting were lifting included estimates including a negative related in univariate analyses (OR 1.4, p<0.05); relationship [Magora 1972], no association no multivariate analyses were conducted. In the [Burdorf et al. 1991], and several positive Walsh et al. [1989] examination of LBP and associations with ORs in the 2.2–10.0 range. work-related factors, LBP was associated with One study found a positive relationship lifting (in jobs just prior to injury) (RR 2.0, 95% between sudden maximal efforts and LBP (OR CI 1.1–3.7), when age, sex, and all exposures 1.7) [Magora 1973]. Punnett et al. [1991] were considered in multivariate analyses. When found a point estimate of 2.16 after adjusting lifetime exposures were considered, lifting for other covariates; Chaffin and Park [1973] remained significantly associated for males (RR found a strong relationship (OR 5) for LSR (but 1.5, 95% CI 1.0–2.4). In Burdorf and not lifting frequency); Marras et al. [1993, Zondervan’s 1990 study, an OR of 5.2 (95% 1995] found that the highest risk of injury was CI 1.1–25.5) was observed for LBP and related to lifting in combination with posture- frequent lifting among crane operators. No related risk factors (OR 10.7). Liles et al. relationship was seen for the referent group of [1984] observed an OR of 4.5 for back injuries noncrane operators from the same plant (OR and the highest JSI. The investigation of school 0.70, 95% CI 0.14–3.5). lunch preparers did not calculate risk estimates [Huang et al. 1988]. In a study that determined exposure status on the basis of job title, Videman et al. [1984] Studies that used subjective measures of found slightly higher rates (not significant) of exposure found point estimates including none LBP in nursing aides than in qualified nurses. [Johansson and Rubenowitz 1994; Kelsey The authors stated that aides had higher 1975a,b; Videman et al. 1984] to a range workloads related to patient handling and lifting. Knibbe and Friele [1996] found that

6-18 including 1.3, 1.4, 2.0, 3.8, and 5.2 [Burdorf outcomes, and exposure assessment methods, and Zondervan 1990; Holmström et al. 1992; they were fairly consistent in demonstrating a Kelsey et al. 1984; Knibbe and Freile 1996; relationship between lifting and low-back Toroptsova et al. 1995; Undeutsch et al. 1982; disorder when objective measures of exposure Walsh et al. 1989]. Although the Kelsey et al. were used to evaluate populations with high [1984] exposure estimates were based on self- exposures. Results were less consistent when report, they showed important relationships subjective exposure measures were utilized. between lifting and posture in multivariate analyses. While the OR for lifting alone was 3.8 A NIOSH review of earlier publications related (for the highest lifting level), the OR rose to 6.1 to patient lifting demonstrated results consistent when postures related to twisting and bent with this review [Jensen 1990]. A knees were included in the model. comprehensive literature search evaluated all studies published between 1967 and 1987 that In summary, the articles reviewed provide contained original research on nursing evidence of a strong positive association personnel and back problems. Of 90 studies, between low-back disorder and lifting. Results six were identified which distinguished between from these and other studies emphasized the two or more groups of nurses with differing importance of awkward postures in the risk of frequencies of patient handling and reported on low-back disorder. back problems for each group. A weighted analysis of results from the six reports Temporal Relationship demonstrated an overall increase in back Two prospective studies assessed exposures problems of 3.7 in those in the higher lifting prior to identification of back disorders. Both frequency category. demonstrated positive associations between exposure and back disorder. Thirteen of the 18 Coherence of Evidence studies were cross-sectional analyses. In two of Lifting and manual materials handling have been these, investigators excluded cases of LBP with studied as risk factors for low back disorder for onset prior to the current job to increase the decades. Studies of workers’ compensation likelihood that exposure preceded disorder. A claims have shown that manual material third cross-sectional study truncated self- handling tasks, including lifting, are associated reported exposures on the birthday preceding with back pain in 25%-70% of injuries [Cust et disorder onset. One case-control study al. 1972; Horal 1969; Snook and Ciriello truncated exposures prior to disorder onset. Of 1991]. Data from the 1994 Bureau of Labor the four cross-sectional and case-control Statistics annual Survey of Occupational studies which attempted to address temporality, Injuries and Illnesses demonstrated that the three found positive relationships between lifting industry with the highest rate of time-loss and back disorder. injuries due to overexertion was nursing and personal care facilities (where employees are

Consistency in Association Although the 18 studies used varying designs,

6-19 required to engage in frequent patient handling 1993]. Other criteria include physiologic and lifting). measures of metabolic stress and muscle fatigue and psychophysical considerations (the During lifting, three types of stress are worker’s perception of his/her lifting capacity, a transmitted through the spinal tissues of the low combination of perceived biomechanical and back: compressive force, shear force, and physiologic attributes of the job). All three torsional force [Waters et al. 1993]. It has been criteria are important in assessing risk across suggested that disc compression is believed to the full spectrum of job and individual worker be responsible for vertebral end-plate fracture, variability. disc herniation, and resulting nerve root irritation [Chaffin and Andersson 1984]. In Exposure-Response Relationships early biomechanical assessments, models Eight studies examined exposure-response showed that large moments are created in the relationships in some form. Of these, four found trunk area during manual lifting. Static dose-response relationships between low-back evaluations of the trunk demonstrated that lifting disorder and objective measures of lifting results in large compressive forces on the spine. [Chaffin and Park 1973; Liles et al. 1984; Marras et al. 1995; Punnett et al. 1991]; More recently, biomechanical investigations another found a dose-response between have focused on spine loading and disc disorder and sudden maximal efforts [Magora tolerances associated with asymmetric loading 1973]. A study of baggage handlers found an of the trunk. In laboratory experiments, association between back disorder and length dynamic trunk motion components of lifting of employment [Undeutsch et al. 1982]. Two have been associated with greater spine studies found no dose-response relationship loading. Increased trunk motion during lifting (using a posture analysis assessment and a activities has been associated with increased manual materials handling index) [Burdorf et al. trunk muscle activity and intra-abdominal 1991; Johansson and Rubenowitz 1994]. measures, among other changes [Marras et al. 1995]. Some laboratory studies have shown The majority of studies which examined that lateral shear forces make trunk motions exposure-response relationships, and in more vulnerable to injury than in a compressive particular those that utilized quantitative loading situation. There is also in vitro exposure measures, demonstrated these trends. evidence that the viscoelastic properties of the spine may cause increased strain during increased speed of motion [Marras et al. Conclusions: Lifting and Forceful 1995]. Movements There is strong evidence that low-back Current models for lifting-related disorders are associated with work-related musculoskeletal injury stress that biomechanical lifting and forceful movements. The five studies considerations comprise only part of the reviewed for this chapter which showed no assessment of risk [Waters et al. association between lifting and back disorder used subjective measures of

6-20 exposure, poorly described exposure BENDING AND TWISTING assessment methodology, or showed little (AWKWARD POSTURES) differentiation of exposure within the study Definition group. The remaining 13 studies were consistent in demonstrating positive Bending is defined as flexion of the trunk, relationships, where those using subjective usually in the forward or lateral direction. measures of exposure showed a range of risk Twisting refers to trunk rotation or torsion. estimates from 1.2 to 5.2, and those using more Awkward postures include non-neutral trunk objective assessments had ORs ranging from postures (related to bending and twisting) in 2.2 to 11. Studies using objective measures to extreme positions or at extreme angles. Several examine specific lifting activities generally studies focus on substantial changes from non- demonstrated risk estimates above three and neutral postures. Risk is likely related to speed found dose-response relationships between or changes and degree or deviation from non- exposures and outcomes. For the most part, neutral position. For the purposes of this higher ORs were observed in high-exposure review, awkward postures also included populations (e.g., one high-risk group kneeling, squatting, and stooping. In most of the averaged 226 lifts per hour with a mean load studies included in this review, awkward weight of 88 N. Evidence from other studies postures were measured concurrently with and reviews has also suggested that groups with other work-related risk factors for back high- frequency exposure to lifting of heavy disorder. loads, such as nursing staff, are at high risk of back disorder. Studies Reporting on the Association Between LBP and Awkward Postures Most of the investigations reviewed for this Twelve studies examined the relationship document adjusted for potential covariates in between low back disorder and bending, analyses: two-thirds of the studies showing twisting, and awkward postures (Table 6-3, positive associations examined effects of age Figure 6-3). Most (nine) also examined the and gender. Nevertheless, some of the effects of occupational lifting. See the previous relatively high ORs that were observed were discussion of lifting and forceful movements. unlikely to be caused by confounding or other Nine studies were cross-sectional in design, effects of lifestyle covariates. Several studies two case-control, and one prospective. suggested that both lifting and awkward postures were important contributors to the risk Participation rates were adequate for 83% of of low-back disorder. The observed the investigations (Table 6-3). Four studies relationships are consistent with biomechanical assessed postures using objective measures and other laboratory evidence regarding the (however, in the study by Magora [1972], effects of lifting and dynamic motion on back details on their observation methods were not tissues. reported; the rest estimated exposures from interview or questionnaire responses). Health outcomes included low-back and sciatic pain symptoms, lumbar-disc prolapse, and back

6-21 injury reports. In four investigations, outcomes analyses that adjusted were defined using both symptoms and medical examination criteria. Only one investigation, the Punnett et al. [1991] case-control study of for a number of covariates (age, gender, length back pain in auto workers, fulfilled the four of employment, recreational activity and evaluation criteria (Table 6-3, Figure 6-3). medical history), time in non-neutral postures mild or severe flexion and bending were Several other studies, while not meeting all of strongly associated with back disorder (OR the four criteria, are particularly notable 8.0, 95% CI 1.4–44). In the same model, lifting because they used objective measures of was also associated (OR 2.16, 95% CI exposure assessment [Burdorf et al. 1991; 1.0–4.7). When the subset with physical Marras et al. 1993, 1995] or met more than medical findings was examined, associations one of the criteria [Holmström et al. 1992; were more pronounced. Although few study Kelsey et al. 1984]. As discussed earlier, the subjects were unexposed to all of the postures physical examination criterion may be less studied, a strong increase in risk was observed important in low-back disorders because of the with both intensity and duration of exposure. It paucity of specific physical findings in most was not possible to determine the relative cases of low-back disorders. contributions of different awkward postures because all were highly correlated. Only Descriptions of five studies which offered the participants’ current jobs (for referents) or jobs most information regarding the effects of when symptoms started (for cases) were bending, twisting, and awkward postures analyzed; the study design thus assumed a follow. Please note that there is some overlap short-term relationship between exposure and with studies that examined lifting effects. outcome. Although length of time in job was Detailed descriptions of the 12 studies appear also included in the models, the authors in Table 6-6. attempted to ensure that exposure preceded disease by identifying time of onset and The Punnett et al. [1991] case-control study measuring exposures in the job held just prior. examined the relationship between back pain The strong associations, after adjustment for and occupational exposures in auto assembly covariates, are notable. workers. Back pain cases (n=95) were determined by symptoms at interview and Burdorf et al. [1991] examined back pain medical examination; controls included those symptoms in a cross-sectional study of male free of back pain. For all participants or proxies concrete fabrication workers and a referent in the same jobs, jobs were videotaped and group of maintenance workers. Back pain work cycles were reviewed using a posture symptoms were assessed by questionnaire. analysis system. Exposures included time spent Exposures were measured using the Ovako in various awkward postures. Peak Working Posture Analysis System, which biomechanical forces were estimated for up to assessed postures for the back and lower nine postures where a load weighing at least 10 limbs, along with lifting load. Information on lb was held in the hands. In multivariate exposures in previous jobs was also collected.

6-22 Concrete workers experienced significantly between high- and low-risk jobs: lifting more back symptoms than referents (OR 2.8, frequency, load moment, trunk lateral velocity, 95% CI 1.3–6.0). trunk twisting velocity, and trunk sagittal angle. The highest combination of exposure measures Univariate results showed associations between produced an OR of 10.7 (95% CI 4.9–23.6) back pain and both posture index and WBV in (in comparison to the lowest combined current job. Correlations were presented measures). The study design was unusual in that showing lifting was not found to be associated the unit of analysis appeared to be job rather with back pain or to vary significantly across than individual. Neither participation rate nor the six job categories examined in the study. In total number of participants was stated. No multivariate analyses adjusting for age, both information appeared regarding the proportions posture index and WBV were significantly of individuals within jobs who were recruited associated with back pain, with ORs of 1.23 for measurement of lifting motions. However, (p=0.04) (for an ordinal scale of 6) and 3.1 the unit of analysis was job, and each was (p=0.001) (dichotomous), respectively. Those characterized by measurement of at least one in the highest posture index category were steel study subject. Effects of other covariates were benders, who spent an average of 47% of their not addressed (multivariate models appeared to time in bent back postures (compared to 12% include only biomechanical variables). The for the lowest exposed group). The posture study results emphasize the multifactorial index and WBV measures were highly etiology of back disorders, including correlated and analyzed separately. Strengths contributions of lifting frequency, loads, and of the study included use of a standardized trunk motions and postures. The study design symptom questionnaire, high participation rates did not allow for examination of temporal and objective measure of exposure, and an relationships. attempt to clarify the temporal relation between exposure and outcome by excluding cases of A case-control study of prolapsed lumbar disc back pain with onset before the present job. was carried out using a hospital population- based design [Kelsey et al. 1984]. Cases Using an unusual cross-sectional study design, (n=232) included individuals diagnosed with Marras et al. [1993, 1995] examined the prolapsed lumbar disc; an equal number of relationship between low-back disorders and controls matched on sex, age, and medical spinal loading during occupational lifting. A total service were selected. Exposure was assessed of 403 jobs from 48 diverse manufacturing using a detailed occupational history (not companies were assessed for risk of low-back described, but presumably obtained by disorder using plant medical department injury interview). An association with work-related reports. Jobs were ranked into three categories lifting, without twisting the body, was observed according to risk then assessed for position, at the highest lifting level (OR 3.8, 95% CI velocity, and acceleration of the lumbar spine 0.7–20.1). Twisting without lifting was during lifting motions in manual materials associated with disc prolapse (OR 3.0, 95% handling using electrogoniometric techniques. A CI 0.9–10.2); a combination of both risk combination of five factors distinguished factors had an OR of 3.1 (95% CI 1.3–7.5).

6-23 The highest risk was observed for simultaneous Strength of Association lifting and twisting with straight knees (OR 6.1, The more informative studies included the 95% CI 1.3–27.9). Despite the fact that Punnett et al’s [1991] case-control exposures were self-reported, these investigation, which fulfilled the four evaluation associations were notably strong. The potential criteria, plus several others that used existed for differential recall bias for cases and independent exposure assessments. In the controls, because study subjects were Punnett et al. study, multivariate analyses that interviewed about work-related factors after adjusted for covariates demonstrated that time case status was established. Interviewers may in non-neutral postures was strongly associated not have been blinded to case/control status. with back disorders (OR 8.09, 95% CI 1.4–44). In the same model, the OR for lifting Holmström et al. [1992] examined the was 2.2. Burdorf et al. [1991] found relationship between LBP and work task associations between posture index and back activities in a cross-sectional study of male symptoms in both univariate and multivariate construction workers. One-year prevalence of analyses: in multivariate analyses adjusting for LBP was ascertained by questionnaire. A age, the OR for posture index was 1.23 sample of workers was clinically examined. (p=0.04), for an ordinal scale of six levels. Exposure relative to lifting, handling, and work Posture index was highly correlated with WBV. postures was obtained by self-report. After However, the Kelsey et al’s [1984] case- adjustment for age, the index for manual control study of prolapsed lumbar discs found material handling, which included lifting, was that twisting without lifting had an OR of 3.0 associated with LBP with a RR of 1.27 (95% (95% CI 0.9–10.2); in combination, the two CI 1.2–1.4). Stooping and kneeling postures had an OR of 3.1 (95% CI 1.3–7.5). The showed a dose-response relationship with highest risk was observed for a combination of LBP, particularly severe LBP (with ORs 1.3, lifting, twisting, and straight knees (OR 6.1, 1.8, and 2.6 in comparison to those with no 95% CI 1.3–27.9). In the Marras et al. [1993, stooping; ORs 2.4, 2.6, and 3.5 in comparisons 1995] cross-sectional study, back injuries were to those with no kneeling, respectively). No associated with spinal loading during lifting, association was observed with sitting. In which included simultaneous exposures to lifting multiple regression analyses, LBP was frequency, load weight, trunk lateral velocity, associated with stooping (p<0.001) and trunk twisting velocity, and trunk sagittal angle. kneeling (p<0.01). While the authors attempted An OR of 10.7 (95% CI 4.9–23.6) was to adjust for some covariates (age, gender, and observed for the highest combination of psychosocial factors) in analyses, they did not exposure measures. Univariate ORs were 1.73 appear to examine simultaneous effects of (95% CI 1.38–2.15) for trunk lateral velocity, physical work-related factors in a single model. 1.66 (95% CI 1.34–2.05) for trunk twisting The cross-sectional design could not ascertain velocity, and 1.60 (95% CI 1.31–193) for the temporal relationships between exposure maximum sagittal flexion when comparing the and disorder. high-and low-risk groups [Marras et al. 1993].

6-24 The other studies showed a range of point disorder, with ORs of 2.6 and 3.5 (p<0.05), estimates. In univariate analyses, Magora respectively. [1972, 1973] found that for bending, the highest rate of LBP was observed for the In summary, three of the four studies using rarely/never category. For twisting and more quantitative exposure assessments reaching, the highest LBP rate was in the showed elevated risk estimates for the sometimes category. Johansson and relationship between low-back disorder and Rubenowitz [1994] found no associations bending, twisting, or awkward postures, with between low-back symptoms and bent or ORs ranging from 1.23 (for a scaled variable) twisted work postures in blue- and white-collar to 8.09; the highest risk estimate, an OR of workers. After adjustment for age and gender, 10.7, was based on combined exposure to however, extreme work postures were lifting and posture risk factors. Most of these significantly associated with the outcome in were based on multivariate analyses that blue-collar workers. Relationships were adjusted for covariates (usually age and presented as partial correlations, thus gender). The remaining studies demonstrate preventing calculation of risk estimates. risk estimates ranging from no association (in Riihimäki et al. [1994] observed that one study), 1.3–1.7 in univariate but not occupational exposure to twisted and bent multivariate analyses, to a high of 3.5 in another postures were associated with incidence of study. Studies utilized a number of definitions sciatic pain in univariate but not multivariate for awkward postures, as noted. analyses. No risk estimates were provided. In Svensson and Andersson’s 1989 study of LBP Temporal Relationship in Swedish women, bending forward was One prospective study assessed exposures associated with lifetime incidence in univariate prior to identification of back disorders. Results (RR 1.3, p<0.05) but not multivariate analyses. demonstrated positive associations in univariate The Masset and Malchaire [1994] univariate but not multivariate analyses. [Riihimäki et al. analyses demonstrated that trunk torsions were 1994]. Nine of 12 studies were cross-sectional associated with LBP in steel workers (OR in design. In one of these, investigators 1.55, p<0.05); no associations were shown in excluded cases of LBP with onset prior to the multivariate analyses. Toroptsova et al. [1995] current job to increase the likelihood that demonstrated that LBP in the past year was exposure preceded disorder. [Burdorf et al. associated with bending (OR 1.7, p<0.01) in 1991]. No association between exposure and univariate analyses (multivariate analyses were back disorder was observed. One case-control not conducted). Riihimäki et al. [1989a] study examined only exposures experienced in observed a dose-response for sciatic pain and the job just prior to disorder onset [Punnett et self-reported twisted or bent postures; the OR al. 1991]. A strong association between for the highest exposure category was 1.5 exposure to awkward postures and back pain [95% CI 1.2–1.9]. Holmström et al. [1992] was observed. observed that stooping and kneeling postures were associated with LBP, particularly severe

6-25 Consistency in Association studies found risk estimates above three and Although the 12 studies used varying designs, dose-response relationships between exposures outcomes, and exposure assessment methods, and outcomes. Many of the studies adjusted for the studies using quantitative exposure potential covariates in their analyses, and a few measures were fairly consistent in examined the simultaneous effects of other demonstrating a moderate relationship between work-related risk factors in analyses. Several awkward postures and low-back disorder. studies suggested that both lifting and awkward postures were important contributors to risk of Coherence of Evidence low back disorder. Nine of the 12 studies which examined posture effects also studied effects of lifting. Therefore, WHOLE BODY VIBRATION (WBV) a discussion of coherence of evidence for the Definition former relationship is similar to that found in the WBV refers to mechanical energy oscillations section on lifting and forceful movements. which are transferred to the body as a whole Forward flexion can generate compressive (in contrast to specific body regions), usually forces on the structures of the low back similar through a supporting system such as a seat or to lifting a heavy object. Similarly, rapid twisting platform. Typical exposures include driving can generate shear or rotational forces on the automobiles and trucks, and operating industrial low back [Marras et al. 1995]. vehicles.

Exposure-Response Relationships Studies Reporting on the Association Six studies examined dose-response Between LBP and Whole Body relationships between posture and low-back Vibration disorder. In one, no dose-response relationship Nineteen investigations addressed WBV as a was found between LBP and estimates for risk factor for back disorder. Fifteen study bending and twisting/reaching. In the other five designs were cross-sectional, two were cohort, studies, relationships were demonstrated one was case-control, and one had both cross- between back injury and spinal loading score, sectional and cohort components. LBP and posture index, sciatic pain and awkward postures, LBP and stooping, and None of the 19 studies fulfilled all of the four low-back symptoms and kneeling. evaluation criteria (Table 6-4, Figure 6-4). Participation rates were over 70% for 13 Conclusions: Awkward Postures investigations. Seven used independent The investigations that were reviewed provided measures of exposure for estimation of WBV; evidence that low-back disorders are in 10 studies, exposure information was associated with work-related awkward obtained by questionnaire or interview. In two postures. Results were consistent in showing studies, exposure to WBV was based on job increased risk of back disorder with exposure, title alone. Health outcomes included symptom despite the fact that studies defined disorders report of LBP, sciatica, or and assessed exposures in many ways. Several

6-26 lumbago, sick leaves or disability retirements with ORs of 4.56 (95% CI 2.6–8.0) for LBP in related to back disorders, and medically the past year and 2.30 (95% CI 1.2–4.5) for confirmed herniated lumbar disc. chronic LBP (for the highest exposure categories). Multivariate analyses adjusted for Five of the nine studies which met two or more age, body mass index, education, sports of the evaluation criteria used similar activity, car driving, marital status, mental methodologies and offered the most information stress, climatic conditions, back trauma and regarding the association between WBV and postural load (or vibration dose, depending back disorder. Detailed descriptions for all 19 upon the exposure examined). investigations can be found in Table 6-6. Bovenzi and Zadini [1992] used a similar Bovenzi and Betta [1994] examined the cross-sectional study design to examine low relationship between WBV and back disorder back symptoms in male bus drivers. Referents in a cross-sectional study of male tractor included maintenance employees who worked drivers. The unexposed group included male for the same company. Back pain symptoms revenue inspectors and administration workers were assessed by questionnaire. WBV was with no vibration exposure. Outcomes included measured for a sample of buses used over the various types of back symptoms reported by relevant time period. Cumulative vibration questionnaire. Vibration measures were exposures were calculated using this obtained from a representative sample of information, along with questionnaire items tractors and linked to individual information on related to work duration, hours, and previous number of hours driven yearly (obtained by exposures. In comparison to referents, bus questionnaire). Self-reported exposures to drivers demonstrated an OR of 2.80 (95% CI postural loads were also obtained. In 1.6–5.0) for lifetime LBP; the OR for LBP in comparison to referents, tractor drivers the past year was 2.57 (95% CI 1.5–4.4). In demonstrated an OR of 3.22 (95% CI multivariate analyses, the ORs for LBP in the 2.1–5.2) for lifetime LBP. For LBP in the past previous year were 1.67, 3.46, and 2.63 for year, the OR was 2.39 (95% CI 1.6–3.7). For three total vibration dose categories. Similar LBP in the past year, ORs ranged from 2.31 to trends were observed for other measures of 3.04 by exposure levels for total vibration dose, vibration (equivalent vibration magnitude and equivalent vibration magnitude, and duration of total duration of exposure), and after exclusion of those with exposure in previous jobs. exposure, after adjustment for covariates. In Statistically significantly increasing trends were multivariate analyses, chronic LBP showed a observed for nearly all types of back symptoms dose-response relationship with total vibration by exposure level (to all three measures of dose (OR 2.00, 95% CI 1.2–3.4, for the vibration) after adjustment for covariates. highest category), equivalent vibration Multivariate analyses adjusted for age, magnitude (OR 1.78, 95% CI 1.04–3.0, for awkward postures, duration of exposure, body the highest category), and duration of exposure mass index, mental workload, education, (OR 2.13, 95% CI 1.2–3.8, for the highest smoking, sports activities, and previous category). Exposure-response relationships exposures. were observed for postural load categories,

6-27 Three studies of WBV effects were conducted 1990a,b]. Two investigations were conducted by the same group of Dutch investigators. The using the same population: a 1986 cross- first examined back pain and WBV exposures sectional study of a cohort identified in 1975, cross sectionally in male helicopter pilots and a cohort analysis of sick leaves and [Bongers et al. 1990]. A referent group of disability retirements due to back disorder nonflying Air Force officers (with through the same time period. For the cross- characteristics similar to pilots) was also sectional analyses, information on back included. Information on back symptoms was symptoms was obtained by questionnaire. obtained by questionnaire. Vibration measures Vibration was measured for a sample of were assessed in two helicopters of each type vehicles and linked with questionnaire used by the study group. Individual exposures information related to types of vehicles driven, were calculated by matching this with hours, and previous employment. Information questionnaire items related to hours of flying regarding exposure to awkward postures was time and types of helicopters flown. Information also collected. Results from the cohort analysis on exposure to bent/twisted postures was also showed an incidence density ratio of 1.47 obtained by questionnaire. In comparison to (95% CI 1.04–2.1) for a comparison of sick controls, ORs for pilots were elevated for a leaves due to back disorders in exposed and number of back symptoms: 9.0 (95% CI referent groups. An increase in sick leaves for 4.9–16.4) for LBP and 3.3 (95% CI 1.3–8.5) disc disorders by vibration dose was observed, for sciatica. All of the above were adjusted for with an OR of 7.2 (95% CI 0.92–179) for the age, height, weight, climate, bent and twisted highest category. Cross-sectional study results postures, and feeling tense at work. In demonstrated increases in LBP symptom multivariate analyses, ORs for LBP were 13.8, prevalence by vibration dose category. 7.5, 6.0, and 13.4 for four categories for total Multivariate ORs increased by vibration dose flight time (in comparison to controls). ORs for (an OR of 2.8, 95% CI 1.6–5.0, for the highest LBP by total vibration dose were 12.0, 5.6, category) and years of exposure (an OR of 3.6, 6.6, and 39.5. By hours of flight time per day, 95% CI 1.2–11, for the highest category) after ORs were 5.6, 10.3, and 14.4 for LBP. adjustment for duration of exposure, age, Although there was some concern that pilots height, smoking, awkward postures, and mental with back pain may have dropped out of workload. employment, risk estimates were high (particularly in analyses by exposure level). Boshuizen et al. [1992] also conducted a Transient back pain appeared to increase with cross-sectional study of back pain in fork-lift daily exposure time, while chronic back pain truck and freight container tractor drivers appeared more associated with total flight time exposed to WBV. Referents included other and total vibration dose. employees working for the same shipping company, but with no vibration exposure. Back In a second study by the same group, WBV pain symptoms were assessed by questionnaire. exposures were examined in male tractor Exposures were estimated by measurement of drivers and a referent group of inspectors and vibration in a sample of vehicles, combined with maintenance technicians [Boshuizen et al. questionnaire responses. Cumulative exposures

6-28 were calculated, truncating at time of symptom [Boshuizen et al. 1992] to 9.5 [Bongers et al. onset. Prevalence of back pain was higher in 1990]. Analyses conducted by exposure level the exposed group than in referents: the RR for demonstrated stronger relationships. In Bovenzi back pain was 1.4 (p<0.05); RRs for LBP and and Betta’s 1994 study of tractor drivers, ORs lumbago were 1.4 (p<0.05) and 2.4 (p<0.05), for lifetime LBP were 3.79 for total vibration respectively, after adjusting for age. Differences dose, 3.42 for equivalent vibration magnitude, in LBP were observed only in younger age and 4.51 for duration of exposure (for the groups after multivariate adjustment for mental highest exposure levels). For LBP in the stress, years of lifting, awkward postures, previous year, ORs were 2.36, 2.29, and 2.74 height, smoking, and hours of sitting. There was for the highest levels of the same three no association between total vibration dose and exposure measures. In Bovenzi and Zadini’s back pain (OR 0.99, 95% CI 0.85–1.2) or 1992 study of urban bus drivers, the highest lumbago (OR 1.14, 95% CI 0.91–1.4). Only ORs for LBP were observed for intermediate vibration in the 5 years immediately preceding rather than the highest exposure categories: symptom onset was significantly associated with 3.46 for total vibration dose, 3.77 for back pain (OR 2.4, 95% CI 1.3–4.2) and equivalent vibration magnitude, and 3.08 for lumbago (OR 3.1, 95% CI 1.2–7.9). It total duration of WBV exposure. The Bongers appeared that a healthy worker selection effect et al. [1990] investigation of back pain in was operating, as differences in back pain were helicopter pilots demonstrated that the highest observed only for those in younger age groups. ORs for LBP were found in the highest categories for total flight time (OR 13.4, 95% Evaluation of the Causal Relationship CI 5.7–32), total vibration dose (OR 39.5, Between Back Disorder and Whole 95% CI 10.8–156) and hours of flight time per Body Vibration day (OR 14.4, 95% CI 5.4–38.4). A study of tractor drivers demonstrated LBP ORs of 2.8 Strength of Association (95% CI 1.6–5.0) for the highest total vibration Recent studies that included quantitative dose and 3.6 (95% CI 1.2–11) for the highest exposure assessments provided the most exposure duration category [Boshuizen et al. information regarding the relationship between 1990a]. In the same population, the OR for all WBV and back disorder [Bongers et al. 1988; sick leaves due to back disorder was 1.47, Boshuizen et al.1990a, b; Bovenzi and Betta comparing exposed (95% CI 1.04–2.1) and 1994; Bovenzi and Zadini 1992]. (Two other referent groups [Boshuizen et al. 1990b]. For recent studies also described quantitative sick leaves related to intervertebral disc exposure assessments, but no results relating to disorders, the highest OR was observed for the these were presented [Burdorf et al. 1993; highest exposure category (OR 7.2, 95% CI Magnusson et al. 1996]). In all five, ORs were 0.92–179). The Boshuizen et al. [1992] study of calculated by levels of vibration exposure, forklift truck and freight container tractor expressed in several ways (usually including drivers showed no association between back magnitude and duration of exposure). In the pain and total vibration dose (OR 0.99, 95% CI five studies, overall ORs comparing back pain 0.85–1.2) but did show an association for in exposed and referent groups ranged from 1.4 vibration in the preceding five years (OR 2.4,

6-29 95% CI 1.3–4.2). In this study the increase in 1.0–2.9) after adjusting for age and other job LBP prevalence in the exposed group was only exposures in multivariate analyses. Burdorf et significant for those in younger age groups (an al. [1991] found that WBV was significantly OR of 5.6 for those age 25-34) in multivariate associated with back pain (OR 3.1, p=0.001) analyses. In all five of these cross-sectional in multivariate analyses that adjusted for age. studies, ORs were calculated by vibration The Kelsey [1975a] case-control study found a exposure category after adjusting for a number significant association between herniated of covariates, as mentioned in the detailed study lumbar disc and time driving (OR 2.75, descriptions, above. p=0.02), and more specifically, working as a truck driver (OR 4.7, p<0.02). Burdorf et al. Other studies assessed both exposure and low- [1993] investigation demonstrated an OR of back disorder by interview or questionnaire. 3.29 (95% CI 1.5–7.1) for crane operators Burdorf and Zondervan [1990] observed no and 2.51 (95% CI 1.5–5.4) for vibration- association between WBV exposure and LBP exposed straddle-carrier drivers after adjusting in crane operators in univariate analyses (OR for a number of covariates. In a study of Danish 0.66, 95% CI 0.14–3.1); no associations were salespeople, annual driving distance was observed in multivariate analyses. Toroptsova associated with low-back symptoms [Skov et et al. [1995] also found no association between al. 1996]. A dose-response relationship was LBP and vibration in their study (no definition observed in multivariate analyses, with an OR for vibration was provided, but WBV was of 2.79 (95% CI 1.5–5.1) for the highest suggested). In the Riihimäki et al. 1994 category. prospective study, sciatic pain was associated with vibration in univariate but not multivariate Four studies assessed exposures primarily by models (no risk estimates were provided). job title. Magnusson et al. [1996] observed an While the definition for “vibration” was not OR of 1.79 (95% CI 1.2–2.8) for bus and clear, the authors suggested it could be truck drivers in comparison to an unexposed interpreted as low-level WBV. The Masset and referent group. In a study of crane operators, Malchaire [1994] cross-sectional study found the exposed group demonstrated ORs of 2.00 that LBP was associated with vehicle driving (95% CI 1.1–3.7) for all intervertebral disc (OR 1.2, p<0.001) in univariate analyses. disorders and 2.95 (95% CI 1.2–7.3) for disc Similar results were observed in multivariate degeneration after adjustment for age and shift analyses (OR 1.2, p<.005). Riihimäki et al. [Bongers et al. 1988]. An examination of risk [1989a] observed an OR of 1.3 (95% CI estimates of disc degeneration by years of 1.1–1.7) for longshoremen and earthmovers in exposure showed the highest OR (5.73) in the comparison to a referent group with no highest exposure category. In the Johanning vibration exposure. In the same study, no [1991] study of subway train operators, an OR association was seen for annual car driving (OR of 3.9 (95% CI 1.7–8.6) was observed for 1.1, 95% CI 0.9–1.4). Walsh et al. sciatica. While not a primary focus of [1989] found that driving (on job held prior to the Magora [1972, 1973] studies of LBP in symptoms) was significantly associated with eight selected occupations, it was observed that low-back symptoms in males (RR 1.7, 95% CI bus drivers had back pain rates similar to those

6-30 of the comparison group of bankers (RR 1.19, 1988; Boshuizen et al. 1990b]. Twelve studies 95% CI 0.8–1.7). had a cross-sectional design that could not directly address temporality. However, three Thus, four out of five studies using quantitative attempted to clarify relationships by excluding exposure assessments demonstrated positive from analysis the cases with disorder onset associations between back disorder outcomes prior to current job [Burdorf et al. 1991, 1993; and vibration exposures, with ORs ranging from Burdorf and Zondervan 1990]. A fourth cross- 1.4 to 39.5. The fifth cross-sectional study sectional study truncated self-reported found no overall association between exposure exposures on the birthday preceding disorder and back disorder but found associations in onset [Walsh et al. 1989]. In these four selected subgroups (which suggested that the investigations, positive relationships between study population was biased, as noted above). back disorder and WBV were also observed. In all of these studies, risk estimates by exposure category were calculated after Consistency in Association adjustment for many covariates. Results with regard to the relationship between low back disorder and WBV were most In the remaining studies, risk estimates varied, consistent in the studies using observational or including no association (n=3), ORs of 1.2, 1.7, measurement approaches to exposure and 2.8 for driving, an OR of 1.8 for truck or assessment. The strength of association was bus driving, an OR of 4.7 for truck driving, an more variable in studies using job titles or OR of 1.3 for machine operation, ORs of 2.0, questionnaires to assess exposures. The 2.95 and 5.73 for crane operation, an OR of variability in the associations does not appear to 3.1 for WBV, and an OR of 3.9 for subway be related to confounding exposures, since train operation. most studies adjusted for age, gender and at least several other confounders. Studies using In summary, the evidence from these more quantitative exposure measures were investigations suggests a positive association fairly consistent in showing the higher risk between WBV and back disorder. estimates. Relationships were particularly strong for high- exposure groups where exposures were In addition to the epidemiologic investigations assessed using observational or measurement that were reviewed for this document, many approaches. more were conducted in the 1960s though the 1980s. Others have summarized this evidence Temporal Relationship in earlier reviews. Hulshof and Veldhuijzen van Three studies had prospective designs in which Zanten [1987] concluded that, although studies temporal relationships between outcome and varied in methodologies and quality, most exposure could be determined [Bongers et al. showed a strong tendency toward a positive 1988; Boshuizen et al. 1990b; Riihimäki et al. association between WBV exposure and LBP. 1994]. In two of these, clear positive Seidel and Heide [1986] stated that the relationships between back disorder and literature they reviewed indicated an increased exposure were demonstrated [Bongers et al. risk of spine disorders after intense long-term

6-31 exposure to WBV. Bongers and Boshuizen peaks may cause ruptures in the superficial [1990] conducted a meta-analysis of studies structure of the disc and changes in the published through 1990 that examined the nutritional balance that lead to degeneration. relationship between WBV and several back Thus, prolonged vibration exposure may cause disorders. The overall OR for WBV exposure spine pathology through mechanical damage and degenerative changes of the spine was 1.5; and/or changes in tissue metabolism. the summary OR for LBP was also 1.5. These conclusions are consistent with the positive In addition to pathology of the vertebrae and associations observed in the evidence reviewed intervertebral discs, vibration exposure has above (although the studies published in the been shown to cause changes in 1990s have tended to report larger ORs). electromyographic (EMG) activity in muscles of the lower back [Wikström et al. 1994]. For Other evidence for the relationship is provided example, EMG experiments have demonstrated by surveillance data. The U.S. population- that lower back muscle exhaustion increases based National Health Interview Survey, during WBV exposure in truck driving. carried out in 1988, found that males Decreased stability of the lower back may employed as truck drivers and tractor result from slower muscle response, perhaps equipment operators had a RR of 2.0 for back increasing the risk of injuring other structures. pain in comparison to all male workers [Guo et al. 1995]. Laboratory investigations have shown that other work-related factors, including prolonged Coherence of Evidence sitting, lifting, and awkward postures, may act Laboratory studies have shown that exposure in combination with WBV to cause back to WBV causes spine changes that may be disorder [Dupuis 1994; Wikström et al. 1994; related to back pain. These include fatigue of Wilder and Pope 1996]. the paraspinal muscles and ligaments, lumbar disc flattening, disc fiber strain, intradiscal Exposure-Response Relationships pressure increases, disc herniation, and Five of six studies which carried out microfractures in vertebral end-plates [Wilder quantitative exposure assessment demonstrated and Pope 1996]. Studies of acute effects have exposure-response relationships between shown that the vertebral end-plate is the WBV and back disorder. structure that is most sensitive to high WBV exposure, followed by the intervertebral disc [Wikström et al. 1994]. Experimental investigations have demonstrated that high exposures to vibration cause injuries such as degeneration and fracturing of the vertebral end-plate. With regard to intervertebral discs, several studies have suggested that vibration causes creep, an increase in intradiscal pressure resulting from compressive loading. Pressure

6-32 Bovenzi and Betta [1994] observed a dose- consistent in demonstrating positive response between chronic LBP and total associations, with risk estimates ranging from vibration dose, equivalent vibration magnitude, 1.2 to 5.7 for those using subjective exposure and duration of exposure. Bovenzi and Zadini measures, and from 1.4 to 39.5 for those using [1992] found statistically significantly increasing objective assessment methods. Most of the trends for nearly all types of back symptoms by studies that examined relationships in high- exposure level, after adjustment for covariates. exposure groups using detailed quantitative Bongers et al. [1990] demonstrated increased exposure measures found strong positive ORs for sciatic pain and transient back pain associations and exposure-response with increasing hours of daily flight time. In relationships between WBV and back pain. their cohort of tractor drivers, Boshuizen et al. These relationships were observed after [1990b] observed an increase in risk of sick adjusting for age and gender, along with several leaves for disc disorder by total vibration dose other covariates (which, depending on the level. study, may have included smoking status, anthropometric measures, recreational activity, In other studies, Bongers et al. [1988] found an and physical and psychosocial work-related increase in risk of disc degeneration by years of factors). This evidence is supported by results exposure to crane operation; Skov et al. observed in many earlier epidemiologic [1996] found an increase in low-back investigations that have been summarized in symptoms with annual driving distance. other reviews. Johanning [1991] found no association between years of employment as a subway train Laboratory studies have demonstrated WBV operator and back pain symptoms. effects on the vertebrae, intervertebral discs, and supporting musculature. Both experimental The majority of studies which examined back and epidemiologic evidence suggests that WBV disorders by exposure level demonstrated may act in combination with other work-related dose-response relationships. factors such as prolonged sitting, lifting, and awkward postures to cause increased risk of Conclusions: Whole Body Vibration back disorder. There is strong evidence of a positive association between exposure to WBV and It is possible that effects of WBV may depend back disorder. Of the 19 studies reviewed for on the source of exposure. For example, in the this chapter, four demonstrated no association studies reviewed for this document, ORs were between WBV and back pain. Possible particularly high for helicopter pilots. It was not explanations for these results included use of possible to determine differences for other subjective exposure assessments that perhaps types of vehicles (automobiles, trucks, and resulted in misclassification of exposure status agricultural, construction, and industrial and, in one cross-sectional study, operation of vehicles). a healthy worker selection effect (where those with higher exposures dropped out of the study group). The remaining 15 studies were

6-33 STATIC WORK POSTURES Below are descriptions of four of the more informative studies. Detailed descriptions for all Definition 10 investigations are found in Table 6-6). Static work postures include isometric positions where very little movement occurs, along with Burdorf and Zondervan [1990] carried out a cramped or inactive postures that cause static cross-sectional study comparing 33 male crane loading on the muscles. In the studies reviewed, operators with noncrane operators from the these included prolonged standing or sitting and same Dutch steel plant, matched on age. sedentary work. In many cases, the exposure Symptoms of LBP and sciatica were assessed was defined subjectively and/or in combination by questionnaire. Activities in current and past with other work-related risk factors. jobs were assessed by questionnaire; exposures were rated according to level of Studies Reporting on the Association heavy work, frequency of lifting, WBV, and Between LBP and Static Work prolonged sedentary posture. Crane operators Postures were significantly more likely to experience Ten studies examined relationships between LBP (OR 3.6, 95% CI 1.2–10.6). Among low back disorder and static work postures, crane operators alone, the OR for heavy work which may have included prolonged sitting, was 4.0 (95% CI 0.76–21.2) after controlling standing, or sedentary work. For none was for age, height, and weight. It was determined static work posture the primary occupational that this heavy work occurred in the past and exposure of interest. Instead, it was often one not in current jobs. Among crane operators of many variables examined in larger studies of alone, the OR for frequent lifting was 5.2 (95% several or many work-related risk factors. Nine CI 1.1–25.5). The frequent lifting in crane of the studies were cross-sectional in design; operators was also determined to be from jobs one was a case-control study. held in the past. Among noncrane operators, history of frequent lifting exposure was not None of the investigations fulfilled the four associated with LBP (OR 0.70, 95% CI research evaluation criteria (Table 6-5, Figure 0.14–3.5). Among crane operators, univariate 6-5). Participation rates were acceptable for ORs for WBV and prolonged sedentary 60%. For four, case definitions included both postures were 0.66 (95% CI 0.14–3.1) and symptoms and medical examination criteria. 0.49 (95% CI 0.11–2.2), respectively. In Health outcomes included symptom report of multivariate analyses controlled for age, height, back pain, sciatica, or lumbago, back pain as weight, and current crane work, associations ascertained by symptoms and medical exam, with specific work-related factors were herniated lumbar disc, and lumbar disc substantially reduced; the high prevalence of pathology. One study claimed to assess job- LBP in crane operators was explained only by related exposures by observation; the nine current crane work. No measures of dose- others obtained information on static work response were examined. Limitations included postures by self-report on interview or a low response rate for crane operators (67%), questionnaire. with some suggestion that those with illness may have been underrepresented (perhaps

6-34 underestimating the OR), and self-report of by questionnaire, as were exposure-related health outcomes and exposures. The items. Physical exposures included lifting, investigators excluded cases of LBP with onset bending, twisting, other work postures, sitting, before the present job to increase the likelihood standing, monotony, and physical activity at that exposure preceded disease. work. Lifetime IRs varied by occupation, with ranges from 61%–83% in younger age groups Kelsey [1975b] carried out a hospital and 53%–75% in older groups. After the study population-based case-control study of was completed, the authors noted that for these herniated lumbar discs and their relationship to women, the highest lifetime incidence of LBP a number of workplace factors, including time was not found in jobs with the highest physical spent sitting, chair type, lifting, pulling, pushing, demands. The measure for “physical activity at and driving. Cases were defined by symptoms, work” was also not significantly associated with medical evaluation, and radiology; exposures LBP in univariate analyses. Bending forward were ascertained by interview (over lifetime job (RR 1.3), lifting (RR 1.2), and standing (RR history). Cases (n=223) and controls (n=494 1.3) were associated with lifetime incidence of unmatched controls) had similar histories of LBP in univariate analyses (p<0.05). Sitting job-related lifting (RR 0.94, p=0.10). Findings was not (OR 0.84, p=0.10). None of the indicated that sedentary work (sitting more than measures of physical workplace factors were half the time at work) was associated with disc associated with lifetime incidence of LBP in herniation, but only for the age group 35 years multivariate analyses. and older (RR 2.4, p=0.01). (The RR for those less than 35 was 0.81). Disc herniation was Videman et al. [1990] studied 86 males who also associated with time spent driving (RR died in a Helsinki hospital to determine the 2.75, p=0.02) and, more specifically, with degree of lumbar spinal pathology. Disc working as a truck driver (RR 4.7, p<0.02), degeneration and other pathologies were suggesting a relationship with WBV. The study determined in the cadaver specimens by design had several potential limitations, discography and radiography. Subjects’ including possible unrepresentativeness of the symptoms and work exposures (heavy physical study population (because the group was work, sedentary work, driving, and mixed) hospital-based). As exposure information was were determined by interview of family obtained retrospectively, cases may have over- members. In comparison to those with mixed reported exposures thought to be associated work exposures, those with sedentary (OR with back problems. Strengths include a well- 24.6, 95% CI 1.5–409) and heavy work (OR defined outcome and consistent results in 2.8, 95% CI 0.3–23.7) had increased risk of comparisons to the two control groups. symmetric disc degeneration. Similar relationships were seen for end-plate defects Svensson and Andersson [1989] examined and facet joint osteoarthrosis. For most LBP in a population-based cross-sectional pathologic changes, study of employed Swedish women. Information on LBP and sciatica was obtained

6-35 sedentary work appeared to have a stronger found that standing was associated with lifetime relationship than heavy work. Back pain incidence of LBP in univariate analyses (OR symptoms were consistently higher in those 1.3, p<0.05), but not in multivariate models. with any form of spinal pathology, although the Sitting was not associated in univariate analyses difference was significant only for anular (OR 0.84, p=0.10). Walsh et al. [1989] found ruptures. This study was unusual in design in that low-back symptoms were associated with that it examined a combination of spinal lifetime occupational exposure to sitting in pathological outcomes, symptoms, and females only (RR 1.7, 95% CI 1.1–2.6) in workplace factors. However, participation in multivariate analyses that considered other the study was dependent on obtaining work exposures. Kelsey’s 1975b case-control information from family members; participation study demonstrated that sedentary work (sitting rates were not stated. While recall bias is often more than half the time at work) was associated a problem in studies of the deceased, in this with lumbar disc herniations, but only for those case it should have been nondifferential, if 35 and older (RR 2.4, p=0.01); the RR for those present. less than 35 was 0.81. In a study of salespeople , a dose-response was observed for sedentary work and low back symptoms. An OR of 2.45 Strength of Association (95% CI 1.2–4.9) was seen for the highest The ten studies were approximately equal in category after adjustment for covariates [Skov terms of information they provided relating to et al. 1996]. The Videman et al’s [1990] study static work postures. Burdorf and Zondervan of cadavers found that those with histories of [1990] observed an OR of 0.49 (95% CI either sedentary or heavy work exposure had 0.11–2.2) for the univariate relationship increased risk of symmetric disc degeneration between prolonged sedentary postures and (OR 24.6, 95% CI 1.5–409 and OR of 2.8, LBP in crane operators. Holmström et al. 95% CI 0.3–23.7, respectively). Similar results [1992] found no association between LBP and were seen for other disc pathologies. For most sitting (in univariate or multivariate analyses). In pathologic changes, sedentary work appeared the Magora [1972, 1973] cross-sectional to have a stronger relationship than heavy investigation, the highest LBP rates were work. observed for those in the “rarely” category for variables related to sedentary postures, sitting, In summary, most (n=6) risk estimates for and standing. No dose responses were variables related to static work postures, observed. In the Toroptsova et al. [1995] study including standing and sitting, were not of machine manufacturing workers, sitting, significantly different from one. Others found standing, and static work postures were not small to moderate significant increases in risk: associated with LBP history in univariate ORs of 1.3 for standing, 1.7 for sitting (females analyses. No details were provided. In only), and 2.4 and 2.5 for sedentary work. The multivariate analyses, Masset and Malchaire Videman et al. [1990] cadaver study found high [1994] found a nonsignificant association risks of disc pathology in those with a history of between LBP and seated posture (OR 1.5, sedentary work. Study quality was similar p=0.09) in multivariate analyses. Svensson and across the range of point estimates observed. Andersson’s 1989 study of Swedish women

6-36 Therefore, an estimate of the strength of disorder. Kelsey [1975a] observed that, in association is difficult to determine. The addition to sedentary work, amount of time magnitude cannot be estimated based on the spent sitting on weekends was associated with available data. herniated discs. The finding that sedentary work was associated with herniated discs only Temporal Relationship in older age groups suggested that duration of Eight of 10 studies were cross-sectional in exposure may be important and that a threshold design. Two of these attempted to use may exist. Toroptsova et al. [1995] observed additional methodologies to increase the that back pain was lower in those who engaged likelihood that exposure preceded disorder by in sports activity, perhaps suggesting that excluding cases with onset prior to current job greater muscle strength prevents back pain. and truncating exposures prior to disorder onset. One found a positive relationship Several authors offered explanations for the between prolonged sitting and LBP symptoms. lack of associations they observed. It was pointed out that perception of “sedentary” is Consistency in Association subjective and that many jobs that investigators The studies showed poor consistency in (or subjects) considered to include prolonged estimation of the relationship between low- static postures may actually have allowed back disorder and static work postures, considerable movement throughout the day perhaps due to considerable differences in (such as office workers). Other “sedentary” definition of exposure. groups (such as industrial sewing machine operators) may be forced by work schedules to Coherence of Evidence maintain static postures for long periods. It is important to have a true range of exposure if As mentioned elsewhere, LBP has been differences in associated disorders are to be associated with mechanical forces causing an detected. increased load on the lumbar spine [Waters et al. 1993]. Increased loading on the spine Exposure-Response Relationships causes increased intervertebral disc pressures, which in turn, may be responsible for herniation Three studies addressed dose-response and back pain. In laboratory experiments, disc relationships, two of which did not demonstrate pressure has been found to be substantially any trends. Magora [1972, 1973] found the greater in unsupported sitting than in standing highest risk of LBP in the lowest exposure positions [Chaffin and Andersson 1984]. categories for sedentary postures, sitting, and standing. Videman et al. [1990] found a high Studies reviewed for this document suggested rate of lumbar disc pathology in those with relationships between back disorder and histories of sedentary and heavy work, with nonwork activities seemed to be consistent with relationships stronger for sedentary work. A the hypothesis that static dose-response for LBP symptoms and sedentary work was observed by Skov et al. work postures might be associated with back [1996].

6-37 Conclusions: Static Work Postures 1992]. The relative contributions of these Ten studies examined the relationship between covariates may be specific to particular low-back disorder and static work postures. In anatomic areas and disorders. For example, a most cases, this exposure was not of primary recent study of identical twins demonstrated interest but was one of many potential that occupational and leisure time physical workplace risk factors that were included in loading contributed more to disc degeneration analyses. Static work posture was defined in of the upper than the lower lumbar region several ways, including sedentary work and [Battié et al. 1995]. For both anatomic areas, work-related sitting and standing. Exposure age and twin effects (genetic influences and information was ascertained by interview for early shared environment) were the strongest nine of 10 studies. The strength of association identifiable predictors for this particular health could not be easily estimated because a large outcome. proportion of point estimates did not differ statistically significantly from unity. As a whole, Psychosocial factors, both work- and the results from these studies provide nonwork-related, have been associated with inadequate evidence that a relationship exists back disorders. These relationships are between static work postures and low-back discussed at length in Chapter 7 and Appendix disorder. B.

ROLE OF CONFOUNDERS In the studies reviewed for this document, As mentioned above, back disorder is gender and age effects were addressed in most multifactorial in origin and may be associated (86% and 74%, respectively). Approximately with both occupational and nonwork-related 40% addressed work-related psychosocial factors and characteristics. The latter may factors. In addition to these, many studies include demographics, leisure time activities, addressed other potential confounders in their history of back disorder, and structural analyses. characteristics of the back [Garg and Moore

6-38 Table 6-1. Epidemiologic criteria used to examine studies of low back MSDs associated with heavy physical work

Investigator Risk indicator blinded to case Basis for assessing back (OR, PRR, IR Participation Physical and/or exposure to heavy physical Study (first author and or p-value)*,† rate $70% examination exposure work year) status

Met at least one criterion:

Åstrand 1987 2.3† Yes Yes No Job titles or self-reports

Bigos 1991b No association No No NR† Observation or measurements

Burdorf 1991 No risk Yes No No Observation or measurements estimate§

Clemmer 1991 2.2†, 4.3† Yes No NR Job titles or self-reports

Heliövaara 1991 1.9, Yes Yes No Job titles or self-reports 2.5† Hildebrandt 1995 1.2† Yes No No Job titles or self-reports

Hildebrandt 1996 No association Yes No No Job titles or self-reports

Johansson 1994 No association Yes No NR Job titles or self-reports

Leigh 1989 1.5† Yes No NR Job titles or self-reports

Masset 1994 No association Yes No NR Job titles or self-reports

Partridge 1968 1.2 Yes Yes No Job titles or self-reports

Riihimäki 1989b 1.0 Yes No NR Job titles or self-reports

Ryden 1989 2.2† Yes No NR Job titles or self-reports

Svensson 1989 No association Yes No NR Job titles or self-reports

Videman 1984 1.1 Yes No NR Job titles or self-reports

Videman 1990 2.8, NR Yes NR Job titles or self-reports 12.1†

Met none of the criteria:

Bergenudd 1988 1.8† No No NR Job titles or self-reports

Burdorf 1990 4.0 No No NR Job titles or self-reports

*Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §Significant associations found in univariate but not multivariate results.

6-39

Table 6-2. Epidemiologic criteria used to examine studies of low back MSDs associated with lifting and forceful movements

Investigator Risk indicator blinded to case Basis for assessing back (OR, PRR, IR Participation Physical and/or exposure to lifting and Study (first author and or p-value)*,† rate $70% examination exposure forceful movements year) status

Met all four criteria:

Punnett 1991 2.2† Yes Yes Yes Observation or measurements

Met at least one criterion:

Burdorf 1991 No association Yes No No Observation or measurements

Chaffin 1973 Approx. 5† NR No NR Observation or measurements

Holmström 1992 1.3§ Yes Yes Yes Job titles or self-reports Huang 1988 No risk estimate Yes No NR Observation or measurements Johansson 1994 No association Yes No NR Job titles or self-reports

Kelsey 1975b 0.94 Yes Yes NR Job titles or self-reports

Kelsey 1984 3.8 Yes Yes NR Job titles or self-reports Knibbe 1996 1.3 Yes No No Job titles or self-reports

Liles 1984 4.5† NR No No Observation or measurements

Magora 1972 No association, NR No NR Observation or 1.7† measurements

Marras 1995 10.7† NR No NR Observation or measurements

Svensson 1989 1.2§ Yes No NR Job titles or self-reports Toroptsova 1995 1.4† Yes Yes NR Job titles or self-reports

Undeutsch 1982 No risk estimate NR Yes NR Job titles or self-reports

Videman 1984 No association Yes No NR Job titles or self-reports Walsh 1989 1.5†, Yes No NR Job titles or self-reports 2.0†

Met none of the criteria:

Burdorf 1990 0.70, No No NR‡ Job titles or self-reports 5.2†

*Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §Significant associations found in univariate but not multivariate results. 6-41

Table 6-3. Epidemiologic criteria used to examine studies of low back MSDs associated with bending, twisting, or awkward postures

Investigator Risk indicator blinded to case Basis for assessing back (OR, PRR, IR Participation Physical and/or exposure to bending, Study (first author and year) or p-value)*,† rate $70% examination exposure twisting, or awkward status postures

Met back criteria:

Punnett 1991 8.09† Yes Yes Yes Observation or measurements

Met at least one criterion:

Burdorf 1991 1.2† Yes No No Observation or measurements Holmström 1992 2.6†, Yes Yes Yes Job titles or self-reports 3.5† Johansson 1994 NR†,‡ Yes No NR Job titles or self-reports

Kelsey 1984 3 Yes Yes NR Job titles or self-reports

Magora 1972, 1973 No association NR No NR Observation or measurements Marras 1993, 1995 10.7† NR No NR Observation or measurements

Masset 1994 No association§ Yes No NR Job titles or self-reports

Riihimäki 1989b 1.5† Yes No NR Job titles or self-reports

Riihimäki 1994 No association§ Yes No NR Job titles or self-reports Svensson 1989 No association§ Yes No NR Job titles or self-reports

Toroptsova 1995 1.7† Yes Yes NR Job titles or self-reports

*Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. If reported with NR, a significant association was reported without a numerical value. ‡Not reported. §Significant associations found in univariate but not multivariate results.

6-43

Table 6-4. Epidemiologic criteria used to examine studies of low back MSDs associated with whole-body vibration

Investigator Risk blinded to case Basis for assessing back indicator (OR, Participation Physical and/or exposure to lifting and Study (first author and PRR, IR rate $70% examination exposure whole-body vibration year) or p-value)*,† status

Met at least one criterion:

Bongers 1988 2.0†–5.7 Yes Yes NR‡ Job titles or self-reports Bongers 1990 3.3–39.5† Yes No NR Observation or measurements

Boshuizen 1990a, 1990b 1.5–3.6† Yes No NR Observation or measurements

Boshuizen 1992 0.99 Yes No NR Observation or measurements Bovenzi 1992 2.6† Yes No NR Observation or measurements

Bovenzi 1994 2.4–4.6† Yes No NR Observation or measurements

Burdorf 1991 3.1† Yes No No Job titles or self-reports

Burdorf 1993 2.5–3.3† Yes No NR Observation or measurements Kelsey 1975b 2.8†, Yes Yes NR Job titles or self-reports 4.7†

Magnusson 1996 1.8† NR No NR Observation or measurements

Magora 1972 1.2 NR No NR Observation or measurements

Masset 1994 1.2† Yes No NR Job titles or self-reports Riihimäki 1989a 1.3† Yes No NR Job titles or self-reports

Riihimäki 1994 No association Yes No NR Job titles or self-reports

Toroptsova 1995 No association Yes Yes NR Job titles or self-reports Walsh 1989 1.7† Yes No NR Job titles or self-reports

Met none of the criteria:

Burdorf 1990 0.66 No No NR Job titles or self-reports

Johanning 1991 3.9† No No NR Job titles or self-reports Skov 1996 2.8† No No NR Job titles or self-reports

*Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported.

6-45

Table 6-5. Epidemiologic criteria used to examine studies of low back MSDs associated with static work postures

Investigator Risk blinded to Basis for assessing back indicator Participation Physical case and/or exposure to static work Study (first author and year) (OR, PRR, IR rate $70% examination exposure postures or p-value)*, status †

Met at least one criterion:

Holmström 1992 No Yes Yes Yes Job titles or self-reports association Kelsey 1975b 0.81, Yes Yes NR Job titles or self-reports 2.4† Magora 1972, 1973 No NR No NR Observation or association measurements

Masset 1994 1.5 Yes No NR Job titles or self-reports

Svensson 1989 1.3§ Yes No NR Job titles or self-reports

Toroptsova 1995 No Yes Yes NR Job titles or self-reports association Videman 1990 24.6† NR Yes NR Job titles or self-reports

Walsh 1989 1.7† Yes No NR Job titles or self-reports (females)

Met none of the criteria:

Burdorf 1990 0.49 No No NR Job titles or self-reports Skov 1996 2.45† No No NR Job titles or self-reports

*Odds ratio (OR), prevalence rate ratio (PRR), or incidence ratio (IR). †Indicates statistical significance. ‡Not reported. §Significant associations found in univariate but not multivariate results.

6-47

Table 6-6. Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Åstrand 1987 Cross- 391 male employees Outcome: Medical, 29.4 % of manual 12.9% of p=0.002 Participation rate: 82.5%. sectional, in a Swedish pulp psychological and social workers reported clerks 1987 and paper industry indicators. Questionnaires back pain in reported back The proportion of backs evaluated as located at one of 4 on social and psychological response to: “Do pain in abnormal by physical examination Åstrand and Retro- sites: Mill 1, Mill 2, Mill factors; medical you often have response to was 16%, similar to U.S. data Isacsson 1988 spective 3, and Head Office. examination of thoracic and back pain?” same collected in 1971. 66% of group with 22 years lumbar spine. question. back abnormalities reported back pain. follow up, 1988 Exposure: Based on the Duration of Psychosocial work factors did not type of work performed at employ- show any significant association with each job site. All mill work ment:1.2 1.0-1.5 back pain. jobs were judged as heavy; all office/clerk jobs Neuro- The working conditions of back pain were judged as light. Some ticism: 2.8 1.4-5.4 sufferers were changed because of worker movement between their reduced working capacity, office/clerk jobs and mill which tends to offset differences in work, based on health prevalence of back pain between status. groups doing heavy work and control populations.

Results support Magora’s findings that heavy work over time is associated with increased back pain.

Back pain was associated with occupation, low education, duration of employment, and neuroticism.

In follow-up study, a “healthy worker effect” was documented.

(Continued)

6-49

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bergenudd Longitudinal 323 males and 252 Outcome: Back pain not Point prevalence: LBP in All: 1.83 1.2-2.7 Participation rate: 67% in and Nilsson females; all tracked by exam. LBP unexposed Females: questionnaire and health survey from 1988 participants in “Attended” for exam but BP males: 28% males: 21.4% 2.03 1.1-3.7 830 individuals living in Malmo. Malmo, Sweden, based only on self females: 30% females: Males: Longitudinal Study assessment and 23.9% 1.76 1.01-3.1 Not controlled for confounders. since 1938. questionnaire, 1983. 5% prevalence of sciatica Exposures rated from job title. Exposure: Exposures and occupations tracked by In heavy or Weak support for occupational questionnaires since 1942. moderate work factors in causation. Some support Work classified into (LBP): for workload causing symptoms. 3 categories of heaviness males: 32.4% based on 10 years work. females: 38.9% Moderate or heavy physical demands had more back pain; then light (1) Light physical work: physical demand group (p<0.01) white collar. statistically significant only in females.

(2) Moderate: Nurses, Those with back pain had fewer shop assistants, bakers, years of education and were less and light industry. satisfied with their working conditions. There was no difference (3) Heavy: Carpenters, in the relationship between family, bricklayers, and heavy relatives, or friends. industry.

(Continued)

6-50

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bigos et al. Retro- Aircraft Outcome: Report of low Highest LB injury Lowest LB Highest to Participation rate: 100% (includes all 1986a spective manufacturer back injury. rates in mechanics injury rates in lowest records). cohort employees in 33 job Rate=38.2 electronic compari- morbidity classifications Exposure: 33 job technicians son is in Exact rates by job titles not reported. (15-month (n=31,200). classifications. and tool range of 5 follow-up) grinders to 7 (exact Authors state that differences by job Rate=NS numbers title are difficult to interpret because not of overlapping confidence intervals. reported)

(Continued)

6-51

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bigos et al. Prospective 3,020 aircraft Outcome: A case was 8% to 9% of N/A Lack of Participation rate: 43% of the original 1991b assembly workers; defined as a subject workers reported enjoyment number of workers solicited 54% of 1,613 involved in reporting an acute industrial an acute industrial of job tasks: participants returned questionnaire work perception and back injury. back injury. OR=1.7 1.3-2.2 with Minnesota Multiphasic Personality psychosocial portion Inventory (MMPI); 75% participated in some part of the study. Of of study. Subjects answered series MMPI: tend volunteers, respondents and non- of questionnaires: On towards respondents were similar. demographic and somatic psychosocial factors, a complaint or Employees’ work exposure not as cardiovascular denial of well documented as psychosocial questionnaire, and a take- emotional factors. home questionnaire on distress: psychosocial and individual OR=1.37 1.1-1.7 Take home questionnaire had 566 question Minnesota Multiphasic factors (see comments). Personality Inventory (MMPI), family Prior back function questionnaire (APGAR), Subjects had physical pain: Health locus of control (HLOC). examination to assess OR=1.7 1.2-2.5 physical attributes: Lifting Other information included medical strength, aerobic capacity, history, previous back discomfort or and flexibility. problem, and previous back injury claims in prior 10 years. Exposure: Based on Study did not investigate actual questionnaire data of work presence of back symptoms or and home activities. Also specific disorders; subjects followed “All jobs employing >19 for three years and became a case if workers analyzed for they: (1) reported to medical heavy and tiring tasks in department, (2) filed an incident or terms of maximal loads.” report, (3) filed an industrial insurance claim. Also analyzed “perceived Authors state that results may not physical exertion” as apply as strongly to cases of severe potential risk factor. symptoms or in work involving heavy job requirements (study performed in a manufacturing industry where “job tasks do not tend to be extremely stressful” for the back.

(Continued)

6-52

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bongers et al. Retro- Dutch, male, steel Outcome: Disability pension Crane operators Floor workers Incidence Participation rate: $70%. 1988 spective workers (n=1,405) for back-related disorder. (n=743) in same Density cohort departments Ratios Adjusted for nationality, shiftwork, (January, Exposure: Job title and (n=662) age, and calendar time. All back 1975- duration of employment. disorders: December Measurements of vibration 1.32 ORs likely are underestimated 1984) in cranes but not used in because of slight vibration exposure this study. Interverte- of the control group and potential bral disc health-based selection of the exposed disorders: group before start of the follow-up 2.00 0.84-2.1 period.

Degenera- The combination of exposure to W.V., tion of unfavorable postures, and adverse interverte- climatic conditions is the probable bral disc: cause of the back disorders. 2.95 1.1-3.7 COX regression: IDR for displace- ment of disc: 2.46 1.2-7.3 IDR for degenera- tion of inteverte- bral disc: 3.28 1.2-12.5

(Continued)

6-53

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bongers et al. Cross- Dutch, male, Outcome: Back symptoms, Dutch helicopter Non-flying air Participation rate: $70%. 1990 sectional helicopter aircrew by questionnaire. pilots and aircrew force officers and non-flying air observers (n=163) (297) Adjusted for age, height, weight, force officers Exposure: Hr of flight time, climate, bending forward, twisted types of helicopters flown, Back pain, 68%; 17% 8.0 4.5-14.3 postures, and feeling tense at work. and time spent in bent or LBP, 55%; 11% 9.0 4.9-16.4 twisted postures were Lumbago, 13%; 9% 2.6 1.1-6.0 Prevalence of transient back pain, in obtained by questionnaire. Sciatica, 12%; 6% 3.3 1.3-8.5 particular, was higher for exposed Vibration measurements Pattern alternating, than referent group. were taken in two 41% 6% 9.5 4.8-18.9 helicopters of each type Prevalence of transient back pain used in the study. increased with daily exposure time. Cumulative exposures were obtained by Chronic back pain increased with total combining questionnaire flight time and total vibration dose. and measurement data. Postures of pilots were constrained due to cockpit conditions.

Selection bias possible in that pilots with back trouble could have dropped out of employment.

(Continued)

6-54

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Boshuizen Cross- Employees of two Outcome: Back pain Sick leave for all Participation rate: 79%. et al. sectional Dutch companies symptoms were obtained back disorders 1.47 1.04-2.1 1990a,b follow-up of performing land by questionnaire in the ORs corrected for duration of a cohort reclamation and cross-sectional study, and LBP prevalence: by RR: 19.1, exposure, age, height, smoking, identified in inspection of roads, back-related sick leave and vibration dose, 4 29.4, 28.03, awkward postures, and mental 1975. Also, dikes, and building disability retirement categories 8.1 workload. includes sites. Several information was collected entire cohort workers operate in the cohort study. By vibration, 3 1.80, 1.78, Association greater with duration of in vehicles. The cross- categories 2.8 exposure than magnitude. examination sectional study Exposure: Vehicle vibration of sick leave included 577 information was combined By years of and disability workers, and the with questionnaire data exposure 3 2.44, 2.50, follow-ups. cohort study 689. regarding vehicle types categories 3.60 driven, awkward postures maintained, hr of work, and Sick leave by previous jobs held. vibration dose, 4 1.0, 0.97, categories 1.51, 1.45

Dose of 5 years, all 1.13 back disorders COX regress. adj. for age

(Continued)

6-55

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Boshuizen Cross- Male employees of Outcome: Back pain Fork-lift truck and Employees of Participation rate: $70%. et al. 1992 sectional six Dutch shipping symptoms by questionnaire freight tractor the same companies (n=452). drivers (n=242). companies Adjusted for age, mental stress, Exposure: Measurement without years lifting > 10 kg and twisting of vibration in sample of vibration spine, height, smoking, looking vehicles combined with exposure backwards, and hr sitting. questionnaire responses (n=210) to calculate cumulative Authors suggested that a healthy- dose (before symptom Prevalence (age worker effect was operating in that onset. standardized: older drivers were subject to health- Back pain, 48% 34% p=<0.05 based selection. LBP, 41% 30% p=<0.05 Lumbago, 19% 8% p=<0.05 Psychosocial factors were not addressed, except for “mental stress Cox regression: from work”. Back pain and total dose: 0.99 0.85-1.2

Lumbago and total dose: 1.14 0.91-1.4

Vibration exposure in last 5 years and back pain: 2.4 1.3-4.2 and lumbago: 3.1 1.2-7.9

Age and prevalence of LBP (multivariate OR): 25-34 5.6 Õ 35-44 1.96 Õ 45-54 0.68 Õ

(Continued)

6-56

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bovenzi and Cross- Male bus employees Outcome: Back-pain 234 bus drivers 125 Participation rate: $70%. Zadini 1992 sectional mail working in Trieste symptoms from maintenance survey questionnaire (rev. Nordic). workers Adjusted for age, awkward posture, working for duration of exposure, BMI, mental Exposure: WBV measured. same bus load, education, smoking, sport Cumulative exposures company activities, previous jobs at risk for estimated from Univariate results: back pain and duration of employment. measurements plus lifetime prevalence questionnaire results of LB symptoms, Does not address sedentary nature of (duration of work, previous 83.8%; 66.4% 3.12 1.8-5.3 work (states sitting is poorly exposures, etc.). LBP, 36.3%; 15.2% 2.80 1.6-5.0 correlated with LBP unless in combination with WBV). Previous 12 months: Psychosocial: adjusted for “mental LB symptoms, load” (no risk estimate provided). 82.9% 65.6% 2.99 1.8-5.1 LBP, 39.7%; 20.0% 2.57 1.5-4.4 Results were similar after excluding those with WBV exposure in previous Dose-response for jobs from analyses. total vibration and lifetime LBP; 4.05 1.8-9.3 Dose-response for 12-mo. LBP. 3.25 1.5-7.0

(Continued)

6-57

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Bovenzi and Cross- Tractor drivers, Outcome: Survey Tractor drivers Revenue Participation rate: 91.2% for exposed Betta 1994 sectional aged 25-65, working questionnaire (modified officers and 92.2% for unexposed. in Italy (n=1155) and Nordic) Univariate: male revenue Back Pain: 86.1% 57.3% 1.83 1.1-3.0 Multivariate analyses adjusted for officers engaged in Exposure: Vibration levels LBP Lifetime: age, BMI, education, sport activity, car inspection and were measured for a 81.3% 42.3% 3.22 2.1-5.2 driving, marital status, mental stress, administrative work representative samples of climatic conditions, back trauma, and (n=220). tractors. Information on 12-month LBP, postural load. awkward postures gained 71.7% 36.8% 2.39 1.6-3.7 from questionnaire. Number Dose-response Relationships reported between of hr operating yearly (highest vibration exposure and back pain, estimated from tractor categories) with clearest dose-responses for maintenance records. Lifetime LBP and chronic LBP outcome. Cumulative exposures tot. vib. dose; 5.49 3.6-8.5 estimated by combining the Chronic LBP and Independent effects observed for information. tot. vib. dose; 2.63 1.7-4.10 postural load and vibration.

Lifetime prevalence Results were similar after excluding LBP and duration of those with WBV exposure in previous exposure: jobs from analyses. 5-15 years 16-25 years 3.08 1.88-5.07 >25 years 3.03 1.80-5.12 4.51 2.43-8.34 Lifetime prevalence LBP and total vibration dose (years m2/s 4) <15 15-30 2.79 1.70-4.58 >30 3.44 2.05-5.77 3.79 2.20-6.53

(Continued)

6-58

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Burdorf and Cross- 33 male crane Outcome: Back pain 61% of crane 27% of 3.6 1.2-10.6 Participation rate: 67% of crane Zondervan sectional operators and 30 assessed by questionnaire operators had controls had operators and 100% of controls. 1990 male non-crane (Nordic). Pain in lower back pain back pain operator control back in the last 12 months. Control workers carried out more subjects matched moderate or heavy work, lifting, for age. Employed Exposure: Defined by job Risk Factors: walking, and standing than crane for $ one year. title and questionnaire operator in past. items: heavy physical Heavy work 4.02 0.76-21.2 work, lifting, WBV, and Physical demands are not significant sedentary postures Frequent lifting 5.21 1.10-25.5 in multivariate analyses. (current and past). Whole body 0.66 0.14-3.1 Controlled for age, height, and weight. vibration Crane operators with long work absences over-represented among non-responders.

Results indicate that the current job of crane operator is associated with reports of onset of back pain.

(Continued)

6-59

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Burdorf et al. Cross- 114 concrete Outcome: Back pain 59% of concrete 31% of 2.80 age 1.31-6.01 Participation rate: 95% concrete 1991 sectional workers compared symptoms assessed by workers had back controls adjusted workers; 91% maintenance males. prevalence to 52 maintenance questionnaire. Back pain pain had back and study engineers (controls). defined as pain which pain controlled Workload related to prevalence of All male. continued for $ a few hr for back back pain. during the past 12 months. pain from previous job Postural load, bending and twisting, Exposure: Assessed by as well as whole body vibration task analysis and OVAKO Model 1 causal factors. working posture analysis Postural system (OWAS) index Questionnaire included previous observation method. OR=1.23 p=0.04 employment history, risk factors in Eleven postures of present and previous jobs. importance for occupational Model 2 strain on the back were Whole body Univariate analysis controlled for used. vibration confounders using Mantel-Haensel OR=3.1 p=0.001 chisquare. Age, height, and weight For each job, two or three not significant factors. workers were chosen at random. Age controlled for in logistic regression. Index for postural load constructed using ordinal 30% with back pain had symptoms scale for rating the average >30 days. proportion of poor back postures. Six jobs were Concrete workers spent significantly ranked by index. more time in bent and/or twisted postures.

Postural index and whole body vibration significantly correlated (0.48, p<0.001). Therefore, authors designed two separate logistic regression models.

Prolonged standing or sitting not found to be risk factors.

(Continued)

6-60

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Burdorf et al. Cross- Crane operators, Outcome: Back pain Crane operators Office Participation rate: $70%. 1993 sectional saddle-carrier symptoms, by (n=94) and saddle- workers drivers and office questionnaire. carrier drivers (n=86) Adjusted for age and confounders workers aged 25- (n=95) (history of heavy work, exposure to 60, working in a Exposure: Postures WBV (y/n), history of work requiring large transport assessed with OWAS, Multivariate prolonged sitting, cold and drafts, company (n=275). WBV measured in sample analyses: working under severe pressure, job of each group, and past satisfaction, height, weight, duration work exposures estimated Crane operators 3.29 1.52-7.12 of total employment were by questionnaire. Straddle-carrier considered). drivers 2.51 1.2-5.4 Risk estimates were not presented by exposure categories, despite quantitative assessment.

Risk estimates reflect simultaneous exposure to WBV, static postures, and awkward postures.

Only persons with no complaints of low back pain before starting their current jobs were included in analyses.

(Continued)

6-61

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Chaffin and Prospective 5 plants in large Outcome: Visit medical Overall back rate, Participation rate: Not reported. Park 1973 with approx. electronics department because of low annual 7.2/100 1 year company. n=411 back complaint. FTEs (25 total back Age, weight, stature not associated follow-up individuals (279 injuries) with low back injuries. males and 132 Exposure: 103 jobs females). evaluated for Lifting A strong positive trend is indicated in Strength Rating (LSR) and the incidence rate data as the LSR lifting frequency. increases.

(Continued)

6-62

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Clemmer et al. Retro- Offshore drilling Outcome: Back-injury 543 cases of low Control room RR=2.2 Participation rate: Not reported. 1991 spective workers. cases reported on back injuries. and cohort standard forms with maintenance Workers performing the heaviest 14,518,845 worker- mention of “rheumatogical 7.5/100 3.18 physical labor had highest number of hr over 1979 to 1985 crux” for which the agent injuries and highest rates. (7,259 FTEs), 4,765 of injury was mechanical Roustabouts, total injuries. energy excluding other floorhands, and Controlling for “job,” age significantly body sites. derrick workers, associated with back strain in low-back strains workers performing heaviest length of Exposure: Based on job rate: 6.92 employment work not associated with title. back pain.

Job was best predictor of lost time.

Back injuries largely from falls. 75% of back strains precipitated by pushing, pulling, or lifting.

(Continued)

6-63

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Deyo and Cross- From the NHANES-II Outcome: Low-back pain Prevalence of LBP Prevalence of Participation rate: Not reported. Bass 1989 sectional national survey of within the past year with in current smokers: LBP current Not 27,801 individuals, $ one episode of near daily 10.7%. non-smokers: significant Lifestyle factors, including smoking 10,404 files of adults pain for $ two weeks. 10.2% and obesity, are risk factors for low- age 25 or older who back pain. had a physical Exposure: Smoking and examination were obesity, personal Ever smoked vs. The attributable risk for smoking was reviewed and 1,134 characteristics. LBP: 10.9% 9.6% 1.13 1.3 cases/100 persons. who met the case Significant definition were 50 pack years vs. Smoking risk increases steadily with selected for this LBP: 14.1% 9.6% 1.47 cumulative exposure and with degree study. The mean Significant of maximal daily exposure. age of the subjects BMI vs LBP, was 48.3 years and Highest quintile: Lowest A stronger association exists half (51.7) were 14.8% quintile: 8.5% 1.70 between back pain and smoking in females. Significant younger subjects than among those Odds ratio >age 45. each increment There is a steady increase in back LOG REGRESSION: pain prevalence with increasing Obesity 1.12 p<0.0006 obesity, but this elevates most Smoking 1.05 p<0.0006 strikingly in the highest 20% of body Chronic cough 1.36 p<0.0006 mass index (levels over 29.0 Activity 1.22 p<0.0006 kg/sq m). Education 0.84 p<0.0006 Age 1.01 p<0.0006 The association between obesity and Working 0.8 NS LBP could be confounded by other unmeasured lifestyle differences between the obese and non-obese so that obesity is just a marker for a true causal factor or factors.

(Continued)

6-64

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Heliövaara Cross- 2,727 males and Outcome: LBP interview Prior traumatic No prior Participation rate: 93% in screening. et al. 1991 sectional 2,946 females (30 to and tests at medical mobile injury increased injury 46 years) with clinic with uniform criteria. risk of LBP and, Physical and mental stress loads history, symptoms, sciatica 2.5 1.9-3.3 related to both sciatica and LBP. or findings indicating Low-back syndrome: and, musculoskeletal Symptoms during the low back syndrome Controlled for age and gender. disease. preceding month and major 2.6 2.1-3.1 pathologic finding on Work load index Body mass index, alcohol physical exam (fingertip- and, consumption, work-related driving, floor distance >25 cm at sciatica parity, and height were not flexion, rotation restricted and, 2.4 1.0-5.7 associated with LBP. to 25 degrees or less, low back syndrome objective signs of scoliosis Diabetes had a significantly of 20 degrees or more, Stress index 3.1 1.7-5.7 decreased prevalence of LBP Lumbar Lordosis, and, (OR=0.4 CI 0.3-0.8). Ladegue’s test positive at sciatica 60 degrees or less, or and, There was no statistical difference in severe abnormality. low back syndrome 2.4 1.7-3.5 LBP between sexes; sciatica significantly more prevalent among Sciatica: Symptoms males. radiating down leg and 2.0 1.5-2.6 findings of Lumbar nerve No association between smoking and root compression. sciatica.

Exposure: Based on self- Significant association between administered questionnaire; smoking and LBP in both older and index for occupational younger males, but only older physical stress and females. occupational mental stress. Significant association between LBP and osteoarthritis, mental disorders, and respiratory disease.

(Continued)

6-65

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Hildebrandt Cross- From the Dutch Outcome: Back pain cases 29.6% (2,327) of 23.9% of p<0.05 Õ Participation rate: “Population 1995 sectional population; a sample defined by symptom heavy workers sedentary sampled was representative of Dutch of 8,748 workers questionnaire ("yes" to reported back pain workers OR=1.2 1.33-1.55 population.” Unable to calculate. from three surveys "back pain quite often") and “quite often.” reported back on successive responses to interviewer. pain “quite Workers performing non-sedentary years. often.” work at highest risk. Exposure: Based on job title classification of work Rates increase with age for males, to demands; four categorical Rates of LBP: age 54, and for females to age 64. exposure variables: trade branch, trade class, Construction: Controlled for age and gender by professional branch, and 35%; stratification. professional class. Truckers: 31%; Professions with high prevalence of back pain on average were Plumbers: 31%. characterized by physically demanding work with dynamic components.

Data originally collected for screening of health and medical consumption, therefore less specific exposure variables—only job titles. However, there may be less potential for information bias because respondents did not then focus exclusively on back pain and work-relatedness.

Conclusion: In non-sedentary work, both males and females have higher prevalence rates than those who work in sedentary jobs.

(Continued)

6-66

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Hildebrandt Cross- 436 male workers in Outcome: Low back pain Prevalence: Reference Õ Õ Participation rate: Varied from 60% to et al. 1996 sectional five maintenance cases defined by symptom 1-year; LBP: 53% group had 80% in different departments. departments of a questionnaire (“yes” to low high physical steel company, back pain in last exposures. Reference group characterized by compared to 396 12 months). high levels of exposure to adverse non-sedentary working conditions. workers also Exposure: Assessed by exposed to heavy questionnaire. Workers Poor selection of referents. workloads. placed into one of 18 groups based tasks Prevalence rates adjusted for age performed “often” or differences between groups. “predominantly.” Tasks assigned a score on four Task groups with high prevalence indices: (1) physical rates of low back symptoms also workload, (2) psychosocial associated with high exposures to workload, (3) poor climate, unfavorable working conditions. and (4) vibration. Rates work groups (within units) according to self-reported exposures but does not cross-tab these with LBP.

(Continued)

6-67

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Holmström Cross- 1,773 randomly Outcome: (1) LBP history 1-year prevalence Participation rate: 76%. et al. 1992 sectional sampled from postal questionnaire. rate LBP 54%; construction Back pain defined as pain, 1-year prevalence Examined medical records for workers (male). ache, or discomfort in for severe LBP 7%. nonrespondents; same as for lower back, including respondents. gluteal regions with or Lifting freq: >1/5 without radiating pain into min <1/5 1.12 p<0.001 Information included individual and leg/s experienced employee-related factors, disorders in sometime, often, or very Stooping: >4 hr locomotor system, physical workload, often during past year, seldom 1.29 1.1-1.5 and psychosocial factors. (2) $ for 1 to 7 days, (3) Kneeling: > 4 hr with any degree of seldom 1.24 1.1-1.4 Examiners blinded to case and functional impairment. Stress: high exposure status. 1.6 1.4-1.8 A sample of workers had Anxiety: high Multiple logistic regression models clinical exam: Active spinal 1.3 1.1-1.4 used; separate models for individual, mobility test, springing test, manual materials handling, and straight leg raising, working postures. interspinal and paraspinal palpation from T11 to S1, In univariate analysis, no relationship combined extension and with daily traveling time, leisure lateral flexion while activity, or height and weight. standing and passive lumbar flexion and Construction tasks such as extension while lying on bricklaying or carpentry did not affect one's side. LBP.

Exposure: Based on Stress index reflected a high achiever questionnaire data person. reporting of task activity. Longer duration of stooping and kneeling was associated with LBP in all age groups (dose-response).

Only severe LBP related to smoking.

(Continued)

6-68

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Huang et al. Cross- Subjects consisted Outcome: Symptoms Consistently N/A Participation rate: All 42 workers 1988 sectional of all 24 female full- relating to upper limbs, constrained completed a symptom, health, and time workers from trunk and lower limbs postures: work history questionnaire and school lunch center during the previous month 17 workers 3 workers p<0.05 20 from each center also participated A and 20 female full- were solicited from a (70.8%) (15%) in a physical examination. Six time workers from questionnaire, while clinical workers from center B declined to center B. findings of pain during Poor equipment participate for personal reasons movements, muscle layout: 18 workers 3 workers p<0.01 unrelated to the purpose of the study. All 42 workers tenderness, signs of CTS, (75%) (15%) completed a signs of epicondylitis, and Center A had a significantly higher symptom, health and signs of tenosynovitis Consult physician: prevalence of musculoskeletal work history were documented in a 17 workers 5 workers p<0.01 complaints, more clinical findings, and questionnaire and physical examination. (70.8%) (25%) greater medical treatment experience 20 from each center than those in center B. also participated in a Exposure: Ergonomic risk Muscle tenderness: physical factors included handling 5.1 +/- 5.6 The ratio of the actual lifting load to examination. Six heavy objects, holding 0.8 +/-2.3 p<0.01 the Action Limit was also larger in workers from center constrained postures, too Signs of center A than in center B. B declined to much stooping, repetitive tenosynovitis: participate for use of arms and hands, 6 workers No significant difference was found personal reasons and poor equipment layout. (30%) 1 workers p<0.05 between the centers for low back unrelated to the NLE used to evaluate (5.0%) pain. purpose of the manual lifting tasks. study. Upper back pain: Study design was ecologic. Health significant outcomes and exposures were examined separately for two centers. Information was not combined for individual participants.

(Continued)

6-69

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Johanning Cross- Employees of the Outcome: Back-pain Subway train Subway Participation rate: Not reported. 1991; sectional mail New York City symptoms in past year, by operators (n=492) control tower Johanning survey transit system questionnaire survey. operators Controlled for age, gender, job title, et al. 1991 (n=584) (n=92) employment duration. Exposure: Job title. Although, WBV measures Any back pain, Study groups are stable working were taken for the 41% 25% PRR=1.11 1.04-1.19 populations with low turnover rates. exposed group, no analyses were presented. Sciatic pain 3.9 1.7-8.6 Exposed and unexposed groups are similar with regard to demographics and job histories.

Workers with a history of back problems or previous WBV exposure were excluded from the study.

Duration of employment not associated with risk.

Exposure data was not associated with outcome data in these articles.

Vibration measures showed high lateral and vertical acceleration levels.

(Continued)

6-70

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Johansson Cross- Subjects were 241 Outcome: Low-back Prevalence of low- Prevalence of PRR=1.76 1.25-2.47 Participation rate: The participation and sectional blue-collar (39% symptoms during the back symptoms LB symptoms rate was approximately 90%. Eighty- Rubenowitz females) and 209 past 12 months as self- =0.43 (CI 0.37- =0.42 (CI 0.35- seven percent of the blue-collar and 1994 white-collar (35% reported on the Nordic 0.50) for blue-collar 0.49) among 95% of the white-collar workers had females) workers Musculoskeletal workers, which wt. collar >2 years experience in their current from eight diversified Questionnaire (NMQ), reduced to p=0.32 workers, jobs. metal industry which was supplemented (CI 0.26-0.39) which companies in with an additional when solely work- reduced to Among blue-collar workers 12 of 15 Sweden. question regarding the related symptoms p=0.18 correlation tests regarding workload work-relatedness of the were considered. (CI 0.11-0.24) factors and work-related symptoms The participation rate symptoms. when solely were not significant. was approximately work-related 90%. Eighty-seven Exposures: Individual and symptoms Among blue-collar workers 10 of 15 percent of the blue- employee-related variables were partial correlation tests (adjusted for collar and 95% of related to the psychosocial considered. the effects of age and sex) regarding the white-collar work environment and the psychosocial job factors and work- workers had physical workload (sitting, related musculoskeletal symptoms >2 years experience manual materials handling, were significant. in their current jobs. lifting). Among blue-collar workers 7 of 15 partial correlation tests regarding psychosocial job factors and musculoskeletal symptoms, according to the NMQ, were significant.

Among white-collar workers none of the relationships between the five psychosocial factors and low-back symptoms were significant, whether or not work-related.

Calculations of associations based on the NMQ, without an effort to determine the work-relatedness of symptoms, could have a powerful effect-masking result.

(Continued)

6-71

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Kelsey 1975b Case-control Cases were Outcome: Herniated lumbar Sitting >half the Participation rate: 79% cases; obtained from a intervertebral time: 77% controls. population in the age discs were the outcomes <35 years Equal RR=0.81 range 20 to 64 years of interest in this study. >35 years Fewer RR=2.40 p=0.01 Results were similar for two control residing in the New Three levels of herniated groups (less strong for unmatched Haven SMSA who disc were classified: Time driving: controls). had lumbar X-rays Surgical cases, probable >half vs. herniation taken during the cases, and possible cases. Fewer RR=2.75 p=0.02 Study design subject to nondifferential period June 1971 Occupation: recall problems (with regard to through May 1973 at Exposure: Occupation, Truck driver vs. case/control status). the three hospitals in years of employment, herniation the area and at the amount of time worked, Fewer RR=4.67 p=0.02 The association between sedentary office of two of the amount of time spent sitting, Lifting vs. occupations, especially those which private radiologists type of chair, lifting, herniation involve driving, and herniated lumbar in New Haven. A pushing, pulling, carrying, Equal RR=0.94 p=0.10 discs exists in both sexes and in total of 217 pairs (89 lifting frequency, and comparisons between cases and females and 128 weight of objects lifted both control groups. males) was obtained were the exposures of for the comparison interest. The strength of this association in of cases and those aged 35 and older and the lack matched controls. of association in those who are under For the analysis of that age suggest that a certain amount cases and of time in sedentary occupations is unmatched controls, necessary for an effect to be seen. there were 223 cases (91 This study gave no evidence of an female and 132 increased risk for herniated lumbar males) and discs among males who did lifting on 494 controls their jobs, and little indication of this (225 females and among the females. Chance could 269 males). explain the slight tendency toward significance in the female subjects.

(Continued)

6-72

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Kelsey et al. Case-control Persons in the age Outcome: Status N/A N/A Lifting: Participation rate: 72% cases; 1984 range of 20 to 64 determined on the basis of >11.3 kg 79% controls. years who had an interview, diagnostic >25/day: lumbar X-ray films or tests performed by OR=3.5 1.5-8.5 All case categories combined in case- myelograms taken interviewers, and data control analyses (same results during 1979 to 1981, recorded in medical Lifting: observed for all categories). in one of three records. Cases classified > 11.3 kg hospitals, one as “surgical” cases, >5/day and Controls matched with cases on sex, neurosurgical “probable” cases, and twisting the age and hospital service. private practice, or “possible” cases. Control body half two orthopaedic group composed of the time: Frequent twisting alone did not affect private practices in persons without known OR=3.1 1.3-7.5 the risk of prolapsed disk, while the New Haven and prolapsed disc admitted to twisting with lifting had a detrimental Hartford, CT areas. the same medical services Lifting: effect. for conditions not related to >11.3 kg 232 matched case- the spine. Cases and while Study design subject to nondifferential control pairs. controls all with recent twisting recall problems (with regard to (within 1 year) disease body with case/control status). onset. the knees almost Exposure: Exposure to straight: activities performed on the OR=6.1 1.3-27.9 current job assessed by interview and Carrying: questionnaire. >11.3 kg 5 to 25/day: OR=2.1 1.0-4.3

Carrying: >11.3 kg >25 per/day: OR=2.7 1.2-5.8

(Continued)

6-73

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Knibbe and Cross- 355 females Outcome: Questionnaire, Lifetime LBP N/A Õ Õ Participation rate: 94%. Males and Friele 1996 sectional employed as developed from Nordic prevalence: pregnant females excluded from (study community nurses or questionnaire for 87% Back pain in sample. intended to community nurse musculoskeletal disorders, last 7 days, provide auxiliaries by the mailed to nurses. 1-year LBP community 89.9% of nurses described situations baseline data home care prevalence: 66.8% nurses vs. they considered physically for organization of the Exposure: Questionnaire community demanding. 82.1% of tasks described longitudinal city of Rotterdam. asked (1) if nurses could Auxiliaries: nurse involved patient transfers. Static load study). describe any work tasks 61.2 auxiliary: on the back was mentioned in 23.2% they considered physically OR=0.84 0.49-1.45 of descriptions. demanding, and 1-week LBP (2) whether the onset of prevalence: Backpain in Prevalence appeared to decrease back pain was related to a 20.6% previous 12 with age. Cross-sectional study specific work situation. months; design prevented investigators from Also job title: Community Prevalence of sick community determining whether observation was nurses vs. Auxiliaries. leave due to back nurses vs. due to selection effect or due to pain in previous 3 community experience. months: 9.7% nurse auxiliary: Rates for community nurses and OR=1.54 0.97-2.47 auxiliaries do not reflect significant differences in hrs worked/week (30.7 vs. 26.2). Adjusted for hrs worked OR is 1.3 (auxiliaries higher).

Authors state that auxiliaries are responsible for more lifting activities.

(Continued)

6-74

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Leigh and Cross- 959 working males Outcome: LBP based on 1-year LBP past Managers Participation rate: Not reported. Sheetz 1989 sectional and 455 working national survey of working prevalence: and (Probably to national survey). females in the United conditions. Question: “Is 19.4% males Professional States employed >20 the trouble with back or 20.7% female Workers in jobs requiring “lots of hr/week. spine in past year?” physical effort and lots of repetitive Occupations: work report more back pain. (U.S. Department of Exposure: Defined by job Farmers Managers 5.17 1.57-17.0 Labor QES Survey title and questionnaire on Clerical Managers 1.38 0.85-2.25 Exposure information based on self respondents.) work conditions, including Operator Managers 2.39 1.09-5.25 report and job title. workload. Service Managers 2.67 1.26-5.69 Health outcome did not distinguish Job demands: between upper and lower back pain. High Low 1.68 1.05-2.90 Gender, race, obesity, height, and Smoker Non smoker 1.48 1.00-2.19 repetitious work are not significantly associated with back pain.

(Continued)

6-75

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Liles and Prospective 28 companies, 63 Outcome: Lifting injury to Total injuries: Total injuries: Participation rate: Not reported (all Deivanayagan jobs in study 1, 38 in back, as recorded or volunteers). 1984 study 2. Selected reported. Injury rate for the Injury rate for jobs with frequent highest job severity the lowest job Dose response for lifting injuries by lifting requirements; Exposure: Jobs rated by index category: severity index JSI. manual handling Job Severity Index for lifting 17.1 injuries/100 category: requirements. (observation, use of FTES 3.8 injuries/ No adjustment for confounders. records for calculation). 100 FTES RR=4.5 Study 1: 220 males; Each individual followed Outcome defined as lifting injuries. 24 females. until job change (up to 2 Disability injury rate Disability injury Not distinct from exposure. years). for the highest job rate for the Study 2: 165 males; severity index lowest job 44 females. Total of 529 FTEs divided category: 11.4 lost severity index equally into 10 SI levels. time injuries/100 category 3.0 FTES lost time injuries/100 FTES RR=3.0

Severe injury rate Severe injury for highest job rate for the severity index lowest job category: 120.8 severity index days lost/number category 3.0 of lost time injuries days lost/number of lost time injuries RR=40

(Continued)

6-76

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Magnusson Cross- Bus drivers, truck Outcome: Back pain Bus drivers Sedentary Participation rate: Not reported. et al. 1996 sectional drivers, and symptoms, by (n=111) and truck workers sedentary workers questionnaire. drivers (n=117) (n=137) ORs do not appear to be from recruited in the state multivariate analyses including other of Vermont and Exposure: Ergonomic Driving 1.79 1.16-2.75 covariates, except as stated. Gothenburg, exposures, by Sweden questionnaire and vibration Freq. lifting 1.55 1.01-2.39 Quantitative exposure measures are level measurements not used in analyses that are according to ISO Heavy lifting 1.86 1.2-2.8 presented. standards. Long-term vibration exposure Long-term vibration calculated as product of exposure 2.0 0.98-4.1 daily exposure and years driving. Vibration and freq. lifting 2.1 0.8-5.7

Vibration and heavy lifting 2.06 1.3-3.3

(Continued)

6-77

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Magora 1972 Cross- A previous article Outcome: The outcome The exposed group The controls NR Participation rate: Not reported. sectional (1970) described the variable, low-back pain, consisted of consisted of process for was defined in a previous workers from 8 2887 The use of two hands to lift a load, selecting 3,316 article [1970]. Symptoms occupations. The individuals and especially holding the load away individuals from 8 by self-report. selection process from 8 from the body, are related to a higher occupations for was described in occupations. incidence of LBP. inclusion into this Exposure: The physical an earlier article by The selection study. activities studied in this the same author process was The lifting risk factors are magnified investigation were sitting, [1970]. described in when completing unaccustomed standing, weight lifting, and an earlier tasks. weight lifting technique. article by the same author Rarely sitting reported to be [1970]. associated with LBP.

Standing less than 4 hr daily reported Sitting > 4 hr day: to be associated with LBP. Often: 0.95 0.8-1.14 Sometimes: 0.09 0.05-0.14 Variable sitting and standing reported Rarely: 3.20 2.69-3.8 to be protective.

Standing Variable: < 4 hr daily 2.38 1.99-2.85

(Continued)

6-78

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Magora 1973 Cross- A previous article Outcome: The outcome The exposed group The controls Sudden Not Participation rate: Not reported. sectional (1970) described the variable, low-back pain, consisted of consisted of maximal reported process for was defined in a previous workers from 8 individuals physical It appears that sudden maximal selecting 3,316 article (1970). occupations. The from 9 efforts efforts, especially if unexpected, play individuals from 8 selection process occupations. were found an important role in the causation of occupations for Exposure: The physical was described in The selection to be LBP. inclusion into this activities studied in this an earlier article by process was related to a study by investigation were bending, the same author described in high Many of the physical causative observation and rotation, reaching, sudden (1970). an earlier incidence of factors, such as bending or rotation, interview. maximal efforts, and the article by the LBP. found by other investigators to be number and type of work same author related to a high incidence of LBP are breaks, by observation, (1970). actually sudden maximal efforts and interview. incidentally carried out at that moment Among those with Among in a certain position of the spine. LBP: controls: While most bending, twisting, and Bending: Bending: reaching motions required by each Often: 14.5% Often: 85.5% occupation are knowingly carried out, Sometimes: Sometimes: sudden maximal physical efforts are 3.4% 96.6% characterized by their Rarely: 23.2% Rarely: 76.8% unexpectedness. This may actually trigger LBP through sudden strain of Spine rotation: soft tissues, possibly caught in a Spine Often: 87.9% condition or posture < optimal for this rotation: Sometimes: kind of effort. Often: 12.1% 78% Sometimes: 22.0% Rarely: 89.7% Rarely: 10.3% Sudden maximal Sudden efforts: maximal Often: 82% Sudden efforts: Sometimes: maximal Often: 18.0% 88.7% efforts and Sometimes: 11.3% Rarely: 89% LBP: 1.65 1.3-2.1 Rarely: 10.9%

(Continued)

6-79

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Marras et al. Cross- 403 industrial jobs Outcome: Existing medical Maximum load Participation rate: Numbers and 1993 sectional from 48 and injury records in each moment: 73.65 Nm proportions of those sampled by job manufacturing industry were examined for 23.64 Nm 5.17 3.19-8.38 group. No information on number of Marras et al. companies: e.g., each job to determine if Sagittal mean individual participants. 1995 automobile workers on those jobs had velocity: assembly, food reported work-related low- 11.74 E/sec Study provides linkage between processing, lumber back disorders. The result 6.55 E/sec 3.33 2.17-5.11 epidemiologic measures of injury (i.e., and wood, yielded an outcome Maximum weight: “probabilities of high-risk LBD group construction, metal measure of “LBD risk,” 104 N 23.3 lb Maximum membership”) and select and paper which was a normalized weight: biomechanical and task factors for production, printing, rate of work-related LBD. 37 N 8.3 lb 3.17 2.19-4.58 repetitive lifting jobs. and rubber production. No data Exposure: A triaxial Study illustrates multi-factored nature provided on the electrogoniometer was of injury risk, but it does not indicate number of workers worn by workers to record the risk of LBD. in study. position, velocity and acceleration of the lumbar Quality and accuracy of injury and spine while workers lifted medical records are unknown. in either “high” or “low” risk Inaccuracies or underreporting would jobs. Workplace and affect the accuracy of the model. individual characteristics were recorded. High risk Exposure assessors may not have exposed was >12% injury been blinded to risk status of jobs rate, yielding 111 high risk they were evaluating. jobs, while 124 jobs were low risk, serving as the control group.

(Continued)

6-80

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Masset and Cross- Steel workers Outcome: Interview-based Lifetime LBP N/A Õ Participation rate: 90%. Malchaire sectional (n=618). checklist and questionnaire: prevalence for all 1994 Back pain defined for three workers: Low back fatigue accounted for 25% All male and all periods: (1) during lifetime, 66% LBP cases. under 40 years of (2) past 12 months, and (3) Vehicle age. past 7 days by the 1-year LBP driving: No objective measure of workload. question, “Did you have prevalence for all 1.15 <0.005 Stratified by age and exposure risk any problems in the lower workers: level. back?” 50% Heavy Ergonomic redesign prior to study, Exposure: Interview-based 1-week LBP efforts of reduced ergonomic hazards. exposure assessment prevalence: the using checklist: postures 25% shoulder: Physical workload, posture, and movements of the 1.62 <0.01 movements of the trunk, repetition, negative perception of working trunk, efforts, physical and Prevalence of environment, exposure to WBV, not psychosocial environment sciatica was low: associated with back pain. (monotony, responsibility), 2-3% Seated vehicular driving and posture: Information obtained included exposure to whole body 1.46 0.09 demographics, height, weight, medical vibration. history, personality, and social status (smoking, sports, satisfaction with family and occupation, abnormal fatigue, temper, headache, depressive tendency, present and past working environment. All long-lasting sick workers excluded from study; may cause survivor bias. Back “fatigue” separated from “back pain.” This cross-sectional study was first part of a prospective study. Heavy efforts with shoulders were strongly correlated with LBP.

(Continued)

6-81

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Partridge and Cross- 206 male civil Outcome: Low-back pain Dockers: current Civil servants: RR=1.27 0.98-1.64 Participation rate: 95.7% for dockers Duthie 1968 sectional servants (clerical (including lumbar disc rheumatic current and 91.0% for civil servants. workers), age 15 disease, pelvic girdle pain, symptoms: 43.2% rheumatic to 64 years, and 171 and leg pain). symptoms: Analyses corrected for age. male dock workers, Low-back pain, 61 34.5% age 25 to 64 years. Participants attended an dockers Overall complaint rates did not differ interview at which time a (Standardized Ratio Low-back between occupations, despite medical and social (SR) by age 106.1) pain, 33 civil differences in physical effort questionnaire was servants (SR requirements. Older civil servants administered and a medical 90.4) complained of more neck/shoulder examination was pain than dockers of a similar age. performed. Difference attributed to static working postures involving the neck and Complaints classified into 8 shoulder. categories. Among civil servants, only 5 weeks Exposure: Based on job (16.1%) of sickness absence in title (civil servant or previous year due to back pain. docker). Among dockers, 75 weeks (68%) of work lost attributed to lumbar disc disease and backache. Authors conclude that there is a positive correlation between the heaviness of work and time lost due to back complaints, even if the complaint rate in different occupations does not vary significantly. Medical examiners probably not blinded to exposure status.

(Continued)

6-82

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Punnett et al. Case- 219 automotive Outcome: Back pain 84% (185) 20 workers Non-neutral Participation rate: 84%. 1991 referent assembly workers. cases: (interview and unexposed postures: (retro- exam) defined as workers 4.9 1.4-17.4 Healthy worker effect. spective) 95 cases compared who filed new reports of to 124 referents back disorders at plant Mild flexion: Of the 124 referents, only 20 workers without back pain. during a 10-month period. 5.7 1.6-20.4 were unexposed to all awkward Back pain in interview postures. defined as history of $ 3 Severe episodes or $ one episode flexion: 5.9 1.6-21.4 Back disorders were found to be lasting $ one week within associated non-neutral trunk the year preceding the date Time in non- postures. of the interview. neutral posture: 69% of subjects in job <5 years. Physical exam consisted 8.09 1.5-44.0 of active, passive, and Questionnaire involved demographics, resisted motions Lift 44.5N: work history, medical history, and concentrating 11 ranges 2.16 1.0-4.7 non-occupational activities. of motion of the back. Age Analyses controlled for gender, age, Referents: No report of (years): length of employment, recreational back disorders. 0.96 0.9-1.0 activity, medical history, and maximum weight lifted in study job. Exposure: Based on video Back injury: analysis of job postures 2.37 1.3-4.3 Exposure variable for non-neutral and bio-mechanical data posture: The sum of the duration spent in non-neutral postures as a continuous variable. A strong trend found for increasing length of exposure and risk of back disorders to both mild and severe trunk flexion. Only current job analyzed: Assumes short-term relationship between outcome and exposure (however, also included duration of employment variables).

(Continued)

6-83

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Riihimäki et al. Cross- Longshoremen, Outcome: Back pain Longshoremen Office Participation rate: $70%. 1989a sectional mail earth moving symptoms, by (n=542), earth workers survey equipment operators questionnaire. movers (n=311), (n=674) Longshoremen and earthmovers (WBV), carpenters and carpenters combined in analysis (machine (heavy physical Exposure: Job title and (n=696) operators). work), and office questionnaire responses workers (sedentary regarding work history, Sciatic pain and After adjustment for age, duration of work) (n=2,223) physical work factors, and machine operators 1.3 1.1-1.7 employment was not associated with work stress. symptoms in any group. Sciatic pain and carpenters 1.0 0.8-1.3 Of the three back symptoms, sciatica, lumbago, and LBP, sciatica Sciatica and discriminated the best among twisted or bent occupational groups. postures 1.5 1.2-1.9 All three exposed groups were Sciatica and annual exposed to $ one work-related risk driving 1.1 0.9-1.4 factor for back disorder.

(Continued)

6-84

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Riihimäki et al. Cross- 216 concrete Outcome: Radiographically Grade 2 to 3 Grade 2 to 3 Participation rate: 84% concrete 1989b sectional workers compared detectable degenerative disc problem: disc problem: N/A p=0.001 workers and 86% house painters. to 201 house changes in lumbar region. painters (all male), 27.8% concrete 15.4% house Occupa-tion Examiners (radiologists) blinded to age-matched. Exposure: Based on job workers painters effect of case or exposure status. Restricted to title (article refers to concrete workers with 5 Wickström [1985] Back problems: Back work: Age, self-reported back accidents, years work evaluation of concrete 55% problems: OR=1.8 1.2-2.5 body mass index, height, and smoking experience and to reinforcement workers). 45% controlled for in analysis. workers <55 years. Age: Sciatic: Sciatic: OR=6.5 1.7-26 Height, weight, smoking no effect on 53% 39% degenerative X-ray changes. Spondy- lophytes Negative bias for occupational factor due to healthy worker effect. Occupa-tion effect of Positive bias due to recall for concrete identifying accidents as risk factors. work: OR=1.6 1.2-2.3 Individual exposure data not available for workers. Age: OR=14.9 2.3-95 Radiographically detectable degenerative changes associated with sciatic pain (1.0, 1.4, 1.9) for three grades of degeneration (not for LBP or lumbage).

No hypotheses regarding specific risk factors. Exposure assessed by job title only.

(Continued)

6-85

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Riihimäki et al. Prospective Machine (heavy Outcome: Based on 2 22% machine 14% office 1.4 0.99-1.87 Participation rate: For follow-up: 81% 1994 (3-years) equipment) Postal questionnaires; operators workers machine operators, 79% carpenters, operators (688), LBP=Low-back symptoms and 89% office workers. Pietri-Taleb carpenters (533), in preceding 7 days, 12 24% carpenters 1.5 1.1-2.1 et al. 1995 and office workers months, and lifetime. Questionnaire included age, level of (591). All males. Sciatic pain = pain radiating education, annual car driving, weekly to leg/s. physical exercise, occupational exposure, and history of other back Exposure: Based on Physical exercise > Maximum problems. specific occupation: once a week physical Machine operators were exercise once Questionnaires administered in 1984 exposed to static loads, a week. 1.26 1.0-1.6 and 1987. low-level, whole body (p<0.06) vibration. Carpenters Separate logistic regression models exposed to dynamic Smokers and ex- Non-smokers 1.29 0.98-1.7 created for specific occupation. physical work. Office smokers p<0.06) workers were sedentary History of other types of low back workers. pain predicted sciatica in all groups. History of lower Questionnaire asked back pain: Monotonous work, problems with co- amount of twisted or bent workers or supervisors, and high- postures, pace of work, Mild LBP; None 2.7 1.7-4.2 paced work were not associated with monotonous work, Severe LBP 4.5 2.7-7.6 sciatica three-year cumulative Incident problems with co-workers (p<0.001) Rate. or superiors, draft, cold, vibration. Article examines only sciatic pain.

(Continued)

6-86

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Ryden et al. Case-control Cases consisted of Outcome: Reported work- Low-back Previous Participation rate: Not reported. 1989 84 employees with related low-back injuries pain: OR=2.27 back injury: back injuries and while employed at the site OR=2.13 1.07-4.24 Disadvantages of the design include: 168 controls of the study during the time a lack of detailed information that (matched triplets). Mean age was 34 period of 1983 through Working could have helped to focus on and 83.3% were 1985. day shift: selected risk factors. For example, female. OR=2.23 1.28-3.89 knowledge of pack-years rather than Exposures: History of only number of cigarettes smoked/day Cases: Employees previous back injury at Low back would have been valuable, if with injuries from work, work shift, heavy pain: available, as would more specific job-related activities that occurred during work, lifting, bending, OR=2.27 1.25-4.12 information on body build, including the working day, slipping, self-reported low- percent body fat and fitness level, not based on back pain or “slipped disc,” Self-report rather than using height/weight and individual lost time and individual risk factors. slip disc: self-reported exercise level. from the job or OR=6.20 2.64-14.57 workers’ Advantages of the design included compensation. The economy, time savings, flexibility, and incidence rate at the work site during the the analysis of a large group of risk study period was factors simultaneously. 29/1,000 in 1983, 29/1,000 in 1984 and Immediate reporting of injuries, 33/1,000 in 1985. including the nature of the injury and pertinent data regarding where and Controls selected from the same how the injuries occurred, is essential population by age, to efforts both to reduce injuries and sex, and to rehabilitate those who are injured. department. For each case, two Cases and controls were (over) controls were matched on occupation risk factors. selected from a list Could not examine these effects. of all employees, stratified by department. Those working day shift felt to have Matching for age greater physical demands. was done within a 5-year span.

(Continued)

6-87

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments Schibye et al. Longitudinal Follow-up of 303 Outcome: Based on Nordic Prevalences of LBP Participation rate: 1985: 94%; 1995 sewing machine Questionnaire: pain in the in Sewing jobs: 1991: 86%. All participants were operators at nine last 12 months in the low females. factories back (last 7 days). 12-month: LBP: representing 1985=38% 77 of 241 workers still operated a different technology Exposure: Assessed by 1991=47% sewing machine in 1991. levels who questions regarding: completed (1) type of machine Prevalences 82 workers had another job in 1991 questionnaire in operated, (2) work 1-week: LBP: among those 35 years or below, 77% 1985. organization, (3) workplace 1985=23% had left job; among those above 35 design, (4) units 1991=25% years 57% left job. In April 1991, 241 of produced/day, (5) payment 279 traced workers system, and (6) time of 20% reported musculoskeletal responded to same employment as a sewing symptoms as the only reason for questionnaire. machine operator. leaving job. Healthy worker effect. Another 13% said symptoms were part of the reason.

No significant changes in prevalences among those employed as sewing machine operators from 1985 to 1991; significant decrease in those who changed employment.

As many as 50% of respondents reported a change in the response to positive or negative symptoms from 1985 to 1991.

This was due to a decrease in the risk factors: e.g., decreased in output and hrs worked/week.

Article examines only neck/shoulder area in detail (no exposure analyses for back outcome).

(Continued)

6-88

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Skov et al. Cross- 1,306 Danish Outcome: Musculoskeletal Danish No unexposed Participation rate: Not reported. 1996 sectional salespersons symptoms, by salespersons group included questionnaire. (n=1,306) Covariates considered in multivariate analyses included age, sex, height, Exposure: Self-reported Annual driving Annual weight, smoking, work-related driving distance, time in distance driving psychosocial variables, lifting, leisure sedentary work, lifting of distance, time sports activities. heavy loads, psychosocial highest job characteristics. category: No unexposed group was included. OR=2.79 1.5-5.1

Sedentary work (% Sedentary of worktime) work (% of worktime) highest category: OR=2.45 1.2-4.9

(Continued)

6-89

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Skovron et al. Cross- 4,000 random- Outcome: Based on back Point prevalence Participation rate: 86%. 1994 sectional stratified sampled pain symptom reporting LBP: 33% adults in Belgium; a from structured interviews. Information included age, gender, bicultural country, Back pain defined by Lifetime social class, habitat, language, uniform health care question “Have you ever prevalence: 59% working status, occupation, work system; 48% male. had back pain?” Cases satisfaction, lifestyle factors, and restricted to those subjects Among workers family history. Population-based currently working. occupation: NS telephone survey. Logistic regression models controlled Exposure: Based on Work for age, and gender; interaction interview data: occupation dissatisfaction: 2.4 p=0.02 tested. and working status, “Are you satisfied with work” First episode of back pain not question.. Female gender: 2.16 p=0.001 associated with work satisfaction. Increasing age: 2.0 p=0.001 Language influence reporting of first time occurrence and history of back pain but not severity of impairment as expressed as daily back pain.

Uniform health care assured equal access and reporting.

Results suggest that work satisfaction is not a cause of LBP, but it intervenes in the expression of LBP.

(Continued)

6-90

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Svensson and Cross- Random sample of Outcome: Low-back pain Univariate Participation rate: Approximately 80% Andersson sectional 1,760 38 to 64-year- (LBP) was defined as all analysis of the final sample of 1,746 females 1989 old females from conditions of pain, ache, found participated in the study. Goteborg, Sweden. stiffness, or fatigue significant At the time of the localized to the lower back. correlations The analysis of correlations between investigation, 14 All episodes of LBP were between the occurrence of LBP and the females could not be included in the study, as LBP and 5 different variables describing work located. determined by exposures history, work environment, and stress questionnaire. in ages 50- was restricted to wage-earning Approximately 80% 64 years: females only (sick-listed included). of the final sample of Exposure: Variables More 1,746 females included working hr, bending, p<0.05 No significant differences existed participated in the working hr/week, amount lifting, p<0.01 between the two age groups study. of overtime, lifting, standing, p<0.01 concerning the incidence and frequency of forward higher prevalence rates of LBP. However, bending and twisting, work degree of several parameters indicated that the posture, possibility to worry, p<0.01 LBP in the older age group was more change work posture, need and severe. to concentrate, monotony, exhaustion satisfaction with work at the end Several of the correlations in the tasks, possibility to take of the work univariate analysis, when tested in rest breaks, worried and day. p<0.0001 the covariate analysis, were found to tense after work, fatigued be dependent on other confounding at the end of the work day, factors. and education. The findings in the present study Exposed and unexposed stress the importance of were determined by psychological factors in relation to questionnaire responses. low-back pain. These factors are probably not only related to the individual’s personality but also to the type of work and the environment at the workplace.

Medical examiners discussed questionnaires with participants—not blinded.

(Continued)

6-91

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Toroptsova Cross- 701 random- Outcome: LBP history from Frequent trunk No trunk 1.66 p<0.01 Participation rate: 88%. et al. 1995 sectional stratified sampled structured interviews. flexion flexion employees of a Back pain defined as pain Analysis did not control for Russian machine lasting in area below 12th confounders. building plant 47% rib and above gluteal folds. Frequent lifting Occasional 1.43 p<0.05 male. All persons with LBP required in job lifting (2 or Information included personal data, complaints examined by less/day) family status, education, profession, rheumatologist. anthropometric data, smoking, sport activity, and professional factors. Exposure: Based on interview data: Work, Lifetime prevalence: 48%. Prevalence sports, and personal higher among older workers and factors. 10 industrial smokers >10/day. factors examined: Lifting, standing, sitting, walking, Back pain decreased in group vibration, static work, >55 years. The year of retirement for postures, repetitive work, females. and bending. No association with sitting or standing postures, walking, vibration, static work postures, and repetitive work.

(Continued)

6-92

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Undeutsch Cross- 366 male cargo Outcome: Standardized Prevalence of N/A N/A N/A Participation rate: Not reported (46% et al. 1982 sectional transport workers at interview administered to all previous back of target population included). a large airport. workers to detect complaints: 56% (Baggage handlers). subjective previous and Current back symptoms positively present back symptoms. Prevalence of correlated with height, age, and length Clinical orthopaedic present back of experience in transport work. examination administered to symptoms: 66% 134 workers to detect Among workers with present objective findings. Prevalence of symptoms, symptoms occurred most objective back frequently during lifting of loads (75%) Exposure: Data on work findings at and while in bended body positions experience in the present examination: 70% (61%). Changing body position (71%) occupation was collected. and absence of work for one or more No other exposure data days were relieving factors for back collected. symptoms.

Comparison of interview and clinical exam results show interview to be a suitable screening method for clinical back pain (sensitivity=86%, specificity=31%).

Significant association between length of transport work and back symptoms (p=0.035) adjusted for age.

No heterogeneity with regard to exposure.

(Continued)

6-93

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Videman et al. Cross- 562 nurses and 318 Outcome: Based on results 85% of aides had $ 79% of $ one “life- Participation rate: 88% nurses; 85% 1984 sectional nursing aides in from a pre-tested one “life-time” nurses had time” nurses aides. Finland, all of them questionnaire and from episode of LBP and experienced episode of females. health information obtained their point $ one “life- LBP: 1.1 1.01-1.14 Workers with back pain were employed in heavy jobs on average 1 from the local Pension prevalence was time” episode year longer than those with no Registers that were used 50% for LBP. of LBP; point previous LBP. to identify nurses who had prevalence been pensioned due to ill was 41% for Musculoskeletal disorders as a cause health during a 4-year LBP of disability increased with age; the period immediately 30-years risk for 25-years old aides was 3.4 times greater than for the preceding the mailing of the Sciatica: 43% life- Sciatica: 38% nurses; similar results for sciatica questionnaire. time prevalence. life-time with a risk of 4.5 times greater for the prevalence aides than nurses. Exposure: Based on self- assessments from data Aides had twice The prevalence of LBP and sciatic obtained using a mailed the lifting, bending symptoms in both nurses and in aides questionnaire that included and rotation. are high and similar to the results found in Britain. nine questions on physical loading factors at work and Physical workload related to patient seven questions on work handling was mainly responsible for history and occupation. the differences in LBP and sciatica rates between the aides and nurses. Jobs were reclassified as The finding was most evident under heavy, intermediate, and the age of 30 years. light based on results of Non work-related factors, such as questionnaire items dealing childbirth, also contributed to the with workload. adverse back conditions. Study lacks a good unexposed population since both nurses and aides were exposed to varying degrees of risk factors for LBP and sciatica. Workers with LBP were in heavier jobs for longer time than those without LBP.

(Continued)

6-94

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Videman et al. Cross- From a Finnish Outcome: Objective 54% of heavy 10% of Heavy vs. Participation rate: Not reported. 1990 sectional workforce of 86 radiologically and workers had LBP sedentary Mixed: males who had discography-based often, and 36% had workers had 2.7 1.1-6.2 Strength: First study linking pathologic worked in four pathologic criteria from the sciatica LBP often, and data with history of occupation and distinct occupational cadaver spines of the 19% had Driving vs. physical loading factors. groups: Sedentary, study population. Degree of sciatica Mixed: Mixed, Driving, and degeneration was outcome 2.3 0.8-6.2 Weakness: Do not know the temporal Heavy. Criteria for measure, i.e., annular 50% of drivers had 29% of mixed pattern in development of the inclusion: Deceased ruptures. Information on LBP often, and group had LBP Sciatica: pathologic changes. below the age of 64 symptoms was obtained 29% of them had often, and NS who had been from family members. sciatica 10% had Possible selection bias due to potential employed before sciatica differential rates between work death and the Exposure: Type of work, Heavy physical groups in leaving jobs because of subjects’ family able based on work history load vs. not: degenerative diseases. to provide working reports from family; OR=2.8 information. classification of work Two important findings: Sedentary or based on heaviness, Sedentary vs. not: heavy work contribute to the Exclusion criteria driving, and sedentary jobs. OR=24.6 development of pathologic findings in were long illnesses Classification based on (symmetric disc spine. Severity of back pain was or a diseased state, physically heaviest degeneration) related to the heaviness of work, i.e., such as cancer or occupation held for $ 5 work factors responsible for infectious disease. years. development of pathologic changes and for the production of pain.

Back pain more common with physically more loading occupations; p<0.001. Similar but weaker trend between loading and sciatica; p=0.03.

General: p<0.01 between groups for back pain; and p<0.07 for sciatica.

Relationships were observed between report of symptoms and disc pathology; also, exposures and disc pathology.

(Continued)

6-95

Table 6-6 (Continued). Epidemiologic studies evaluating back musculoskeletal disorders

MSD prevalence Study Study Outcome and Exposed Referent RR, OR, Study design population exposure workers group or PRR 95% CI Comments

Walsh et al. Cross- A postal Outcome: Self-reported Lifetime incidence Participation rate: 436 questionnaires 1989 sectional questionnaire was low-back pain, by of LBP was 63%. were returned, giving an overall sent to a random interview. response rate of 81%. sample of 267 males Recent Occup. The association with use of vibrating and 268 females in Exposure: Standing or Activity: machinery among females (repetitive the age range of 20 walking for > 2 hr; sitting Males risk=5.7) was based on only one to 70 who lived in for > 2 hr; driving a car or Driving>4hr/d RR=1.7 1.0-2.9 exposed case. Whitchurch, van for > 4 hr; driving a Lifting 25kg RR=2.0 1.3-3.1 Cases of low-back pain were England. truck, tractor or digger; Females ascertained solely on the basis of lifting or moving weights of Lifting 25kg RR=2.0 1.1-3.7 reported symptoms. Four hundred, thirty- 25kg or more by hand; or Successive birth cohorts reported the six questionnaires using hand held vibrating Lifetime Occup. development of low-back pain at any were returned, machinery were the Activity: given age with increasing frequency. giving an overall exposures of interest. Males Driving a car for >4 hr a day was response rate of Lifting 25kg RR=1.5 1.0-2.4 associated with low-back pain in 81%. Lifetime occupational Females males but not with low-back pain in history obtained by Sit >2hr/d RR=1.7 1.1-2.6 females. interview. Vib. machine RR=5.7 1.1-29.3 Authors believe the data give strong support for a role of regular heavy Risk of unremitting lifting in the etiology of low-back pain LBP: and add weight to the evidence Males implicating occupational driving as a risk factor. At the same time, Lifting 25kg RR=5.3 1.3-20.9 however, they suggest that such Females activities account for only a small Lifting 25kg RR=2.9 0.8-10.2 proportion of the total burden of low- back pain in the general population. Author’s estimates of the fraction of disease attributable to heavy lifting and car driving are 14 and 4%, respectively, leaving a substantial proportion of cases unexplained. Authors attempted to recreate a retrospective cohort design; asked participants to remember dates and jobs and LBP. Questionable recall for temporal relationships.

6-96 CHAPTER 7 Work-Related Musculoskeletal Disorders and Psychosocial Factors

SUMMARY While the etiologic mechanisms are poorly understood, there is increasing evidence that psychosocial factors related to the job and work environment play a role in the development of work-related musculoskeletal disorders (MSDs) of the upper extremity and back. Though the findings of the studies reviewed are not entirely consistent, they suggest that perceptions of intensified workload, monotonous work, limited job control, low job clarity, and low social support are associated with various work-related musculoskeletal disorders.

As some of these factors are seemingly unrelated to physical demands, and a number of studies have found associations even after adjusting for physical demands, the effects of these factors on MSDs may be, in part or entirely, independent of physical factors. It is also evident that these associations are not limited to particular types of jobs (e.g., video display terminal work [VDT]) or work environments (e.g., offices) but, rather, seem to be found in a variety of work situations. This seems to suggest that psychosocial factors may represent generalized risk factors for work-related MSDs. These factors, while statistically significant in some studies, generally have only modest strength.

At present, two of the difficulties in determining the relative importance of the physical and psychosocial factors are: (1) psychosocial factors are usually measured at the individual level, while physical factors are more often measured at the group (e.g., job or task) level and often by methods with limited precision or accuracy and (2) “objective measures" of aspects of the psychosocial work environment are difficult to develop and are rarely used, while objective methods to measure the physical environment are more readily available. Until we can measure most workplace and individual variables with more comparable techniques, it will be hard to determine precisely their relative importance.

INTRODUCTION There is considerable confusion regarding the contribution of psychosocial factors to that fall within three separate domains: musculoskeletal illness and injury. Because of (1) factors associated with the job and work this, it is examined in this separate section of the environment, (2) factors associated with the report. Unlike the more finite (and generally extra-work environment, and more familiar) range of physical factors (e.g., (3) characteristics of the individual worker. force, repetition, and posture), the concept of Interactions among factors within each of these psychosocial factors includes a vast array of domains constitute what is referred to as a conditions. Indeed, the term “psychosocial” is “stress process,” the results of which are commonly used in the occupational health arena thought to impact upon both health status and as a catchall term job performance [Bongers and deWinter 1992; to describe a very large number of factors ILO 1986; Sauter and Swanson 1996; WHO 1989].

7-1 Included in the domain of job and work conditions. In particular, both personal and environment are a host of conditions, situational characteristics may lead to sometimes referred to as “work organization differences in the way individuals exposed to factors,” which include various aspects of job the same job and work environment perceive content (e.g., workload, repetitiveness, job and/or react to the situation [Hurrell and control, mental demands, job clarity, etc.); Murphy 1992]. Recent theoretical models of organizational characteristics (e.g., tall versus the relationship between psychosocial factors flat organizational structures, communications and MSDs [Bongers et al. 1993; Sauter and issues); interpersonal relationships at work Swanson 1996] clearly reflect the complexity (e.g., supervisor-employee relationships, social and multifactorial nature of the problem. support); temporal aspects of the work and task (e.g., cycle time and shift work); financial In general, four plausible types of explanations and economic aspects (e.g., pay, benefit, and have been suggested to account for equity issues); community aspects (e.g., associations between work-related occupational prestige and status). These work psychosocial factors and MSDs [Bergqvist and job environment factors are often thought 1984; Bongers et al. 1993; Bernard et al. of as demands, or “risk factors,” that may pose 1993; Sauter and Swanson 1996; Sauter et al. a threat to health [Hurrell and Murphy 1992]. 1983; Ursin et al. 1988]. First, psychosocial Extra-work environment parameters typically demands may produce increased muscle include factors associated with demands arising tension and exacerbate task-related from roles outside of work, such as biomechanical strain. Second, psychosocial responsibilities associated with a parent, demands may affect awareness and reporting of spouse, or children. Finally, individual worker musculoskeletal symptoms, and/or perceptions factors are generally of three types [Payne of their cause. Within this second explanation 1988] corresponding to: genetic factors (e.g., may fall the “perverse incentive” view, in which gender and intelligence); acquired aspects (e.g., societies may provide workers with systems social class, culture, educational status); and (such as workers' compensation) that may lead dispositional factors (e.g., personality traits, and to overreporting of MSD symptoms [Frank et characteristics and attitudes such as life and job al. 1995]. Third, initial episodes of pain based satisfaction). on a physical insult may trigger a chronic nervous system dysfunction, physiological as PSYCHOSOCIAL PATHWAYS well as psychological, which perpetuates a The purpose of this discussion is to summarize chronic pain process. Finally, in some work research evidence linking work-related situations, changes in psychosocial demands psychosocial factors, as described above, to may be associated with changes in physical MSDs of the neck, shoulder, elbow, demands and biomechanical stresses, and thus hand/wrist, and back. It should be recognized associations between psychosocial demands at the outset, however, that the linkages and MSDs occur through either a causal or between work-related psychosocial factors and effect-modifying relationship. health outcomes of all varieties are often complex and influenced by a multitude of

7-2 The research evidence reviewed in the UPPER-EXTREMITY DISORDERS following discussion is organized into two (NECK, SHOULDER, ELBOW, HAND separate sections. The first section includes AND WRIST) studies of disorders of the neck, shoulder, Individual and Extra-Work elbow, hand and wrist which are discussed Environment Factors under the rubric of “upper extremity disorders.” This convention was adopted because many of A variety of psychosocial factors associated the studies utilize measures which combine with both the individual worker and extra-work symptoms associated with several upper environment have been linked to upper extremity body areas (e.g., neck and shoulder), extremity MSDs [Sauter and Swanson 1996; and it is therefore not possible in reviewing Bongers and deWinter 1992; Bongers et al. these studies to isolate the effects of the 1993]. These factors have included such psychosocial variables under consideration on conditions as depression and anxiety [Helliwell more specific areas. The second section et al. 1992], symptoms of psychological examines studies of back disorders. distress [Leino 1989], and home problems Associations reported in this review are [Karasek et al. 1987]. The connection between statistically significant in nearly all cases (at the factors of this nature and the job and work p<0.05 level and frequently also at the p<0.01 environment, however, is unclear. While level). Where possible, odds ratios (ORs) are affective problems (such as anxiety and also reported. depression) and symptoms of distress may certainly be a consequence of the work The studies examined in this review are situation, they may also be causally related to summarized in Tables 7-1 and 7-2. In non-work circumstances only. Likewise, while interpreting the studies reviewed, it is necessary extra-work environment conditions (e.g., to be aware that, in general, researchers have “home problems") may be exacerbated by the not used standardized methods for assessing work situation (e.g., shift work) their “work- psychosocial factors in relationship to MSDs. relatedness” remains unclear. Because of the Thus, individual psychosocial factors assessed uncertainty regarding the work-relatedness of by investigators vary from study to study. these individual and extra-work environment Moreover, even when work-related factors (and because discussions can be found psychosocial factors (e.g., workload , job in other sources), only the individual control, social support, job satisfaction, etc.) psychosocial factor, job dissatisfaction, is included by various investigators are the same examined here. or similar, they may be measured by different Job Dissatisfaction methods and different kinds of scales which can vary in psychometric quality. These A number of studies suggest associations methodological limitations complicate the between low levels of satisfaction with work process of drawing definitive conclusions and upper extremity musculoskeletal symptoms regarding the literature as a whole and when and disorders. Tola et al. [1988], for example, comparing results between studies, one must in a study of 1,174 machine operators, 1,054 take these differences into account. carpenters, and 1,013 office workers, found an

7-3 association (OR 1.2) between job Positive associations with upper extremity dissatisfaction and neck and shoulder physical disorders have also been found in studies using findings or symptoms, after adjusting for measures of perceived work pressure and confounders. Likewise, Hopkins [1990] workload. High levels of perceived workload, reported a positive association between job for example, were found to be positively dissatisfaction and musculoskeletal symptoms. associated with musculoskeletal symptoms in However, low job satisfaction was not found to the Pot et al. [1987] and Theorell et al. [1991] predict neck and shoulder problems one year studies (which adjusted for physical demands later in a study of 154 Finnish workers [Viikari- such as lifting and awkward postures) reported Juntura et al. 1991a]. Likewise, in a study of above. Kvarnström and Halden [1983], in a 273 nursing aids employed in a geriatric case control study of 112 cases and 112 age- hospital [Dehlin and Berg 1977] job satisfaction and sex-matched controls from an engineering was found to be unrelated to reports of ever firm, found sick leave due to fatigue or shoulder having cervical pain. muscle soreness to be positively associated with high perceived workload. Karasek et al. Job and Work Environment Factors [1987], in a study of 8,700 full-time members

Intensified Workload of the Swedish white collar labor union federation, found perceived workload to be One of the factors most consistently associated positively associated with musculoskeletal with upper extremity MSDs has been the aches as measured by a combination of several perception of an intensified workload, as questions (OR 1.1 for males, 1.2 for females). measured by indices of perceived time Likewise, Sauter et al. [1983], in a study of pressure, workload, work pressure, and 248 VDT users, found perceived workload and workload variability. Pot et al. [1987], for demands for attention to be associated with example, in a cross-sectional study of 222 neck, back, and shoulder discomfort after VDT operators, found high levels of perceived adjusting for a wide variety of variables time pressure associated with the reporting of denoting physical demands. Bernard et al. upper extremity musculoskeletal complaints. [1993], in a study of 1,050 newspaper Kompier [1988] found perceived time pressure employees, found perceived increased to be associated with upper extremity workload demands (increased time working complaints (in the preceding 12 months) among under deadline and increased job pressure) to some 158 male bus drivers. Likewise, Takala be positively associated with neck, shoulder, et al. [1991], in a longitudinal study of 351 and hand-wrist symptoms. Similarly, Hales et female bank cashiers, reported a positive al. [1994], in a study of 553 association between perceived time pressure telecommunications workers, found increased and symptoms of the neck and shoulder after work pressure to be associated with neck (OR adjusting for postural load. Theorell et al. 1.2) and upper extremity [1991], however, in a sample of some 206 (OR 1.1) disorders, as defined by physical workers from six occupations, found that examination and questionnaire. Ryan and perceived time pressure was not significantly Bampton [1988], using a total sample of 143 correlated with neck or shoulder symptoms. data processors, compared 41 individuals

7-4 reporting a number of neck symptoms to 28 that self-reports of “being bored most of the reporting very few neck symptoms (middle time” were highly (OR 7.7) associated with group left out) and found a positive association neck symptoms. Likewise, Linton [1990], in a between symptom reports and reports of study of approximately 22,200 Swedish having to push themselves (OR 3.9). Ekberg workers undergoing a screening examination by et al. [1994] compared 109 workers who the occupational health care service, found that consulted a physician for new musculoskeletal monotonous work was positively associated neck and shoulder disorders with 637 controls with neck/shoulder pain (OR 2.3) during the and found a positive association (OR 3.5) with preceding year. Ekberg et al. [1994], in the rushed work pace. Houtman et al. [1994], in a study described above, found an association representative sample of 5,865 workers in the between “low quality work” (lacking stimulation Netherlands, found reported high work pace and variation) and neck and shoulder problems associated with muscle or joint symptoms (OR (OR 2.6). Similarly, Kvarnström and Halden 1.3) after adjusting for physical stressors and [1983] in the case control study described modifying personal characteristics. However, above, found monotonous work to be Dehlin and Berg [1977] in the study described associated with sick leave due to fatigue or above, found no relationship between reports tenderness in the shoulder muscles. Finally, of high perceived physical and psychological Hopkins [1990] in a study of around 280 demands and reports of ever having pain in the clerical workers found high levels of boredom cervical region. Finally, Houtman et al. [1994], to be associated with musculoskeletal in a representative sample of 5,865 workers in symptoms (in any part of the body) during the Netherlands, found reported high work work hours. pace associated with muscle or joint symptoms (OR 1.29) after adjusting for physical stressors Job Control and modifying personal characteristics. Numerous studies have reported positive associations between limited job control or Variability in workload (surges in workload) autonomy at work and upper extremity has also been linked to upper extremity problems. These include neck symptoms [Ryan disorders. The studies by Hales et al. [1994] of and Bamptom 1988, OR 3.9; Hales et al. 553 telecommunication workers and Hoekstra 1994, OR 1.6], neck/back/shoulder symptoms et al. [1994] of some 108 teleservice [Sauter et al. 1983; Theorell et al. 1991], representatives, found perceived workload musculoskeletal aches [Karasek et al. 1987], variability to be associated with elbow (OR and muscle/joint symptoms [Hopkins 1990; 1.2) and neck (OR 1.2) disorders, but not with Houtman et al. 1994]. The study by Pot et al. shoulder or hand disorders. [1987], however, failed to support this relationship. Monotonous Work Monotonous work has been positively linked to Job Clarity the prevalence of upper extremity symptoms in A number of studies, including those of Ryan various studies. In a study of 143 data and Bamptom [1988], Karasek et al. [1987], processors, Ryan and Bamptom [1988] found

7-5 and Ekberg et al. [1994], have shown positive Extremities associations between reports of role ambiguity Overall, the epidemiologic studies of upper (uncertainty about job expectations) and upper extremity disorders suggest that certain extremity disorders (particularly neck psychosocial factors (including intensified disorders). Similarly, uncertainty regarding job workload, monotonous work, and low levels of future was found to be predictive of neck and social support) have a positive association with shoulder discomfort [Sauter et al. 1983] and these disorders. Lack of control over the job elbow, neck, and hand/wrist symptoms [Hales and job dissatisfaction also appear to be et al. 1994]. positively associated with upper extremity MSDs, although the data are not as supportive. Social Support Limited social support from supervisors and The evidence for the relationship between coworkers has been found to be positively psychosocial factors and upper extremity associated with a variety of upper extremity disorders appears to be stronger for symptoms. The studies by Pot et al. [1987], neck/shoulder disorders or musculoskeletal Kompier [1988], Hopkins [1990], Sauter et al. symptoms in general than for hand/wrist [1983], and Hales et al. [1994], all support a disorders. This stronger association for positive association. Linton [1990] reported a neck/shoulder disorders may be due to the positive association between neck symptoms following reasons: the large number of studies and limited support from supervisors. Ryan and performed in the Nordic countries which have Bampton [1988] reported an effect of limited focused more on the neck/shoulder MSD support from coworkers (OR 6.7), but not health outcome than a hand/wrist outcome; supervisors, on neck symptoms, while many of the neck/shoulder studies included Kvarnström and Hagberg [1983] reported an numerous psychosocial variables in their effect of limited support from supervisors but models, whereas studies of hand/wrist MSDs not coworkers on sick leave due to shoulder have not, as a rule, included as extensive muscle symptoms. Dehlin and Berg [1977], psychosocial variable testing (therefore the however, found no effect of social support on variables are absent from the risk factor neck/shoulder symptoms, while Theorell et al. models); and the fact that most of the studies [1991] found no effect of social support at with extensive psychosocial scales were in work on neck and shoulder symptoms or office settings, where physical factors may be symptoms of the other joints (with or without less important than psychosocial factors in their adjustment for physical load). Likewise, relationship with MSDs. This finding can be Karasek et al. [1987] found no significant contrasted with studies in heavy industrial association between musculoskeletal aches and settings, where higher exposure to physical social support at work. factors may have

Summary and Conclusions for Upper

7-6 played a greater role than psychosocial factors While there are a number of prospective in the development of MSDs. Also, studies of low back pain and individual physical pathophysiologic processes resulting from factors, there appear to be only a few adverse psychosocial and work organization prospective studies that incorporate individual factors may exert a greater effect on the and extra-work environment psychosocial neck/shoulder musculature to produce factors. Bigos et al. [1991b] defined, in a 4- increased muscle tension and strain than on the year study of 3,020 hourly wage earners at an hand/wrist region. aircraft manufacturing plant, an outcome as reporting a back pain complaint to the company BACK DISORDERS medical department, filing a back-related incident report, or filing an industrial insurance Individual and Extra-Work claim. The psychosocial assessment included Environment Factors personality traits, as measured by the As with upper extremity disorders, a host of Minnesota Multiphasic Personality Inventory psychosocial factors associated with the (MMPI), and limited information on family individual worker (e.g., personality and support, health locus of control, and work psychological status) and extra-work social support. One question about enjoyment environment (e.g., living alone) have been of tasks in the job was also included. Of the 37 linked to back pain and disability [Bongers et variables used to evaluate the role of social al. 1993]. As the “work-relatedness” of these support, health locus of control, and personality factors is unclear and because they have been traits, three were found to be significant in a examined by others (e.g., Bongers [1993]), multivariate analysis. They were Scale 3 of the with the exception of job dissatisfaction MMPI [tendencies towards somatic complaints discussed above, they will not be extensively or denial of emotional distress (relative risk reviewed in this report. In general, these studies [RR]=1.4), dissatisfaction with work (RR=1.7), show clear associations between measures of and prior back pain (RR=1.7)]. Although psychological distress or dysfunction and self- significant, these variables explained only a reported back pain. However, the temporal small fraction of the back pain reports in this relationship between psychological factors and population. The number of back pain reports musculoskeletal symptoms/ disorders remains was three times higher in the group with the unclear. One possibility is that psychological highest scores on these three variables distress is simply a consequence of chronic low compared with the group with the lowest back pain, with no etiologic role in the scores, although only 9% of the work force development of the disorder. Alternatively, it is was in the highest risk group. Because this possible that psychological factors may have study focused on the reporting of back pain some etiologic role in the transition from an complaint and not the actual development of employee with a history of back pain to the back pain, it would be a mistake to generalize status of an unemployed patient with chronic the results to workers developing back pain. back pain, due to fear of re-injury, or other This study suggests factors which would make it impossible to perform the job [Feyer et al. 1992].

7-7 that individual premorbid personality traits only The cross-sectional study by Dehlin and Berg explain a small fraction of work-related lower [1977] of nursing aids described earlier found back problems. an association between dissatisfaction and self- reported back symptoms. However, this study Job Dissatisfaction did not adjust for confounders. Likewise, Job dissatisfaction has been associated with Magora [1973] in a mailed survey study of back disorders in both longitudinal and cross- Israeli workers in 8 occupational categories sectional investigations. Bergenudd and Nilsson found job satisfaction to be associated with [1988], studying some 575 residents of Malmo reports of sick leave due to low back pain. This for over 19 years, found job dissatisfaction to study also did not adjust for potential be associated with self-reported back pain. As confounders. Svensson and Anderson [1989], described above, Bigos et al. [1991b] found a in a cross-sectional study of 1,746 Swedish positive association between job dissatisfaction residents, found an association after adjustment. and workers filing compensation claims for However, in a cross-sectional study by Åstrand back injury. Here, subjects who stated that they [1987] of 391 male Swedish paper company “hardly ever” enjoyed their job tasks were 2.5 workers (clerks and manual workers), no times more likely to report a back injury than association was found between dissatisfaction those who “almost always” enjoyed their job and back disorders, as assessed by symptoms tasks. However, as Frank et al. [1995] point and physical examination after confounder out, some reviewers have argued that the adjustment. airplane manufacturing jobs with the highest levels of dissatisfaction were also the most Job and Work Environment Factors physically demanding. Frank et al. [1995] also Intensified Workload noted that, unfortunately, the extent of the A number of studies have reported associations interaction is difficult to assess because of the between perceptions of intensified workload, as limited measurement of workplace measured by reports of time pressure and high biomechanical exposures in the Bigos et al. work pace, and self-reports of back pain. studies [1986a,b; 1991a,b]. While Heliövaara psychosocial and psychological factors were et al. [1991] in a study of approximately 5,600 assessed at the individual level, workplace Finns, found a composite measure (containing biomechanical factors were assessed only at items on perceived time pressure at work, the group level. Biering-Sorensen et al. [1989], monotony, and fear of mistakes) to be in a one-year follow-up mail survey study of associated (OR 2.0) with back disorders some 928 inhabitants of Denmark (which (defined by interview and physical examination) adjusted for confounders such as previous back after adjusting for potential confounders, pain), also found no association of back pain including physical load and previous back pain. with job dissatisfaction. Because information Lundberg et al. [1989] found perceived time was limited to the use of mailed survey pressure to be associated with perceived back questionnaires, no workplace biomechanical load among 20 workers on a Swedish factors were measured in this study either. assembly line. In a similar vein, Houtman et al.

7-8 [1994], in the study of 5,865 Dutch workers monotonous work is often work which is also across all occupations reported above, found either short-cycled or involves a high static an association (OR 1.21) between reporting (postural) load. high work pace and self-reported back pain (but not chronic back pain problems, defined as Job Control back pain for more than three months or at In the study of teleservice operators cited least three times in the study period) (OR above, Hoekstra et al. [1994], after controlling 1.2). Magora [1973], in the study of Israeli for a number of individual and work-related workers described above, found high levels of factors, found perceived job control at work to concentration to be associated with reports of be inversely associated with back disorders sick leave due to low back pain (OR 2.9). (OR 0.6), that is, the less perceived job control However, Åstrand [1987], found no at work, the higher the odds of back disorders. association between “hustling” and “nerve Likewise, as noted above, Sauter et al. [1983] wracking work” and back pain in male paper found that low job control was related to neck, company workers. back, and shoulder discomfort.

Monotony Social Support Several studies described above [Heliövaara et Bigos et al. [1991b] found a significant al. 1991; Houtman et al. 1994] have reported univariate relationship between limited social associations between perceived monotony and support at work and back trouble. However, reports of back complaints. Svensson and this association was found to be nonsignificant Anderson [1983], in a study of 940 male by the investigators when included in a residents of Goteborg, Sweden, between the multivariate analysis. ages of 40 and 47, similarly found monotonous work (rated “absolutely” or “unacceptably” Summary and Conclusions for boring) to be associated with back complaints. Back Disorders This relationship remained after adjusting for In general, the studies reviewed suggest an several physical factors. However, Svensson association between back disorders and and Anderson [1989] found no relationship perceptions of intensified workload as between monotony and back pain complaints measured by indices of both perceived time among Swedish women in a multivariate pressure and workload. Despite the analysis which included measures of job and considerable differences in the types of task satisfaction. Similarly, in the Houtman et al. methods used to assess both the independent [1994] study, controlling for a combination of and dependent variables, four of the five studies physical stressors (dangerous work, heavy that explicitly included measures of intensified physical load, noise at work, dirty work, and workload found significant associations. It is bad smell at work) reduced the magnitude of also noteworthy that all four of these studies the relationship (for back complaints, the OR attempted to control or adjust for potential decreased from 3.90 to 3.46.) The authors covariates. Five of the seven studies that assess suggest that this may be because job dissatisfaction

7-9 also found positive associations with back of these factors on MSDs may be, in part or disorders. While this evidence is clearly entirely, independent of physical factors. It is suggestive, Biering-Sorensen et al. [1989] also evident that these associations are not found no association in a large-scale limited to particular types of jobs (e.g., VDT one-year follow-up study; while Åstrand work) or work environments (e.g., offices) but, [1987] likewise found no evidence of an rather, seem to be found in a variety of work association among 391 paper workers. Limited situations. This observation seems to suggest support for an association between back that psychosocial factors may represent disorders and low job control is also evident, generalized risk factors for work-related while the evidence for a relationship between MSDs. These factors, while statistically monotonous work and back disorders is mixed. significant in some studies, generally have only Only one study examined the relationship modest strength. between social support and back disorders and found only weak evidence for an association. At present, two of the difficulties in determining the relative importance of the physical and Overall Conclusions psychosocial factors are the following: (1) While the etiologic mechanisms are poorly psychosocial factors are usually measured at understood, there is increasing evidence that the individual level, while physical factors are psychosocial factors related to the job and more often measured at the group (e.g., job or work environment play a role in the task) level and often by methods with limited development of work-related MSDs of the precision or accuracy, and (2) “objective upper extremity and back. Though the findings measures” of aspects of the psychosocial work of the studies reviewed are not entirely environment are difficult to develop and are consistent, they suggest that perceptions of rarely used, while objective methods to intensified workload, monotonous work, limited measure the physical environment are more job control, low job clarity, and low social readily available. Until we can measure most support are associated with various work- workplace and individual variables with more related MSDs. As some of these factors are comparable techniques, it will be hard to seemingly unrelated to physical demands, and a determine precisely their relative importance in number of studies have found associations even the causation of MSDs. after adjusting for physical demands, the effects

7-10

Table 7–1. Summary of studies examining psychosocial factors and upper extremity disorders (neck, shoulder, elbow, hand, and wrist)

Methods Associations with UE outcomes

Worker group Psychosocial MSD Covariate Low Low Low (particip. factor outcome adjust- Job/task Int. Mono. job job social Study rate) Design assessment assessment ments dissat. wkld. work control clarity supp.

Bernard et 1,050 Cross- Self-report MSD case + + al. 1993 newspaper sectional questionnaire definition based workers with job stress on (93%) scales questionnaire

Dehlin and 233 nursing Cross- Self-report Interviews— o o o Berg 1977 aides (85%) sectional questionnaire— pain/ache 7 scales symptoms

Ekberg et al. 109 Cross- Self-report— MD consults for + + 1994 workers vs. sectional modified Nordic MSD disorders 637 controls (case- questionnaire control)

Hales et al. 553 Cross- Self-report Disorders Controlled + + + 1994 telecom- sectional questionnaire based on for extra munications with job stress symptom job factors workers scales questionnaire and MD exam

Hoekstra et 108 Cross- Self-report job MSD case + al. 1994 teleservice sectional stress definition based workers questionnaire on self-report (95%) questionnaire

Hopkins 291 Cross- Self-report Questionnaire + + + + 1990 keyboard sectional questionnaire— symptoms operators items from habits and other of living clerical questionnaire groups

See footnotes at end of table. (Continued)

7-11

Table 7–1(Continued). Summary of studies examining psychosocial factors and upper extremity disorders (neck, shoulder, elbow, hand, and wrist)

Methods Associations with UE outcomes

Worker group Psychosocial MSD Covariate Low Low Low (particip. factor outcome adjust- Job/task Int. Mono. job job social Study rate) Design assessment assessment ments dissat. wkld. work control clarity supp.

Houtman et 5,865 Cross- Self-report work- Symptoms Physical + + al. 1994 workers— sectional living questionnaire stressors general questionnaire — population personal character- istics

Karasek et 8,700 white Cross- Self-report Questionnaire— + + + + al. 1987 collar labor sectional questionnaire musculoskeletal union (random aches members sample) (87%)

Kompier 158 male Cross- Self-report Self report + + 1988 bus drivers sectional questionnaire questionnaire— (73%) complaints and sick leave

Kvarnstrom 224 Cross- Structured Disorders from + + +/o and Halden fabrication sectional interview medical and sick 1983 workers (case- questionnaire absence control) records

Linton 1990 22,200 Cross- Self-report work Pain + + workers— sectional environment general questionnaire population and habits of living questionnaire

Pot et al. 222 VDT Cross- Structured Complaints— +/+ o + 1987 operators sectional interview structured questionnaire interview

See footnotes at end of table. (Continued)

7-12

Table 7–1(Continued). Summary of studies examining psychosocial factors and upper extremity disorders (neck, shoulder, elbow, hand, and wrist)

Methods Associations with UE outcomes

Worker group Psychosocial MSD Covariate Low Low Low (particip. factor outcome adjust- Job/task Int. Mono. job job social Study rate) Design assessment assessment ments dissat. wkld. work control clarity supp.

Ryan and 143 data Cross- Self-report Symptoms + + + + +/o Bampton processors sectional questionnaire— based on MD 1988 (high vs. items from work interview and low environment exam symptoms) scale

Sauter et al. 248 VDT Cross- Self-report Questionnaire— Physical + + + + 1983 users and sectional questionnaire— discomfort work 85 non- work scale demands users (90%) environment (adj.) scale items

Takala et al. 351 bank Longi- Self-report Questionnaire— Postural + 1991 cashiers tudinal questionnaire muscle load (adj.) symptoms

Theorell et 207 Cross- Self-report Questionnaire— Physical +/o + o al. 1991 workers in 6 sectional questionnaire muscle tension load (adj.) occupations symptoms

Tola et al. 1,174 Cross- Mailed Symptoms in o 1988 machinists; sectional questionnaire— last 12 months; 1,034 worker questionnaire carpenters; characteristics and interview 1,013 office workers (67% to 76%)

+ = Significant association found. o = No significant association found. +/+ = Two different measures of factor (e.g., time pressure and workload) found significant. +/o = Mixed results (on factor significantly associated; second factor not significantly associated).

7-13

Table 7–2. Summary of studies examining psychosocial factors and back disorders

Methods Associations with back disorders

Worker group Psychosocial MSD Covariate Low Low (participation factor outcome adjust- Job Int. Mono. job social Study rate) Design assessment assessment ments dissat. wkld. work control supp.

Åstrand 391 workers in Cross- Questionnaire— Interview and MD o o 1987 paper-pulp sectional questions on exam—back pain industry work conditions abnormalities

Bergenudd 575 55-year-old Longi- Interview and Interview reports + and Nilsson city residents tudinal mailed of back pain 1988 (96%) questionnaire

Biering- 928 persons— Longi- Mail Questionnaire— o Sorenson et general tudinal questionnaire back pain in last al. 1989 population 12 months (82%)

Bigos et al. 3,020 male Longi- Questionnaire— Back problems— Control for + o 1991b aircraft plant tudinal Personality medical reports, prior back employees Inventory (MMPI), insurance claims problems (54% with all other questions data)

Dehlin and 233 nursing Cross- Questionnaire— Interview— + Berg 1977 aides (85%) sectional 7 scales, reported pain/ache 52 items symptoms

Heliövaraa 5,600 Cross- Questionnaire— MD exam and Physical Combined et al. 1987 workers— sectional scale assessing interview—back load, prior + + general combined hurried disorders back population work, problems (92%) monotonous work, tight work schedules

See footnotes at end of table. (Continued)

7-127-14

Table 7–2 (Continued). Summary of studies examining psychosocial factor and back disorders

Methods Associations with back disorders

Worker group Psychosocial MSD Covariate Low Low (participation factor outcome adjust- Job Int. Mono. job social Study rate) Design assessment assessment ments dissat. wkld. work control supp.

Hoekstra et 108 teleservice Cross- Job stress MSD case Individual + al. 1994 workers (95%) sectional questionnaire definition based on work questionnaire data factors

Houtman et 5,865 Cross- Questionnaire— Questionnaire Physical + + al. 1994 workers— sectional work living symptoms stressors; general questionnaire personal population survey character- istics

Lundberg et 20 male Cross- Ratings of time Back load ratings + al. 1989 assembly line sectional pressure during during 2-hr work workers 2-hr work period period

Magora 3,316 workers Cross- Questionnaire— Questionnaire— Analyses + + 1973 in 8 sectional ratings of job reports of low- stratified by occupations (low pain aspects and back pain and sick occupation vs. satisfaction leave due to low- controls) back pain

Sauter et al. 248 VDT users; Cross- Questionnaire— Questionnaire— Physical + 1983 85 non-users sectional work reports of work (90%) environment discomfort demands scale survey

Svensson 940 males— Cross- Questionnaire— Interview report of Physical + and general sectional perceptions of back pain work Anderson population stress, boredom demands— 1983 life and job satisfaction

See footnotes at end of table. (Continued)

7-137-15

Table 7–2 (Continued). Summary of studies examining psychosocial factor and back disorders

Methods Associations with back disorders

Worker group Psychosocial MSD Covariate Low Low (participation factor outcome adjust- Job Int. Mono. job social Study rate) Design assessment assessment ments dissat. wkld. work control supp.

Svensson 1,746 females Cross- Questionnaire— Interview— Physical + o and ages 38–64— sectional items on job and reports of back workload Anderson general task satisfaction pain 1989 population

+ = Significant association found. o = No significant association found.

7-147-16 REFERENCES

Åaras A [1994]. Relationship between trape- manufacturing assembly line workers and a zius load and the incidence of musculoskele-tal cross-sectional factor-control study on illness in the neck and shoulder. Int J Ind occupational cervicobrachial disorders. Jpn J Ergonomics 14(4):341–348. Ind Health 30:3–12.

Abenhaim L, Suissa S [1988]. Economic scope Andersen JH, Gaardboe O [1993a]. Muscu- of occupationally-induced back pain. Institute loskeletal disorders of the neck and upper limb de recherche en sante et en securite du travail among sewing machine operators: a clinical du Quebec. Quebec, Canada: IRSST, investigation. Am J Ind Med 24(6): 689–700. Research Summary No. 66. Andersen JH, Gaardboe O [1993b]. Preva- AFL-CIO [1997]. Stop the pain. Washing-ton, lence of persistent neck and upper limb pain in DC: AFL-CIO. a historical cohort of sewing machine operators. Am J Ind Med 24:677–687. Alexander and Alexander [1996]. 1996 U.S. Andersson GBJ [1981]. Epidemiologic aspects risk management survey. New York, NY: of low back pain in industry. Spine 6:53–60. Alexander and Alexander, Inc. Andersson GBJ [1996]. The epidemiology of Allender E [1974]. Prevalence, incidence, and spinal disorders. In: Frymoyer JW, ed. The remission rates of some common rheumatic adult spine: principles and practice. 2nd ed. diseases or syndromes. Scand J Rheumatol New York, NY: Raven Press Ltd. 3:145–153. Andersson GBJ, Pope MH, Frymoyer JW, Alterman T, Shekells RB, Vernon SW, Burau Snook S [1991]. Epidemiology and cost. In: KD [1994]. Decision latitude, psy-chologic Pope M, ed. Occupational low back pain. demand, job strain and coronary heart disease St. Louis, MO: Mosby Year Book. in the Western Electric study. Am J Epidemiol 139(6):620–627. Andersson JAD [1984]. Shoulder pain and tension neck and their relation to work. Scand J Amadio PC [1995]. De Quervain’s disease Work Environ Health 10(6): 435–442. and tenosynovitis. In: Gordon SL, Blair SJ, Armstrong RB, Warren GL, Warren JA Fine LJ, eds. Repetitive motions disorders of [1991]. Mechanisms of exercise-induced the upper extremity. Rosemont, IL: American muscle fibre injury. J Sports Med 12(3): Academy of Orthopaedic Surgeons, pp. 184–207. 435–448. Armstrong TJ, Chaffin DB [1979]. Carpal Amano M, Gensyo U, Nakajima H, Yatsuki K tunnel syndrome and selected personal [1988]. Characteristics of work actions of shoe

R-1 attributes. J Occup Med 21(7):481–486. cross-sectional study of male employees in a Swedish pulp and paper industry. Br J Ind Med Armstrong TJ, Foulke JA, Joseph BS, 44(5):327–336. Goldstein SA [1982]. Investigation of cumulative trauma disorders in a poultry Åstrand NE, Isacsson SO [1988]. Back pain, processing plant. Am Ind Hyg Assoc J back abnormalities, and competing medical, 43(2):103–116. psychological, and social factors as predictors of sick leave, early retirement, unemployment, Armstrong TJ, Fine LJ, Goldstein SA, Lifshitz labour turnover and mortality: a 22-year YR, Silverstein BA [1987a]. Ergonomic follow-up of male employees in a Swedish pulp considerations in hand and wrist tendinitis. J and paper company. Br J Ind Med Hand Surg 12A(5):830–837. 45(6):387–395.

Armstrong TJ, Fine LJ, Radwin RG, Silverstein Atterbury MR, Limke JC, Lemasters GK, BS [1987b]. Ergonomics and the effects of Li Y, Forrester C, Stinson R, Applegate H vibration in hand-intensive work. Scand J [1996]. Nested case-control study of hand Work Environ Health 13:286–289. and wrist work-related musculoskeletal disorders in carpenters. Am J Ind Med Armstrong TJ, Punnett L, Ketner P [1989]. 30:695–701. Subjective worker assessments of hand tools used in automobile assembly. Am Ind Hyg Barnhart S, Demers PA, Miller M, Longstreth Assoc J 50(12):639–645. WT Jr., Rosenstock L [1991]. Carpal tunnel syndrome among ski manufacturing workers. Armstrong TJ, Buckle P, Fine LJ, Hagberg M, Scand J Work Environ Health 17(1):46–52. Jonsson B, Kilbom Å [1993]. A conceptual model for work-related neck and upper-limb Baron S, Milliron M, Habes D, Fidler A musculoskeletal disorders. Scand J Work [1991]. Hazard evaluation and technical Environ Health 19(2):73–84. assistance report: Shoprite Supermarkets, New Jersey—New York. Cincinnati, OH: U.S. Aro S, Leino P [1985]. Overweight and Department of Health and Human Services, musculoskeletal morbidity: a ten-year follow- Public Health Service, Centers for Disease up. Int J Obes 9:267–275. Control, National Institute for Occupational Safety and Health, NIOSH Report No. HHE Asbury FD [1995]. Occupational repetitive 88–344–2092. strain injuries and gender in Ontario 1986–1991. J Occup Environ Med 37(4): Battié MC, Bigos SJ, Fisher LD, Hansson TH, 479–485. Jones ME, Wortley MD [1989]. Isometric lifting strength as a predictor of industrial back Åstrand NE [1987]. Medical, psychological, pain. Spine 14(8):851–856. and social factors associated with back abnormalities and self-reported back pain: a Battié MC, Bigos SJ, Fisher LD, Spengler

R-2 DM, HanssonTH, Nachemson AL, et al. Bergqvist U [1989]. Possible health effects of [1990]. The role of spinal flexibility in back working with VDUs (editorial). Br J Ind Med pain complaints within industry: a prospective 46:217–221. study. Spine 15(8):768–793. Bergqvist U, Wolgast E, Nilsson B, Voss M Battié MC, Videman T, Gibbons LE, [1995a]. Musculoskeletal disorders among Fisher L, Manninen H, Gill K [1995]. visual display terminal workers: individual, Determinants of lumbar disc degeneration: ergonomic, and work organizational factors. a study relating lifetime exposures and magnetic Ergonomics 38(4):763–776. resonance imaging findings in identical twins. Spine 20:2601–2612. Bergqvist U, Wolgast E, Nilsson B, Voss M [1995b]. The influence of VDT work on Behrens V, Seligman P, Cameron L, Mathias musculoskeletal disorders. Ergonomics CGT, Fine L [1994]. The preva-lence of back 38(4):754–762. pain, hand discomfort, and dermatitis in the U.S. working population. Am J Public Health Bergqvist-Ullman M, Larsson N [1977]. Acute 84(11):1780–1785. low back pain in industry. Acta Orthop Scand 170(Suppl):1–117. Bennett GE [1946]. Shoulder and elbow lesions distinctive of baseball players. Ann Surg Berkeley M, Bennett J, Woods W [1985]. 126:107–110. Elbow problems: injuries to the throwing arm. Zarins B, Andrews J, Carson W, eds. Berg M, Sanden Å, Torell G, Järvholm B Phildelphia, PA: W.B. Saunders Publishing [1988]. Persistence of musculoskeletal symp- Company, pp. 235–249. toms: a longitudinal study. Ergonomics 31(9):1281–1285. Bernard B, Sauter S, Fine LJ [1993]. Hazard evaluation and technical assistance report: Los Bergenudd H, Nilsson B [1988]. Back pain in Angeles Times, Los Angeles, CA. Cincinnati, middle age. Occupational workload and OH: U.S. Department of Health and Human psychologic factors: an epidemiologic survey. Services, Public Health Service, Centers for Spine 13(1):58–60. Disease Control and Prevention, National Institute for Occupational Safety and Health, Bergenudd H, Lindgärde, Nilsson B, Petersson NIOSH Report No. HHE 90–013–2277. C [1988]. Shoulder pain in middle age: a study of prevalence and relation to occupational Bernard B, Sauter S, Fine LJ, Petersen M, work load and psychosocial factors. Clin Hales T [1994]. Job task and psychosocial risk Orthop Res 23:234–238. factors for work-related musculoskeletal disorders among newspaper employees. Scand Bergqvist U [1984]. Video display terminals J Work Environ Health 20(6):417– 426. and health. Scand J Work Environ Health 10(2):68–77. Biering-Sørensen F [1983]. A prospective

R-3 study of low-back pain in a general population. Bjelle A, Hagberg M, Michaelsson G [1979]. I. Occurrence, recurrence, and aetilogy. Scand Clinical and ergonomic factors in prolonged J Rehabil Med 15:81–88. shoulder pain among industrial workers. Scand J Work Environ Health 5(3):205–210. Biering-Sørensen F [1984]. Physical measurements as risk indicators for low back Bjelle A, Hagberg M, Michaelson G [1981]. trouble over a one-year period. Spine Occupational and individual factors in acute 9(2):106–119. shoulder-neck disorders among industrial worker. Br J Ind Med 38:356–363. Biering-Sørensen F, Thomsen CE, Hilden J [1989]. Risk indicators for low back trouble. Blåder S, Barck-Holst U, Danielsson S, Ferhm Scand J Rehabil Med 21:151–157. E, Kalpamaa M, Leijon M, et al. [1991]. Neck and shoulder complaints among sewing- Bigos SJ, Spengler DM, Martin NA, Zeh J, machine operators: a study concerning Fisher L, Nachemson A, et al. [1986a]. Back frequency, symptomatology and dysfunction. injuries in industry: a retrospective study. II. Appl Ergonomics 22(4):251– 257. Injury factors. Spine 11:246–251. Bleecker ML, Bohlman M, Moreland R, Bigos SJ, Spengler DM, Martin NA, Zeh J, Tipton A [1985]. Carpal tunnel syndrome: role Fisher L, Nachemson A [1986b]. Back injuries of carpal canal size. Neurol 35(11): in industry: a retrospective study. III. 1599–1604. Employee-related factors. Spine 11:252–256. BLS [1995]. Workplace injuries and ill-nesses Bigos SJ, Battié MC, Fisher LD [1991a]. in 1994. Washington, DC: U.S. Department of Methodology for evaluating predictive factors Labor, Bureau of Labor Statistics, USDL for the report of back injury. Spine 95–508. 16(6):669–670. Bongers PM, Boshuizen HC [1990]. Back Bigos SJ, Battié MC, Spengler DM, disorders and whole-body vibration at work Fisher LD, Fordyce WE, Hansson TH, et al. [Dissertation]. Amsterdam, The Nether-lands: [1991b]. A prospective study of work per- University of Amsterdam. ceptions and psychosocial factors affecting the report of back injury. Spine 16(1):1–6. Bongers PM, de Winter CR [1992]. Psy- chosocial factors and musculoskeletal disease: Bigos S, Bowyer O, Braen G, et al. [1994]. a report of the literature. The Netherlands: Acute low back problems in adults. Clinical TNO Institute of Preventive Health Care, practice guideline, quick reference guide No. Report No. 92–082. 14. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Bongers PM, Boshuizen HC, Hulshof CTJ, Agency for Health Care Policy and Research, Koemeester AP [1988]. Back disorders in AHCPR Pub. No. 95–6043. crane operators exposed to whole-body

R-4 vibration. Int Arch Occup Environ Health vibration syndrome and dose-response relation 60:129–137. for vibration-induced white finger among quarry drillers and stonecarvers. Occup Environ Med Bongers PM, Hulshof CTJ, Dijkstra L, 51(9):603–611. Boshuizen HC [1990]. Back pain and exposure to whole-body vibration in helicopter pilots. Bovenzi M, Zadini A [1992]. Self-reported low Ergonomics 33:1007– 1026. back symptoms in urban bus drivers exposed to whole-body vibration. Spine Bongers PM, de Winter CR, Kompier MAJ, 17(9):1048–1059. Hildebrandt VH [1993]. Psychosocial factors at work and musculoskeletal Bovenzi M, Betta A [1994]. Low-back disease. Scand J Work Environ Health disorders in agricultural tractor drivers exposed 19(5):297–312. to whole-body vibration and postural stress. Appl Ergonomics 25:231–241. Boshuizen HC, Bongers PM, Hulshof CTJ [1990a]. Self-reported back pain in tractor Bovenzi M, Franzinelli A, Strambi F [1988]. drivers exposed to whole-body vibration. Int Prevalence of vibration-induced white finger Arch Occup Environ Health 62:109–115. and assessment of vibration exposure among travertine workers in Italy. Int Arch Occup Boshuizen HC, Hulshof CTJ, Bongers PM Environ Health 61:25–34. [1990b]. Long-term sick leave and disability pensioning due to back disorders of tractor Bovenzi M, Zadini A, Franzinelli A, Borgogni F drivers exposed to whole-body vibration. Int [1991]. Occupational mus-culoskeletal Arch Occup Environ Health 62:117–122. disorders in the neck and upper limbs of forestry workers exposed to Boshuizen HC, Bongers PM, Hulshof CTJ hand-arm vibration. Ergonomics 34(5): [1992]. Self-reported back pain in fork-lift 547–562. truck and freight-container tractor drivers exposed to whole-body vibration. Spine Bovenzi M, Franzinelli A, Mancini R, Cannava 17:59–65. MG, Maiorano M, Ceccarelli F [1995]. Dose- response relation for vascular disorders Boshuizen HC, Verbeek JHAM, Broersen JPJ, induced by vibration in the fingers of forestry Weel ANH [1993]. Do smokers get more workers. Occup Environ Med 52:722–730. back pain? Spine 18(1):35–40. Boyce RW, Jones GR, Hiatt AR [1991]. Bovenzi M [1990]. Medical aspects of the Physical fitness capacity and absenteeism of hand-arm vibration syndrome. Int J Ind police officers. J Occup Med 33(11):1137– Ergonomics 6(1):61–73. 1143.

Bovenzi M, Italian study group on physical Brammer AJ, Taylor W, eds. [1981]. Vibration hazards in the stone industry [1994]. Hand-arm effects on the hand and arm in industry. New

R-5 York, NY: John Wiley and Sons. Burdorf A, Zondervan H [1990]. An epi- demiological study of low-back pain in crane Brattberg G, Thorslund M, Wikman A [1989]. operators. Ergonomics 33(8):981–987. The prevalence of pain in a general population: the results of a postal survey in a county of Burdorf A, Monster A [1991]. Exposure to Sweden. Pain 37:215–222. vibration and self-reported health complaints of Brinckmann P [1988]. Stress and strain of riveters in the aircraft industry. Ann Occup Hyg human lumbar discs. Clin Biomechanics 35(3):287–298. 3(4):232–235. Burdorf A, Govaert G, Elders L [1991]. Broberg E [1984]. Ergonomic injuries at work: Postural load and back pain of workers in the ISA-information system on occupa-tional manufacturing of prefabricated concrete injuries. Swedish National Board of elements. Ergonomics 34(7):909–918. Occupational Safety and Health, Report No. 1984:3E. Burdorf A, Naaktgeboren B, deGroot HC [1993]. Occupational risk factors for low back Bross JD [1966]. Spurious effects from an pain among sedentary workers. extraneous variable. J Chron Dis 19:736–747 J Occup Med 35:1213–1220.

Brubaker R, Mackenzie C, Bates D [1983]. Burry HC, Gravis V [1988]. Compensated Vibration white finger disease among tree back injury in New Zealand. N Z Med J fellers in British Columbia. J Occup Med 101(852):542–544. 25(5):403–408. Burt S, Hornung R, Fine L [1990]. Hazard Brubaker RL, Mackenzie CJG, Hertzman C, evaluation and technical assistance report: Hutton SG, Slakov J [1987]. Longitudinal Newsday, Inc., Melville, NY. Cincinnati, OH: study of vibration-induced white finger among U.S. Department of Health and Human coastal fallers in British Columbia. Scand J Services, Public Health Service, Centers for Work Environ Health 13(4):305– 308. Disease Control, National Institute for Occupational Safety and Health, NIOSH Buckwalter JA, Woo SL-Y, Goldberg VM, Report No. HHE 89–250–2046. Hadley EC, Booth F, Oegema TR, et al. [1993]. Current concepts review: soft-tissue Byström S, Hall C, Welander T, Kilbom Å aging and musculoskeletal function. J Bone [1995]. Clinical disorders and pressure-pain Joint Surg [Am] 75A(10):1533–1548. threshold of the forearm and hand among automobile assembly line workers. J Hand Surg Burdorf A [1992]. Exposure assessment of risk [Br] 20B(6):782–790. factors for disorders of the back in occupational epidemiology. Scand J Work Cady LD Jr., Bischoff DP, O'Connell ER, Environ Health 18(1):1–9. Thomas PC, Allan JH [1979]. Strength and fitness and subsequent back injuries in fire-

R-6 fighters. J Occup Med 21(4):269–272. Occupational Safety and Health, DHEW (NIOSH) Publication No. 77–163. Cady LD Jr., Thomas PC, Karwasky RJ [1985]. Program for increasing health and Chatterjee D [1992]. Workplace upper limb physical fitness of firefighters. J Occup Med disorders: a prospective study with inter- 27(2):110–114. vention. J Occup Med 42(3):129–136.

Cannon LJ, Bernacki EJ, Walter SD [1981]. Chatterjee DS, Barwick DD, Petrie A [1982]. Personal and occupational factors associated Exploratory electromyography in the study of with carpal tunnel syndrome. J Occup Med vibration-induced white finger in rock drillers. 23(4):255–258. Br J Ind Med 39:89–97.

CDC (Centers for Disease Control) [1989]. Chiang H, Chen S, Yu H, Ko Y [1990]. The Occupational disease surveillance: carpal tunnel occurrence of carpal tunnel syndrome in frozen syndrome. MMWR 38(28):485–489. food factory employees. Kao Hsiung J Med Sci 6:73–80. Chaffin DB [1973]. Human strength capability and low back pain. Unpublished paper Chiang H, Ko Y, Chen S, Yu H, Wu T, Chang presented at the American Industrial Health P [1993]. Prevalence of shoulder and upper- Conference, Denver, CO, April 16–19, 1973. limb disorders among workers in the fish- processing industry. Scand J Work Environ Chaffin DB [1979]. Manual materials handling Health 19(2):126–131. the cause of overexertion injury and illness in industry. J Environ Pathol Toxicol 2(5):67–73. Christensen H, Pedersen MB, Sjøgaard G [1995]. A national cross-sectional study in the Chaffin DB, Park KS [1973]. A longitudinal Danish wood and furniture industry on working study of low-back pain as associated with postures and manual materials handling. occupational weight lifting factors. Am Ind Hyg Ergonomics 38(4):793–805. Assoc J 34:513–525. Clemmer DI, Mohr DL, Mercer DJ [1991]. Chaffin DB, Andersson GBJ [1984]. Occu- Low-back injuries in a heavy industry: I. pational biomechanics. New York, NY: John Worker and workplace factors. Spine Wiley and Sons. 16(7):824–830.

Chaffin DB, Herrin GD, Keyserling WM, Cohen S, Pyrrell DAJ, Smith AP [1991]. Foulke JA [1977]. Preemployment strength Psychological stress and susceptibility to the testing in selecting workers for materials common cold. New Engl J Med 325(9):606– handling jobs. Cincinnati, OH: U.S. 612. Department of Health, Education, and Welfare, Health Services and Mental Health Cornfield J [1956]. A statistical problem arising Administration, National Institute for from retrospective studies. Pro-ceedings of the

R-7 third Berkely Symposium, pp. 135–148. logical perception of work and low back symptoms in nursing aides. Scand J Rehabil Cornfield J, Haenszel W, Hammond EC, Med 10(4):201–209. Lilienfeld AM, Shimkin MB, Wynder EL [1959]. Smoking and lung cancer: recent Dehlin O, Berg S, Andersson GBJ, Grimby G evidence and a discussion of some questions. J [1981]. Effect of physical training and Natl Cancer Inst 22:173–203. ergonomic counseling on psychological perception of work and on the subjective Cust G, Pearson JCG, Mair A [1972]. The assessment of low-back insufficiency. Scand J prevalence of low back pain in nurses. Int Nurs Rehabil Med 13:1–9. Rev 19(2):169–179. deKrom MCT, Kester ADM, Knipschild PG, CWCI [1993]. Report on CTDs. San Spaans F [1990]. Risk factors for carpal tunnel Francisco, CA: California Workers’ syndrome. Am J Epidemiol Compensation Institute. 132(6):1102–1110.

Cyriax JH [1936]. The pathology and treatment Deyo RA, Bass JE [1989]. Lifestyle and low of tennis elbow. J Bone Joint Surg Am back pain: the influence of smoking and obesity. 18(4):921–940. Spine 14(5):501–506.

Damkot DK, Pope MH, Lord J, Frymoyer JW Dieck GS, Kelsey JL [1985]. An epidemio- [1984]. The relationship between work history, logic study of the carpal tunnel syndrome in an work environment and low-back pain in men. adult female population. Prev Med Spine 9(4):395–399. 14(1):63–69.

Dammeskiold-Samsøe B, Christiansen E, Lund Dimberg L [1987]. The prevalence and B, Andersen RB [1982]. Regional muscle causation of tennis elbow (lateral humeral tension and pain (“fibrositis”). Scand J Rehabil epicondylitis) in a population of workers in an Med 15:17–20. engineering industry. Ergonomics 30(3):573–580. Dehlin O, Berg S [1977]. Back symptoms and psychological perception of work: a study Dimberg L, Oden A [1991]. White finger among nursing aides in a geriatric hospital. symptoms: a cross-sectional study. Aviat Space Scand J Rehabil Med 9:61–65. Environ Med 62:879–883

Dehlin O, Hedenrud B, Horal J [1976]. Back Dimberg L, Olafsson A, Stefansson E, Aagaard symptoms in nursing aides in a geriatric H, Oden A, Andersson GBJ, et al. [1989]. The hospital. Scand J Rehabil Med 8(2):47. correlation between work environment and the occurrence of cervicobrachial symptoms. J Dehlin O, Berg S, Hedenrud B, Anderson G, Occup Med 31(5): 447–453. Grimby G [1978]. Muscle training, psycho-

R-8 Dixon SJ [1973]. Progress and problems in tunnel syndrome in butchers. Scand J Work back pain research. Rheumatol Rehabil 12:165. Environ Health 9(3):291–297.

Dupuis H [1994]. Medical and occupational Färkkilä M, Aatola S, Starck J, Korhonen O, preconditions for vibration-induced spinal Pyykkö I [1986]. Hand-grip force in lumber- disorders: occupational disease No. 2110 in jacks: two-year follow-up. Int Arch Occup Germany. Int Arch Occup Environ Health Environ Health 58:203–208. 66:303–308. Färkkilä M, Pykkö I, Jäntti V, Aatola S, Starck J, Korhonen O [1988]. Forestry Earman WA, Andersson GB, Leavitt F, workers exposed to vibration: a neurological McNeill TW, Durudogan H, Reagan J [1996]. study. Br J Ind Med 45:188–192. Factors influencing the cost of chronic low back injuries: an analysis of data from Feldman RG, Travers PH, Chirico-Post J, independent medical examinations. Keyserling WM [1987]. Risk assessment in J Occup Rehabil 6(1):5–16. electronic assembly workers: carpal tunnel syndrome. J Hand Surg [Am] 12A(5): Ekberg K, Björkqvist B, Malm P, Bjerre-Kiely 849–855. B, Karlsson M, Axelson O [1994]. Case- control study of risk factors for disease in the Ferguson D [1976]. Posture, aching and body neck and shoulder area. Occup Environ Med build in telephonists. J Hum Ergol 5:183–196. 51(4):262–266. Feyer A, Williamson A, Mandryk J, de Silva I, Ekberg K, Karlsson M, Axelson O [1995]. Healy S [1992]. Role of psychosocial risk Cross-sectional study of risk factors for symp- factors in work-related low back pain. Scand J toms in the neck and shoulder area. Work Environ Health 18(6):368–375. Ergonomics 38(5):971–980. Finkelstein MM [1995]. Back pain and Eklund J, Odenrick P, Zettergren S, Johansson parenthood. Occup Environ Med 52(1): H [1994]. Head posture measure-ments among 51–53. work vehicle drivers and implications for work and workplace design. Ergonomics Fishbein M, Middlestadt S, Ottati V, 37(4):623–639. Straus S, Ellis A [1988]. Medical problems among ICSOM musicians: overview of a English CJ, Maclaren WM, Court-Brown C, national survey. Med Prob Perform Art Hughes SPF, Porter RW, Wallace WA 1:136–139. [1995]. Relations between upper limb soft tissue disorders and repetitive movements at Flodmark BT, Aase G [1992]. Musculo-skel- work. Am J Ind Med 27(1):75–90. etal symptoms and type A behaviour in blue collar workers. Br J Ind Med 49(10): Falck B, Aarnio P [1983]. Left-sided carpal 683–687.

R-9 Frank JW, Pulcins IR, Kerr MS, Shannon HS, low-back pain in industry. Occup Med: State Stansfield SA [1995]. Occupational back of the Art Rev 7(4):593–608. pain—an unhelpful polemic. Scand J Work Environ Health 58:203–208. Gelberman RH, Herginroeder PT, Hargens AR, Lundborg GN, Akeson WH [1981]. The Franklin GM, Haug J, Heyer N, Checkoway carpal tunnel syndrome: a study of carpal canal H, Peck N [1991]. Occupational carpal tunnel pressures. J Bone Joint Surg 63A(3): 380–383. syndrome in Washington State, 1984–1988. Am J Public Health 81(6): 741–746. Gemne G [1982]. Pathophysiology and multifactorial etiology of acquired vaso-spastic Fransson-Hall C, Byström S, Kilbom Å disease (Raynaud Syndrome) in vibration [1995]. Self-reported physical exposure and exposed workers. Scand J Work Environ musculoskeletal symptoms of the forearm-hand Health 8:243–249. among automobile assembly-line workers. J Occup Environ Med 37(9): 1136–1144. Gemne G, Lundström R, Hansson J [1993]. Disorders induced by work with hand-held Frymoyer JW, Pope MH, Costanza MC, vibrating tools: a review of current knowledge Rosen JC, Goggin JE, Wilder DG [1980]. for criteria documentation. Arbete och Hälsa Epidemiologic studies of low-back pain. Spine 6:1–83. 5(5):419–423. Gerr F, Letz R [1992]. Risk factors for carpal Frymoyer JW, Pope MH, Clements JH tunnel syndrome in industry: blaming the victim? [1983]. Risk factors in low back pain. [Letter to the editor.] J Occup Med J Bone Joint Surg [Am] 65A:213. 34(11):1117–1119.

Frymoyer JW, Cats-Baril WL [1991]. An Gerr F, Letz R, Landrigan PJ [1991]. Upper- overview of the incidence and costs of low extremity musculoskeletal disorders of back pain. Orthop Clin North Am 22: occupational origin. Ann Rev Public Health 262–271. 12:543–566.

Futatsuka M, Uneno T [1985]. Vibration Graham NM, Douglas RM, Ryan P [1986]. exposure and vibration induced white finger Stress and acute respiratory infection. Am J due to chain saw operation. J Occup Med Epidemiol 124(3):389–401. 27:257–264. Grandjean E, Hunting W, Nishiyama K [1984]. Futatsuka M, Uneno T [1986]. A follow-up Preferred VDT workstation settings, body study of vibration-induced white finger due to posture and physical impairments. Appl chain saw operation. Scand J Work Environ Ergonomics 15(2):99–104. Health 12(4):304–306. Guo HR, Tanaka S, Cameron LL, Seligman Garg A, Moore SJ [1992]. Epidemiology of PJ, Behrens VJ, Ger J, et al. [1995]. Back pain

R-10 among workers in the United States: national London, England: Taylor and Francis. estimates and workers at high risk. Am J Ind Med 28(5):591–602. Hägg G, Suurküla J, Kilbom Å [1990]. Predictors for work-related shoulder/neck Hadler NM [1993]. Occupational musculo- disorders: a longitudinal study of female skeletal disorders. New York, NY: Raven assembly workers. Arbete och Hälsa 10:1–78. Press. Hägg GM, Öster J, Byström S [1996]. Hadler N [1997]. Repetitive upper-extremity Forearm muscular load and wrist angle among motions in the workplace are not hazardous. J automobile assembly line workers in relation to Hand Surg [Am] 22A(1):19–29. symptoms. Appl Ergonomics 28(1):41–47.

Hagberg M [1981]. Work load and fatigue in Hales TR, Fine L [1989]. Hazard evaluation repetitive arm elevations. Ergonomics and technical assistance report: Cargill Poultry 24:543–555. Division, Buena Vista, CA. Cincinnati, OH: U.S. Department of Health and Human Hagberg M [1984]. Occupational musculo- Services, Public Health Service, Centers for skeletal stress and disorders of the neck and Disease Control, National Institute for shoulder: a review of possible pathophysio- Occupational Safety and Health, NIOSH logy. Int Arch Occup Environ Health Report No. HHE 89–251–1997. 53:269–278. Hales TR, Sauter S [1992]. Hazard evalua-tion Hagberg M, Kvarnström S [1984]. Muscular and technical assistance report: U.S. West endurance and electromyographic fatigue in Communications, Phoenix, AZ; Minneapolis, myofascial shoulder pain. Arch Phys Med MN; Denver, CO. Cincinnati, OH: U.S. Rehabil 65:522–525. Department of Health and Human Services, Public Health Service, Centers for Disease Hagberg M, Wegman DH [1987]. Preva-lence Control and Prevention, National Institute for rates and odds ratios of shoulder-neck diseases Occupational Safety and Health, NIOSH in different occupational groups. Br J Ind Med Report No. HHE 89–299–2230. 44(9):602–610. Hales TR, Habes D, Fine L, Hornung R, Hagberg M, Morgenstern H, Kelsh M [1992]. Boiano J [1989]. Hazard evaluation and tech- Impact of occupations and job tasks on the nical assistance report: John Morrell and Co., prevalence of carpal tunnel syndrome. Scand J Sioux Falls, SD. Cincinnati, OH: U.S. Work Environ Health 18(6):337– 345. Department of Health and Human Services, Public Health Service, Centers for Disease Hagberg M, Silverstein B, Wells R, Control, National Institute for Occupational Smith M, Hendrick H, Carayon P [1995]. Safety and Health, NIOSH Report No. HHE Work-related musculoskeletal disorders 88–180–1958. (WMSDs): a reference book for prevention.

R-11 Hales TR, Sauter SL, Peterson MR, Fine LJ, Wright V [1992]. Work-related upper limb Putz-Anderson V, Schleifer LR, et al. [1994]. disorders: the relationship between pain, Musculoskeletal disorders among visual display cumulative load, disability, and psychological terminal users in a telecom-munications factors. Ann Rheum Dis 51(12): 1325–1329. company. Ergonomics 37(10): 1603–1621. Herberts P, Kadefors R [1976]. A study of Harms-Ringdahl K, Ekholm J [1986]. Intensity painful shoulders in welders. Acta Orthop and character of pain and muscular activity Scand 47:381–387. levels elicited by maintained extreme flexion position of the lower-cervical-upper thoracic Herberts P, Kadefors R, Andersson G, spine. Scand J Rehabil Med 18:117–126. Petersen I [1981]. Shoulder pain in industry: an epidemiological study on welders. Acta Orthop Hart DA, Frank CB, Bray RC [1995]. Scand 52:299–306. Inflammatory processes in repetitive motion and overuse syndromes: potential role of Herberts P, Kadefors R, Högfors C, Sigholm neurogenic mechanisms in tendons and G [1984]. Shoulder pain and heavy manual ligaments. In: Gordon SL, Blair SJ, Fine, LJ, labor. Clin Orthop 191:166–178. eds. Repetitive motions disorders of the upper Herrin GD, Jaraiedi M, Anderson CK [1986]. extremity. Rosemont, IL: American Academy Prediction of overexertion injuries using of Orthopaedic Surgeons, pp. 247–262. biomechanical and psychophysical models. Am Ind Hyg Assoc J 47(6):322– 330. Hartz CR, Linscheid RL, Gramse RR, Daube JR [1981]. The pronator teres syndrome: Heyer N, Checkoway H, Daniell, et al. [1990]. compressive neuropathy of the median nerve. J Self-reported musculoskeletal symptoms Bone Joint Surg 63(6): 885–890. among office video display terminal operators. In: Sakurai H, Okazaki I, Omae I, eds. Heliövaara M [1987]. Occupation and risk of Occupational epidemiology. Amsterdam, The herniated lumbar intervertebral disc or sciatica Netherlands: Elsevier Science Publishers, B.V., leading to hospitalization. J Chronic Dis pp. 255–258. 40(3):259–264. Higgs P, Young VL, Seaton M, Edwards D, Heliövaara M, Knekt P, Aromaa A [1987]. Feely C [1992]. Upper extremity impairment in Incidence and risk factors of herniated lumbar workers performing repetitive tasks. Plast intervertebral disc or sciatica leading to Reconstr Surg 90(4):614–620. hospitalization. J Chronic Dis 40:251– 258. Hildebrandt VH [1987]. A review of epi- demiological research on risk factors of low Heliövaara M, Mäkelä M, Knekt P, Impivaara back pain. In: Buckle P, ed. Musculoskeletal O, Aromaa A [1991]. Determi-nants of sciatica disorders at work. Proceedings of a con- and low back pain. Spine 16(6):608–614. ference held at the University of Surrey, Guilford, April, 1987. London, England: Taylor Helliwell PS, Mumford DB, Smeathers JE, and Francis, pp. 9–16.

R-12 Hildebrandt VH [1995]. Back pain in the interview and clinical examination. Scand J working population: prevalence rates in Dutch Rehabil Med 23:119–125. trades and professions. Ergonomics 38(6):1283–1298. Holmström EB, Lindell J, Moritz U [1992]. Low back and neck/shoulder pain in Hildebrandt VH, Bongers PM, Dul J, van Dijk construction workers: occupational work-load FJH, Kemper HCG [1996]. Identifica-tion of and psychosocial risk factors. Spine high-risk groups among maintenance workers 17(6):663–671. in a steel company with respect to musculoskeletal symptoms and workload. Holmström E, Lindell J, Moritz U [1993]. Ergonomics 39(2):232–242. Healthy lower backs in the construction industry in Sweden. Work and Stress Hill AB [1965]. The environment and disease: 7(3):259–271. association or causation. Proc Royal Soc Med 58:295–300. Hopkins A [1990]. Stress, the quality of work, and repetition strain injury in Australia. Work Hill AB [1966]. Principles of medical statistics. and Stress 4(2):129–138. 8th ed. London, England: Lancet. Horal J [1969]. The clinical appearance of low Hill AB [1971]. Principles of medical statistics. back disorders in the city of Gothen-burg, 9th ed. New York, NY: Oxford University Sweden. Acta Orthop Scand Press. 118(Suppl):7–37.

Hoekstra EJ, Hurrell JJ, Swanson NG [1994]. Houtman ILD, Bongers PM, Smulders PGW, Hazard evaluation and technical assistance Kompier MAJ [1994]. Psychosocial stressors report: Social Security Admin-istration at work and musculoskeletal problems. Scand J Teleservice Centers, Boston, MA; Fort Work Environ Health 20(2):139–145. Lauderdale, FL. Cincinnati, OH: U.S. Department of Health and Human Services, Hrubec Z, Nashold BS [1975]. Epidemio-logy Public Health Service, Centers for Disease of lumbar disc lesions in the military in World Control and Prevention, National Institute for War II. Am J Epidemiol 102(5):366– 376. Occupational Safety and Health, NIOSH Report No. 92–0382–2450. Huang J, Ono Y, Shibata E, Takeucki Y, Hisanaga N [1988]. Occupational musculo- Holmström E [1992]. Musculoskeletal skeletal disorders in lunch centre workers. disorders in construction workers. Lund, Ergonomics 31(1):65–75. Sweden: Lund University, Department of Physical Therapy, pp. 1–175. Hughes RE, Silverstein BA, Evanoff BA [1997]. Risk factors for work-related Holmström E, Moritz U [1991]. Low back musculoskeletal disorders in an aluminum pain correspondence between questionnaire smelter. Am J Ind Med 32:66–75.

R-13 Hulshof C, Veldhuijzen van Zanten B [1987]. Jensen RC [1988]. Epidemiology of work- Whole-body vibration and low-back pain: a related back pain. Top Acute Care Trauma review of epidemiologic studies. Int Arch Rehabil 2(3):1–15. Occup Environ Health 59:205–220. Jensen RC [1990]. Back injuries among nursing Hunting KL, Welch LS, Cuccherini BA, Seiger personnel related to exposure. Appl Occup LA [1994]. Musculoskeletal symptoms among Environ Hyg 5:38–45. electricians. Am J Ind Med 25(2):149–163. Johanning E [1991]. Back disorders and health Hünting W, Läubli TH, Grandjean E [1981]. problems among subway train operators Postural and visual loads at VDT workplaces. exposed to whole-body vibration. Scand J I. Constrained postures. Ergonomics Work Environ Health 17(6):414– 419. 24(12):917–931. Johanning E, Wilder DG, Landrigan PJ, Pope Hurrell JJ, Murphy LR [1992]. Psycho-logical MH [1991]. Whole-body vibration exposure in job stress. In: Rom WN, ed. Environmental and subway cars and review of adverse health occupational medicine. 2nd ed. New York, effects. J Occup Med 33(5):605–612. NY: Little, Brown and Company, pp. 675–684. Johansson JA [1994]. Work-related and non- work-related musculoskeletal symptoms. Appl Hurrell JJ Jr., Murphy LR, Sauter SL, Cooper Ergonomics 25(4):248–251. CL [1988]. Occupational stress: issues and developments. New York, NY: Taylor and Johansson JA, Rubenowitz S [1994]. Risk Francis. indicators in the psychosocial and physical work environment for work-related neck, Ignatius Y, Yee T, Yan L [1993]. Selfreported shoulder and low back symptoms: a study musculoskeletal problems amongst typists and among blue and white collar workers in eight possible risk factors. J Hum Ergol 22:83–93. companies. Scand J Rehabil Med 26:131–142.

ILO [1986]. Psychosocial factors at work: Johnson JV, Stewart W, Hall EM, recognition and control. Geneva, Switzer-land: Fredlund P, Theorell T [1996]. Long-term International Labour Office. psychosocial work environment and cardiovascular mortality among Swedish men. Iwata H, Makimo S [1987]. Prevalence of Am J Public Health 86(3):324–331. Raynaud’s phenomenon in individuals not using vibrating tools. Sangyo Igaku 29: 500–503. Jonsson B [1988]. The static load component in muscle work. Eur J Appl Physiol Jensen RC [1987]. Disabling back injuries 57:305–310. among nursing personnel: research needs and justification. Res Nurs Health 10(1): 29–38. Jonsson BG, Persson J, Kilbom Å [1988]. Disorders of the cervicobrachial region among

R-14 female workers in the electronics industry: a tennis players with lateral epicondylitis. Am J two-year follow-up. Int J Ind Ergo 3(1):1–12. Sports Med 22(3): 359–363.

Kaji H, Honma H, Usui M, Yasuno Y, Saito K Kellgren JH, Lawrence JS [1958]. Osteo- [1993]. Hypothenar hammer syndrome in arthrosis and disc degeneration in an urban workers occupationally exposed to vibrating population. Ann Rheum Dis 17(4):338–397. tools. J Hand Surg [Br] 18B(6):761–766. Kelsey CA, Mettler FA [1990]. Flexible Kamwendo K, Linton, SJ, Moritz U [1991]. protective gloves: the Emperor's new clothes? Neck and shoulder disorders in medical Radiology 174(1):275–276. secretaries. Part I. Pain prevalence and risk factors. Scand J Rehabil Med 23(3):127–33. Kelsey JL [1975a]. An epidemiological study of acute herniated lumbar inter-vertebral disc. Karasek RA, Baker D, Marxer F, Ahlbom A, Rheumatol Rehabil 14:144. Theorell T [1981]. Job decision latitude, job demands, and cardiovascular disease: a Kelsey JL [1975b]. An epidemiological study prospective study of Swedish men. Am J Public of the relationship between occupations and Health 71(7):694–705. acute herniated lumbar intervertebral discs. Int J Epidemiol 4(3):197–205. Karasek RA, Gardell B, Lindell J [1987]. Work and non-work correlates of illness and Kelsey JL, Hardy RJ [1975]. Driving of motor behaviour in male and female Swedish white vehicles as a risk factor for acute herniated collar workers. J Occup Behav 8(3):187– 207. lumbar intervertebral disc. Am J Epidemiol 102(1):63–73. Karasek RA, Theorell T, Schwartz JF, Schnall PL, Pieper CF, Michela JL [1988]. Job Kelsey JL, White AA [1980]. Epidemiology characteristics in relation to the preva-lence of and impact on low back pain. Spine 5(2): myocardial infarction in the U.S. Health 133–142. Examination Survey (HES) and the Health and Nutrition Examination Survey (HANES). Am J Kelsey JL, Golden AL [1988]. Occupational Public Health 78(8):910– 918. and workplace factors associated with low back pain. Occup Med: State of the Art Rev Katz JN, Punnett L, Simmons BP, Fossel AH, 3(1):7–16. Mooney N, Keller RB [1996]. Workers compensation recipients with carpal tunnel Kelsey JL, Githens PB, White AA, Holford syndrome: the validity of self-reported health TR, Walter SD, O’Connor T, et al. [1984]. An measures. Am J Public Health 86(1):52–56. epidemiologic study of lifting and twisting on the job and risk for acute pro-lapsed lumbar Kelley JD, Lombardo SJ, Pink M, Perry G, intervertebral disc. J Orthop Res 2(1):61–66. Giangarra CE [1994]. Electromyographic and cinematographic analysis of elbow function in Kelsey JL, Golden AL, Mundt DJ [1990].

R-15 Low back pain/prolapsed lumbar inter- Kivekäs J, Riihimäki H, Husman K, Hänninen vertebral disc. Epidemiol Rheum Dis K, Härkönen H, Kuusela T, et al. [1994]. 16(3):699–716. Seven-year follow-up of white- finger symptoms and radiographic wrist findings Kennedy JC, Hawkins R, Krissoff WB [1978]. in lumberjacks and referents. Scand J Work Orthopaedic manifestations of swimming. Am J Environ Health 20(2):101–106. Sports Med 6:309–322. Kivi P [1984]. Rheumatic disorders of the Keyserling WM, Herrin GD, Chaffin DB upper limbs associated with repetitive [1980]. Isometric strength testing as a means of occupational tasks in Finland, 1975–1979. controlling medical incidents on stren-uous Scand J Rheumatol 13:101–107. jobs. J Occup Med 22(5):332–336. Klein BP, Jensen RC, Sanderson LM [1984]. Kiken S, Stringer W, Fine LJ [1990]. Hazard Assessment of workers’ compen-sation claims evaluation and technical assistance report: Perdue Farms, Lewiston, NC; Roberson-ville, for back strains/sprains. NC. Cincinnati, OH: U.S. Department of J Occup Med 26(6):443–448. Health and Human Services, Public Health Service, Centers for Disease Control, National Kleinbaum DG, Kupper LL, Morgenstern H Institute for Occupational Safety and Health, [1982]. Epidemiologic research: principles and NIOSH Report No. HHE 89–307–2009. quantitative methods. London, England: Lifetime Learning Publications. Kilbom Å [1994]. Repetitive work of the upper extremity. Int J Ind Ergonomics 14(1/2):51–86. Knave BG, Wibom RI, Voss M, Hedström LD, Bergqvist UO [1985]. Work with video Kilbom Å, Persson J [1987]. Work technique display terminals among office employees. I. and its consequences for musculo-skeletal Subjective symptoms and discomfort. Scand J disorders. Ergonomics 30(2): 273–279. Work Environ Health 11(6):457–466.

Kilbom Å, Horst D, Kemfert K, Richter A Knibbe JJ, Friele RD [1996]. Prevalence of [1986]. Observation methods for reduction of back pain and characteristics of the physical load and strain on the human body—a review. workload of community nurses. Ergonomics Abetarskyddsstyrelsen Publikation Service 39(2):186–198. 171(84):92.

Kim SH, Chung MK [1995]. Effects of Kompier MAJ [1988]. Arbeid en gezondheid posture, weight and frequency on trunk van stadsbus-chauffeurs. Delft Eburon muscular activity and fatigue during repetitive (Proefschrift Rijksuniversitiet Groningen). lifting tasks. Ergonomics 38(5):853–863. Kopf T, Von Feder T, Karmaus W [1988]. King JW, Brelsford HJ, Tullos HS [1969]. Risks for musculoskeletal disorders of the low Analysis of the pitching arm of the pro-fessional back, the shoulders, the elbows, and the wrists baseball pitcher. Clin Orthop 67:116–123. in bricklayers. In: Hogstedt C, Reuterwall C,

R-16 eds. Progress in occupational epidemiology. Kvarnström S [1983a]. Incidence of Amsterdam: Elsevier Science Publishers B.V., musculoskeletal disorders in a mechanical pp. 219–222. engineering plant, with emphasis on occupat- ional shoulder injuries. Abetarskyddssty-relsen Koskimies K, Färkkilä M, Pyykkö I, Jäntti V, Publikation Service 171(84):79. Aatola S, Starck J, et al. [1990]. Carpal tunnel syndrome in vibration disease. Br J Ind Med Kvarnström S [1983b]. Occurrence of 47(6):411–416. musculoskeletal disorders in a manufac-turing industry with special attention to occupational Koskimies K, Pyykkö I, Starck J, Inaba R shoulder disorders. Scand J Rehabil Med [1992]. Vibration syndrome among Finnish (Suppl 8):1–114. forest workers between 1972 and 1990. Int Arch Occup Environ Health 64:251–256. Kvarnström S, Halden M [1983]. Occupa- tional cervicobrachial disorders in an Kukkonen R, Luopajärvi T, Riihimäki V engineering company. Scand J Rehabil Med [1983]. Prevention of fatigue amongst data 8:1–114. entry operators. In: Kvalseth TO, ed. Ergonomics of workstation design. London, Lapidus PW, Guidotti FP [1970]. Lateral and England: Buttersworth, pp. 28–34. medical epicondylitis of the humerus. Ind Med Surg 39(4):30–32; 171–173. Kumar S [1994]. The back compressive forces during maximal push-pull activities in the sagittal Larsson SE, Bengtsson A, Bodegard L, plane. J Hum Ergol 23(2):133– 150. Henriksson KG, Larsson J [1988]. Muscle changes in work-related chronic myalgia. Acta Kuorinka I, Koskinen P [1979]. Occupa-tional Orthop Scand 59(5):552–556. rheumatic diseases and upper limb strain in manual jobs in a light mechanical industry. Larsson SE, Bodegard L, Henriksson KG, Scand J Work Environ Health 5(Suppl Öberg PA [1990]. Chronic trapezius myalgia: 3):39–47. morphology and blood flow studied in 17 patients. Acta Orthop Scand 61(5):394–398. Kuorinka I, Forcier L [1995]. Work-related musculoskeletal disorders (WMSDs): Lawrence JS [1955]. Rheumatism in coal a reference book for prevention. London, miners. Part III: Occupational factors. Br J Ind England: Taylor & Francis, p. 1. Med 12(3):249–261.

Kurppa K, Viikari-Juntura E, Kuosma E, Leboeuf-Yde C, Klougart N, Lauritzen T Huuskonen M, Kivi P [1991]. Incidence of [1996]. How common is low back pain in the tenosynovitis or peritendinitis and epicondy-litis Nordic population? Spine 21(13):1518– 1526. in a meat-processing factory. Scand J Work Environ Health 17(1):32–37. Leigh JP, Sheetz RM [1989]. Prevalence of

R-17 back pain among full-time United States Liira JP, Shannon HS, Chambers LW, Haines workers. Br J Ind Med 46:651–657. TA [1996]. Long-term back problems and physical work exposures in the 1990 Ontario Leigh JP, Markowitz S, Fahs M, et al. [1996]. Health Survey. Am J Public Health Costs of occupational injuries and illnesses in 86(3):382–387. 1992. NIOSH cooperative agreement with ERC, Inc., contract no. U60/CCU902886. Liles DH, Deivanayagam S, Ayoub MM, Mahajan P [1984]. A job severity index for the Leino P [1989]. Symptoms of stress predict evaluation and control of lifting injury. Hum musculoskeletal disorders. J Epidemiol Com- Factors 26(6):683–693. munity Health 43(3):293–300. Lindman R, Hagberg M, Ängqvist KA, Leino P, Lyyra A [1990]. The effect of mental Söderlund K, Hultman E, Thornell L [1991]. stress and social support on the development of Changes in muscle morphology in chronic musculoskeletal morbidity in the engineering trapezius myalgia. Scand J Work Environ industry. In: Sakurai H, Okazaki L, Omae I, Health 17(5):347–355. eds. Occupational epidemiology. Amsterdam, The Nether-lands: Elsevier Science Publishers Linton SJ [1990]. Risk factors for neck and B.V., pp. 267–272. back pain in a working population in Sweden. Work and Stress 4(1):41–49. Leino P, Aro S, Hasan J [1987]. Trunk muscle function and low back disorders: a ten-year Linton SJ, Kamwendo K [1989]. Risk factors follow-up study. J Chronic Dis 40(4):289–296. in the psychosocial work environ-ment for neck and shoulder pain in secre-taries. J Occup Med Leino P, Hasan J, Karppi SL [1988]. 31(7):609–613. Occupational class, physical workload, and musculoskeletal morbidity in the engineer-ing Liss GM, Jesin E, Kusiak RA, White P [1995]. industry. Br J Ind Med 45(10):672–681. Musculoskeletal problems among Ontario dental hygienists. Am J Ind Med Letz R, Cherniack MG, Gerr F, Hershman D, 28(4):521–540. Pace P [1992]. A cross-sectional epide- miological survey of shipyard workers exposed Loslever P, Ranaivosoa A [1993]. Biome- to hand-arm vibration. Br J Ind Med chanical and epidemiological investigation of 49(1):53–62. carpal tunnel syndrome at workplaces with high risk factors. Ergonomics 36(5): 537–555. Levitz CL, Iannotti JP [1995]. Overuse injuries of the shoulder. In: Gordon SL, Blair SJ, Fine Louis, Harris & Associates [1996]. Pain and LJ, eds. Repetitive motions disorders of the absenteeism in the workplace. New York, NY: upper extremity. Rosemont, IL: American Louis, Harris & Associates. Academy of Orthopaedic Surgeon. Lukaš E [1970]. Lesion of the peripheral

R-18 nervous system due to vibration. Work Environ at an increased risk for developing Health 7(1):67–79. musculoskeletal disorders? Spine 6:710–717.

Lundberg U, Granqvist M, Hansson T, Magora A [1970]. Investigation of the relation Magnusson M, Wallin L [1989]. Psycho- between low back pain and occupation. Ind logical and physiological stress responses Med 39:31–37. during repetitive work at an assembly line. Work and Stress 3(2):143–153. Magora A [1972]. Investigation of the relation between low back pain and occupation. III. Lundborg G, Dahlin LB, Danielsen N, Hansson Physical requirements: sitting, standing, and HA, Necking LE, Pyykkö I [1987]. Intraneural weight lifting. Ind Med 41(12):5–9. edema following exposure to vibration. Scand J Work Environ Health 13(4):326–329. Magora A [1973]. Investigation of the relation between low back pain and occupation. IV. Lundborg G, Dahlin LB, Hansson HA, Kanje Physical requirements: bending, rotation, M, Necking LR [1990]. Vibration exposure reaching, and sudden maximal effort. Scand J and peripheral nerve fiber damage. J Hand Rehabil Med 5:191–196. Surg [Am] 15A(2):346–351. Magora A, Schwartz A [1978]. Relation Luopajärvi T, Kuorinka I, Virolainen M, between the low back pain syndrome and Holmberg M [1979]. Prevalence of teno- x-ray findings. I. Degenerative osteoarthritis. synovitis and other injuries of the upper Scand J Rehabil Med 10:135–145. extremities in repetitive work. Scand J Work Environ Health 5(Suppl 3):48–55. Maizlish N, Rudolph L, Dervin K, Sankaranarayan M [1995]. Surveillance and Maeda K, Hirayama H, Takamatsu M [1977]. prevention of work-related carpal tunnel Occupational cervicobrachial disorders in syndrome: an application of the Sentinel Events workwomen in assembly lines of a cigarette Notification System for Occupational Risks. factory. Jpn J Ind Health 19:8–21. Am J Ind Med 27(5):715–729.

Maeda K, Horiguchi S, Hosokawa M [1982]. Mäkelä M, Heliövaara M, Sievers K, History of the studies on occupa-tional Impivaara O, Knekt P, Aromaa A [1991]. cervicobrachial disorders in Japan and Prevalence, determinants, and consequences of remaining problems. J Human Ergol 11:17–29. chronic neck pain in Finland. Am J Epidemiol 134(11):1356–1367. Magnusson M, Hanson T [1994]. The effect of seat back inclination on spine height changes. MaKinnon SE, Novak CB [1997]. Repeti-tive Appl Ergonomics 25(5):294–298. strain in the workplace. J Hand Surg [Am] 22A(1):2–18. Magnusson ML, Pope MH, Wilder DG, Areskoug B [1996]. Are occupational drivers Marcus M, Gerr F [1996]. Upper extremity

R-19 musculoskeletal symptoms among female office Masset D, Malchaire J [1994]. Low back pain: workers: associations with video display epidemiologic aspects and work- terminal use and occupational psychosocial related factors in the steel industry. Spine stressors. Am J Ind Med 29(2):161–170. 19(2):143–146.

Marley RJ, Fernandez JE [1995]. Psycho- Matheson LN, Brophy RG, Vaughan KD, physical frequency and sustained exertion at Nunez C, Saccoman KA [1995]. Workers’ varying wrist postures for a drilling task. compensation managed care: preliminary Ergonomics 38(2):303–325. findings. J Occup Rehabil 5(1):27–36.

Marras WS, Schoenmarklin RW [1991]. Mathiowetz V, Kashman N, Volland G, Quantification of wrist motion in highly Weber K, Dowe M, Rogers S [1985]. Grip repetitive, hand-intensive industrial jobs. and pinch strength: normative data for adults. Washington, DC: U.S. Department of Health Arch Phys Med Rehabil 66(2):69–72. and Human Services, Public Health Service, Centers for Disease Control, National Institute Matthews KA, Cottington EM, Talbott E, for Occupational Safety and Health. Kuller LH, Siegel JM [1987]. Stressful work conditions and diastolic blood pressure among Marras W, Schoenmarklin RW [1993]. Wrist blue collar factory workers. Am J Epidemiol motions in industry. Ergonomics 126(2):280–290. 36(4):341–351. Mayer TG, Gatchel RJ, Kishino N, Keeley J, Marras WS, Lavender SA, Leurgans SE, Capra P, Mayer H, et al. [1985]. Objective Rajulu SL, Allread WG, Fathallah FA, et al. assessment of spinal function following [1993]. The role of dynamic three-dimen-sional industrial injury: a prospective study with com- trunk motion in occupationally- parison group and one-year follow-up. Spine related low back disorders: the effects of work- 10(6):482–493. place factors, trunk position, and trunk motion characteristics on risk of injury. Spine McCormack RR Jr., Inman RD, Wells A, 18(5):617–628. Berntsen C, Imbus HR [1990]. Prevalence of tendinitis and related disorders of the upper Marras WS, Lavender SA, Leurgans SE, extremity in a manufacturing workforce. J Fathallah FA, Ferguson SA, Allread WG, et al. Rheumatol 17:958–964. [1995]. Biomechanical risk factors for occupationally-related low back disorders. McKenna K, McGrann S, Blann A, Allen J Ergonomics 38(2):377–410. [1993]. An investigation into the acute vascular effects of riveting. Br J Ind Med 50:160–166. Martin DK, Dain SJ [1988]. Postural modifications of VDU operators wearing McNeil T, Warwick D, Andersson G, Schultz bifocal spectacles. Appl Ergonomics A [1980]. Trunk strengths in attempted flexion, 19(4):293–300. extension, and lateral bending in healthy

R-20 subjects and patients with low back disorders. Moon SD, Sauter SL [1996]. Beyond Spine 5(6):529–538. biomechanics: psychosocial aspects of musculoskeletal disorders in office work. Merriam WP, Burwell RG, Mulholland RC, London, England: Taylor and Francis. Pearson JCG, Webb JK [1980]. A study revealing a tall pelvis in subjects with low back Moore JS [1992]. Carpal tunnel syndrome. pain. J Bone Joint Surg [Br] 65B:153–156. Occup Med: State of the Art Rev 7(4):741– 763. Milerad E, Ekenvall L [1990]. Symptoms of the neck and upper extremities in dentists. Moore JS, Garg A [1991]. Determination of Scand J Work Environ Health 16(1): 129–134. the operational characteristics of ergonomic exposure assessments for prediction of Mintz G, Fraga A [1973]. Severe osteo- disorders of the upper extremities and back. In: arthritis of the elbow in foundry workers. Arch Proceedings of the 11th Congress of the Environ Health 27:78–80. International Ergonomics Association. London, England: Taylor & Francis, pp. 144–146. Mirbod SM, Inaba R, Iwata H [1992a]. A study on the vibration-dose limit for Japanese Moore JS, Garg A [1994]. A comparison of workers exposed to hand-arm vibration. Ind different approaches for ergonomic job Health 30:1–22. evaluation for predicting risk of upper extremity disorders. IEA 94. Occup Health Saf 2. Mirbod SM, Yoshida H, Nagata C, Inaba R, Komura Y, Iwata H [1992b]. Hand-arm Morgenstern H, Kelsh M, Kraus J, Margolis vibration syndrome and its prevalence in the W [1991]. A cross-sectional study of present status of private forestry enterprises in hand/wrist symptoms in female grocery check- Japan. Int Arch Occup Environ Health ers. Am J Ind Med 20(2):209–218. 64:93–99. Morris HH, Peters BH [1976]. Pronator Mirbod SM, Yoshida H, Komura Y, syndrome: clinical and electrophysiological Fujita S, Nagata C, Miyashita K, et al. [1994]. features in seven cases. J Neurol Neurosurg Prevalence of Raynaud’s phenome-non in Psychiatry 39:461–464. different groups of workers operating hand- held vibrating tools. Int Arch Occup Environ Mostardi RA, Noe DA, Kovacik MW, Health 66:13–22. Porterfield JA [1992]. Isokinetic lifting strength and occupational injury: a prospective study. Miyashita K, Morioka I, Iwata H, Takeda S Spine 17(2):189–193. [1992]. Symptoms of construction workers exposed to whole-body vibration and local Musson Y, Burdorf A, van Drimmelen D vibration. Int Arch Occup Environ Health [1989]. Exposure to shock and vibration and 64:347–351. symptoms in workers using impact power tools. Ann Occup Hyg 33(1):85–96.

R-21 Nachemson A, Lindh M [1969]. Measure- Nathan PA, Keniston RC, Meadows KD, ment of abdominal and back muscle strength Lockwood RS [1994a]. The relationship with and without low back pain. Scand J between body mass index and the diagnosis of Rehabil Med 1:60–65. carpal tunnel syndrome [letter to the editor]. Muscle Nerve, December 1994, Nachemson A, Elfstrom G [1970]. Intravital pp. 1491–1492. dynamic pressure measurements in lumbar discs. Scand J Rehabil Med 1(Suppl):1–40. Nathan PA, Takigawa K, Keniston RC, Meadows KD, Lockwood RS [1994b]. Nagata C, Yoshida H, Mirbod S, Komura Y, Slowing of sensory conduction of the median Fujita S, Inaba R, et al. [1993]. Cutaneous nerve and carpal tunnel syndrome in Japanese signs (Raynaud’s phenomenon, sclerodacty-lia, and American industrial workers. J Hand Surg and edema of the hands) and hand-arm [Am] 19B(1):30–34. vibration exposure. Int Arch Occup Environ Health 64:587–591. Necking LE, Dahlin LB, Fridin J, Lundborg G, Lundström R, Thornell LE Nathan PA, Meadows KD, Doyle LS [1988]. [1992]. Vibration-induced muscle injury: an Occupation as a risk factor for impaired experimental model and preliminary findings. J sensory conduction of the median nerve at the Hand Surg [Br] 17B:270–274. carpal tunnel. J Hand Surg [Br] 13B(2): 167–170. Nicholas JA [1990]. Clinical observations on sports-induced soft-tissue injuries in sports- Nathan PA, Keniston RC, Myers LD, induced inflammation. In: Leadbetter WB, Meadows KD [1992a]. Longitudinal study of Buckwalter JA, Gordon SL, eds. Rosemont, median nerve sensory conduction in industry: IL: American Academy of Orthopedic relationship to age, gender, hand dominance, Surgeons, p. 143. occupational hand use, and clinical diagnosis. J Hand Surg [Am] 17A(5):850–857. Nilsson T, Burström L, Hagberg M [1989]. Risk assessment of vibration exposure and Nathan PA, Keniston RC, Myers LD, white fingers among platers. Int Arch Occup Meadows KD [1992b]. Obesity as a risk Environ Health 61(7):473–481. factor for slowing of sensory conduction of the median nerve in industry: a cross-sectional and Nilsson T, Hagberg M, Burström L [1990]. longitudinal study involving 429 workers. J Prevalence and odds ratio of numbness and Occup Med 34(4):379–383. carpal tunnel syndrome in different exposure categories of platers. In: Okada A, Dupuis Nathan PA, Keniston RC, Meadows KD, WTH, eds. Hand-arm vibration. Kanazawa, Lockwood RS [1993]. Validation of Japan: Kyoei Press Co., pp. 235–239. occupational hand use categories. J Occup Med 35(10):1034–1042. Nilsson T, Hagberg M, Burström L, Kihlberg S [1994]. Impaired nerve conduction in the

R-22 carpal tunnel of platers and truck assemblers Nirschl RP [1975]. The etiology and treatment exposed to hand-arm vibration. Scand J Work of tennis elbow. J Sports Med 2(6):308–319. Environ Health 20:189–199. Nirschl R [1993]. Muscle and tendon trauma: Nilsson T, Hagberg M, Burström L, Lundström tennis elbow. In: Morrey B, ed. The elbow and R [1995]. A five-year follow-up study of nerve its disorders. 2nd ed. Philadel-phia, PA: W.B. conduction across the carpal tunnel among Saunders Company, pp. 552–573. workers at a heavy engineer-ing production plant. In: International Conference on the Nordgren B, Schek R, Linrotn K [1980]. Prevention of Musculo-skeletal Disorders Evaluation and prediction of back pain during (PREMUS). Montreal, Canada: International military field service. Scand J Rehabil Med Committee on Occupational Safety and Health, 12(1):1–8. pp.81–83. Ohara H, Nakagiri S, Itani T, Wake K, NIOSH [1989]. NIOSH criteria for a Aoyama H [1976]. Occupational health recommended standard: occupational exposure hazards resulting from elevated work rate to hand-arm vibration. Cincinnati, OH: U.S. situations. J Hum Ergol 5(2):173–182. Department of Health and Human Services, Public Health Service, Centers for Disease Ohlsson K, Attewell R, Skerfving S [1989]. Control, National Institute for Occupational Self-reported symptoms in the neck and upper Safety and Health, DHHS (NIOSH) limbs of female assembly workers. Scand J Publication No. 89–106. Work Environ Health 15(1):75–80.

NIOSH [1991]. Epidemiological basis for Ohlsson K, Hansson GA, Balogh I, Strömberg manual lifting guidelines. In: Scientific support U, Pålsson B, Nordander C, et al. [1994]. documentation for the revised 1991 NIOSH Disorders of the neck and upper limbs in lifting equation: technical contract reports. women in the fish processing industry. Occup Springfield, VA: National Technical Information Environ Med 51:826–832. Service (NTIS), NTIS No. PB–91–226274. Ohlsson K, Attewell R, Paisson B, NIOSH [1996]. National occupational Karlsson B, Balogh I, Johnsson B, et al. research agenda. Cincinnati, OH: U.S. [1995]. Repetitive industrial work and neck Department of Health and Human Services, and upper limb disorders in females. Am J Ind Public Health Service, Centers for Disease Med 27(5):731–747. Control and Prevention, National Institute for Occupational Safety and Health, DHHS Olafsdottir H, Rafnsson V [1995]. New tech- (NIOSH) Publication No. 96–115. niques increases the prevalences of musculoskeletal symptoms from upper limbs Nirschl RP [1973]. Tennis elbow. Orthop Clin among women in fish fillet plants. In: Second North Am 4(3):787–800. International Scientific Conference on Prevention of Work-Related Musculoskeletal

R-23 Disorders. Montreal, Canada, pp. 248–249. Pietri-Taleb F, Riihimaki H, Viikari-Juntura E, Lindström K, Moneta GB [1995]. The role of Ollivierre CO, Nirschl RP, Pettrone FA psychological distress and personal-ity in the [1995]. Resection and repair for medical tennis incidence of sciatic pain among working men. elbow: a prospective analysis. Am J Sports Am J Public Health 85:541– 545. Med 23(2):214–221. Pope MH, Rosen JC, Wilder DG, Frymoyer Onishi N, Namura H, Sakai K, Yamamoto T, JW [1980]. The relation between biome- Hirayama K, Itani T [1976]. Shoulder muscle chanical and psychological factors in patients tenderness and physical features of female with low back pain. Spine 5(2): 173–178. industrial workers. J Hum Ergol 5(2):87–102. Pope MH, Frymoyer JW, Andersson GBJ, et Osorio AM, Ames RG, Jones J, Castorina J, al. [1984]. Occupational low back pain. New Rempel D, Estrin W, et al. [1994]. Carpal York, NY: Praeger. tunnel syndrome among grocery store workers. Am J Ind Med 25(2):229–245. Pope MH, Andersson GBJ, Frymoyer JW, et al. [1991]. Occupational low back pain: Owen B, Damron C [1984]. Personal assessment, treatment, and prevention. St. characteristics and back injury among hospital Louis, MO: Mosby-Year Book, pp. 105–108. nursing personnel. Res Nurs Health 7:305–313. Porter RW [1987]. Does hard work prevent Parker KG, Imbus HR [1992]. Cumulative disc protrusion? Clin Biomechanics 2(4): trauma disorders. Current issues and 196–198. ergonomic solutions: a system approach. Chelsea, MI: Lewis Publishers. Pot F, Padmos P, Brouwers A [1987]. Determinants of the VDU operator’s well- Partridge REH, Duthie JJR [1968]. Rheu- being. In: Knave B, Wideback PG, eds. Work matism in dockers and civil servants: a with display units 86. Selected papers from the comparison of heavy manual and sedentary International Scientific Conference on Work workers. Ann Rheum Dis 27:559–568. with Display Units, Stockholm, Sweden, May 12–15, 1986. Amsterdam: Elsevier Science Payne R [1988]. Individual differences in the Publishers B.V., pp. 16–25. study of occupational stress. In: Cooper CL, Payne R, eds. Causes, coping, and Priest JD, Braden V, Gerberich SG [1980]. consequences of stress at work. Chichester, The elbow and tennis, part 1: an analysis of England: John Wiley & Sons, pp. 209–232. players with and without pain. Physician Sports Med 8:81–91. Pecina M, Krmpotic-Nemanic J, Markiewitz A [1991]. Tunnel syndromes. Boston, MA: CRC Priest JD, Hones HH, Tichenor CJ, Nagel DA Press. [1977]. Arm and elbow changes in expert tennis players. Minn Med 60:399– 404.

R-24 Punnett L, Robins JM, Wegman DH, vascular pattern of the rotator cuff. J Bone Joint Keyserling WM [1985]. Soft-tissue disorders Surg [Br] 52B(3):540–553. in the upper limbs of female assembly workers: impact of length of employment, work pace, Rempel D [1995]. Musculoskeletal loading and and selection. Scand J Work Environ Health carpal tunnel pressure. In: Gordon SL, Blair SJ, 11(6): 417–425. Fine LJ, eds. Repetitive motions disorders of the upper extremity. Rosemont IL: American Punnett L, Fine LJ, Keyserling WM, Herrin Academy of Orthopaedic Surgeons, pp. GD, Chaffin DB [1991]. Back disorders and 123–133. nonneutral trunk postures of automobile assembly workers. Scand J Work Environ Rempel D, Harrison R, Barnhart S [1992]. Health 17(5):337–346. Work-related cumulative trauma disorders of the upper extremity. JAMA 267(6): 838–841. Pyykkö I, Starak J [1986]. Pathophysiologic and hygienic aspects of hand-arm vibration. Rempel D, Dennerlein J, Mote CD, Armstrong Scand J Work Environ Health 12(4):237– 241. T [1994a]. A method of mea-suring fingertip loading during keyboard use. J Biomechanics Queinnec Y, Daniellou F, eds. [1991]. 27(8):1101–1104. Shoulder symptoms and headache in light sedentary work. In: Designing for everyone. Rempel D, Manojlovic R, Levinsohn DG, Proceedings of the 11th Congress of the Bloom T, Gordon L [1994b]. The effect of International Ergonomics Association, Paris, wearing a flexible wrist splint on carpal tunnel France, July 15–20, 1991. London, England: pressure during repetitive hand activity. J Hand Taylor and Francis, pp. 129–131. Surg [Am] 19A(1):106–110.

Radecki P [1995]. Variability in the median Riihimäki H [1991]. Low-back pain, its origin and ulnar nerve latencies: implications for and risk indicators. Scand J Work Environ diagnosing entrapment. J Occup Med 37: Health 17(2):81–90. 1293–1299. Riihimäki H [1995]. Hands up or back to Radwin RG, VanBergeijk E, Armstrong TJ work—future challenges in epidemiologic [1989]. Muscle response to pneumatic hand research on musculoskeletal diseases. Scand J tool torque reaction forces. Ergonomics Work Environ Health 21(6):401–403. 32(6):655–673. Riihimäki H, Tola S, Videman T, Hänninen K Ranney D, Wells R, Moore A [1995]. Upper [1989a]. Low-back pain and occupation: a limb musculoskeletal disorders in highly cross-sectional questionnaire study of men in repetitive industries: precise anatomical physical machine operating, dynamic physical work, and findings. Ergonomics 38(7):1408– 1423. sedentary work. Spine 14: 204–209. Rathburn JB Macnab I [1970]. The micro-

R-25 Riihimäki H, Wickström G, Hänninen K, Advances in industrial ergonomics and safety, Luopajärvi T [1989b]. Predictors of sciatic IV, Proceedings of the Annual International pain among concrete reinforcement workers Industrial Ergonomics and Safety Conference, and house painters: a five-year follow-up. Denver, CO, June 10–14, 1992. London, Scand J Work Environ Health 15(6):415– 423. England: Taylor and Francis, pp. 303–310.

Riihimäki H, Wickström G, Hänninen K, Rossignol AM, Morse EP, Summers VM, Mattson T, Waris P, Zitting AG [1989c]. Pagnotto LD [1987]. Video display terminal Radiographically detectable lumbar use and reported health symptoms among degenerative changes as risk indicators of back Massachusetts clerical workers. J Occup Med pain. Scand J Work Environ Health 29(2):112–118. 15:280–285. Rothfleisch SD, Sherman S [1978]. Carpal Riihimäki H, Viikari-Juntura E, Moneta G, tunnel syndrome: biomechanical aspects of Kuha J, Videman T, Tola S [1994]. Incidence occupational occurrence and implications of sciatic pain among men in machine operating, regarding surgical management. Orthop Rev dynamic physical work and sedentary work. 7(6):107–109. Spine 19(2):138–142. Rothman KJ, ed. [ 1986]. Modern epidemio- Ringel SP, Treihaft M, Carry M, Fisher R, logy. Boston, MA: Little, Brown, and Com- Jacobs P [1990]. Suprascapular neuropathy in pany. pitchers. Am J Sports Med 18(1):80–86. Roto P, Kivi P [1984]. Prevalence of epicon- Ritz BR [1995]. Humeral epicondylitis among dylitis and tenosynovitis among meatcutters. gas- and waterworks employees. Scand J Scand J Work Environ Health 10(3):203– 205. Work Environ Health 21:478–486. Rowe ML [1965]. Disc surgery and chronic Rodgers SH, The Ergonomics Group, Health low back pain. J Occup Med 7(5):196–202. and Environment Laboratories, Eastman Kodak Co. [1986]. Ergonomic design for Rowe ML [1985]. Backache at work. Fair- people at work. Vol. II: The design of jobs, port, NY: Perington Press. including work patterns, hours of work, manual materials handling tasks, methods to evaluate Ruberman W [1992]. Psychosocial influences job demands, and the physiological basis of on mortality of patients with coronary heart work. New York, NY: Van Nostrand Reinhold disease. JAMA 267(4): 559–560. Co., Inc. Ruberman W, Weinblatt E, Goldberg JD, Rose L, Ericson M, Glimskar B, Nordgren B, Chaudhary BS [1984]. Psychosocial Ortengren R [1992]. Ergo-index: a model to influences on mortality after myocardial determine pause needs after fatigue and pain infarction. N Engl J Med 311(9):552–559. reactions during work. In: Kumar S, ed.

R-26 Ryan GA, Bampton M [1988]. Comparison of Grandjean E, ed. Ergonomics and health in data process operators with and without upper modern offices. London, England: Taylor and limb symptoms. Community Health Stud Francis, pp. 129–135. 12(1):63–68. Sauter SL, Swanson NG [1996]. Psycho- Ryan GA, Hage B, Bampton M [1987]. logical aspects of musculoskeletal disorders in Postural factors, work organization and office work. In: Moon S, Sauter S, eds. musculoskeletal disorders. In: Buckle P, ed. Psychosocial factors and musculoskeletal Musculoskeletal disorders at work. disorders. London, England: Taylor and Proceedings of a conference held at the Francis. University of Surrey, Guilford, April, 1987. London, England: Taylor and Francis, Sauter SL, Gottlieb MS, Jones KC, Dodson pp. 251–253. VN, Rohrer KM [1983]. Job and health implications of VDT use: initial results of the Ryden LA, Molgaard CA, Bobbitt S, Wisconsin-NIOSH study. Commun of the Conway J [1989]. Occupational low-back ACM 26(4):284–294. injury in a hospital employee population: an epidemiologic analysis of multiple risk factors of Schibye B, Skov T, Ekner D, Christiansen J, a high-risk occupational group. Spine Sjøgaard G [1995]. Musculoskeletal symptoms 14:315–320. among sewing machine operators. Scand J Work Environ Health 21:427–434. Sairanen E, Brüshaber L, Kaskinen M [1981]. Felling work, low back pain and osteoarthritis. Schierhout GH, Meyers JE, Bridger RS Scand J Work Environ Health 7(1):18–30. [1995]. Work-related musculoskeletal disorders and ergonomic stressors in the South Saito K [1987]. Prevention of the hand-arm African workforce. Occup Environ Med vibration syndrome. Scand J Work Environ 52(1):46–50. Health 13:301–304. Schlesselman J [1978]. Assessing the effects of Sakakibara H, Masaru M, Kondo T, Yamada confounding variables. Am J Epidemiol S [1987]. Relation between overhead work 108:3–8. and complaints of pear and apple orchard workers. Ergonomics 30(5):805–815. Schlesselman J [1982]. Monographs in epidemiology and biostatistics. Case-control Sakakibara H, Miyao M, Kondo T, Yamada S studies: design, conduct, analysis. New York, [1995]. Overhead work and shoulder-neck NY: Oxford University Press. pain in orchard farmers harvesting pears and apples. Ergonomics 38(4):700–706. Schnall PL, Pieper C, Schwartz JE, Karasek RA, Schlussel Y, Devereux RB, et al. [1990]. Sauter SL [1984]. Predictors of strain in VDT The relationship between job strain, workplace users and traditional office workers. In: diastolic blood pressure, and left ventricular

R-27 mass: results of a case-control study. JAMA Hand wrist cumulative trauma disorders in 263(14):1929–1935. industry. Br J Ind Med 43(11):779–784.

Schnall PL, Schwartz JE, Landsbergis PA, Silverstein BA, Fine LJ, Armstrong TJ [1987]. Warren K, Pickering TG [1992]. The Occupational factors and the carpal tunnel relationship between job strain, alcohol, and syndrome. Am J Ind Med 11(3):343– 358. ambulatory blood pressure. Hypertension 19:488–494. Silverstein B, Grossman E, Kalat J, Nelson N [in press]. Upper limb and low back Schnall PL, Landsbergis PA, Baker D [1994]. musculoskeletal disorders: State and national Job strain and cardiovascular disease. Ann Rev estimates based on workers' compensation Public Health 15:381–411. accepted claims. Proceedings for managing ergonomics in the 90's. Alexandria, VA: Center Schottland JR, Kirschberg GJ, Fillingim R, for Office Technology and the American Davis VP, Hogg F [1991]. Median nerve Automobile Manufacturer's Association. latencies in poultry processing workers: an approach to resolving the role of industrial Sinclair A [1964]. Tennis elbow in industry. Br "cumulative trauma" in the development of J Ind Med 22:144–148. carpal tunnel syndrome. J Occup Med 33(5):627–631. Sjøgaard G [1988]. Intramuscular changes during long-term contraction. In: Corlett N, Scott V, Gijsbers K [1981]. Pain perception in Wilson J, Manenica I, eds. The ergonomics of competitive swimmers. Br Med J 283: 91–93. working postures: models, methods and cases. Proceedings of the First International Seidel H, Heide R [1986]. Long-term effects of Occupational Ergonomics Symposium, Zadar, whole-body vibration: a critical survey of the Yugoslavia, April 18–20, 1985, London, literature. Int Arch Occup Environ Health England: Taylor and Francis, pp. 136–143. 58:1–26. Sjøgaard G, Savard G, Juel C [1988]. Muscle Shinev V, Komleva L, Zhuravlev A, Druzhinin blood flow during isometric activity and its V [1992]. Study of the physical characteristics relation to muscle fatigue. Eur J Appl Physiol of percussive vibration of hand-held machines 57:327–335. and its effect on the body. Noise and Vibration Bulletin, pp. 138–140. Skie M, Zeiss J, Ebraheim NA, Jackson WT [1990]. Carpal tunnel changes and median Silverstein BA [1985]. The prevalence of nerve compression during wrist flexion and upper extremity cumulative trauma disorders in extension seen by magnetic resonance imaging. industry [Dissertation]. Ann Arbor, MI: The J Hand Surg [Am] 15A(6):939. University of Michigan. Skov T, Borg V, Orhede E [1996]. Psycho- Silverstein BA, Fine LJ, Armstrong TJ [1986]. social and physical risk factors for musculo-

R-28 skeletal disorders of the neck, shoulders, and disability. Clin Orthop 221:77–84. lower back in salespeople. Occup Environ Med 53:351–356. Snook SH, Ciriello VM [1991]. The design of manual handling tasks: revised tables of Skovron ML, Mulvihill MN, Sterling RC, maximum acceptable weights and forces. Nordin M, Tougas G, Gallagher M, et al. Ergonomics 34(9):1197–1213. [1987]. Work organization and low back pain in nursing personnel. Ergonomics Snook SH, Campanelli RA, Hart JW [1978]. 30(2):359–366. A study of three preventive approaches to low back injury. J Occup Med 20(7):478– 487. Skovron ML, Szpalski M, Nordin M, Melot C, Cukier D [1994]. Sociocultural Sokal RR, Rohlf FJ [1981]. Biometry: the factors and back pain: a population-based principles and practice of statistics in biological study in Belgian adults. Spine 19(2):129– 137. research. 2nd ed. San Francisco, CA: WH Freeman and Co. Slocum D [1960]. Overuse syndromes of the lower leg and foot in athletes. AAOS Sommerich CM, McGlothin JD, Marras WS Instructional Course Lectures, pp. 359–367. [1993]. Occupational risk factors associated with soft tissue disorders of the shoulder: a Smith MJ, Stammerjohn LW, Cohen BGF, review of recent investigations in the literature. Lalich NR [1977]. Job stress in video display Ergonomics 36(6):697–717. operations. In: Grandjean E, Vigliani E, eds. Ergonomic aspects of video display terminals. Spengler DM, Bigos S, Martin NA, Zeh J, London, England: Taylor and Francis, pp. Fisher L, Nachemson A [1986]. Back injuries 201–210. in industry: a retrospective study. I. Overview and cost analysis. Spine 11: 241–245. Snook SH [1978]. The design of manual handling tasks. Ergonomics 21(12):963– 985. Starck J, Jussi P, Ilmari P [1990]. Physical characteristics of vibration in relation to Snook SH [1982]. Low back pain in industry. vibration-induced white finger. Am Ind Hyg In: White AA, Gordon SL, eds. American Assoc J 51(4):179–184. Academy of Orthopaedic Surgeons Sym- posium on Idiopathic Low Back Pain. St. Stenlund B [1993]. Shoulder tendinitis and Louis, MO: C.O. Mosby Company, osteoarthritis of the acromioclavicular joint and pp. 23–38. their relation to sports. Br J Sports Med 27:125–128. Snook SH [1987]. Approaches to preplace- ment testing and selection of workers. Stenlund B, Goldie I Hagberg M, Hogstedt C, Ergonomics 30(2):241–247. Marions O [1992]. Radiographic osteo- arthrosis in the acromioclavicular joint resulting from manual work or exposure to vibration. Br Snook SH, Webster BS [1987]. The cost of

R-29 J Ind Med 49:588–593. 648.

Stenlund B, Goldie I, Hagberg M, Hogstedt C Svensson H [1982]. Low back pain in 40- to [1993]. Shoulder tendinitis and its relation to 47-year-old men: II. Socioeconomic factors heavy manual work and exposure to vibration. and previous sickness absence. Scand J Re- Scand J Work Environ Health 19: 43–49. habil Med 14:55–60.

Stetson DS, Keyserling WM, Silverstein BA, Svensson H, Andersson GBJ [1983]. Low- Leonard JA [1991]. Observational analysis of back pain in 40- to 47-year old men: work the hand and wrist: a pilot study. Appl Occup history and work environment factors. Spine Environ Hyg 6(11):927–937. 8:272–276. Stetson DS, Silverstein BA, Keyserling WM, Wolfe RA, Albers JW [1993]. Median sensory Svensson H, Andersson GBJ [1989]. The distal amplitude and latency: comparisons relationship of low-back pain, work history and between nonexposed managerial/professional work environment, and stress: a retrospective employees, and industrial workers. Am J Ind cross-sectional study of 38- to 64-year old Med 24:175–189. women. Spine 14(5):517–522.

Stevens JC, Sun MD, Beard CM, O’Fallon Sweeney MH, Petersen M, O’Neill V [1994]. WM, Kurland LT [1988]. Carpal tunnel Hazard evaluation and technical assistance syndrome in Rochester, Minnesota, 1961 to report: Registry of Interpreters for the Deaf, 1980. Neurology 38:134–138. Ohio Chapter. Cincinnati, OH: U.S. Department of Health and Human Services, Stevens PH, Chase DO, Brownlie AW [1966]. Public Health Service, Centers for Disease Industrial design of a narrow aisle sit-down lift Control and Prevention, National Institute for truck. Hum Factors 8:317–325. Occupational Safety and Health, NIOSH Report No. HHE 92–0268–2477. Stock SR [1991]. Workplace ergonomic factors and the development of musculo- skeletal disorders of the neck and upper limbs: Szabo RM, Chidgey LK [1989]. Stress carpal a meta-analysis. Am J Ind Med 19(1):87–107. tunnel pressures in patients with carpal tunnel syndrome and normal patients. J Hand Surgery Stubbs DA [1986]. Back pain and its [Am] 14A(4):624–627. relationship to manual handling. In: Morrison D, Hartley L, Kemp D, eds. Trends in the Takala E, Viikari-Juntura E [1991]. Muscle ergonomics of work. Proceed-ings of the 23rd force endurance and neck-shoulder symp-toms Annual Conference of the Ergonomics Society of sedentary workers: an experimental study on of Australia, pp. 118–128. bank cashiers with and without symptoms. Amsterdam, The Netherlands: Elsevier Science Susser M [1991]. What is a cause and how do Publishers. B.V., pp. 123–132. we know one?: a grammar for pragmatic epidemiology. Am J Epidemiol 133(7):635– Takala EP, Viikari-Juntara E, Moneta G,

R-30 Saarenmaa K, Kaivanto K [1991]. Predic-tors factors and symptoms from the locomotor for the natural course of neck–shoulder system: a multicausal analysis. Scand J Rehabil symptoms and headache in light sedentary Med 23(3):165–173. work. In: Queinnec Y, Daniellou F, eds. Designing for everyone. Proceedings of the Theorell T, Ahlberg-Hulten G, Jodko M, Sigala 11th Congress of the International Ergono-mics F, de la Torre B [1993]. Influence of job strain Association, Paris, France, July 15–20, 1991. and emotion on blood pressure levels in female London, England: Taylor and Francis, pp. hospital personnel during work hours. Scand J 129–131. Work Environ Health 19(5):313–318. Takeuchi T, Futatsuka M, Imanishi H, Yamada S [1986]. Pathological changes observed in the Tola S, Riihimäki H, Videman T, Viikari- finger biopsy of patients with vibration-induced Juntura E, Hänninen K [1988]. Neck and white finger. Scand J Work Environ Health shoulder symptoms among men in machine 12:280–283. operating, dynamic physical work and sedentary work. Scand J Work Environ Health Tanaka S, Wild DK, Seligman PJ, Halperin 14(5):299–305. WE, Behrens VJ, Putz-Anderson V [1995]. Prevalence of work-related carpal tunnel Toomingas A, Hagberg M, Jorulf L, Nilsson T, syndrome among U.S. workers: analysis of the Burström L, Kihlberg S [1991]. Outcome of occupational health supplement data of 1988 the abduction external rotation test among National Health Interview Survey. Am J Ind manual and office workers. Am J Ind Med Med 27(4):451–470. 19(2):215–227.

Tanaka S, Wild DK, Cameron L, Freund E [in Toomingas A, Theorell T, Michélsen H, et al. press]. Association of occupational and [1992]. Associations between preceived non-occupational risk factors with the psychosocial job factors and prevalence of prevalence of self-reported carpal tunnel musculoskeletal disorders in the neck and syndrome in a national survey of the working shoulder regions. In: Hagberg M, Kilbom Å, population. Am J Ind Med. eds. International Scientific Conference on Tanii K, Kogi K, Sadoyama T [1972]. Prevention of Work-Related Musculo-skeletal Spontaneous alteration of the working arm in Disorders PREMUS, Sweden, May 12–14, static overhead work. J Hum Ergol 1:143–155. 1992. Arbete och Hälsa 17:289.

Tauber J [1970]. An unorthodox look at Toroptsova NV, Benevolenskaya LI, Karyakin backaches. J Occup Med 12(4):128–130. AN, Sergeev IL, Erdesz S [1995]. Cross-sectional study of low back pain among Taylor W [1982]. Vibration white finger in the workers at an industrial enterprise in Russia. workplace. J Soc Occup Med 32:159–166. Spine 20(3):328–332.

Theorell T, Harms-Ringdahl K, Ahlberg- Torrell G, Sanden A, Järvholm B [1988]. Hultén G, Westin B [1991]. Psychosocial job Musculoskeletal disorders in shipyard workers.

R-31 J Soc Occup Med 38(4):109–112. related risk factors for trapezius myalgia. In: Hagberg M, Kolbom Å, eds. International Troup JDG [1978]. Driver’s back pain and its Scientific Conference on Prevention of Work- prevention—a review of the postural, vibratory, Related Musculoskeletal Disorders PREMUS, and muscular factors, together with the problem Sweden, May 12–14, 1992. Arbete och Hälsa of transmitted road shock. Appl Ergonomics 17:307–309. 9(4):207–214. Veiersted KG, Westgaard RH [1994]. Subjectively assessed occupational and Troup JDG, Martin JW, Lloyd DC [1981]. individual parameters as risk factors for Back pain in industry: a prospective survey. trapezius myalgia. Int J Ind Ergo Spine 6(1):61–69. 13(3):235–245.

Troup JDG, Foreman TK, Baxter CE, Brown Veiersted KB, Westgaard RH, Andersen P D [1987]. Musculoskeletal disorders at work, [1993]. Electromyographic evaluation of tests of manual working capacity and the muscular work pattern as a predictor of prediction of low back pain. In: Buckle P, ed. trapezius myalgia. Scand J Work Environ Musculoskeletal disorders at work. London, Health 19:284–290. England: Taylor and Francis, pp. 165–170. Vessey MP, Villard-Mackintosh L, Yeates D Uhl JE, Wilkinson WE, Wilkinson CS [1987]. [1990]. Epidemiology of carpal tunnel Aching backs? A glimpse into the hazards of syndrome in women of childbearing age. nursing. AAOHN J 35(1):13–17. Findings in a large cohort study. Int J Epidemiol 19(3):655–659. Ulin SS, Armstrong TJ, Snook SH, Keyserling WM [1993]. Perceived exertion and Videman T, Nurminen T, Tola S, Kuorinka I, discomfort associated with driving screws at Vanharanta H, Troup JDG [1984]. Low-back pain in nurses and some loading factors of various work locations and at different work work. Spine 9(4):400–404. frequencies. Ergonomics 36(7):833–846. Videman T, Rauhala H, Asp S, Lindstrom K, Undeutsch K, Gärtner KH, Luopajärvi T, Cedercreutz G, Kamppi M, et al. [1989]. Küpper R, Karvonen MJ, Löwenthal I, et al. Patient-handling skill, back injuries, and back [1982]. Back complaints and findings in pain: an intervention study in nursing. Spine transport workers performing physically heavy 14(2):148–156. work. Scand J Work Environ Health 8(Suppl 1):92–96. Videman T, Nurminen M, Troup JDG [1990]. Lumbar spine pathology in cadaveric material in Ursin H, Endresen IM, Ursin G [1988]. relation to history of back pain, occupation, and Psychological factors and self-reports of muscle physical loading. Spine 15(8):728–740. pain. Eur J Appl Physiol 57(3):282–290. Vihma T, Nurminen M, Mutanen P [1982]. Veiersted KG, Westgaard RH [1992]. Work- Sewing machine operators’ work and

R-32 musculoskeletal complaints. Ergonomics [1986]. Generation of muscle tension related to 25(4):295–298. a demand of continuing attention. In: Knave B, Wideback PG, eds. Work with display units. Viikari-Juntura E [1995]. The role of physical Amsterdam, The Netherlands: Elsevier Science stressors in the development of hand/wrist, and Publishers, B.V. North-Holland, pp. 288–293. elbow disorders. In: Gordon SL, Blair SJ, Armstrong, et al., eds. Pathophysiology and Walsh K, Varnes N, Osmond C, Styles R, clinical factors in upper extremity repetitive Coggin D [1989]. Occupational causes of low back pain. Scand J Work Environ Health motion syndromes. Bethesda, MD: American 15(1):54–59. Academy of Orthopedic Surgeons. Waris P, Kuorinka I, Kurppa K, Luopajärvi T, Viikari-Juntura EJ, Vuori J, Silverstein B, Virolainen M, Pesonen K, et al. [1979]. Kalimo R, Kuosma E, Videman T [1991a]. A Epidemiologic screening of occupational neck life-long prospective study on the role of and upper limb disorders: methods and criteria. psychosocial factors in neck-shoulder and low Scand J Work Environ Health 5(Suppl back pain. Spine 16:1056–1061. 3):25–38.

Viikari-Juntura E, Kurppa K, Kuosma E, Washington State Department of Labor and Huuskonen M, Kuorinka I, Ketola R, et al. Industries [1996]. Work-related musculo- [1991b]. Prevalence of epicondylitis and elbow skeletal disorders. Washington State summary pain in the meat-processing industry. Scand J 1992–1994. Olympia, WA: Washington State Work Environ Health 17(1):38–45. Department of Labor and Industries, Consultation and Compliance Services Viikari-Juntura E, Riihimäki H, Tola S, Division. Videman T, Mutanen P [1994]. Neck trouble in machine operating, dynamic physical work Waters TR, Putz-Anderson V, Garg A, Fine LJ [1993]. Revised NIOSH equation for the and sedentary work: a prospective study on design and evaluation of manual lifting tasks. occupational and individual risk factors. J Clin Ergonomics 36:749–776. Epidemiol 47(12):1411– 1422. Webster BS, Snook SH [1994a]. The cost of Virokannas H [1995]. Dose-response relation compensable upper extremity cumulative between exposure to two types of hand-arm trauma disorders. J Occup Med 36(7):713– vibration and sensorineural per-ception of 727. vibration. Occup Environ Med 52:332–336. Webster BS, Snook SH [1994b]. The cost of 1989 workers’ compensation low back pain Virokannas H, Tolonen U [1995]. Responses claims. Spine 19(10):1111–1116. of workers exposed to vibration in autono-mic nervous function tests. Int Arch Occup Environ Weiss ND, Gordon L, Bloom T, So Y, Rempel Health 67(3):201–205. DM [1995]. Position of the wrist associated with the lowest carpal-tunnel pressure: Waersted M, Bjorklund R, Westgaard RH implications for splint design.

R-33 J Bone Joint Surg 77A (11):1695–1699. muscle load. Ergonomics 30(7):911–923.

Welch LS, Hunting KL, Kellogg J [1995]. White AA, Gordon SL [1982]. Synopsis: Work-related musculoskeletal symptoms workshop on idiopathic low back pain. Spine among sheet metal workers. Am J Ind Med 7(2):141–149. 27(6):783–791. WHO [1985]. Identification and control of Wells JA, Zipp JF, Schuette PT, McEleney J work-related diseases. Geneva, Switzerland: [1983]. Musculoskeletal disorders among World Health Organization. WHO Technical letter carriers. J Occup Med 25(11):814– Report Series 714. 820. WHO [1989]. Work with visual display Werner RA, Franzblau A, Albers JW, Buchele terminals: psychosocial aspects and health. J H, Armstrong TJ [1997]. Use of screening Occup Med 31(12):957–968. nerve conduction studies for predicting future carpal tunnel syndrome. Occup Environ Med Wickström BO, Kjellberg A, Landström U 54:96–100. [1994]. Health effects of long-term occupa- tional exposure to whole-body vibration: a Werner RA, Franzblau A, Johnston E [1994a]. review. Int J Ind Ergonomics 14:273–292. Quantitative vibrometry and electro- physiological assessment in screening for carpal Wieslander G, Norbäck D, Göthe CJ, tunnel syndrome among industrial workers: a Juhlin L [1989]. Carpal tunnel syndrome (CTS) comparison. Arch Phys Med and exposure to vibration, repetitive wrist Rehabil 75:1228–1232. movements, and heavy manual work: a case-referent study. Br J Ind Med 46(1): Werner S, Albers JW, Franzblau A, Armstrong 43–47. TJ [1994b]. The relationship between body mass index and the diagnosis of carpal tunnel Wiktorin C, Karlqvist L, Winkel J [1993]. syndrome. Muscle Nerve December:1492. Validity of self-reported exposures to work postures and manual materials handling. Scand Westerling D, Jonsson BG [1980]. Pain from J Work Environ Health 19:208–214. the neck-shoulder region and sick leave. Scand J Soc Med 8(3):131–136. Wilder DG, Pope MH [1996]. Epidemio- logical and aetiological aspects of low back Westgaard R, Aaras A [1985]. The effect of pain in vibration environments—an update. Clin improved workplace design on the develop- Biomechanics 11:61–73. ment of work-related musculoskeletal disorders. Appl Ergonomics 16(2):91–97. Wilson RN, Wilson S [1957]. Tenosynovi-tis in industry. Practitioner 178:612–615. Westgaard R, Bjorklund R [1987]. Genera- Winkel J, Westgaard R [1992]. Occupational tion of muscle tension additional to postural and individual risk factors for shoulder-neck

R-34 complaints. Part II: the scientific basis (literature Eastman Kodak, Windsor, Colorado. review) for the guide. Int J Ind Ergonomics Cincinnati, OH: U.S. Department of Health, 10:85–104. Education, and Welfare, Public Health Service, Center for Disease Control, National Institute Winkel J, Westgaard R [1992]. Occupational for Occupational Safety and Health, DHEW and individual risk factors for shoulder-neck (NIOSH) Report No. TA 76–0131. complaints. Part II. The scientific basis (literature review) for the guide. Int J Ind Yu ITS, Wong TW [1996]. Musculoskeletal Ergonomics 10:85–104. problems among VDU workers in a Hong Kong bank. Occup Med 46(4):275–280. Winn FJ, Habes DJ [1990]. Carpal tunnel area as risk factor for carpal tunnel syndrome. Zwerling C, Ryan J, Schootman M [1993]. A Muscle Nerve 13:254–258. case-control study of risk factors for industrial low back injury: the utility of preplacement Wisseman C, Badger D [1976]. Hazard screening in defining high-risk groups. Spine evaluation and technical assistance report: 18(9):242–247.

R-35 APPENDIX A Epidemiologic Review

Various investigators have used different occupational epidemiologic methods to identify the patterns of work-related MSD occurrence in different working groups, as well as the factors that influence these disease patterns. The following section briefly summarizes these study designs and then addresses the most common biases (such as misclassification or selection) that can affect the results of these studies.

TYPES OF EPIDEMIOLOGIC STUDY DESIGNS REVIEWED The NIOSH reviewers have first addressed studies that use a prospective approach. Prospective cohort studies, identify groups of subjects (exposed and nonexposed) and observe them over a period of time to compare the number of new work-related MSD cases in the two groups. All subjects are initially disease-free. The rate (or risk) of new cases (the incidence) is calculated for both groups, and the ratio of these two incidences (the relative risk or rate ratio, RR) can be used to assess the association of the exposure with the occurrence of the MSD. A RR greater than 1.0 implies that the incidence of cases was higher in the exposed group than in the nonexposed group and that an association has been observed between the exposure and the disease. A confidence interval (CI) is derived, which is an estimated range of values within which the true RR is likely to fall. The CI reflects the precision of the effect observed in the study. Ordinarily, if the CI includes 1.0, the association between the exposure and the MSD could be due to chance alone and the elevated odds ratio (OR) is not considered statistically significant.

The cohort study ensures that the exposure to work-related factors occurs before the observation of the MSD, thereby allowing a causal interpretation of the observed association. Cohort studies are often done prospectively; they follow a group of current workers forward in time. The length of time required for a prospective study depends on the problem studied. With adverse health conditions that occur as a result of long-term exposure to some factor in the workplace, many years may be needed. Extended time periods make prospective studies costly. Arguing causation is more difficult with extended time periods because other events may affect outcome. Prospective studies that require long periods of time are especially vulnerable to problems associated with worker follow-up, particularly worker attrition (workers discontinue participation in the study) and worker migration (diseased workers move to other employment before investigators ascertain their disease).

A-1 The second type of epidemiologic study evaluated for this document is the case-control study, which is retrospective and examines differences in exposures among workers with (cases) and without (controls) MSDs. In such studies, cases should be all incident (new) cases in a given population over a defined period or a representative sample of the cases. Controls should be a representative sample of non-cases from the same population. The ratio of the odds of exposed cases to the odds of exposed controls is called the OR. An OR above 1.0 indicates an association between the exposure and the work-related MSD, and a 95% CI indicates the probable range of the true OR. Case control studies are useful for evaluating rarely occurring conditions or small numbers of cases. One limitation of case control studies is the difficulty of obtaining accurate information about past exposures. In occupational studies of MSDs, a further limitation of case-control studies is the difficulty of identifying cases who are representative of all cases that occurred in a defined period (many of these workers will have left the workforce). Another problem with case-control studies is the selection of an inappropriate control group.

Third, the reviewers considered cross-sectional studies. Cross-sectional studies provide a “snapshot in time” of a disease process; that is, they measure both health outcomes and exposures at a single point in time. These studies usually identify occupations with differing levels of exposure and compare the prevalences of MSDs in each group. Cross-sectional studies are most useful for identifying risk factors of a relatively frequent disease with a long duration that is often undiagnosed or unreported [Kleinbaum et al. 1982]. Typically, cross-sectional studies do not provide the evidence of the correct temporal relationship between exposure and disease inherent in prospective studies, but they nevertheless can be valuable. Some cross-sectional studies discussed here had inclusion criteria such as working at a specific job for a defined period of time before onset of symptoms. This condition adds a dimension of temporality to the studies. A common problem with cross-sectional studies that use surveys is obtaining sufficiently large response rates; many people who are asked to participate decline because they are busy, not interested, etc. The conclusions are therefore based on a subset of workers who agree to participate, and these workers may not be representative of or similar to the entire population of workers. Furthermore, cross-sectional studies are often confined to current workers who may not be representative of true prevalence rates if workers with disease have left the workforce. (The problem of representativeness is not confined to cross-sectional studies and may occur in the other study designs mentioned whenever subjects are selected, decline, or drop out.) Either ORs or prevalence ratios (PRs) (proportion of diseased in exposed divided by the proportion of diseased in unexposed) may be used to report results in cross-sectional studies.

The last type of observational study used is the case-series study, in which certain characteristics of a group (or series) of cases (or patients) are described. The simplest design is a set of case reports for which the author describes some interesting or intriguing observations that occurred in a small number of patients. Cases included in case series have usually been drawn from a single patient population, whose makeup may have influenced the observations noted because of selection bias. Case-series studies frequently lead to a generation of hypotheses that are subsequently investigated in a cross- sectional, case-control, or prospective study. Because case-series do not involve comparison groups

A-2 (who do not have the condition or exposure to the risk factors being studied), some investigators would not consider them epidemiologic studies because they are generally not planned studies and do not involve any research hypotheses.

BIASES AND OTHER ISSUES IN EPIDEMIOLOGIC STUDIES In interpreting the validity of epidemiologic studies to provide evidence for work-relatedness of MSDs, several assumptions and sources of bias must be considered when analyzing the findings from such studies.

1. Selection bias (internal validity). In occupational health studies, at least two types of selection bias may occur: (a) a selection of “healthy workers” in the work population studied, and (b) an exclusion of “sick” workers who leave the active workforce. Both of these biases tend to cause an underestimate of the true relationship between a workplace risk factor and an observed health effect because the workers who are in better health tend to be those in the workforce and available for study.

A basic assumption underlying the analysis of these studies is that the selected cases of work- related MSDs in the specific studies are representative of all workers at that worksite with work- related MSDs. In a single study, representativeness generally increases with increasing population size and participation rate. A parallel assumption is that the nondiseased groups are representative of the entire nondiseased population. The fact that some cases leave the workforce causes the disease prevalence among currently employed workers to be underestimated. However, if cases are missing from the current workforce in equal proportion for both nonexposed and exposed workers, the underestimate of prevalence will not affect the internal validity of the study.

2. Generalizability (external validity). Some studies are based on a single population, occupation, or restricted data base (individual insurance companies, specific industrial settings) and, therefore, the sample may not be representative of the general population. Another assumption is that MSD cases in one study are comparable to cases in another study. This assumption needs particular scrutiny in work-related MSD studies because no standardized case definitions may exist for the particular illnesses.

3. Misclassification bias. Misclassification bias may be introduced during selection of cases and determination of their exposure. Erroneous diagnoses may result in work-related MSD cases misclassified as noncases, and similarly, noncases may be misclassified as cases. The calculated RR or OR would usually underestimate the true association because of a dilutional effect if both exposed and nonexposed cases are equally misclassified. Similarly, misclassification can occur when determining the exposure factor of interest. Again, such misclassification will create a bias towards finding no association if equal misclassification is assumed for cases and noncases.

A-3 4. Confounding and effect modification. Other factors may explain the supposed relationship between work and disease. Confounding is a situation in which the relationship (in this case with MSDs) appears stronger or weaker than it truly is as a result of something (the confounder) being associated with both the outcome and the apparent causal factor. In other words, the risk estimate is distorted because symptoms of exposed and nonexposed workers differ because of some other factors that cause disease. For example, diabetes might result in abnormal nerve conduction testing, a sign of CTS. If a higher proportion of exposed workers than nonexposed workers were diabetic, diabetes would act as a positive confounder, causing an apparent exposure-disease association.

An effect modifier is a factor that alters the effect of exposure on disease. For example, it is possible that repetitive motion causes tendinitis only in older workers; in this case, age would be an effect modifier. Although effect modification is not a bias per se, if an investigator has failed to analyze old and young workers separately, the investigator might have missed a true work/disease association.

5. Sample size, precision, and CIs. The CI around an estimated measure of effect (such as a RR) is an estimated range of values in which the true effect is likely to fall. It reflects the precision of the effect observed in the study. Large studies generally have smaller CIs and can estimate effects more precisely. In studies that are “statistically significant” the CI excludes the null value for no effect (for example, a RR of 1.0). Small studies are generally less precise, lead to wider CIs, and less likely to be “statistically significant” even if the exposed have a greater prevalence of disease than the nonexposed.

A-4 APPENDIX B Individual Factors Associated with Work- Related Musculoskeletal Disorders (MSDs)

Although the purpose of this document is to examine the weight of evidence for the contribution of work factors to MSDs, the multifactorial nature of MSDs requires a discussion of individual factors that have been studied to determine their association with the incidence and prevalence of work-related MSDs. These factors include age [Guo et al. 1995; Biering-Sorensen 1983; English et al. 1995; Ohlsson et al. 1994]; gender [Hales et al. 1994; Johansson 1994; Chiang et al. 1993; Armstrong et al. 1987a]; anthropometry [Werner et al. 1994b; Nathan et al. 1993, Heliövaara 1987]; and cigarette smoking [Finkelstein 1995; Owen and Damron 1984; Svensson and Andersson 1983; Kelsey et al. 1990; Hildebrandt 1987], among others. Nonoccupational physical activities, such as nonoccupational VDT use, hobbies, second jobs, and household activities that might increase risk for MSDs are described in the detailed tables for those studies in which they were analyzed as risk factors.

A worker's ability to respond to external work factors may be modified by his/her own capacity, such as tissue resistance to deformation when exposed to high force demands. The level, duration, and frequency of the loads imposed on tissues, as well as adequacy of recovery time, are critical components in whether increased tolerance (a training or conditioning effect) occurs, or whether reduced capacity occurs which can lead to MSDs. The capacity to perform work varies with gender and age, among workers, and for any worker over time. The relationship of these factors and the resulting risk of injury to the worker is complex and not fully understood.

Certain epidemiologic studies have used statistical methods to take into account the effects of these individual factors (e.g., gender, age, body mass index), that is, to control for their confounding or modifying effects when looking at the strength of work-related factors. Studies that fail to control for the influence of individual factors may either mask or amplify the effects of work-related factors. The comments column of the detailed tables notes whether studies have adjusted for potential confounders.

A number of factors can influence a person's response to risk factors for MSDs in the workplace and elsewhere. Among these are the following:

AGE

B-1 The prevalence of MSDs increases as people enter their working years. By the age of 35, most people have had their first episode of back pain [Guo et al. 1995; Chaffin 1979]. Once in their working years (ages 25 to 65), however, the prevalence is relatively consistent [Guo et al. 1995; Biering-Sorensen 1983]. Musculoskeletal impairments are among the most prevalent and symptomatic health problems of middle and old age [Buckwalter et al. 1993]. Nonetheless, age groups with the highest rates of compensable back pain and strains are the 20–24 age group for men, and 30–34 age group for women. In addition to decreases in musculoskeletal function due to the development of age-related degenerative disorders, loss of tissue strength with age may increase the probability or severity of soft tissue damage from a given insult.

Another problem is that advancing age and increasing number of years on the job are usually highly correlated. Age is a true confounder with years of employment, so that these factors must be adjusted for when determining relationship to work. Many of the epidemiologic studies that looked at populations with a wide age variance have controlled for age by statistical methods. Several studies found age to be an important factor associated with MSDs [Guo et al. 1995; Biering-Sorensen 1983; English et al. 1995; Ohlsson et al. 1994; Riihimäki et al. 1989a; Toomingas et al. 1991] others have not [Herberts et al. 1981; Punnett et al. 1985]. Although older workers have been found to have less strength than younger workers, Mathiowetz et al. [1985] demonstrated that hand strength did not decline with aging; average hand pinch and grip scores remained relatively stable in their population with a range of 29 to 59 years. Torell et al. [1988] found no correlation between age and the prevalence of MSDs in a population of shipyard workers. They found a strong relationship between workload (categorized as low, medium, or heavy) and symptoms or diagnosis of MSDs.

Other studies have also reported a lack of increased risk associated with aging. For example, Wilson and Wilson [1957] reported that the age and gender distribution of 88 patients with tenosynovitis from an ironworks closely corresponded to that of the general population of that plant. Similarly, Wisseman and Badger [1976] reported that the median age of workers with chronic hand and wrist injuries in their study was 23 years, while the median age of the unaffected workers was 24 years. Riihimäki et al. [1989a] found a significant relationship between sciatica and age in machine operators, carpenters, and sedentary workers. Age was also a strong risk factor for neck and shoulder symptoms in carpenters, machine operators and sedentary workers [Riihimäki et al. 1989a]. Some authors may have incorrectly attributed age as the sole cause of their findings in their analysis, when data presented suggested a relationship with work [Schottland et al. 1991].

An explanation for the lack of an observed relationship between an increased risk for MSDs and aging may be “survivor bias” (this is different from the “healthy worker effect”). If workers who have health problems leave their jobs, or change jobs to one with less exposure, the remaining population includes only those workers whose health has not been adversely affected by their jobs. As an example, in a study of female plastics assembly workers, Ohlsson et al. [1989] reported that the degree of increase in the odds of neck and shoulder pain with the duration of employment depended on the age of the worker. For the younger subjects, the odds increased significantly as the duration of employment

B-2 increased (p=0.01), but for the older ones no statistical change was found with length of employment. The older women who had been employed for shorter periods of time had more reported symptoms than the younger ones, while older workers with longer employment times reported fewer symptoms than younger workers. Ohlsson et al. [1989] interviewed 76 former assembly workers and found that 26% reported pain as the cause of leaving work. This finding supports the likely role of a survivor bias in this study, the effect of which is to underestimate the true risk of developing MSDS, in this case in the older workers.

GENDER Some studies have found a higher prevalence of some MSDs in women [Bernard et al. 1994; Hales et al. 1994; Johansson 1994; Chiang et al. 1993]. A male to female ratio of 1:3 was described for carpal tunnel syndrome (CTS) in a population study in which occupation was not evaluated [Stevens et al. 1988]. However, in the Silverstein [1985] study of CTS among industrial workers, no gender difference could be seen after controlling for work exposure. Franklin et al. [1991] found no gender difference in workers compensation claims for CTS. Burt et al. [1990] found no gender difference in reporting of neck or upper extremity MSD symptoms among newspaper employees using video display terminals (VDTs). Nathan et al. [1988, 1992a] found no gender differences for CTS. In contrast, Hagberg and Wegman [1987] reported that neck and shoulder muscular pain is more common among females than males, both in the general population and among industrial workers. Whether the gender difference seen with some MSDs in some studies is due to physiological differences or differences in exposure is unclear. One laboratory study, Lindman et al. [1991], found that women have more type I muscle fibers in the trapezius muscle than men, and have hypothesized that myofascial pain originates in these Type I muscle fibers. Ulin et al. [1993] noted that significant gender differences in work posture were related to stature and concluded that the lack of workplace accommodation to the range of workers' height and reach may, in part, account for the apparent gender differences. The reporting bias may exist because women may be more likely to report pain and seek medical treatment than men [Armstrong et al. 1993; Hales et al. 1994]. The fact that more women are employed in hand-intensive jobs and industries may account for the greater number of reported work-related MSDs among women. Byström et al. [1995] reported that men were more likely to have deQuervain’s disease than women; they attributed this to more frequent use of hand tools. Some studies have reported that workplace risk factors account for increased prevalence of MSDs among women more than personal factors (e.g., Armstrong et al. [1987a], McCormack et al. [1990]). In a recent evaluation of Ontario workers compensation claims for “RSI,” Asbury [1995] reported a RR for female to male claims ranging from 1.3 to 1.6 across industries. Within 5 different broad occupational categories, females were approximately 2–5 times as likely to have a lost-time RSI claim. No information on gender differences in hand intensive jobs was reported. May researchers have noted that men and women tend to be employed in different jobs.

In order to separate the effect of work risk factors from potential effects that might be attributable to biological differences, researchers must study jobs that men and women perform relatively equally.

B-3 SMOKING Several papers have presented evidence that a positive smoking history is associated with low back pain, sciatica, or intervertebral herniated disc [Finkelstein 1995; Owen and Damron 1984; Frymoyer et al. 1983; Svensson and Anderson 1983; Kelsey et al. 1984]; whereas in others, the relationship was negative [Kelsey et al. 1990; Riihimäki et al. 1989b; Frymoyer 1993; Hildebrandt 1987]. Boshuizen et al. [1993] found a relationship between smoking and back pain only in those occupations that required physical exertion. In their study, smoking was more clearly related to pain in the extremities than to pain in the neck or the back. Deyo and Bass [1989] observed that the prevalence of back pain increased with the number of pack-years of cigarette smoking and with the heaviest smoking level. Heliövaara et al. [1991] only observed a relationship in men and women older than 50 years. Two studies did not find a relationship between sciatica and smoking among concrete reinforcement workers and house painters [Heliövaara et al. 1991; Riihimäki et al. 1989b].

In the Viikari-Juntura et al. [1994] prospective study of machine operators, carpenters, and office workers, current smoking (OR 1.9 1.0–3.5), was among the predictors for change from “no neck trouble” to “severe neck trouble.” In a study of Finnish adults ages 30–64, [Mäkelä et al. 1991], neck pain was found to be significantly associated with current smoking (OR 1.3, 95% CI 1–1.61) when the logistic model was adjusted for age and gender. However, when the model included mental and physical stress at work, obesity, and parity, then smoking (OR 1.25, 95% CI 0.99–1.57) was no longer statistically significant [Mäkelä et al. 1991]. With univariate analysis, Holmström [1992] found a PRR of 1.2 (95% CI 1.1–1.3) for neck-shoulder trouble in “current” smokers versus “never” smokers. But using multiple logistic regression, when age, individual and employment factors were in the model, only “never smoked” contributed significantly to neck-shoulder trouble. Toomingas et al. [1991] found no associations between multiple health outcomes (including tension neck, rotator cuff tendinitis, CTS or problems in the neck/scapula or shoulder/upper arm) and nicotine habits among platers, assemblers and white collar workers. In a case/referent study, Wieslander et al. [1989] found that smoking or using snuff was not related to CTS among men operated on for CTS .

Several explanations for the relationship have been postulated. One hypothesis is that back pain is caused by coughing from smoking. Coughing increases the abdominal pressure and intradiscal pressure and puts strain on the spine. A few studies have observed this relationship [Deyo and Bass 1989; Frymoyer et al. 1980; Troup et al. 1987]. The other mechanisms proposed include nicotine-induced diminished blood flow to vulnerable tissues [Frymoyer et al. 1983], and smoking-induced diminished mineral content of bone causing microfractures [Svensson and Andersson 1983]. Similar associations with diminished blood flow to vulnerable tissues have been found between smoking and Raynaud's disease.

PHYSICAL ACTIVITY The relationship of physical activity and MSDs is more complicated than just “cause and effect.” Physical activity may cause injury. However, the lack of physical activity may increase susceptibility to injury, and after injury, the threshold for further injury is reduced. In construction workers, more

B-4 frequent leisure time was related to healthy lower backs [Holmström et al. 1993] and severe low back pain was related to less leisure time activity [Holmström et al. 1992]. On the other hand, some standard treatment regimes have found that musculoskeletal symptoms are often relieved by physical activity. Having good physical condition may not protect workers from risk of MSDs. NIOSH [1991] stated that persons with high aerobic capacity may be fit for jobs that require high oxygen uptake, but will not necessarily be fit for jobs that require high static and dynamic strengths and vice versa.

When physical fitness is examined as a risk factor for MSDs, results are mixed. For example, some early case series reported an increased risk of MSDs associated with playing professional sports [Bennet 1946; Nirschl 1993], or with physical fitness and exercise [Kelsey 1975b; Dehlin et al. 1978, 1981] while other studies indicate a protective effect and reduced risk [Cady et al. 1979; Mayer et al. 1985; Åstrand et al. 1987; Biering-Sorensen 1984]. Boyce et al. [1991] reported that only 7% of absenteeism could be explained by age, sex, and physical fitness among 514 police officers 35 years or older. Cady et al. [1979, 1985], on the other hand, found that physical capacity was related to musculoskeletal fitness. Cady defined fitness for most physical activities as combinations of strength, endurance, flexibility, musculoskeletal timing and coordination. Cady et al. [1979] evaluated male fire fighters and concluded that physical fitness and conditioning had significant preventive effects on back injuries (least fit 7.1% injured, moderately fit 3.2% injured and most fit 0.8% injured). However, the most fit group had the most severe back injuries. Low cardiovascular fitness level was a risk factor for disabling back pain in a prospective longitudinal study among aerospace manufacturing workers by Battie et al. [1989]. Good endurance of back muscles was found to be associated with low occurrence of low back pain [Biering-Sorensen 1984].

Few occupational epidemiologic studies have looked at non-work-related physical activity in the upper extremities. Most NIOSH studies [Hales and Fine 1989; Kiken et al. 1990; Burt et al. 1990; Baron et al. 1991; Hales et al. 1994; Bernard et al. 1994] have excluded MSDs due to sports injury or other nonwork-related activity or injury and have not included these factors in analyses. However, many of the risk factors that are important in occupational studies occur in sports activities—forceful, repetitive movements with awkward postures. A combination of high exposure to load lifting and high exposure to sports activities that engage the arm was a risk factor for shoulder tendinitis, as well as osteoarthritis of the acromioclavicular joint [Stenlund et al. 1993]. Kennedy et al. [1978] found that 15% of competitive swimmers with repetitive overhead arm movements had significant shoulder disability primarily due to impingement from executing butterfly and freestyle strokes. Epicondylitis in professional athletes has been well documented, and many of the biomechanical and physiological studies of epicondylitis have been conducted in professional tennis players and baseball pitchers [King et al. 1969; Nirschl 1993]. One prospective study of healthy baseball players has found slowing of the suprascapular nerve function as the season progresses [Ringel et al. 1990]. Scott and Gijsbers [1981] found an association between athletic

B-5 performance and pain tolerance, and suggested that physically fit persons may have a higher threshold for injury.

In summary, although physical fitness and activity is generally accepted as a way of reducing work- related MSDs, the present epidemiologic literature does not give such a clear indication. The sports medicine literature, however, does give a better indication that sports involving activities of a forceful, repetitive nature (such as tennis and baseball pitching) are related to MSDs. It is important to note that professional sports activities usually provide players (i.e., workers) with more substantial breaks for recovery and shorter durations for intense tasks as compared with more traditional work settings in which workers are required to perform repetitive, forceful work for 8 hours per day, 5 days per week.

STRENGTH Some epidemiologic support exists for the relationship between back injury and a mismatch of physical strength and job tasks. Chaffin and Park [1973] found a sharp increase in back injury rates in subjects performing jobs requiring strength that was greater or equal to their isometric strength-test values. The risk was three times greater in the weaker subjects. In a second longitudinal study, Chaffin et al. [1977] evaluated the risk of back injuries and strength and found the risk to be three times greater in the weaker subjects. Keyserling et al. [1980] strength-tested subjects, biomechanically analyzed jobs, and assigned subjects to either stressed or non-stressed jobs. Following medical records for a year, they found that job matching based on strength criteria appeared to be beneficial. In another prospective study, Troup et al. [1981] found that reduced strength of back flexor muscles was a consistent predictor of recurrent or persistent back pain, but this association was not found for first time occurrence of back pain.

Other studies have not found the same relationship with physical strength. Two prospective studies of low back pain reports (or claims) of large populations of blue collar workers [Battie et al. 1989; Leino 1987] failed to demonstrate that stronger (defined by isometric lifting strength) workers are at lower risk for low back pain claims or episodes. One study followed workers for ten years after strength testing and the other followed workers for a few years. Neither of these studies included precise measurement of exposure level for each worker, so the authors could not estimate the degree of mismatch between workers' strength and tasks demands. Battie et al. [1990] compared workers with back pain with other workers on the same job (by isometric strength testing) and did not find that workers with back pain were weaker. In two studies of nurses [Videman et al. 1989; Mostardi et al. 1992] lifting strength was not a reliable predictor of back pain.

When examined together, these studies reveal the following: The studies that found a significant relationship between strength/job task and back pain used more thorough job assessment or analysis and have focused on manual lifting jobs. However, these studies only followed workers for a period of one year, and whether this same relationship would hold over a much longer working period remains unclear. Studies that did not find a relationship, although they followed workers for a longer period of time, did not include precise measurements of exposure level for each worker, so they could not assess

B-6 the strength capabilities that were important in the individual jobs. Therefore, they could not estimate the degree of mismatch between workers' strength and task demands.

ANTHROPOMETRY Weight, height, body mass index (BMI) (a ratio of weight to height squared), and obesity have all been identified in studies as potential risk factors for certain MSDs, especially CTS and lumbar disc herniation.

Few studies examining anthropometric risk factors in relationship to CTS have been occupational epidemiologic studies; most have used hospital-based populations who may differ substantially from working populations. Nathan et al. [1989, 1992, 1994] have published several papers on the basis of a single industrial population and have reported an association between CTS and obesity; however, the methods employed in their studies have been questioned in a number of subsequent publications [Gerr and Letz 1992; Stock 1991; Werner et al. 1994b]. Several investigators have reported that their industrial study subjects with CTS were shorter and heavier than the general population [Cannon et al. 1981; Dieck and Kelsey 1985; Falk and Aarnio 1983; Nathan et al. 1992; Werner et al. 1994b; Wieslander et al. 1989]. In the Werner et al. [1994b] study of a clinical population requiring electrodiagnostic evaluation of the right upper extremity, patients classified as obese (BMI>29) were 2.5 times more likely than slender patients (BMI<20) to be diagnosed with CTS. Werner et al. [1994b] developed a multiple linear regression CTS model (with the difference between median and ulnar sensory latencies as the dependent variable) that demonstrated that BMI was the most influential variable, but still only accounted for 5% of the variance in the model. In Nathan's [1994a] logistic model, body mass index accounted for 8.6% of the total risk; however, this analysis used both hands from each study subject as separate observations, although they are not independent of each other. Falck and Aarnio [1983] found no difference in BMI among 17 butchers with (53%) and without (47%) CTS. Vessey et al. [1990] found that the risk for CTS among obese women was double for that of slender women. The relationship of CTS and BMI has been suggested to relate to increased fatty tissue within the carpal canal or to increased hydrostatic pressure throughout the carpal canal in obese persons compared with slender persons [Werner 1994b].

Carpal tunnel canal size and wrist size has been suggested as a risk factor for CTS, however, some studies have linked both small and large canal areas to CTS [Bleecker et al. 1985; Winn and Habes 1990].

For back MSDs, Hrubec and Nashold [1975] found that height and weight were predictive of herniated disc disease among World War II U.S. army recruits compared with age-matched controls. Some studies have reported that people with back pain, are, on the average, taller than those without it [Rowe 1965; Tauber 1970; Merriam et al. 1980; Biering-Sorensen 1983]. Heliövaara et al. [1987], in a Finnish population study, found that height was a significant predictor of herniated lumber disc in both sexes, but a moderately increased BMI was predictive only in men. Severe obesity (exceeding 30 kg/m2) involved less risk than moderate obesity. Kelsey [1975a] and Kelsey et al. [1984] failed to

B-7 reveal any such relationships between height or BMI among patients with herniated lumber discs and control subjects. Magora and Schwartz [1978] found an association between obesity and radiological disc degeneration, but Kellgren and Lawrence [1958] did not. A study of Finnish white collar and blue collar workers found no association between overweight (relative weight (>120%) and lumbosacral disorders either cross-sectionally or in a 10-year follow-up [Aro and Leino 1985].

Schierhout et al. [1995] found that short stature was significantly associated with pain in the neck and shoulder among workers in 11 factories, but not in the back, forearm, hand and wrist. Height was not a factor for neck, shoulder or hand and wrist MSDs among newspaper employees [Bernard et al. 1994]. Kvarnström [1983a] found no relationship between neck/shoulder MSDs and body height in a Swedish engineering company with over 11,000 workers.

Anthropometric data are conflicting, but in general indicate that there is no strong correlation between stature, body weight, body build and low back pain. Obesity seems to play a small but significant role in the occurrence of CTS.

B-8 APPENDIX C Summary Tables

Appendix C contains summary tables of articles reviewed in this document. These tables provide a concise overview of the studies reviewed relative to the evaluation criteria, risk factors addressed, and other issues.

C-1

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Andersen 1993a Andersen 1993b Baron 1991 Bergqvist 1995a Bergqvist 1995b Bernard 1994 Ferguson 1976 Hales 1989 Study type CS CS CS CS CS CS CS CS Participation Y Y N Y Y Y Y Y rate $70% Outcome S S and PE S and PE S and PE S and PE S S S and PE Exposure Job title Categorization by Observation, Questionnaire, Questionnaire, Observation, Measurements, Observation, video categorization job duration video analysis, observation observation questionnaire observation, taping, job measurement of questionnaire categorization, items, (assessment was (assessment was for hand/wrist, not for hand/wrist, not neck) neck) Covariates Age, having Age, having Age, gender, Age, gender Adjustments made Age, gender, Height, weight Age, duration of considered children, not children, not duration of work for confounders height, employment exercising, exercising, environment psychosocial smoking, SES, smoking, SES factors marital status Investigators Y Y Y Y Y Y NR Y blinded Repetition Combined Combined Combined Repeated work Combined Time spent typing: Õ Combined movements: 3.6 NS (0.4-29.6)

Force Combined Combined Combined Õ Õ Õ Õ Combined Extreme Combined Õ Combined Too highly placed Õ Time spent on NR, sig. Õ posture keyboard: 4.4 telephone: 1.4 (1.1-17.0) (1.0-1.8)

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (Continued)

C-2

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Andersen 1993a Andersen 1993b Baron 1991 Bergqvist 1995a Bergqvist 1995b Bernard 1994 Ferguson 1976 Hales 1989 Risk factors Sewing operators Current high Checkers vs. Õ VDT work >20 hr Õ High exposure vs. (combined) vs. referents: 4.9 exposure: noncheckers: 2.0 and eye glasses at Low exposure (2.0-12.8) 1.6 (0.7-3.6) (0.6-6.7) VDT: 6.9 (1.1-42) jobs 8 to 15 years: 6.8 (estimated crude (1.6-28.5) OR): 3.7 (0.4-164) Outcome, neck symptoms: RR=1.64 (0.4-3.9)

Duration of 0 to 7 years: 1.9 0 to 7 years: 2.3 NS Õ Õ NS Õ Adjusted for in employment (1.3-2.9) (0.5-11) analysis 8 to 15 years: 3.8 8 to 15 years: 6.8 (2.3-6.4) (1.6-28.5) >15 years: 5.0 >15 years: 16.7 (2.9-8.7) (4.1-67.5)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload Psychosocial Õ Õ Job satisfaction: Limited break Deadline hr: 1.7 Õ Õ factors NS opportunity: 7.4 work variance: 1.7 (3.1-17.4) management issues: 1.9 Individual/oth Age at least 40 Age $ 40 years: Age, gender, Females with Smoking, stress Age, gender, Õ Age er factors years: 1.5 1.9 (0.9-4.1); hobbies controlled children: 6.4; reaction, height, considered (1.1-2.2); having having children: for in analysis smoking, stress stomach-related psychosocial children: 1.3 0.5 (0.1-1.7); reaction, stress, use of factors; VDT use (0.8-2.0); SES: exercise: 1.4 stomach-related spectacles, peer outside of work 1.29 (0.7-2.3); (0.6-2.9); stress, use of contacts, rest smoking: 1.39 smoking: 1.5 spectacles, peer breaks, work task (0.99-1.9) (0.7-3.3) contacts, rest flexibility, overtime, breaks, work static work task flexibility, position, nonuse of overtime, static lower arm support, work position, hand in nonuse of lower non-neutral arm support, posture, high hand in visual angle to non-neutral VDT, glare on VDT posture, high visual angle to VDT, glare on VDT

Dose/respon Years worked: Õ Õ Õ Õ Õ Õ Õ se Sig.

See footnotes at end of table. (Continued)

C-3

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Hales 1994 Hunting 1994 Kamwendo 1991 Kiken 1990 Knave 1985 Kukkonen 1983 Kuorinka 1979 Linton 1990 Study type CS CS CS CS CS Prospective, CS CS intervention

Participation Y Y Y Y Y NR Y Y rate $70% Outcome S and PE S S S and PE S S and PE S and PE S Exposure Observation, Questionnaire Questionnaire Observation, Observation, gaze Observation, Observation, job Questionnaire questionnaire (assessment was direction interview analysis, video for hand/wrist, instrument, job title taping not neck) or self-report (assessment was for hand/wrist, not neck) Covariates demographics, Years worked, Age, length of Age, gender Age, gender, Gender, Age, duration of Age, gender, considered work practices, age, current work employment, smoking, prospective employment, BMI, exercise, eating age, gender, as electrician, psychosocial educational design metabolic disease, regularly, smoking, hobbies gender work environment status, drinking hobbies, “extra alcohol work” consumption, psychosocial variables Investigators Y NR NR Y NR Y NR NR blinded Repetition Õ Õ Combined Combined Combined Combined Scissor makers Õ vs. Referents: 4.1 (2.3-7.5) Short cycle tasks vs. long cycle tasks: 1.64 (0.7-3.8)

Force Õ Õ Õ Combined Õ Õ Combined Õ Extreme Use of bifocals: Õ Combined Combined Combined Combined Combined Uncomfortable posture 3.8 (1.5-9.4) posture and poor psychosocial environment: 3.5 (2.7-4.5) Vibration Õ Õ Õ Õ Õ Õ Õ Univariate analysis showed elevated OR for vibration

See footnotes at end of table. (Continued)

C-4

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Hales 1994 Hunting 1994 Kamwendo 1991 Kiken 1990 Knave 1985 Kukkonen 1983 Kuorinka 1979 Linton 1990 Risk factors Õ Õ Work with office High exposure vs. Typing hr: Sig. Intervention Scissor-makers Õ (combined) machines >5 low exposure group: PRR=3.6 vs. department hr/day: 1.65 jobs: 1.3 (0.2-11) (2.2-5.9) No store shop (1.02-2.67) intervention 1.0 assistants: OR=4.1 (2.3-7.5) Duration of NS 1 to 3 years: 1 Length of Õ Õ Õ Controlled for Õ employment 4 to 5 years: 1.3 employment: Sig. 6 to 10 years: 1.6 >10 years: 1.3

Physical Õ Õ Being given too Õ Õ Õ Õ Õ workload much to do: Sig. Psychosocial Decision making: Ability to influence Õ Interest in work, Õ Monotonous work factors 4.2; productivity work, cooperative positive attitude SS, work content, standard: 3.5; spirit between work load, social fear of co-workers: sig. support replacement by computer: 3.0; higher information processing demands: 3.0; job task variety: 2.9; work pressure: 2.4

Individual/oth Electronic Age group, Sitting 5 or more Õ Õ Õ Extra work, Exercise, eating, er factors performance current work as hr/day: 1.6 hobbies, outside smoking, alcohol considered monitoring, electrician: NS (0.9-2.8); age: activities: NS consumption keystrokes, Sig. hobbies, recreational activities: NS Dose/respon Õ Õ Õ Õ Between Õ Õ Õ se registered work duration and musculoskeletal complaints

See footnotes at end of table. (Continued)

C-5

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Liss 1995 Luopajärvi 1979 Milerad 1990 Ohlsson 1989 Ohlsson 1995 Onishi 1976 Ryan 1988 Sakakibara 1987 Study type CS CS CS CS CS CS CS CS Participation N Y Y NR Y NR Y Y rate $70% Outcome S S and PE S S S and PE S and PE S and PE S Exposure Questionnaire Observation, Questionnaire Questionnaire Videotaping, Observation, then Observation Observation job video analysis, observation, job categorization measurements at analysis and neck interviews analysis of work stations angle posture, flexion of measurements neck, questionnaire

Covariates N Age, gender, Gender, age, Age, gender, Age , gender, Õ Age, height, length Õ considered social leisure-time duration of psychosocial of training time background, exposure, employment scales hobbies, amount systemic disease of housework

Investigators N Y NR NR Blinded to NR Y NR blinded exposure information but “Not possible to completely blind the examiners.”

Repetition Combined Combined Combined Combined Combined Combined Õ Combined Force Combined Combined Õ Õ Industrial workers Combined Õ Õ exposed to repetitive tasks vs. referents: 3.6 (1.5-8.80)

Extreme Combined Combined Combined Combined Õ Combined Significant Combined posture difference in mean elbow angle and shoulder flexion of left arm

Vibration Õ Õ NS for exposure Õ Õ Õ Õ Õ to vibration

See footnotes at end of table. (Continued)

C-6

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Liss 1995 Luopajärvi 1979 Milerad 1990 Ohlsson 1989 Ohlsson 1995 Onishi 1976 Ryan 1988 Sakakibara 1987 Risk Factors Dental hygienists Assembly Dentists compared Assemblers vs. Õ Film rolling Õ Pear work vs. (Combined) vs. dental workers vs. shop to pharmacists: referents pain in workers: 3.8 apple work right assistants: 1.7 assistants: 1.6 2.1 (1.4-3.1) last 12 months: Lamp assemblers: side: p<0.05 (1.1-2.6) (0.9-2.7) 1.9 (0.9-3.7) 3.8 (2.1-6.6) Pear work vs. Teachers and Apple work at left nurses: 1.5 side: p<0.01 (0.7-3.2)

Duration of NS Õ NS Employees Õ Õ NS Õ employment <35 years: Sig. Physical Õ Õ Õ Õ Õ Õ Õ Õ workload Psychosocial Õ Õ Õ Increased OR for Õ Insufficient rest, Õ factors medium and fast break time, more paced work boredom, more compared to slow stress, lower peer paced but OR cohesion, lower lower for very antonomy, lower fast paced work job clarity, higher staff support, higher work pressure

Individual/oth Gender (99% Õ Leisure time Õ Õ Age, height, Õ er factors females in study exposure, marital and considered group); had to smoking systemic parental status, modify work or disease handedness, unable to work at length of training some point: 2.4 time (1.1-5.4)

Dose/respon Õ Õ Õ Õ Õ Õ Õ Õ se

See footnotes at end of table. (Continued)

C-7

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Sakakibara 1995 Schibye 1995 Veiersted 1994 Viikari-Juntuna Welch 1995 Wells 1983 Yu 1996 1994 Study type CS Cohort Cohort Cohort CS CS CS Participation Y Y N (55%) Y Y (83%) Y Y rate $70% Outcome S and PE S S and PE/ pain diaries S S S S Exposure Observation, Questionnaire EMG, interviews Questionnaire, Questionnaire Questionnaire, Questionnaire measurements every 10 weeks observation interview Covariates Õ Subjects served as Metabolic or other All male, smoking, Smoking, years of Age, gender, number Age, gender, “other considered their own controls diseases, gender age, physical employment of years on job, covariates” exercise, occupation, previous work duration of work, car experience, driving education, marital status, quetelet ratio

Investigators NR NR NR Y N NR NR blinded Repetition Õ Combined Õ Õ Combined Õ Frequent VDT use: 28.9 (2.8-291.8)

Force Õ Combined Strenuous previous Combined Õ Combined Õ work: 6.7 (1.6-28.5)

Extreme Combined Combined Strenuous postures: No neck pain to Percent of time Combined Inclining neck at posture 7.2 (2.1-25.3) severe, machine hanging duct: 7.5 work: 784.4 operators vs. office (0.8-68) (33.2-18,630) workers: 3.9 (2.3-6.9) Persistently severe: 4.2 (2.0-9.0) Vibration Õ Õ Vibration (floor or Combined (machine Õ Õ Õ machine) operators)

See footnotes at end of table. (Continued)

C-8

Appendix C Table C-1. Summary table for epidemiologic studies evaluating work-related neck musculoskeletal disorders Components of study Sakakibara 1995 Schibye 1995 Veiersted 1994 Viikari-Juntuna Welch 1995 Wells 1983 Yu 1996 1994 Risk factors Pear vs. Apple Other employment Physical Occupation Sig. from Õ All letter carriers vs. Frequent video (combined) bagging: 1.5 group vs. garment environment: 0.9 no neck trouble to Clerks and readers: display terminal use: (0.99-2.35) workers: 3.3 (0.5-1.7) moderate neck 2.57 (1.13-6.2) 28.9 (2.8-291.8) (1.4-7.7) trouble; occupation Sig. from no neck to severe neck trouble Carpenters vs. Office workers persistently severe: 3.0 (1.4-6.4) Duration of Õ NS Õ Õ Õ Controlled for in Õ employment analysis Physical Õ Õ Õ Õ Õ workload Psychosocial Õ Õ Psychosocial Job satisfaction: NS Õ Õ Õ factors` factors: 3.3 (0.8-14.2)

Individual/oth Õ Age Anthropometrics, Current smoking and Õ Education, marital General health er factors general health, age Sig. in model of status, quetelet ratio considered previous employment “no neck trouble to variables, draft, severe neck trouble” noise, personality

Dose/respon Õ Õ Õ Õ Õ Õ Õ se Õ Not studied. BMI Body mass index. CS Cross-sectional. EMGElectromyography. hrs Hours. MSDMusculoskeletal disorders MVCMaximum voluntary contraction. N No. NR Not reported. NS Not statistically significant. OR Odds ratio. PE Physical examination. PRR Prevalence rate ratio. S Symptoms. SES Socioeconomic status. Sig. Statistically significant. VDTVideo display terminal. vs. Versus. Y Considered (yes).

C-9

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components of study Åaras 1994 Andersen 1993a Andersen 1993b Bergqvist 1995a Bergqvist 1995b Bjelle 1981 Blåder 1991 Ekberg 1994 Study type Prospective CS CS CS CS Case Control CS Case Control Participation NR Y Y Y Y NR Y Y rate $70% Outcome S and Records S S and PE S S and PE S and PE S and PE S Exposure Observation and Job title Categorization by Observation, Job title and Observation, Questionnaire Questionnaire EMG categorization job duration measurements questionnaire videotape analysis

Covariates Õ Age, having Age, having Age, gender, Age, gender, Age, Age, nationality, Age, gender, considered children, children, smoking, rest smoking anthropometric employment time, smoking, having education, marital education, marital breaks, stress data working hr/week preschool children status, smoking, status, smoking, not exercising not exercising Investigators NR Y Y Y Y Y; Videotape N NR blinded analysis blinded to case status

Repetition Õ Combined Combined For intensive <20 hr/week VDT No sig difference Combined Precise repetitive neck/shoulder use: 1.2 (0.4-3.7) in cycle time movements discomfort: 3.6 >20 hr/week VDT High: 15.6 (0.4-29.6) use: 0.7 (0.3-1.5) (2.2-113.0)

Force Static trapezius Combined Combined Õ Õ Cases had Õ Õ load dropped from significantly 4.1 to 1.4% higher shoulder NR, Sig. loads than controls

Extreme Intervention Õ Õ For tension neck Õ Cases with longer Combined Work with lifted posture consisted of syndrome: too duration and arms 4.8 (1.3-18); equipment and highly placed higher frequency uncomfortable tool adjustment to VDT: 4.4 of abduction or sitting posture: 3.6 create relaxed (1.1-17.6) forward flexion (1.4-9.3) position of than referents: shoulders and NR, Sig. neck: NR, Sig.

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (Continued)

C-10

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components of study Åaras 1994 Andersen 1993a Andersen 1993b Bergqvist 1995a Bergqvist 1995b Bjelle 1981 Blåder 1991 Ekberg 1994

Risk factors Õ Sewing machine Current high Õ VDT work >20 hr Õ Working >30 hr Õ (combined) operators vs. exposure (yes vs. and stressful per referents: no): 1.6 (0.7-3.6) stomach week: p<0.05 4.6 (2.2-10.2) reactions: 3.9 (1.1-13.8) VDT work $ 20 hr and bifocals or progressive glasses: 6.9 (1.1-42.1)

Duration of Õ Years as sewing Years as sewing Õ Õ Õ Working >30 Õ employment machine machine hr/week and operators 0 to 7 operators tension neck years: 3.2 0 to 7 years: 2.3 syndrome: p<0.05 (0.6-16.1) (0.5-11) 8 to 15 years: 8 to 15 years: 6.8 11.2 (2.4-52) (1.6-28.5) >15 years: 36.7 >15 years: 16.7 (7.1-189) (4.1-67.5)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ For cervical Combined Õ Smaller High work pace: factors diagnoses: randomized study 3.5 (1.3-9.4); Stressful stomach group interviewed Low work reactions: 5.4 by sociologist and content: 2.6 (1.6-17.6) psychologist for (0.7-9.4); psychosocial Work role history ambiguity: 16.5 (6.0-46); Demands on attention: 3.8 (1.4-11) Individual/ Median sick days Age >40 yrs: 1.96 Age $ 40 years: Children at home, Children at home, Age-isometric Cervical Female: 11.4 other factors decreased from (0.8-5); 1.9 (0.9-4.1); negative, negative, testing syndrome (4.7-28); considered 22.9 to 1.8 exercise: children: 0.5 affectivity, peer affectivity, peer correlated with immigrant status: 1.28 (0.5-3.4); (0.1-1.7); contacts, contacts, age 4.9 (1.8-14); smoking: exercise: 1.4 overtime, work overtime, work current smoker: 2.3 (0.9-6.1); (0.6-2.96); task flexibility, task flexibility, 8.2 (2.3-29) children: 0.35 smoking: 1.5 visual angle to visual angle to (0.1-1.9) (0.7-3.3) VDT VDT

Dose/respon Õ Duration of Duration of Õ Õ Õ Õ Repetitive se employment as employment precision sewing machine movements, work operator pace

See footnotes at end of table. (Continued)

C-11

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components Kilbom 1986, of study Ekberg 1995 Holmström 1992 Hünting 1981 Jonsson 1988 1987 Linton 1989 Maeda 1982 Milerad 1990 Study type CS CS CS Cohort CS CS CS CS Participation Y Y NR Y Y Y NR Y rate $70% Outcome S S S and PE S and PE S and PE S S S Exposure Questionnaire Questionnaire Observation, Observation, Observation, video Questionnaire Observation, Questionnaire questionnaire video taping, job taping, job dealing with measurement analysis, MVC of analysis, MVC of psychosocial forearm forearm issues

Covariates Age, smoking, Age, physical Psychosocial Used prospective Age, spare time Õ Gender, leisure considered exercise habits, factors, factors cohort design physical activities, time, smoking, family situations psychosocial with same study hobbies, systemic disease with preschool stress scales sample psychosocial children, immigrant stress, muscle status, gender strength Investigators NR Y NR Y Y NR NR NR blinded

Repetition Repetitive Õ Combined Combined Combined Õ Õ Combined movements demanding precision: 1.2 (1.0-1.3)

Force Õ Õ Õ Combined Combined Õ Õ Õ

Extreme Hand above Combined/head Combined Combined Õ Constrained tilted Combined posture shoulder: <1 inclination >56E head posture: hr/day: 1.1 Sig. for p<0.05 (0.8-1.5) neck/shoulder 1 to 4 hr/day: 1.5 MSDs (1.2-1.9) >4 hr/day: 2.0 (1.4-2.7)

Vibration Õ Õ Õ Õ Õ Õ Õ NS

See footnotes at end of table. (Continued)

C-12

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components Kilbom 1986, of study Ekberg 1995 Holmström 1992 Hünting 1981 Jonsson 1988 1987 Linton 1989 Maeda 1982 Milerad 1990

Risk factors Õ Roofers: 1.6 Data entry At third year, 38 Average time/work Õ Õ Dentists vs. (combined) Plumbers: 1.5 workers vs. workers cycle in neck pharmacists: Floor workers: 1.3 non-keyboard- reallocated had flexion sig, Upper 2.1 (1.3-3.0); using office improved, 26% arm abducted males: 2.6 workers: 9.9 with unchanged 0-30E: NR, Sig. (1.2-5.0); females (3.7-26.9) conditions 2.0 (1.3-3.1) deteriorated further: NR, Sig.

Duration of Õ Õ Õ Õ NS Õ Õ NS employment

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Qualitative Job satisfaction; Job satisfaction, Productivity, work Poor work content: Õ Õ factors demands: 1.4 relationship with productivity satisfaction, 2.5 (1.3-4.9) (1,2) supervisors, perceived stress: Lack of social Quantitative colleagues; NS support: 1.6 demands: 3.0 decision making, (0.9-2.8) (2.1-4) use of skills all NS Work demand Solitary work: 1.5 social support at (1.2-1.8) work Anxiety: 3.2 (2.5-4)

Individual/ Immigrant status: Psychosomatic: Medical findings in Õ Age, muscle Õ Age Leisure time, other factors 1.3 (1.1-1.5) 5.0 (3.6-6.9) neck and shoulder strength, rest smoking NS considered Social work Psychological: 4.7 significant for pauses: NS climate, work (3.6-6) typists with head planning, job Stress: 3.4 rotation security, job (2.6-4.2) >20E compared to constraints Discretion, < 20E support, under stimulation, anxiety, job satisfaction, quality of life

Dose/respon Õ Stress index and Õ Õ Õ Õ Õ Õ se neck-shoulder MSDs

See footnotes at end of table. (Continued)

C-13

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components Viikari-Juntura of study Ohara 1976 Ohlsson 1995 Punnett 1991 Rossignol 1987 Ryan 1988 Tola 1988 Vihma 1982 1991a Study type CS and Cohort CS CS CS CS CS CS Cohort Participation CS study: NR; Y Y N to Y (6 Y Y overall: NR Y rate $70% industries) 67% carpenters Cohort: Y 67% office workers

Outcome S and PE S and PE S S S S S S and PE Exposure Observation Observation, Observation, Questionnaire Observation, Occupation title Observation, Questionnaire video, analysis, questionnaire workstation interview muscle strength measurement, testing questionnaire Covariates Used prospective Age, gender, Age, gender Age, cigarette Height, weight, Years in Age, duration of Physical hobbies, considered cohort design psychosocial smoking, industry, gender, age, occupation, age, employment creative hobbies with same study scales education, VDT marital status, leisure time sample training parental status activities, car driving, general health

Investigators NR Y to exposure NR NR Y NR NR NR blinded information, no for physical

Repetition Combined Repetitive work: Combined Combined Õ Õ Combined Õ 4.6 (1.9-12)

Force Õ Õ Combined Õ Õ Õ Õ Õ Extreme Combined Significant time Associated with Combined More non-cases Use of twisted or Combined Sitting in a posture spent in neck extended duration trained in bent postures forward posture flexion <60°: NR of and lifting adjustment of during work: Little Sewing machine 1-3 hr/day: 10.7 weight in furniture than (referent): 1.0 operator with (0.4-291); abduction/flexion cases: NR, Sig. Moderate: 1.2 significantly >3 hr/day: 1.5 and extension of (1.0-1.5) greater static (0.7-29.5) the shoulder Rather much: 1.6 work compared to (1.4-1.9) seamstresses Very much: 1.8 (1.5-2.2)

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (continued)

C-14

Appendix C Table C-2. Summary table for evaluating work-related neck/shoulder disorders Components Viikari-Juntura of study Ohara 1976 Ohlsson 1995 Punnett 1991 Rossignol 1987 Ryan 1988 Tola 1988 Vihma 1982 1991a

Risk factors Operators hired Industrial workers Male: 1.8 (1.0-3.2) ½ to 3 hr of VDT Õ Machine operators Sewing machine Õ (combined) post-intervention vs. referents: 2.7 Female: 0.9 use: 1.8 (0.5-6.8) vs. office operators vs. had less reports (1.2-6.3) (0.5-1.9) 4 to 6 hr of VDT workers: 1.7 seamstresses: of MSDs use: 4.0 (1.5-2.0) 1.6 (1.1-2.3) (1.1-14.8) 7 $ hr Carpenters vs. of VDT use: 4.6 office workers: (1.7-13.2) 1.4 (1.1-1.6)

Duration of Õ Õ Õ Õ Õ Õ Õ employment

Physical Õ Õ Õ Õ Õ Õ Cases had Õ workload significantly higher shoulder loads

Psychosocial Õ Stress/worry Õ Õ Adequate rest Job satisfaction, Õ Social confidence, factors tendency: 1.9 breaks, boredom, poor vs. very much fear vs. (1.1-3.5) work stress job good: 1.2 (1.1-1.4) none: 1.4 pressure, (0.05-42.2); autonomy, peer Sense of cohesion, role coherence: 0.95 ambiguity, staff (0.9-0.99) support

Individual/oth Õ Muscle tension Õ Smoking, industry, Õ Working in a draft: Õ Alexithymia er factors tendency: 2.3 education 1.1 (1.0-1.3) 1.02 (0.97-1.1) considered (1.3-4.9)

Dose/respon Õ Õ Õ Hours of VDT use Õ Use of twisted or Õ Õ se bent posture

Õ Not studied S Symptoms CI Confidence interval Sig. Statistically significant CS Cross-sectional VDT Video display terminal EMG Electromyography vs. Versus hr Hours Y Considered (yes) Med. Medium MSDSMusculoskeletal disorders MVC Maximum voluntary contraction N No NR Not reported NS Not statistically significant OR Odds ratio PE Physical examination

C-15

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Andersen 1993a Andersen 1993b Baron 1991 Bergenudd 1988 Bernard 1994 Bjelle 1979 Bjelle 1981 Burdorf 1991 Study type CS CS CS CS CS Case control Case control CS Participation Y Y N N Y NR NR Y for riveters; rate $70% N for referents Outcome S S and PE S and PE S and PE S S and PE PE S

Exposure Job title, Job title, Observation and Questionnaire, job Questionnaire and Observation, Measurement, Observation, categorization by categorization by videotape classification observation measurement, videotape measurement of job duration job duration analysis, weight (light, moderate, analysis, vibration of scanned items, heavy physical EMG on 15 cases, observation, EMG job category demands) open muscle on 3 subjects and biopsies on 11 2 healthy cases volunteers

Covariates Age, having None for the Age, gender, Gender Age, race, Age, gender, and Age, gender, and Height, weight, considered children, not shoulder analysis hobbies, duration gender, height, workshop place of work smoking status exercising, of work, second medical duration of job, metabolic conditions, employment, disease, duration psychosocial socioeconomic of employment factors, typing hr status, smoking away from work status, current neck/shoulder exposure

Investigators Y Y Y NR N N Y NR blinded

Repetition for Combined Combined Combined Õ R no surrogate for Combined Combined Õ shoulder hand used: number of hr typing

Force Combined Combined Combined Õ Õ Combined Cases had Sig. Õ higher shoulder loads than controls

Extreme Combined Combined Combined Õ Õ Combined Combined Õ posture

See footnotes at end of table. (Continued)

C-16

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Andersen 1993a Andersen 1993b Baron 1991 Bergenudd 1988 Bernard 1994 Bjelle 1979 Bjelle 1981 Burdorf 1991

Vibration Õ Õ Õ Õ Õ Õ Õ 1.5 (no confidence limits)

Risk factors Increasing years Chi sq test for Checkers vs. Õ Õ Work at or above Cases had Sig. Õ (combined) of experience: trend using others 3.9 shoulders, cases longer duration 1.38-10.25 (Sig.) exposure time in (1.4-11.0) (65%) vs. and higher years for rotator Checkers using referents (15%): frequency of cuff syndrome: scanners vs. 10.6 (2.3-54.9) abduction or 9.51; p<0.01 others 8.6 forward flexion (1.0-72.2) than controls, p<0.001

Duration of See under See under “Risk Number of hr per Õ Years at Õ Õ Years of riveting: employment “Physical factors combined” week as a newspaper: 1.4 0.05# p<0.10 workload” checker Sig. (1.2-1.8)

Physical 0 to 7 years: 1.56 Õ Õ Prevalence of Õ Õ Õ Õ workload (0.76-3.75) occupational 8 to 15 years: workload in 4.28 (2.14-10.0) subjects with >15 years: 7.27 shoulder pain: (3.82-16.3) Heavy, 11%; Moderate, 49%; Light, 40%

Psychosocial Õ Õ Õ Females showed Lack of decision Õ Õ Õ factors Sig. association making with shoulder pain participation: 1.6 and (1.2-2.1) dissatisfaction job pressure: 1.5 (1.0-2.2)

Individual/oth Age-matched Age-matched Age, gender, Gender Gender, race, Age, gender Age, gender; Age er factors controls controls metabolic disease height median number of considered sick-leave days Sig. different between cases and controls, p=0.01

Dose/respon Y with years of Y with years of Õ Õ Õ Õ Õ se employment exposure

See footnotes at end of table. (Continued)

C-17

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Burt 1990 Chiang 1993 English 1995 Flodmark 1992 Hales 1989 Hales 1994 Herberts 1981 Herberts 1984 Study type CS CS Case control CS CS CS CS CS Participation Y Y Y Y Y Y NR NR rate $70% Outcome S S and PE S and PE S S and PE S and PE S and PE S and PE

Exposure Observation, Observation and Self-reports Õ Observation Observation and Analyses by job Analyses by job questionnaire, job recording of walk-through, job questionnaire title title sampling representative categorization jobs, hand F estimation High vs. low exposure (hand/wrist exposure) Covariates Age, gender, Age, gender, Age, height, Age, headache, Age and duration Age, race, Age, job duration Controls matched considered psychosocial metabolic gender, weight, tiredness, medical of employment gender, work for age and factors, metabolic diseases injury, study problems, practices, work gender disease duration center, hobbies, sleeping problems organization of employment sporting activities, or lack of factors, individual average hr of concentration, factors, electronic driving, sleep performance compensation monitoring, claim made recreational activities, hobbies

Investigators Õ Y Y Õ Y Y NR NR blinded

Repetition for Typing speed fast Repetitive Combined Õ Combined No Combined Combined shoulder compared to movement of slow: 4.1 upper limb: 1.6 (1.8-9.4) (1.1-2.5)

Force Õ Sustained forceful Õ Õ Combined Õ Welders vs. office Welders vs. office movement of workers: 15-18 workers: 15-18 upper limb: 1.8 (1.2-2.5)

Extreme Õ Õ Combined Õ Combined Number of times Combined Combined posture arising from chair: 1.9 (1.2-15.5)

See footnotes at end of table. (Continued)

C-18

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Burt 1990 Chiang 1993 English 1995 Flodmark 1992 Hales 1989 Hales 1994 Herberts 1981 Herberts 1984

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

Risk factors Õ Repetition Repeated Õ Any symptom of Õ Welders vs. office ST results of 23 (combined) multiplied by shoulder rotation shoulder: 49% vs. workers: shoulder welders called force: 1.4 with elevated arm: 43%; symptoms: 15.2 back for clinical (1.0-2.0) 2.3, p<0.05 1.2 (0.7-2.0) (2.1-108) follow-up exams: 16 had ST; 18.3 Period Shoulder (13.7-22.1) prevalence: 19% Tendinitis: 8.3 (90% CI) vs. 4%; 3.8 (NS) (0.6-22.8) ST results of 30 plate-workers Point prevalence: called back for 7% vs. 4%;0.9 clinical follow-up (0.1-7.3) exams: 15 plate-workers had ST: 16.2 (10.9-21.5) (90% CI)

Duration of NS Õ Õ Õ Õ Õ Õ Õ employment

Physical Õ Õ Õ Õ Õ Õ NS Õ workload

Psychosocial Job Õ Õ Type A Behavior: Õ Fear of Õ Õ factors dissatisfaction: p<0.001 replacement by 2.3 (1.2-4.3) computers: 1.5 (1.1-2.0)

Individual/oth Pre-existing Plant effect age: Per 5 years of Õ Õ Typing outside of Õ Õ er factors arthritis: 2.3 1.0 (0.9-1.1) age: 1.4 (1.2-1.5) work considered (1.2-4.4) Gender: 1.1 (0.7-1.7)

Dose/respon Õ Dose response Õ Õ Õ Õ Õ Õ se found for shoulder diagnosis as exposure status increased from Group 1 to Group 3

See footnotes at end of table. (Continued)

C-19

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Component Kilbom 1986, of study Hoekstra 1994 Hughes 1997 Ignatius 1993 Jonsson 1988 Kiken 1990 1987 Kvarnström McCormack 1983 1990 Study type CS CS CS Prospective CS CS CS and Case CS control Participation Y N N Y Y Y NR Y rate $70% Outcome S S and PE S S and PE S and PE S and PE S and PE S and PE Exposure Analyses based Observation and Observation, Observation, Observation Observation, Observation, Observation on questionnaire, job analysis questionnaire, measurement of (exposure based measurement, interview, self-reports weight of mail exertion, on repetitive and videotaping, questionnaire bags videotaping forceful hand observation motions, not shoulder)

Covariates Age, seniority, Controlled for age, Age, duration of Age, hobbies, Age and gender Age, years of Õ Age, gender, considered gender smoking status, employment, bag spare time, employment, race, job sports, hobbies weight, walking physical action, productivity, category, duration time psychosocial muscle strength of employment, factors, breaks, general health rest pauses history Investigators Y NR NR Y Y Y N N blinded

Repetition for Õ Õ Combined Combined Combined Fewer total Combined Combined shoulder number of upper arm flexions/hr. (p<0.05)

Force Õ Õ Combined Õ Combined Õ Combined Õ Extreme Non-optimally Years of forearm Combined Relative time Combined Greater Combined Combined posture adjusted desk twist: 46.0 spent with percentage of height work: 5.1 (3.8-550) shoulder elevated work cycle time (1.7-15.5) negatively related with upper arm to ‘remaining abducted 0-30° healthy ‘ after (p<0.05) both 1 and 2 years: Sig.

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (Continued)

C-20

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Component Kilbom 1986, of study Hoekstra 1994 Hughes 1997 Ignatius 1993 Jonsson 1988 Kiken 1990 1987 Kvarnström McCormack 1983 1990

Risk factors Center B Õ Letter delivery 38 subjects who Plant #1 Õ Die casting Boarding workers (combined) compared to postal workers were reallocated Any symptom for machine vs. knitting Center A: 4.0 compared to other to more varied shoulder: 46% vs. operators: 5.4; workers: 2.1 (1.2-13.1) postal workers tasks improved 28%; 1.6 (0.9-2.9) plastic workers: (0.6-7.3) Recurrent: 1.8 2.2; spray (1.5-2.2) Period prevalence: painters: 3.7; 13% vs. 3%; 4.0 surface treatment Severe joint pain: (0.6-29) operators: 4.7; 2.2 (1.5-3.1) Plant #2 assembly line Any symptom for workers: 5.2 shoulder: 50% vs. 30%; 1.7 (0.8-3.3) Period prevalence: 14%vs. 5%; 2.8 (0.4-19.6)

Duration of Õ Õ Õ Õ Õ Years of Õ NS employment employment in electronics: p<0.05

Physical Õ Õ Õ Low muscle Õ Õ Õ Õ workload strength no a predictor for shoulder MSD

Psychosocial Job Low decision Õ Strong negative Õ Õ 9 cases and 1 Õ factors dissatisfaction, latitude: 4.0 relationship control reported exhaustion (not (0.8-19) between poor relationship for shoulder) remaining health with supervisor. and satisfaction Sig. differences in with colleagues group piece rate, shift work, heavy work, monotonous work, stressful work,

Individual/oth Location Age: 0.93 Age, work Predictors of Õ Shorter stature: Sig. differences in Õ er factors (0.8-1.0); good experience, bag deterioration, p<0.05, heavy lifting and considered health: 0.35 weight, walking previously productivity: NS, unsuitable (0.1-0.87) time physically heavy muscle strength: working job, high NS conditions productivity, and sick leave

Dose/respon Õ Õ Õ Õ Õ Õ Õ Õ se

See footnotes at end of table. (Continued)

C-21

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Milerad 1990 Ohara 1976 Ohlsson 1989 Ohlsson 1994 Ohlsson 1995 Onishi 1976 Punnett 1985 Rossignol 1987 Study type CS CS and CS CS CS CS CS CS Prospective Participation Y NR (CS), NR Y Y NR Y Y: clerical rate $70% Y (Prospective) workers N: industry groups Outcome S S and PE S S and PE S and PE S, PE, and S and PE S measurement Exposure Questionnaire Observation Job categorization Observation, Observation, Observation Observation and Observation and questionnaire, video analysis, questionnaire questionnaire video analysis measurement

Covariates Age, gender, Õ Age, gender Sports activities, Age, employment Body height, Age, number of Age, cigarette considered leisure time (females only) age, gender status weight, grip years employed, smoking, industry, exposure, (females only) strength native language VDT educational smoking, systemic psychosocial training disease, duration factors of employment

Investigators NR NR NR Y Yes, to exposure NR NR Õ blinded information Repetition for Combined Combined Combined Combined Combined Combined Combined 4-6 hrs. VDT use: shoulder 4.0 (1.0-16.9) >7 hrs. VDT use: 4.8 (1.6-17.2)

Force Combined Combined Combined Combined Combined Combined Combined Õ

Extreme Combined Combined Combined Combined Combined Combined Combined Õ posture

Vibration NS Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (Continued)

C-22

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Milerad 1990 Ohara 1976 Ohlsson 1989 Ohlsson 1994 Ohlsson 1995 Onishi 1976 Punnett 1985 Rossignol 1987

Risk factors Dentists vs. Shoulder Assemblers vs. Supraspinatus, Assembly work Shoulder Garment workers Õ (combined) pharmacists: stiffness: referents shoulder infraspinatus, or compared to tenderness: vs. hospital males: 2.4 cashiers (81% vs. pain last 7 days: bicipital tendinitis referent 5.0 assemblers vs. employees 2.2 (1.0-5.4), office workers 3.4 (1.6-7.1) working in the fish (2.2-11.0) ref.: 1.1 (0.6-1.9); (1.0-4.9) females: 2.4 (72%), 1.7 industry: OR=3.03 film rollers vs. (1.5-3.7) (1.0-2.8) (2.5-7.2) ref.: 6.0 Shoulder dullness (3.0-12.2); and pain: Shoulder tendinitis teachers vs. ref.: cashiers (49%) alone: PRR=3.5 1.6 (0.7-3.3) vs. other workers (2.0-5.9) (68%), 2.0 Shoulder (1.4-2.8); vs. stiffness: reservationists office workers vs. ref: 2.5 (30%), 2.2 (1.1-5.6); (1.4-3.5) assemblers vs. ref.: 3.7 (2.0-7.0); film rollers vs. ref.: 2.7 (1.5-4.9); teachers vs. ref.: 2.1 (0.9-4.6)

Duration of NS Õ Sig. with duration For age <45 <10 years: 9.6 Õ NS Õ employment of employment years, duration of (2.8-33.0) (p=0.03) for employment 10-19 years: 4.4 younger workers showed dose- (1.5-13.0) but not older response with >20 years: 3.8 workers shoulder MSDs (1.4-10.0)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Increasing work Stress, worry Control, Õ Õ Õ factors pace factors, stimulation, tendencies psychosocial towards muscle climate, work tension Sig. strain, social support, psychosomatic symptoms

Individual/oth Õ Sports activities: Employment Body height and Õ er factors 4-9 status weight: NS considered

Dose/respon Õ Õ Reported pain For age <45 Õ Õ Õ As VDT use se increased with years, duration of increased, increasing work employment and shoulder pace except for shoulder MSDs symptoms very high paces increased

See footnotes at end of table. (Continued)

C-23

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Sakakibara 1987 Sakakibara 1995 Schibye 1995 Stenlund 1992 Stenlund 1993 Sweeney 1994 Wells 1983 Study type CS CS Cohort CS CS CS CS Participation Y Y Y (But there was a Y Y N Y rate $70% significant dropout of work as a sewing machine operator in those >35 years Outcome S S and PE S S and PE S and PE S and PE S Exposure Observation and Observation and Questionnaire Questionnaire, Questionnaire and Questionnaire Questionnaire, job measurement of measurement of self-reports, weight self-reports categorization postures representative of tools workers or job titles job title, duration of employment

Covariates Gender, age Õ Cohort study: Age, smoking, Age, handedness, Õ Age, number of considered followed same dexterity, ethnicity smoking, sports years on job, workers over time activities, duration of quetelet ratio, employment previous work experience, education

Investigators Õ NR NR Y Y Yes NR blinded

Repetition for Õ Combined Combined Õ Õ Combined Õ shoulder

Force Õ Õ Combined Combined Manual work: Õ Combined right side: 1.1 (0.7-1.8) left side: 1.9 (1.0-3.4)

Extreme Thinning out, bagging Combined Combined Õ Õ Combined Combined posture pears had significantly more forward shoulder flexion than bagging apples

Vibration Õ Õ Õ Right side: 2.2 Right side 1.7 Õ Õ (1.0-4.6) (1.1-2.6) Left side: 3.1 left side 1.8 (1.1-3.1) (1.4-6.9)

See footnotes at end of table. (Continued)

C-24

Appendix C Table C-3. Summary table for evaluating work-related shoulder musculoskeletal disorders Components of study Sakakibara 1987 Sakakibara 1995 Schibye 1995 Stenlund 1992 Stenlund 1993 Sweeney 1994 Wells 1983

Risk factors Õ Pear baggers Development of Rockblasters vs. Rock blasters >20 hrs./ week Letter carriers with (combined) compared to apple shoulder symptoms Foremen: 4.0 compared to signing: 2.5 (0.8-8.2) increased shoulder baggers: 1.7 not related to work (1.8-9.2) foremen: load vs. postal (1.1-2.9) exposure but Bricklayers right side: 1.7 clerks: 5.7 (2.1-17.8) Posture: NR, Sig. significant dropout of (0.7-4.0) workers >35 years compared to left side: 3.3 foremen: right side: 2.2 (1.2-9.3) (1.0-4.7)

Physical Õ Õ Õ Right side: 1.0 Õ Õ workload (0.6-1.8) left side: 1.8 (0.9-3.4)

Psychosocial Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Õ Õ Rock blasters Õ Õ er factors compared to considered foremen: Right side: 2.1 (0.9-4.6) Left side: 4.0 (1.8-9.2)

Duration of Õ Õ Õ Right side: 2.9 Õ NS employment (1.2-7.4) Left side: 2.5 (1.0-5.9)

Dose/respon Õ Õ None for increasing As length of High vibration Õ Õ se piece work in employment and compared to low previous years exposure to vibration vibration and amount lifted increased, osteoarthritis of shoulder increased

Õ Not studied. Ref. Referents. EMGElectromyography. S Symptoms. F Force. Sig. Significant. MSDMusculoskeletal disorders. ST Supraspinatus tendinitis. N Considered (no). PE Physical examination. NR Not reported. VDT Video display terminals. NS Not statistically significant. Y Considered (yes). R Repetition.

C-25

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Andersen 1993a Baron 1991 Bovenzi 1991 Burt 1990 Byström 1995 Chiang 1993 Dimberg 1987 Dimberg 1989 Study type CS CS CS CS CS CS CS CS Participation Y N NR Y Y Y Y Y rate $70% Outcome S S and PE S and PE S S and PE S and PE S and PE S and PE Exposure Job categorization Observation Observation, Questionnaire Observation, Observation Observation job Observation, job by job duration videotape, checklist, vibration videotape videotape analysis analysis, questionnaire measured analysis, EMG of analysis, EMG categorization categorization forearm muscle load collected, however, job title used for analysis Covariates Age, number of Age, gender, Age, ponderal Age gender, Gender, age >40 Age, gender, Gender, age, Ponderal index, considered children, smoking, hobbies, second index years on job, years, psycho- metabolic disease employee gender, age, time socioeconomic jobs, height, psychosocial social variables category, degree in present job, status systemic disease factors and potential of stress, tennis height, weight, confounders playing smoking, house addressed by ownership, Fransson-Hall et racquet sports al. 1995

Investigators Y Y Y Y Y to Y NR NR blinded questionnaire responses, No to exposure status

Repetition Combined Combined Õ 80% of time Combined Combined Õ Õ reported typing vs. 0-19% of time: 2.8 (1.4-5.7)

Force Combined Combined Õ Combined Combined Combined Combined Combined

Extreme Combined Combined Õ Combined Combined Combined Combined Combined posture

See footnotes at end of table. (Continued)

C-26

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Andersen 1993a Baron 1991 Bovenzi 1991 Burt 1990 Byström 1995 Chiang 1993 Dimberg 1987 Dimberg 1989

Vibration Õ Õ Vibration-exposed Õ Õ Õ Õ p<0.01 forestry workers vs. referents: 4.9 (1.27-56.0)

Risk factors Sewing machine Checkers vs. Õ Reporters Assembly line Group III vs. Group Force and posture: Force and (combined) operators vs. Noncheckers: compared to workers vs. I (females): 1.44 NR, Sig. posture: NR, NS general population 2.3 (0.5-11.0) others: 2.5 population referen (0.3-5.6) 1.7 (0.9-3.3) (1.5-4.0) ts: 0.74 High force/high (0.04-1.7) repetition vs. low force/low repetition: (males) 6.75 (1.6-32.7)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Job satisfaction: Õ Job control and Addressed by Õ Õ Mental stress at factors NS satisfaction: NS Fransson-Hall et the onset of al. 1995 symptoms: p<0.001

Individual/oth Õ Õ Õ Sick leave more Õ Õ “Work” the cause Ponderal index er factors common among in 35% of elbow associated with considered strenuous jobs problems, most elbow symptoms than white collar nonstrenuous jobs

Duration of Õ NS Õ Õ Õ Õ Õ Õ employment

Dose/respon Õ Õ Õ Y for time spent Õ Y for males with Õ Õ se typing increasing force/repetition

See footnotes at end of table. (Continued)

C-27

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Fishbein 1988 Hales 1994 Hoekstra 1994 Hughes 1997 Kopf 1988 Kurppa 1991 Luopajärvi 1979 McCormack 1990 Study type CS CS CS CS CS Cohort CS CS Participation N Y Y N N Y Y Y rate $70% Outcome S S and PE S S and PE S S and PE S and PE S and PE Exposure Questionnaire Observation and Observation and Observation, Questionnaire, job Observation, Observation, Observation, job Questionnaire Questionnaire checklist, formal categories measurements, interviews, categories based job analysis categorized by job videotape analysis on manual titles exposure Confounders Age, gender Age, gender, Age, gender, Age, smoking Age, job Workers used as Age, gender, Gender, age, considered stratification, metabolic location, seniority status, sports, satisfaction, job their own social race, job smoking status, disorder, hobbies, hobbies, metabolic security, controls; age, background, category, years alcohol, beta recreation diseases, acute moistness, gender, duration hobbies, amount of employment blockers, other traumatic injuries, vibration, of employment of housework, drugs smoking Scheuerman’s (with exceptions) length of Disease employment

Investigators NR Y Y NR NR NR Y NR blinded

Repetition Combined Number of key- Õ Õ Combined Combined Combined Combined strokes per day: NS

Force Õ Õ Õ Number of years Combined Combined Combined Combined handling >2.5 kg/hand: NS

Extreme Combined Õ Non optimally Wrist flexion/ Combined Combined Combined Õ posture adjusted chair: 4.0 extension: NS; (1.2-13.1) years of ulnar deviation: NS; years of forearm twisting: 37 (3.0-470.0)

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table. (Continued)

C-28

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Fishbein 1988 Hales 1994 Hoekstra 1994 Hughes 1997 Kopf 1988 Kurppa 1991 Luopajärvi 1979 McCormack 1990

Risk factors Female musicians Õ Õ Õ Bricklayers Workers in Assembly Boarding vs. Non- (combined) compared to compared to strenuous vs. workers vs. shop office workers: males: 2.04 manual workers: nonstrenuous assistants: 0.5 (0.09-2.1) (1.6-2.6) 2.8; Increasing job jobs: 6.7 for epicondylitis: Knitting vs. Non- demands OR (3.3-13.9) 2.7 (0.66-15.9) office workers: increased from 1.2 (0.5-3.4) 1.8 to 3.4

Physical Õ Õ Õ Push/pull; lift Sig Õ Õ Õ workload carry: NS

Psychosocial Õ Fear of Job Low decision Õ Õ Õ Õ factors replacement by dissatisfaction; latitude: computers: 2.9 exhaustion 3.5 (0.6-19.0) (1.4-6.1); decision making: 2.8 (1.4-5.7); surge in workload: 2.4 (1.2-5.0)

Individual/oth Õ Race (non-white): Õ Age: 0.96 Õ Õ Õ Age, race Sig er factors 2.4 (1.2-5.0) (0.9, 1.2) considered

Duration of Õ Õ Õ Õ Õ Õ Õ Y, Sig, with <6 employment months and >13 years

Dose/respon Õ Õ Õ Õ Yes, increasing Õ Õ No se levels of job demands

See footnotes at end of table. (Continued)

C-29

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Moore 1994 Ohlsson 1989 Punnett 1985 Ritz 1995 Roto 1984 Viikari-Juntura 1991b

Study type CS CS CS CS CS CS

Participation Y NR Y for cases NR Y Y rate $70% N for referents

Outcome PE records S S S and PE S and PE S and PE

Exposure Observation, videotape Questionnaire, job Questionnaire, job Observation and record Job categorization Observation, job analysis, job strain index categorization category review and employee analysis; weights of interviews items

Confounders Age, gender, duration of Age, gender, duration of Age, number of years Age, age-squared, and Gender, other work Age, gender, duration of considered employment employment employed, native “history of cervical spine tasks employment, leaving the language symptoms”. Having ever company, changing the played tennis, squash, task, being on sick leave other racquet sports, rowing, bowling,

Investigators Y NR NR Y Y NR blinded

Repetition Õ Combined Combined Õ Combined Combined

Force 5.5 (1.5-62) Õ Combined 10 years of high Combined Combined exposure to elbow straining work: 1.7 (1.0-2.7)

Extreme NR: was not found to be Combined Combined Õ Combined Õ posture sig. associated with “hazardous” jobs.

Vibration Õ Õ - Õ Õ Õ

Risk factors Õ Non significant pain in Garment workers vs. Õ Meatcutters vs. Strenuous vs. (combined) last year assembly vs. hospital employees: 2.4 construction workers: nonstrenuous: NS; referents: 1.5 (0.6-3.4) (1.2-4.2) 6.4 (0.99-40.9), p=0.05 difference: 0.88 Work inability in last year (0.27-2.8) assembly vs. Referents: 2.8 (0.8-10.7)

See footnotes at end of table. (Continued)

C-30

Appendix C Table C-4. Summary table for evaluating elbow musculoskeletal disorders Components of study Moore 1994 Ohlsson 1989 Punnett 1985 Ritz 1995 Roto 1984 Viikari-Juntura 1991b

Physical Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Not associated with Age; Non-English Õ Õ Õ er factors work pace speakers sig. less likely considered to report symptoms

Duration of Õ No association Õ Increased duration of All with epicondylitis had Õ employment current exposure >15 years of employment increased risk of epicondylitis

Dose/respon Õ Õ Õ Õ Õ Õ se

Õ Not studied. CS Cross-sectional. EMGElectromyography. F force. Hrs Hours. MSDMusculoskeletal disorders. N no. NR Not reported. NS Not statistically significant. PE Physical examination. R Repetition. Sig. Statistically significant. S Symptoms. Y Considered (yes).

C-31

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Armstrong 1979 Barnhart 1991 Baron 1991 Bovenzi 1991 Bovenzi 1994 Cannon 1981 Chatterjee 1982 Chiang 1990 Study type CS CS CS CS CS Case control Case control CS Participation NR N N NR Y NR Y Y rate $70% Outcome S or surgery or PE PE and NCS S and PE S and PE S and PE Industry medical S and PE and NCS S and PE and NCS findings records Exposure Observation, Observation Observation, Observation, Observation, Medical records, Observation, Observation video, EMG videotape measurement vibration, job category Measurement analysis, job measurement category

Covariates Gender, metabolic Age, gender Age, gender, Age, gender, Age, smoking, Age, gender, Age, gender Age, gender, considered or soft tissue hobbies, past weight alcohol, upper limb race, weight, length of disease employment, injuries occupation, years employment, years on job employed, history of workers metabolic disease compensation status, history of metabolic disease, hormonal status, gynecologic surgery

Investigators N Y, but clothing Y Y N NR Y Y blinded may have biased observation

Repetition Õ Repetitive ski Combined Õ Õ 2.1 (0.7-5.3) Õ 1.87 manufacturing vs. (p<0.018) others NCS: 1.9 (1.0-3.6) PE+NCS: 4.0 (1.0-15.8) S+PE+NCS: 1.6 (0.8-3.2)

See footnotes at end of table. (Continued)

C-32

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Armstrong 1979 Barnhart 1991 Baron 1991 Bovenzi 1991 Bovenzi 1994 Cannon 1981 Chatterjee 1982 Chiang 1990

Force Pinch F: 2.0 Õ Combined Õ Õ Õ Õ Õ (1.6-2.5) Hand F: 1.05 (1.0-1.2)

Extreme Pinch force Õ Õ Õ Õ Õ Õ Õ posture exertion: 2.0 (1.6-2.5)

Vibration Õ Õ Õ 23.1 (no Quarry drillers and 7.0 (3.0-170.0) 10.89 Õ confidence limits) stone carvers vs. (1.02-524.0) p=0.002 polishers and machine operators: 3.4 (1.4-8.3)

Risk factors Õ Õ Grocery checkers Chain saw Õ Õ Õ High cold/ high (combined) vs. other grocery operators vs. repetition: 11.66 workers: 3.7 maintenance (2.92-46.6) (0.7-16.7) workers: 18.8 (2.7-795)

Duration of Õ Õ Y, Sig. Õ Õ 0.09 (0.8-10) Õ NS employment

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Õ Õ Õ Õ Õ Õ Õ er factors considered

Dose/respon Õ Õ Y, Sig. Õ Y, NS Õ Õ Õ se

See footnotes at end of table. (Continued)

C-33

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Chiang 1993 deKrom 1990 English 1995 Färkkilä 1988 Feldman 1987 Franklin 1991 Koskimies 1990 Liss 1995 Study type CS CS Case control CS CS for symptoms Retrospective CS CS and cohort for cohort NCS Participation Y Y Y NR Y Y NR No rate $70% Outcome S and PE S and PE and NCS S and PE S and PE and NCS S and in some PE Records review S and PE and NCS Mailed survey and NCS of workers’ compensation cases Exposure Observation, Questionnaire Questionnaire Interview Observation, Job title and Records of Mailed survey measurement, biomechanical industry vibration exposure EMG analysis, videotaping

Covariates Age, gender, Age, gender, Gender, height, Alcohol Gender, past None NR Gender, age considered metabolic disease, weight, slimming weight medical history, hormonal status courses cigarette smoking, hobbies (No analyses performed to take these into account)

Investigator Y NR, participants Y NR NR Y NR N blinded blinded

Repetition Repetitive fish Õ CTS patients vs. Õ Combined Combined Õ Combined processing vs. other patients: 0.4 other: 1.1 (0.2-0.7) (0.7-1.8)

Force Repetitive fish Õ Õ Õ Combined Combined Õ Õ processing vs. other: 1.8 (1.1-2.9)

Extreme Õ Reported 20 to 40 CTS patients vs. Õ Õ Combined Õ Combined posture hrs./week Flexed other patients: 1.8 wrist: 8.7 (1.2-2.8) (3.1-24.1) Extended 5.4 (1.1-27.4)

See footnotes at end of table. (Continued)

C-34

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Chiang 1993 deKrom 1990 English 1995 Färkkilä 1988 Feldman 1987 Franklin 1991 Koskimies 1990 Liss 1995

Vibration Õ Õ Õ Vibration: Õ Õ Vibration Õ p< 0.05 exposure time and NCS Sig. Right hand: r=-0.27; p=0.01 Left hand r=-0.12 p=NS

Risk factors Repetitive and Õ Õ Õ Year 2 vs. Year 1, Oyster and crab Õ CTS symptoms, (combined) forceful fish numbness and packers vs. dental hygienists processing vs. tingling in fingers: industry-wide vs. dental others: 1.1 2.26 (1.14-4.46) rates: 14.8 assistants: 3.7 (0.7-1.8) (11.2-19.5) (1.1-11.9) Female poultry Responder told workers hi R/hi F that they had CTS: vs. low R F: 2.6 5.2 (0.9-32.0) (1.0-7.3)

Duration of Y,<12 months; No Õ Õ Õ Õ Õ Exposure time Sig. Õ employment for 12 to 60 months and >60 months

Physical Y Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Õ Õ Õ Õ Õ Õ Õ er factors considered

Dose/respons Y, Sig. Y, Sig. Õ Õ Õ Õ Õ Õ e

See footnotes at end of table. (Continued)

C-35

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components Morgenstern of study Loslever 1993 Marras 1991 McCormack 1991 Moore 1994 Nathan 1988 Nathan 1992a Nathan 1992b 1990 Study type CS CS CS CS Retrospective CS Cohort Longitudinal cohort Participation Jobs selected due NR Y Y Y NR N Y=Japanese rate $70% to CTS N=Overall occurrence

Outcome S Records and S and PE S PE and NCS from NCS S and NCS S and NCS medical records records Exposure Observation; Observation; Observation, job Survey Observation, Observation Observation Questionnaire measurements, measurements title videotape, videotaping measurement

Covariates Gender, age, Age, gender, Age, gender, Age, gender, None Age, gender Age, gender, hand Gender, hand considered years on the job, handedness, job race, job pregnancy status, dominance, dominance, hand orientation satisfaction category, years of work history job duration of occupational hand employment tasks, use of employment and use, duration of selected drugs, industry employment, history of wrist industry, leisure injury exercise, heavy lifting, keyboard use, coffee, tea, alcohol

Investigator N NR NR N Y NR NR NR blinded

Repetition Õ Number of wrist Combined 1.88 (0.9-3.8) Combined Group II vs. Group Combined Found to be movements: NS 1:1.0 (0.05-2.0) “protective”

Force Combined Grip forces three Combined Õ Combined Combined Combined times as great in high-risk jobs

Extreme Combined Radial/ulnar ROM: Õ Õ Combined Õ Õ Õ posture 1.52 (1.1-2.1); Flexion/extension ROM: 1.3 (1.0-1.7); Pronation/supinati on ROM: 1.2 (0.9-1.6)

See footnotes at end of table. (Continued)

C-36

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components Morgenstern of study Loslever 1993 Marras 1991 McCormack 1991 Moore 1994 Nathan 1988 Nathan 1992a Nathan 1992b 1990

Vibration Õ Õ Õ Õ Õ Õ Õ Õ

Risk factors High force with Flexion/extension Boarding vs. Õ Meat processors Group I vs. Group Group V vs. Americans with (combined) high flexion: velocity: 3.8 non-office: 0.5 in hazardous vs. III: 1.7 (1.3-2.3) Group I: 1.0 significantly r=0.62; high force (1.5-9.6) (0.05-2.9) safe jobs: 2.8 Group I vs. Group (0.5-2.2) greater and high Flexion/extension Packing vs. Non- (0.2-36.7) V: 2.2 (1.3-3.3) Group IV vs. prevalence of CTS extension: r=0.29 acceleration: 6.1 Group I: 1.4 compared to office 0.4 (1.7-22) (0.9-2.1) Japanese (0.04-2.4) Group III vs. Group Sewing vs. Non- I: 1.5 (1.0-2.2) office 0.9 (0.3-2.9)

Duration of Õ Sig. Prevalence higher >34 hrs./week: Õ Õ Õ Duration of employment in workers with 1.9 (1.1-3.1) employment found <3 years >9 years: 1.7 to be protective employment (1.0-3.2)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Job satisfaction: Õ Õ Õ Õ Õ Õ factors NS

Individual/oth Õ trunk depth: Sig. Õ Õ Õ Õ Age, hand Mean age, body er factors dominance sig. mass index and considered leisure exercise Sig., cigarettes Sig .

Dose/respons Õ Õ Õ Õ Õ Y, Sig. Õ Õ e

See footnotes at end of table. (Continued)

C-37

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Osorio 1994 Punnett 1985 Schottland 1991 Silverstein 1987 Stetson 1993 Tanaka (In Press) Weislander 1989 Study type CS CS CS CS CS CS Case control Participation Y Y for cases; N for NR Y Y Y Y rate $70% comparison group Outcome S and PE, NCS S and PE NCS S and PE S and PE and NCS S S and PE and NCS Exposure Job title, observation Observation, Job title Observation, Observation, Questionnaire Telephone interview questionnaire videotape analysis, questionnaire, job EMG analysis

Covariates Age, gender, alcohol, Age, gender, Age, gender Age, gender, plant, Age, height, skin Age, gender, race, Age, gender, year of considered medical history hormonal status, years on job temperature, cigarettes, income, operation native language, dominant index finger education, BMI history of metabolic circumference disease

Investigator Y NR NR Y NR No No blinded

Repetition Combined Combined Combined Repetition: 5.5 p<0.05 NS Õ 2.7 (1.3-5.4)

Force Combined Combined Combined Combined Y, Sig. combined Õ Õ

Extreme Õ Õ Combined Ulnar deviation and Combined (pinch Bending/twisting of Õ posture pinching, elevated grip) the wrist: 5.9 but NS (3.4-10.2)

Vibration Õ Õ Õ 5.3 Õ Vibration: 1.85 Vibrating tool use 3.3 (no confidence limits) (1.2-2.8) (1.6-6.8)

Risk factors NCS: 6.7 (0.8-52.9) Force, repetition, Workers vs. High force/high Y, Sig. median Õ Õ (combined) Super-market posture: 2.7 (1.2-7.6) applicants: repetition vs. low sensory amplitudes workers, high vs. females, right hand: force/low repetition: Sig. smaller (p < low exposure 2.86 (1.1-7.9); 15.5 (1.7-142.0) 0.01) and latencies symptoms: 8.3 males, right hand: longer (p<0.05) with (2.6-26.4) 1.87 (0.6-9.8) exposure to high pinch grip forces

See footnotes at end of table. (Continued)

C-38

Appendix C Table C-5a. Summary table for evaluating work-related carpal tunnel syndrome (CTS) Components of study Osorio 1994 Punnett 1985 Schottland 1991 Silverstein 1987 Stetson 1993 Tanaka (In Press) Weislander 1989

Duration of Y NS Õ 0.9 Õ Õ Õ employment p>0.09

Physical Y Õ Õ Õ Õ Õ Loads on wrist 1.8 workload (1.0-3.5)

Psychosocial Õ Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Õ Õ Õ Õ Female gender: 2.4 Õ er factors (1.6-3.8); BMI $25: considered 2.1 (1.4-3.1); white race: 4.2 (1.9-15.6) Cigarettes: 1.6 (1-2.5); annual income $$20,000: 1.5 (1-2.4)

Dose/respons Y, Sig. Õ Õ Y, Sig. Õ Õ Õ e

Õ Not studied BMI Body Mass Index CS Cross-sectional CTS Carpal tunnel syndrome EMGElectromyography F Force hrs Hours NCS Nerve conduction studies NR Not reported NS Not statistically significant PE Physical examination R Repetition Sig. Statistically significant S Symptoms Y Considered (yes)

See footnotes at end of table. (Continued)

C-39

Appendix C Table C-5b. Summary table for evaluating work-related hand/wrist tendinitis Components of study Amano 1988 Armstrong Byström 1995 Kuorinka 1979 Kurppa 1991 Luopajärvi 1979 McCormack Roto 1984 1987a 1990 Study type CS CS CS CS Cohort CS CS CS Participation NR Y Y Y Y Y Y Y rate $70% Outcome S and PE S and PE S and PE S and PE S and PE S and PE S and PE S and PE Exposure Job titles or self- Observation, Questionnaire, Records, Observation, Observation, Observation, job Job title reports measurements, observation, observation, measurements, measurements, category video analysis, measurements, measurements, video analysis. video analysis EMG videotape videotape Reader referred analysis, EMG analysis to methods found in previous publications Covariates Age, gender Age, gender, Age, gender, Age, gender, Age, gender Gender (only Race, age, gender Rheumatoid considered years on job, and psychosocial body mass index, females in study arthritis industrial plant factors “muscle-tendon” groups), age, (addressed by syndrome hobbies, Fransson-Hall housework, et al. 1995) medical conditions

Investigator NR Y No NR NR Y NR Y=occupation s blinded No=occupation of meat processing subjects No=construction foremen (referent)

Repetition Combined Combined Combined Combined Combined Combined Combined Combined Force Combined Combined Combined Combined Combined Combined Combined Combined Extreme Combined Significant Combined Combined Combined Combined Combined Õ posture differences between males and females Vibration Õ Õ Õ Õ Õ Õ Õ Õ

See footnotes at end of table (Continued)

C-40

Appendix C Table C-5b. Summary table for evaluating work-related hand/wrist tendinitis Components of study Amano 1988 Armstrong Byström 1995 Kuorinka 1979 Kurppa 1991 Luopajärvi 1979 McCormack Roto 1984 1987a 1990 Risk factors Right index finger Comparison De Quervain’s Scissor makers Meat cutter Assembly line Textile workers Meat cutters vs. (combined) flexor: 3.67 between low tendinitis among vs. shop compared to workers vs. shop compared to non- construction (1.85-7.27) R/low F and high among auto assistants: 1.38 office workers: assistants: 4.13 office workers: workers: 3.09 Left index finger R/high F: assembly (0.76-2.51) risk ratio: 14.0 (2.63-6.49) 3.0 (1.4-6.4) (1.43-6.67) flexor: 6.17 4.8 (0.6-39.7) workers vs. (5.7-34.4); Overall group (2.72-13.97) 5.5 (0.7-46.3) general Meat packers exposed: 1.75 17.0 (2.3-126.2) population: 2.5 compared to (0.9-3.39) (1.00-6.23) office workers: risk ratio: 38.5 (11.7-56.1); sausage makers compared to office workers: risk ratio: 25.6 (19.2-77.5)

Physical Õ Õ Õ Õ Õ Õ Õ Õ workload Psychosocial Õ Analyzed by Õ Õ Õ Õ factors Fransson-Hall et al. 1995

Individual/ Õ Õ Õ Pieces handled Õ NS for age, Female gender Rheumatoid other factors over the years: a hobbies, or significant for arthritis found not considered nonsignificant housework tendinitis at to be a trend with p=0.01; confounder increasing number job category of pieces handled significant at p=0.001 Duration of Õ Õ Õ Õ Õ No association Õ Õ employment Dose/respon Õ With increasing Õ Õ Õ Õ Õ Õ se combination of R and F Õ Not studied. NS Not statistically significant. CS Cross-sectional PE Physical examination. EMG Electromyography. R Repetition. F Force. S Symptoms. HAVSHand-arm vibration syndrome Y Considered (yes). NR Not reported.

C-41

Appendix C Table 5c. Summary table for evaluating hand-arm vibration syndrome Components Brubaker Brubaker Dimberg Koskimies McKenna of study Bovenzi 1988 Bovenzi 1994 Bovenzi 1995 1983 1987 1991 Kivekäs 1994 1992 Letz 1992 1993 Study type CS CS CS CS Cohort CS Cohort Cohort CS CS Participation NR Y Y Y N Y Y Y Y NR rate $70% Outcome S and PE; S and PE S and PE; S and PE; S and PE; S S and PE S and PE S S and cold cold cold cold PE; provocatio provocation provocatio provocatio cold n n n provocati on Exposure Observatio Observatio Questionnair Question- Observati Questionn Questionna Measureme Questionnai Question n; n, e, naire data on; aire ire nt of the re, naire measurem interview, observation, measure tools measureme ents of the measurem measureme ments of nts of the tool ents of the nts of the the tool tool used tool tool from previous studies

Covariates Õ Age, Age, Smoking, Age, Õ Age Õ Age, race, Age, considered smoking, smoking, age, gender, smoking, smoking, alcohol drinking height, psychoso alcohol, only consumpti habits, weight cial medical males on, upper cardiovascul scales conditions studied, limb ar, those injuries; neurologic, with leisure previous injury to activities, musculoskel the neck, systemic etal injuries, upper diseases use of limbs medicines excluded. Investigator NR N Y NR NR NR Y NR No N s blinded

Repetition Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ

Force Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ

Extreme Õ Õ Õ Õ Õ Õ Õ Õ Õ posture

See footnotes at end of table (Continued)

C-42

Appendix C Table 5c. Summary table for evaluating hand-arm vibration syndrome Components Brubaker Brubaker Dimberg Koskimies McKenna of study Bovenzi 1988 Bovenzi 1994 Bovenzi 1995 1983 1987 1991 Kivekäs 1994 1992 Letz 1992 1993 Vibration Stone Stone Forestry NR 15% of Vibrating Lumberjack Decrease Full-time Riveters drillers and workers workers and fellers tool use es vs. in vibration vs. cutters vs. vs. 2.6% in ship- reported sig. referents: prevalence workers vs. referents: quarry and polishers yard new Correlated for 1978: in forest referents: 24 mill and referents: symptom with HAVS 3.4, workers 5.0 (3.1-510) workers: machine OR = 11.8 s of VWF symptom (1.7-6.9) from 1972 (2.1-12.1) 6.06 operators: (4.5-31.1) from 1979 prevalance Cumulative to 1990, Full-time (2.0-19.6) 9.33 For workers to 1985; incidence attributed to vibration (4.9-17.8) only using 28% HAVs reduction in workers vs. antivibration increase (7-years) weight of Controls: saws: OR = in 14.7% vs. saws, 40.6 6.2 prevalenc 2.3%: 6.5 increase in (11-177) (2.3-17.1) e of VWF (2.4-17.5) vibration For those in workers frequency, using non- using reduction in antivibration antivibrati acceleratio saws: OR = on chain- n 32.3 saws (11.2-93)

Risk factors Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ (combined)

Physical Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ factors

Individual/ Õ See See Age Õ Vibrating Õ Õ Smoking Õ other factors “Covariate “Covariates significantl tool use Sig. considered s considered” y different significantl considered above between y ” above cases and correlated controls, with HAVS height and symptoms weight prevalence were not.

Duration of Õ Õ Õ Õ Õ Õ No Õ Õ Õ employment differece in lumberjack s with <15 years of exposure, but then increased with duration of exposure

Dose/respon Õ Õ Y, between Õ Õ Increased Õ Sig. for Õ se increasing HAVS with reported vibration duration of exposure to exposure exposure vibratory and tools in “vibration workers white finger” with <17,000 hours of exposure

See footnotes at end of table (Continued) C-43

Appendix C Table 5c. Summary table for evaluating hand-arm vibration syndrome Components Mirbod Mirbod Miyashita Virokannas of study 1992a, 1994 1992b 1992 Musson 1989 Nagata 1993 Nilsson 1989 Saito 1987 Shinev 1992 Starck 1990 1995 Study type CS CS CS CS CS CS Cohort CS CS CS Participation NR NR NR N NR Y for N NR NR NR rate $70% platers; NR for office workers Outcome S S and PE S S S and PE S and PE S and S and PE S S and PE PE

Exposure Questionn Questionn Job Title Postal Based on Questionnai Question Measurem Measurem Interview aire; aire; questionnai years of re, naire ent of tool ent of interviews, measurem re, exposure measureme tools measurem ents of the measurem since nt of tool, ents of the workers ent of employme exposure workers and the representat nt time and the tools ive tools tools

Covariates Age Õ Õ Age, height, Age Age Follow- Age, N Age, considered weight, up of cigarette duration of smoking, cohort smoking, employme time industry, nt pressure, education working VDT posture training

Investigator NR N N NR N NR NR NR N NR s blinded

Repetition Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ

Force Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ

Extreme Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ posture Vibration Male chain Symptom Male Exposure For >20 Office NR Percussiv High NR saw severity Constructi duration not years workers e vibration prevalenc operators positively on related to vibration with no had a e of HAVS vs. correlated workers HAVS exposure: vibration greater among referents: with compared symptoms 7.1 exposure to effect on workers 3.77 exposure to male (2.5-19.9) former muscle using (2.1-6.8) duration office exposure: and bone vibrating workers: 14 (5-38) pathology tools 0.5 Office than (0.1-11.8) workers constant with no high- exposure: frequency 85 (15-486) vibration

See footnotes at end of table (Continued) C-44

Appendix C Table 5c. Summary table for evaluating hand-arm vibration syndrome Components Mirbod Mirbod Miyashita Virokannas of study 1992a, 1994 1992b 1992 Musson 1989 Nagata 1993 Nilsson 1989 Saito 1987 Shinev 1992 Starck 1990 1995

Risk factors Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ (combined)

Physical Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ workload

Psychosocial Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ factors

Individual/ Õ Õ Õ Õ Õ Õ Age Sig. Õ Poor Õ other factors Correlat correlation considered ed to between recovery vibration rates exposure from and HAVS 1978 to when 1983 tools were highly impulsive

Duration of Õ Õ Õ Õ Õ Õ Õ Õ Õ Õ employment

Dose/respon Õ HAVS Õ Õ Õ OR increased by Õ Õ Õ Õ se symptom 11% for each severity year of positively exposure correlated with exposure duration

Õ Not studied. CS Cross-sectional. CTS Carpal tunnel syndrome. EMGElectromyography. F Force. Hrs Hours. NCS Nerve conduction studies. NR Not reported. NS Not statistically significant. OR Odds ratio. PE Physical examination. R Repetition. S Symptoms. Sig Statistically significant. VPT Vibration perception threshold. Y considered (yes).

C-45

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components Boshuizen 1990a, of study Åstrand 1987, 1988 Bergenudd 1988 Bigos 1991b Bongers 1988 Bongers 1990 1990b Study type 1987: CS; Cohort Cohort Retrospective cohort CS CS 1988: Cohort Cohort Participation Y N N Y Y Y rate $ 70% Outcome S and PE S S Physical exam from S CS: S disability records Cohort: records Exposure Questionnaire Questionnaire Questionnaire; For jobs Job title and records; Questionnaire; vibration Questionnaire; vibration with >19 workers: job vibration measurements measurements measurements analysis obtained but not used

Covariates Education level, Years of education, Medical history, previous Nationality, shift-work, Age, height, weight, Duration of exposure, considered psychosocial factors psychosocial factors episodes of back pain, age, and calendar time climate, bending forward, age, height, smoking, (including neuroticism) “individual” factors, twisted postures and awkward postures, and psychosocial factors feeling tense at work mental workload (from MMPI)

Investigators N NR NR NR NR NR blinded

Heavy Combined Workers in moderate and No association Õ Õ Õ physical work heavy physical demand work groups vs. light physical demand group: 1.8 (1.2-2.7)

Lifting and Combined Õ Õ Õ Õ Õ forceful movements

See footnotes at end of table. (Continued)

C-46

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components Boshuizen 1990a, of study Åstrand 1987, 1988 Bergenudd 1988 Bigos 1991b Bongers 1988 Bongers 1990 1990b

Awkward Õ Õ Õ Õ Õ Õ postures

Whole body Õ Õ Õ All back disorders: 1.32 LBP in exposed vs. LBP by vibration dose vibration (0.84-2.1); referents: 9.0 (4.9-16.4), category: ORs=19.1, Intervertebral disc Sciatica: 3.3 (1.3-8.5); 29.4, 28.0, 38.1; disorders: 2.00 (1.1-3.7); LBP by total vibration By vibration dose: Disc degeneration by dose: ORs=12.0, 5.6, ORs=1.80, 1.78, 2.8; years of exposure: 5.7 6.6, 39.5 years of exposure: 3.6 (for highest exposure LBP by hours of flight (1.2-11) category) time per day: 5.6, 10.3, 14.4;

Static work Õ Õ Õ Õ Õ Õ postures

Risk factors Mill workers vs. clerical Õ Õ Õ Õ Õ (combined) workers: 2.3 p=0.002

Psychosocial Neuroticism and back Those with back pain MMPI: tend towards Õ Õ Õ factors pain: 2.8 (1.4-5.4) less satisfied with somatic complaint or working conditions; no denial of emotional difference in social distress and reporting support injury: 1.37 (1.1-1.7)

Individual/oth Õ Õ Does not enjoy job tasks Õ Õ Õ er factors and reporting injury: 1.7 considered (1.3-2.2)

Duration of Duration of employment Õ Prior back pain and Õ Õ Õ employment and back pain: 1.2 reporting injury: 1.7 (1.0-1.5) (1.2-2.5)

Dose/respon Õ Õ Õ Õ Õ Õ se

See footnotes at end of table. (Continued)

C-47

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Boshuizen 1992 Bovenzi 1992 Bovenzi 1994 Burdorf 1990 Burdorf 1991 Burdorf 1993 Study type CS CS mail survey CS CS CS CS Participation Y Y Y N Y Y rate $70% Outcome S S S S S S Exposure Questionnaire; vibration Questionnaire, Questionnaire, Questionnaire, Questionnaire, task Questionnaire, measurements measurement of WBV measurement of vibration job title, and expert analysis and OWAS measurements of WBV, levels knowledge Postures assessed with OWAS

Covariates Mental stress, years Age, awkward posture, Age, BMI, education, Age, height, and weight Age, height, and weight Age, history of heavy considered lifting >10 kg and twisting duration of exposure, sport activity, car driving, work, exposure to WBV, spine, height, smoking, BMI, mental load, marital status, mental work requiring prolonged looking backwards, education, smoking, stress, climatic sitting, cold, drafts, hours sitting sport activities and conditions, back trauma, working under severe previous jobs at risk for and postural load (or pressure, job back pain total vibration dose) satisfaction, height, weight, duration of total employment

Investigators NR NR NR NR N NR blinded

Heavy Õ Õ Õ Heavy work: 4.02 Heavy physical work sig Õ physical work (0.76-21.2) in univariate but not multivariate model

Lifting and Õ Õ Õ Frequent lifting: 5.21 No association Õ forceful (1.10-25.5) movements

Awkward Õ Õ Õ Õ Postural Index and LBP: Õ postures 1.23 p=0.04 Whole body Total vibration dose and Low back: LBP in the past year: WBV: 0.66 (0.14-3.1) WBV and LBP, 3.1 Combined vibration back pain: 0.99 Previous 12 months OR=2.39 (1.6-3.7) p=0.001 (0.85-1.2); In younger prevalence of LBP, bus Postural load category: workers: vibration in drivers vs. controls: 2.57 OR=4.56 (2.6-8.0) (for past 5 years and (1.5-4.4) the highest exposure lumbago, 3.1 (1.2-7.9) Multivariate: category) LBP symptoms in previous. 12 months: and total vibration dose: OR’s= 1.67, 3.46, 2.63

See footnotes at end of table. (Continued)

C-48

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Boshuizen 1992 Bovenzi 1992 Bovenzi 1994 Burdorf 1990 Burdorf 1991 Burdorf 1993

Static work Õ Õ Õ For univariate analysis: Posture index based on Õ postures sedentary postures in time spent in a working crance operators: 0.49 posture with the back in (0.11-2.2) a bent and/or twisted position: 1.23 p=0.04

Risk factors Õ Õ Õ Job title: 3.6 (1.2-10.6) Õ Crane operators vs. (combined) office workers: 3.29 (1.52-7.12) Straddle-carrier drivers vs. office workers: 2.5 (1.2-5.4)

Psychosocial Õ Õ Õ Õ Õ Õ factors

Individual/oth Õ Õ Õ Õ Postural load, bending, Õ er factors and twisting are causal considered factors. Standing and sitting are not found to be risk factors.

Duration of Õ Õ Õ Õ Õ Õ employment

Dose/respon Õ Univariate analysis, total Dose/response of Õ Õ Õ se vibration dose: combined effects to total lifetime LBP symptoms: vibration dose and 4.05 (1.8-9.3); postural load, highest 12 months LBP combination of symptoms: 3.25 categories: 4.58. (1.5-7.0).

See footnotes at end of table. (Continued)

C-49

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Chaffin 1973 Clemmer 1991 Deyo 1989 Heliövaara 1991 Hildebrandt 1995 Hildebrandt 1996 Study type Cohort CS CS CS CS CS Participation NR Y NHANES-ll data Y Y Y, but varied from 60% rate $70% to 80% by department Outcome S Injury report Data base S and PE S S (LBP) Exposure Observation and Job title Data base Questionnaire Questionnaire Questionnaire measurement (smoking, obesity, personal characteristics) Covariates Age, weight, stature, Age, job, length of Age, gender, smoking, Age and gender Age and gender Age considered number of prior back employment obesity, exercise level, episodes, isometric lifting employment status strengths Investigators NR NR N N N N blinded

Heavy Õ Roustabouts vs. control Õ Combined Heavy physical work Nonsedentary steel physical work room operator: 4.3 (no ORs=1.9, 2.5 vs. sedentary work: 1.2, workers vs. referents: confidence limits) p<0.05 No association

Lifting and Approx. 5 Õ Õ Õ Õ Õ forceful movements

Awkward Õ Õ Õ Õ Õ Õ postures

Whole body Õ Õ Õ Õ Õ Õ vibration

Static work Õ Õ Õ Õ Õ Õ postures

See footnotes at end of table. (Continued)

C-50

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Chaffin 1973 Clemmer 1991 Deyo 1989 Heliövaara 1991 Hildebrandt 1995 Hildebrandt 1996

Risk factors Lifting of loads in Job was best predictor Õ LBP and physical stress: Õ NS, (combined) positions which create a of lost time. 2.5 (1.4-4.7) Reference group had Lifting Strength Rating $ high exposure to was considered adverse working potentially hazardous to conditions some people

Psychosocial Õ Õ Ever smoked vs. LBP: Stress load index: 2.4 Õ Õ factors 1.13, Sig. 50 pack years (1.7-3.5) vs. LBP: 1.47, Sig. Body mass index vs. LBP: 1.70, Sig.

Individual/ Age, weight, and stature 75% of back strains Õ Body mass index, alcohol Rates of LBP: Õ other factors did not correlate with precipitated by pushing, , work-related driving, construction: 35%; considered increased incidence of pulling, or lifting. parity, height not truckers: 31%; LBP associated with LBP. plumbers: 31% Smoking sig in both older and younger males, but only older females. Prior traumatic injury increased risk of LBP: 2.5 (1.9-3.3); and sciatica: 2.6 (2.1-3.1)

Duration of Õ Õ Smoking risk increases Õ Õ Õ employment steadily with cumulative exposure and with degree of maximal daily exposure. There is a steady increase in LBP with increasing obesity.

Dose- Õ Õ Õ Õ Õ Õ response

See footnotes at end of table. (Continued)

C-51

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Holmström 1992 Huang 1988 Johanning 1991 Johansson 1994 Kelsey 1975b Kelsey 1984 Knibbe 1996 Study type CS CS CS mail survey CS Case control Case control CS Participation Y Y N Y Y Y Y rate $70% Outcome S; (A sample had PE S S S Medical records: S S and PE S for purposes of and PE required validation) Exposure Postal questionnaire Ergonomic Job title, Questionnaire Questionnaire Interview and Questionnaire assessment including measured WBV in questionnaire NLE exposed group but results not presented Covariates Daily traveling time, Age, height, length of Age, gender, job Age and gender. Non Age, gender Age, gender, medical Age considered leisure activity, height employment, title, employment work-related S could service and weight olecranon height, duration have an effect weight masking result, if not identified.

Investigators Y NR NR NR NR NR N blinded

Heavy Õ Õ Õ Blue collar workers Õ Õ Õ physical work vs. white collar workers: no association

Lifting and One year prevalence The workers in the Õ No association Lifting vs. herniation: Lifting >25 lb or more, Registered nurses vs forceful of BP and manual center with higher 0.94, p=0.10 without twisting the nursing aides: movements materials handling: rates had greater body: 3.8 (0.7-20.1) Unadjusted OR=1.2, 1.3 (1.2-1.4); lifting compared to p=0.04; after Lifting frequency: >1 the referent center: adjusting for hr per 5 min vs.<1 per 5 no risk estimate worked, aides had min: 1.12, p<0.001 higher rate: 1.3

Awkward Stooping and More awkward Õ Extreme work Combined Twisting without Õ postures kneeling with severe postures found in postures sig lifting: 3.0 (0.9-10.2) LBP compared to no center A than B, associated with stooping: 2.6; in p=0.05. outcome in blue collar comparison to no workers kneeling: 3.5

Whole body Õ Õ WBV and sciatica Õ Combined Õ Õ vibration pain: 3.9 (1.7-8.6)

See footnotes at end of table. (Continued)

C-52

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Holmström 1992 Huang 1988 Johanning 1991 Johansson 1994 Kelsey 1975b Kelsey 1984 Knibbe 1996

Static work No association Õ Õ Õ Sedentary work and Õ Õ postures disc herniation for workers 35 years and older: 2.4, p=0.01; for those < 35 years, 0.81

Risk factors Õ Õ Õ Õ Time sitting, >35 Lifting >25 lb >5 times Physically demanding (combined) years old: 2.4 per day, and twisting work vs. lifetime LBP, p=0.01; More than the body half the prevalence: 87%; half time driving vs. time: 3.1 (1.3-7.5); 1-year LBP, herniation: 2.75, prevalence: 67%; p=0.02; Simultaneous lifting 1-week LBP, Truck driver vs. and twisting with prevalence: 21%; herniation: 4.67, straight knees: 6.1 Prevalence of sick Chi-sq.=5.88, p=0.02 (1.3-27.9) leave due to back pain in previous 3 months: 9.7%

Psychosocial High stress and LBP: Õ Blue collar workers In blue-collar Õ Õ Õ factors 1.6 (1.4-1.8); were less satisfied workers, 10 of 15 high anxiety: 1.3 with “influence on psychosocial job (1.1-1.4). and control of work, factors sig; in supervisor climate, white-collar workers, stimulus from work none of the five itself, and relations psychosocial factors with fellow workers sig

Individual/oth Severe LBP related Õ Gastrointestinal Õ Õ Carrying >11.3 kg, Õ er factors to smoking; problems: subway 5-25 per day: 2.1 considered construction tasks train operators vs. (1.0-4.3) such as brick laying, referents: 1.6 carpentry, etc. did (1.1-2.5) Carrying >11.3 kg , not affect LBP. >25/day: 2.7 (1.2-5.8)

Duration of Õ Õ Õ Õ Õ Õ Õ employment

Dose/respon Õ Õ Õ Õ Õ Õ Õ se

See footnotes at end of table. (Continued)

C-53

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Leigh 1989 Liles 1984 Magnusson 1996 Magora 1972, 1973 Marras 1993, 1995 Masset 1994 Partridge 1968 Study type CS Cohort CS CS CS CS CS Participation Y NR NR NR NR Y Y rate $70% Outcome S Records S S Records review S S and PE Exposure Questionnaire Observation, use of Questionnaire, Observation, Observation, Interview, Questionnaire, (job title) records vibration interview, measurements self-reports job title measurements questionnaire

Covariates Gender, race, Õ Õ Õ Õ Gender (males only), Age considered obesity, height, and age (all participants repetitious work younger than 40). General health status, social, demographic, psychologic factors

Investigators NR N NR NR NR NR N blinded

Heavy Self reporting: “Job Õ Õ Õ Õ No association Combined physical work requires a lot of physical effort”: 1.5 (1.0-2.2)

Lifting and Õ Injury rate for highest Heavy lifting: 1.86 1973: Sudden Combined Heavy efforts of the Õ forceful job severity index (1.2-2.8) maximal efforts and shoulder, 1.62, movements category vs lowest : Frequent lifting: 1.55 LBP: 1.65 (1.3-2.1) p<0.01 4.5 (1.01-2.39)

Awkward Õ Õ Õ No association: Õ Univariate analysis Õ postures highest rate of back showed trunk pain found in the torsions associated “rarely/never bend” with LBP in steel category workers; no association seen in multivariate

Whole body Õ Õ Bus and truck drivers Bus drivers Õ Vehicle driving: 1.2 Õ vibration compared to compared to (p<0.001) referents: 1.8 bankers: 1.2 (1.2-2.8) (0.8-1.7)

See footnotes at end of table. (Continued)

C-54

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Leigh 1989 Liles 1984 Magnusson 1996 Magora 1972, 1973 Marras 1993, 1995 Masset 1994 Partridge 1968

Static work Õ Õ Õ No association Õ Seated posture: 1.5, Õ postures p<0.09

Risk factors High vs. low physical Õ Driving: 1.79 Sudden maximal Max. load moment, Õ Rheumatic S: (combined) demands: 1.68 (1.16-2.75) physical efforts; max. lateral velocity, dockers vs. civil (1.05-2.90) Vibration plus prolonged sitting or ave. twisting servants: 1.2 frequent lifting: 2.1 standing, inability to velocity, lifting (0.98-1.64); (0.8-5.7) sit during the working frequency, and max. LBP: dockers vs. civil Vibration plus heavy day, and poor lifting sagital trunk angle servants: NS lifting: 2.06 (1.3-3.3) technique related to related to high-risk LBP LBP groups: 10.7(4.9-23.6)

Psychosocial Õ Õ Õ Õ Õ Negative perception Õ factors of the work environment: NS.

Individual/oth Smoker vs. Õ Õ Õ Maximum load Physical work load Õ er factors nonsmoker and LBP: moment: 73.65 Nm (no objective considered 1.48 (1.0-2.19) vs. 23.64 Nm: 5.17, measurement) and (3.19-8.38); repetition were NS. Sagittal mean Final logistic model velocity: 11.74 included “whole set degrees/sec. vs. of variables from 6.55 degrees/ general health status, sec: 3.33 social, demographic, (2.17-5.11); and psychologic Max. weight: 104 N characteristics.” vs. 37 N: 3.17 (2.19-4.58)

Duration of Õ Õ Õ Õ Õ Õ Õ employment

Dose/respon Õ Õ Õ Õ Õ Õ Õ se

See footnotes at end of table. (Continued)

C-55

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components Riihimäki 1994; of study Punnett 1991 Riihimäki 1989a Riihimäki 1989b Pietri-Taleb 1995 Ryden 1989 Schibye 1995 Skov 1996 Study type Case referent CS mail survey CS Prospective Case control Cohort CS (retrospective) Participation Y Y Y Y Y Y N rate $70% Outcome S and PE S X-ray confirmed S Records S S Exposure Observation and Job title and Questionnaire and Postal questionnaire Work injury reports Questionnaire Questionnaire, measurements, questionnaire job title and self-reports self-reports Videotape analysis Covariates Gender, age, length Age, previous back Age, self-reported Age, gender (only Age Subjects served as Age, gender, height, considered of employment, accidents, awkward back accidents, body males were studied, their own controls weight, smoking, recreational activity, postures at work, mass index, height, previous history of work-related medical history, and and annual car and smoking back accidents, psychosocial maximum weight driving mental distress, variables, lifting, lifted in study job general state of leisure time sports health, smoking, activities lifestyle factors, education Investigators Y NR Y NR NR NR NR blinded

Heavy Õ Combined Õ Õ Combined Õ Õ physical work

Lifting and Lift 44.5 N: 2.16 (1.0- Õ Õ Õ Õ Õ Õ forceful 4.7) movements

Awkward Time in non-neutral Sciatica and twisted Õ Association found Õ Õ Õ postures postures, mild or or bent postures: 1.5 between twisted and severe bending: 8.09 (1.2-1.9) bent postures with (1.4-44) sciatica in univariate, but not multivariate analysis

See footnotes at end of table. (Continued)

C-56

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components Riihimäki 1994; of study Punnett 1991 Riihimäki 1989a Riihimäki 1989b Pietri-Taleb 1995 Ryden 1989 Schibye 1995 Skov 1996

Whole body Õ Longshoremen and Õ No association Õ Õ In Danish vibration earthmovers salespeople, the compared to annual driving referents: 1.3 distance for highest (1.1-1.7) category: 2.8 (1.5-5.1)

Static work Õ Õ Õ Õ Õ Õ Sedentary work postures (% of worktime): 2.45 (1.2-4.9) Risk factors Time in non-neutral Sciatic pain and Concrete vs. painting Machine operators Job title or shifts No sig differences in Annual driving (combined) posture: 8.09 machine operators: work and disc space vs. office workers: requiring heaviest back pain in garment distance: 2.79 (1.5-44.0) 1.3 (1.1-1.7) narrowing: 1.8 1.4 (0.99-1.87); physical efforts: 2.2 workers versus (1.5-5.1) Sciatic pain and (1.2-2.5); carpenters vs. office (1.28-3.89) other employment carpenters: 1.0 Spondylophytes: 1.6 workers: 1.5 group upon follow-up (0.8-1.3) (1.2-2.3) (1.1-2.1)

Psychosocial Õ Õ Õ Monotonous work, Õ Õ Õ factors problems with co-workers or supervisors, and high paced work were NS.

Individual / Age: 0.96 (0.09-1.0) Õ Age and disc space Physical exercise >1 Previous back injury: Of 82 workers with Õ other factors back injury: 2.37 narrowing: 6.5 time per week vs. 1 2.13 (1.07-4.24); another job in 1991, considered (1.3-4.3) (1.7-26.0) time per week: 1.26 Working day shift: 20% reported MSDs (1.0-1.6) 2.23 (1.28-3.89); a s the reason for Spondylophytes: Smokers vs. Self-reported LBP: change. 14.9 (2.3-95.0) non-smokers: 1.29 1.25 (1.25-4.12); (0.98-1.7) Self-reported slipped Severe back pain disc: 6.20 and later sciatica: 4.5 (2.64-14.57) (2.7-7.6)

Duration of Analysis controlled Õ Õ Õ Õ Sig Õ employment for length of employment.

Dose/respon A strong trend found Dose/response is Õ Õ Õ Õ Dose/response is se for increasing length observed for twisted observed for annual of exposure and risk or bent postures driving and of back disorders to (see above) sedentary work (see both mild and severe above) trunk flexion.

See footnotes at end of table. (Continued)

C-57

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Skovron 1994 Svensson 1989 Toroptsova 1995 Undeutsch 1982 Videman 1984 Videman 1990 Walsh 1989 Study type CS CS CS CS CS CS and lab study CS (retrospective)

Participation Y Y Y NR Y NR Y rate $70% Outcome S S S; then S and PE S and PE (Clinical S X-ray confirmed S orthopaedic exam given to 134 of the 366 subjects)

Exposure Interview Questionnaire Interview Interview and Postal questionnaire Questionnaire, Postal questionnaire Reports from family questionnaire members

Covariates Age and gender Age, gender (only Analysis did not Age, height, weight, Age, gender (only Age, gender (only Age, year of onset considered females studied), control for nationality, years of females studied), male cadavers used) of symptoms, level of education, confounders experience in menstruation, physical exercise, gender psychosocial transport work pregnancy, exercise heaviness of factors, work occupation breaks, demand on concentration

Investigators NR NR NR NR NR NR NR blinded

Heavy Õ No association Õ Õ Sig. difference in Heavy vs. mixed Õ physical heavy occupational work: 2.8 (0.3-23.7) work workload category among ages 20-29 Heaviest work year olds but not category: 12.1 other age groups: 1.1 (1.4-107)

Lifting and Õ Lifetime incidence of Frequent lifting and Combined No association - no Õ Lifting in jobs just forceful LBP and Lifting: 1.2, LBP: 1.43, p<0.05 sig difference prior to injury: 2.0 movements p<0.01 found in between qualified (1.1-3.7) univariate analysis nurses and nursing but not in multivariate aides analysis

Awkward Õ LBP and bending Trunk flexion and Õ Õ Õ Õ postures forward: 1.3, p<0.05 LBP: 1.7 p<0.01 in univariate; not sig in multivariate analysis

See footnotes at end of table. (Continued)

C-58

Appendix C Table C-6. Summary table for evaluating back musculoskeletal disorders Components of study Skovron 1994 Svensson 1989 Toroptsova 1995 Undeutsch 1982 Videman 1984 Videman 1990 Walsh 1989

Whole body Õ Õ No association Õ Combined Õ Driving on job held vibration prior to symptoms in males: 1.7 (1.0-2.9)

Static work Õ “Standing” No association Õ Õ Sedentary work and Sitting and LBP: postures associated with LBP: disc degeneration: females: 1.7 1.3 in univariate 24.6 (1.5-409) (1.1-2.6) analysis, not sig in multivariate

Risk factors Occupation: NS Õ Õ In workers with Õ Driving vs. Mixed Driving and LBP: (combined) present S, they work: 2.3 (0.8-6.2) males: 1.7 (1.0-2.9) occurred most frequently while lifting loads and while in bended postures: no risk estimate

Psychosocial Work dissatisfaction: LBP and worry and Õ Õ Õ Õ Õ factors 2.4, p=0.02 fatigue at end of work day: p<0.0001 Dissatisfaction with work tasks: p<0.05

Individual / Female gender: 2.16, LBP and standing: NS for sitting, Current back S Õ Õ other factors p=0.001; p<0.01 standing, walking, or positively correlated considered increasing age: 2.0, repetitive work with height and age. p=0.001

Duration of Õ Õ Õ Current back S Õ Õ Õ employment positively correlated with length of experience in transport work.

Dose/respon Õ Õ Õ Õ Õ Õ Õ se

Õ Not studied. N No. Y Considered (yes). ADL Activities of daily living. NHANESNational Health and Nutrition Examination Survey. CS Cross-sectional. NR Not reported. F Force. NS Not statistically significant. Hrs Hours. OWASOVAKO working posture analysis system. LBP Low-back disorders. PE Physical examination. LBP Low-back pain. R Repetition. LBS Low-back symptoms. S Symptoms. MMPI Minnesota Multiphasic Personality Inventory. Sig. Statistically significant. MS Musculoskeletal. WBV Whole body vibration.

C-59