Work ability and sickness absence
A follow-up study in general practice
Harald Reiso
Section of Occupational Health and Social Insurance Medicine,
Department of General Practice and Community Medicine, University of Oslo
Aust-Agder County Office of the National Insurance Service, Norway © Harald Reiso
ISBN: 82-8080-094-8
Series of dissertations submitted to the Faculty of Medicine, University of Oslo No. 206
All rights reserved. No part of this publication may be reproduced or transmitted, in any form or by any means, without permission.
Printed in Norway: AiT AS e-dit, Oslo
Publisher: Unipub AS, Oslo 2004
Unipub AS is a subsidiary company of Akademika AS owned by The University Foundation for Student Life (SiO) To Nina - my best critic
3 Errata In Paper III, reference 24 should be written as reference 129 in the list of references on page 53.
In Paper VI, in the second paragraph in column 2 on page 1469, all figures in superscript after “Pain intensity on the VAS scale was recoded into three groups.” should be deleted.
4 Contents
Preface 7
Acknowledgements 8
English summary 9
Norsk sammendrag 11
LIST OF PAPERS 13
GENERAL INTRODUCTION 14
Background 14
Function 19
Work ability 21
Work ability and sickness absence 23
AIMS 25
MATERIAL AND METHODS 25
Definitions 25
Variables 25
Collection of data 27
Missing data 31
Statistical analyses 31
Ethical aspects 32
MAIN RESULTS 33
Assessments of reduced work ability 33
Prediction of sickness absence duration 34
GENERAL DISCUSSION 35
Methodological considerations 35
Assessments of reduced work ability 37
Prediction of sickness absence duration 39
Temporal aspects 41
5 Implications of results 41
CONCLUSIONS 42
REFERENCES 43
PAPER I
Work ability assessed by patients and their GPs in new episodes of sickness certification.
PAPER II
Arbeidsevne og kjønn – legers vurderinger av sykmeldte.
PAPER III
Work ability and duration of certified sickness absence.
PAPER IV
To ask the right question at the right time:
When is the patient’s self-assessed work ability most accurate as a predictor of the remaining duration of certified sickness absence?
PAPER V
Doctors’ prediction of certified sickness absence.
PAPER VI
Back to Work:
Predictors of Return to Work Among Patients With Back Disorders Certified As Sick.
A Two-Year Follow-up Study.
APPENDIX
Explanatory notes
Participating doctors
Questionnaire I – patients/doctors
Questionnaire II – patients
6 Preface After finishing medical school in Bergen in 1981, I have mostly worked as a general practitioner. In 1985 – 93 I lived in Vågå, a rural community 140 kilometres north of Lillehammer. My research interest began in that period. My colleague Nils Homb was a great inspiration. In 1993 I began as a part-time senior medical officer at the Aust-Agder County Office of the National Insurance Service. Initially I wanted to explore the research possibilities of insurance databases and look at geographical differences in utilization of insurance benefits in the county. That field of interest was narrowed to assessments of work ability and the possible development of decision-making instruments in sickness absence.
In discussions with general practitioners, medical insurance officers, friends and family it is striking how they all have strong opinions about work ability and sickness absence based on their own work experience. This includes concerns about peoples’ working moral and about doctors who are too liberal or restrictive in their sickness certification practice. The National Insurance is considered to be too rewarding or rigid, and politicians the same. Some say that employers demand too much from their employees and that the National Insurance has to handle problems generated in the workplace.
The task of doctors in sickness absence may be difficult and conflicting. A colleague of mine stated when I planned this study: “Of all in medicine this is the most difficult part.” The study is about the relation between functional consequences of having a health problem and health-related absence from work, i.e. work ability and sickness absence. The main focus is on how work ability reduction is assessed in general practice consultations by the patients certified sick and their doctors, with follow-up of sickness absence duration.
In the context of health-related absence from work it is easy to forget that most people are conscientious about their work and many are working despite of disability. My father-in-law had to amputate his right leg after a shooting accident at the age of 21. That did not keep him from running a physically demanding butcher’s shop for almost 40 years.
Arendal/Oslo, August 2004
Harald Reiso
7 Acknowledgements Many people have helped me in the process of making this study. First my two mentors, professor Gunnar Tellnes at the University of Oslo, the Department of General Practice and Community Medicine, Section of Social Insurance Medicine, and chief medical officer, Sören Brage, at the National Insurance Administration. Thank you for your support, encouragement and wise guidance, and for never giving up the hope that I would reach my goal. I also thank researcher Jan F Nygård at the Cancer Registry of Norway for his patience and help with statistics, and director Pål Gulbrandsen at the Centre for Health Services Research (“HELTEF”) for clear and good advice.
I am grateful to professor Per Fugelli at the University of Oslo for inspiring me to start this research project, after my letter to him in December 1993. He invited me to Oslo and brought me in contact with Gunnar Tellnes.
Other research fellows at the Section of Social Insurance Medicine, at the University of Oslo, need to be thanked. Doctor Inger Sandanger for initiating useful discussions and social activities, the doctors Migle Gamperiene, Bård Natvig, and Hans Magnus Solli, professor Bjørgulf Claussen, and professor Dag Bruusgaard for guidance and co-operation, and secretary Gunn Tønder for practical help.
I appreciate the support of the General Practitioner Research Group in Arendal, especially doctor Mark Fagan for advice and language help, and professor emeritus Christian Borchgrevink for focusing on what really matters. I am grateful for the co-operation with physiotherapists Gudrun Sandbu Jørgensen and Rune Holanger, and doctor Dag Soldal about patients with back disorders in Aust-Agder. I also thank doctor Gudmund Waaler at the Aust-Agder Hospital, and my superiors Sven Arild Hansen and Lisbeth Skranes at the Aust-Agder County Office of the National Insurance Service.
My gratitude is extended to Steinar Mathisen, Hans Thomas Meinich, Thi Hong Trihn Vu and Ola Thune at the National Insurance Administration for assistance in the collection and preparation of sickness absence data, and to professor emeritus Tor Bjerkedal for advice.
Special thanks to doctor John Anderson for language help, to statistician Håkon Gjessing for important comments on Paper IV, and to my family for keeping up with me while not always being present minded and crowding our home with piles of printouts, articles, books, and computers.
The study was given financial support by the Norwegian Ministry of Health and Social Affairs, the National Insurance Administration, and the Aust-Agder County Office of the National Insurance Service. Without the help of the participating doctors (see Appendix) and patients this study could not have been conducted.
8 English summary
Background Disease, reduced function and reduced work ability are key concepts in sickness benefit legislation. There is a growing interest for assessments of function and work ability, because they probably can supplement information about disease in decisions on granting sickness benefits, rehabilitation actions and disability pension, or in predicting duration of sickness absence. Reduced work ability caused by disease is the point of interest in certified sickness absence. How patients certified sick and their doctors assess work ability has been little investigated.
Aims This study deals with the relation between work ability and certified sickness absence. The aims are: - To examine how reduced work ability in patients certified sick is assessed by themselves and by their doctors, and to identify factors associated with the assessments. - To determine the predictive power of assessed work ability and other factors on the duration of sickness absence.
Methods Two sets of data were generated by questionnaires, and merged with benefit registers. Dataset 1 was collected among doctors and their patients certified sick in the county of Aust-Agder, southern Norway in 1996 (Papers I – V). Dataset 2 was collected from 190 patients certified sick and referred to a back disorder outpatient clinic in Aust-Agder in 1997/1998 (Paper VI). Dataset 1 was supplied by 52 (57%) of 91 invited primary health care doctors on 913 patients certified sick. 652 (71%) of these patients completed similar questionnaires. The sample was representative. Patients certified sick assessed their work ability by the question: “To what degree is your ability to perform your ordinary, remunerative work reduced today?” Five answer options ranged from “Very much reduced” to “Hardly reduced at all”. The doctors assessed the patients’ work ability by a similar question. Information about sickness certification diagnoses, medical and demographic factors, and work demands was collected. The duration of sickness absence was followed-up in National Sickness Benefit Registers. A method for estimating predictive accuracy was developed (Paper IV). Logistic and Cox regression analyses were performed.
Results The patients and the doctors concurred on their work ability assessments, mostly at the start of episodes of sickness absence (Paper I, III). They assessed work ability as very much or much
9 reduced in 2/3, and as moderately reduced in 1/3 of the cases (Paper I). The doctors’ assessments were mirrored in full-time/part-time certification. The patients assessed their work ability as more reduced the more stressful or physically strenuous their jobs were (Paper I). Patients with back disorders assessed their work ability as more reduced than other patients certified sick. The doctors assessed the patients’ work ability as more reduced when the doctors’ assessments were based on clinical findings, and as more reduced in depression than in other diagnoses (Paper I). In prolongation of ongoing episodes the doctors assessed fewer female than male patients to have very much or much reduced work ability (Paper II). At the start of episodes of certified sickness absence, the duration of absence was predicted by diagnoses and the doctors’ assessments of who would return to work four weeks ahead (Paper III, IV). In prolongation of ongoing episodes of sickness absence, the duration was predicted by the patients’ own assessments of reduced work ability, and by the doctors’ assessments of who would still be certified sick four weeks ahead (Paper III, IV). Back disorders with radiating pain predicted shorter duration of absence than back disorders without radiating pain (Paper VI). Patients with very much reduced self-assessed work ability had twice the duration, and twice the probability of not returning to work, compared with patients with moderately reduced work ability (Paper III, VI). Diagnoses were most accurate as predictors in the first weeks of absence (Paper IV). The patients’ self-assessed work ability was most accurate as a predictor in the interval between eight and 20 weeks from the start of absence, with a maximum at 13 weeks. The patients’ work demands and the doctors’ work ability assessments were not associated with the duration of sickness absence (Paper III).
Interpretation The patients’ self-assessed work ability is associated with work demands and diagnoses. The doctors’ assessments are associated with clinical findings and diagnoses. At the start of certified sickness absence, diagnoses predict duration of absence. In prolongation of ongoing episodes of absence, the patients’ self-assessed work ability predict duration of sickness absence. The doctors’ assessments of certified sickness absence status four weeks ahead predict duration at the start of, and in prolongation of episodes. This information can be used to differentiate between patients certified sick with high probability of short, intermediate and long-term sickness absence.
10 Norsk sammendrag
Bakgrunn
Sykdom, redusert funksjonsevne og redusert arbeidsevne er nøkkelbegrep i lovgivning om sykefravær. Det er økt interesse for funksjons- og arbeidsevnevurderinger, siden de antagelig kan supplere opplysninger om sykdom i avgjørelser om sykepenger, rehabiliteringstiltak og uførepensjon, eller predikere varighet av sykefravær. Ved sykmelding er det redusert arbeidsevne som følge av sykdom som er interessant. Hvordan sykmeldte og legene deres vurderer arbeidsevne er lite undersøkt.
Formål Dette doktorgradsarbeidet ser på sammenhenger mellom redusert arbeidsevne og sykefravær. Formålet er: - Å undersøke hvordan nedsatt arbeidsevne vurderes av de sykmeldte selv og legene deres, og å påvise faktorer som har sammenheng med vurderingene. - Å bestemme i hvilken grad vurderinger av arbeidsevne og andre faktorer kan predikere varighet av sykefravær.
Metode To datasett ble generert ved bruk av spørreskjema. Dataene ble koblet mot trygderegistre. Datasett 1 ble innsamlet blant alle sykmeldte i Aust-Agder i 1996 (Artiklene I - V), datasett 2 fra 190 sykmeldte henvist til en ryggpoliklinikk i Aust-Agder i 1997/1998 (Artikkel VI). Datasett 1 kom fra 52 (57%) av 91 inviterte leger i primærhelsetejenesten, angående 913 sykmeldte. 652 (71%) av de sykmeldte fylte ut et lignende spørreskjema. Utvalget var representativt. De sykmeldte vurderte sin nedsatte arbeidsevne ved følgende spørsmål: “I hvilken grad er din evne til å utføre ditt vanlige arbeid nedsatt i dag?” Det var fem svaralternativ, fra ”Svært mye nedsatt” til ”Ubetydelig nedsatt”. Legene vurderte de sykmeldtes arbeidsevne ved et tilsvarende spørsmål. Informasjon om sykmeldingsdiagnoser, medisinske og demografiske faktorer, og jobbkrav ble innsamlet. Varigheten av sykefraværet ble fulgt i nasjonale registre. En metode for å beregne prediktiv riktighet (predictive accuracy) ble utviklet (Artikkel IV). Logistiske og Cox regresjonsanalyser ble utført.
Resultat De sykmeldte og legene var i stor grad enige i sine vurderinger av arbeidsevne, mest ved starten av sykmeldinger (Artikkel I, III). Arbeidsevne ble vurdert som svært mye eller mye nedsatt i 2/3, og som middels nedsatt i 1/3 av tilfellene (Artikkel I). Legenes vurderinger ble gjenspeilet i sykmeldingsprosenten. De sykmeldte vurderte sin arbeidsevne som mer redusert jo mer stressende eller fysisk anstrengende deres jobber var (Artikkel I). Sykmeldte med
11 rygglidelser vurderte sin arbeidsevne som mer redusert enn andre sykmeldte. Legene vurderte de sykmeldtes arbeidsevne som mer redusert jo mer legenes vurderinger var basert på kliniske funn, og som mer redusert ved depresjon enn ved andre diagnoser (Artikkel I). Ved forlengelse av sykmeldinger vurderte legene færre kvinnelige enn mannlige sykmeldte til å ha svært mye eller mye nedsatt arbeidsevne (Artikkel II). Ved starten av sykmeldinger ble varigheten av sykefraværet predikert av diagnoser og legenes vurderinger av hvilke sykmeldte som trolig ville være helt friskmeldt om fire uker (Artikkel III, V). Ved forlengelse av sykmeldinger ble varigheten predikert av de sykmeldtes egne vurderinger av arbeidsevne, og av legenes vurderinger av hvilke sykmeldte som trolig fortsatt ville være sykmeldt om fire uker (Artikkel III, V). Rygglidelser med utstrålende smerter predikerte kortere fraværsvarighet enn rygglidelser uten utstrålende smerter (Artikkel VI). Sykmeldte med svært mye nedsatt egenvurdert arbeidsevne hadde dobbelt så langt fravær, og dobbel sannsynlighet for ikke å komme tilbake i jobb, sammenlignet med sykmeldte med middels nedsatt egenvurdert arbeidsevne (Artikkel III, VI). Diagnoser hadde høyest prediktiv riktighet i de første ukene av sykmeldingene (Artikkel IV). De sykmeldtes egenvurderte arbeidsevne hadde høyest prediktiv riktighet mellom åtte og 20 uker fra starten av sykmeldingene, med maksimum ved 13 uker. De sykmeldtes jobbkrav, og legenes vurderinger av de sykmeldtes arbeidsevne viste ikke sammenheng med sykefraværets varighet (Artikkel III).
Fortolkning De sykmeldtes egenvurderte arbeidsevne har sammenheng med jobbkrav og diagnoser. Legenes vurderinger har sammenheng med kliniske funn og diagnoser. Ved starten av sykmeldinger forutsier diagnoser varighet av sykefravær. Ved forlengelse av sykmeldinger forutsier de sykmeldtes egenvurderte arbeidsevne varighet. Legenes vurderinger av sykmeldingsstatus om fire uker forutsier varighet av sykefravær ved start og ved forlengelse av sykmeldinger.
Denne informasjonen kan brukes til å skille mellom sykmeldte med høy sannsynlighet for å bli kortvarig, middels eller langvarig sykemeldt.
Key words Epidemiology, functional ability, functional assessments, predictive accuracy, sickness absence, sickness certification, work ability.
Annotations Some Norwegian words are given within “ “ and in parenthesis in the text, and some Norwegian text as notes in the Appendix. Italics are used when introducing a concept, or to emphasise a word for other reasons.
12 List of papers
I Reiso H, Nygård JF, Brage S, Gulbrandsen P, Tellnes G.
Work ability assessed by patients and their GPs in new episodes of sickness certification.
Fam Pract 2000; 17: 139–44.
II Brage S, Reiso H.
Arbeidsevne og kjønn – legers vurderinger av sykmeldte.
[Work ability and gender – physicians’ assessment of sick-listed patients]
(In Norwegian, English summary)
Tidsskr Nor Lægeforen 1999; 119: 3737–40.
III Reiso H, Nygård JF, Brage S, Gulbrandsen P, Tellnes G.
Work ability and duration of certified sickness absence.
Scand J Public Health 2001; 29: 218–25.
IV Reiso H, Nygård JF, Gulbrandsen P, Brage S, Tellnes G.
To ask the right question at the right time: When is the patient’s self-assessed work ability most accurate as a predictor of the remaining duration of certified sickness absence?
Norsk Epidemiologi 2003; 13: 297–302.
V Reiso H, Gulbrandsen P, Brage S.
Doctors’ prediction of certified sickness absence.
Fam Pract, 2004; 21: 192–8.
VI Reiso H, Nygård JF, Jørgensen GS, Holanger R, Soldal D, Bruusgaard D.
Back to Work: Predictors of Return to Work Among Patients With Back Disorders Certified As Sick. A Two-Year Follow-up Study.
Spine 2003; 28: 1468–73.
13 General introduction Background
Socio-economic importance of sickness absence There is concern in Norway, as in other welfare states, about the increase in sickness absence over the last decades (67, 89, 97, 111). The number of days of certified sickness absence paid by the Norwegian National Insurance, per employee, was 8.4 in 1980, and 13.5 in 2002 (97). The number of persons with disability pension in Norway, in percent of the population between 16-67 years of age, was 6.1% in 1980, and 10.2% in 2002. These trends are illustrated in Figure 1.
Figure 1 The number of days of certified sickness absence paid by the Norwegian National Insurance per employee (five days per week), and the number of persons with disability pension in percent of inhabitants 16-67 years, Norway 1980 – 2002 (97)
Days %
16.0 12.0
14.0 10.0 12.0 8.0 10.0
8.0 6.0
6.0 4.0 4.0 2.0 2.0
0.0 0.0
2 8 4 0 2 8 4 0 980 98 984 986 98 990 992 99 996 998 00 002 80 8 84 86 8 90 92 9 96 98 0 02 1 1 1 1 1 1 1 1 1 1 2 2 19 19 19 19 19 19 19 19 19 19 20 20
In 2002, the State budget in Norway was approximately NOK 580 billion (73 billion €), whereof 191 billion NOK (33%) was accounted for by social insurance benefits (24, 97). Disability benefits accounted for 39 billion (20% of 191), certified sickness absence for 26 (14%), and medical and vocational rehabilitation benefits for 11 billion (6%) (97). Similar proportions of costs are found in other European countries (31, 89, 111). Norwegian employers have an estimated cost due to certified sickness absence of about 13 billion NOK (11). Health-related absence from work is thus a field of research interest from the perspectives of health, economy, law, and politics (107).
The roles of doctors in sickness absence Doctors are gatekeepers for sickness benefits, in accordance with sickness benefit legislation (13, 14, 70, 95, 107, 119). The Sickness Benefit Scheme Act in Norway states: “Cash benefit
14 compensation can be given to a person incapacitated for work due to a reduced function that clearly is caused by disease or injury. Incapacitation due to social or economic problems does not give cash benefit rights” (30, §8-4).1 Disease and reduced function are also used as key concepts in sickness benefit legislation in other European countries (14, 31, 95, 107). Doctors’ roles as assessors of reduced function in sickness certification consultations include assessing the medical problems of the patients, and the consequences these problems may have for the patients’ work ability, possible sickness absence, and duration of sickness absence. Doctors have to interpret health problems in the context of social insurance legislation, and may choose to assume the role of patients’ advocate for insurance benefit rights, or the role of gatekeeper, trying to restrict the use of benefits. The roles may conflict (22, 119).
Patients have certain expectations when they decide to consult their doctors. Discussions between patients and doctors about sickness certification take place in the “closed room” at consultations. The doctors most often meet their patients’ initial demands for sickness certification. 95% of patients receive certifications by their doctors when the initiative is taken by the patient (64). Even in cases where general practitioners do not recommend sickness certification, certificates are issued in 87% of the cases (26). Thus, the strength of the patients’ demands influences the decisions of doctors (2). The doctor may be “captured” by the patients’ wishes. However, some patients do not want sickness certifications despite of need, or do not even contact their doctors for that purpose.
How doctors perform clinical diagnostic work include elements such as “pattern recognition”, and global assessments (100, 137). This probably parallels how doctors carry out work ability assessments. Such assessments are presumably also global and holistic, based on the doctors’ clinical experience, practical judgement, knowledge about diseases, treatment, prognosis and the patients’ work, the doctors’ intuition, the doctors’ personal knowledge of the patients, and other factors concerning the patients’ life situations (38).
There is a growing interest for functional assessments of persons with health problems (52, 111). In Norway, the government and the labour and employers unions agreed to increase their efforts to reduce sickness absence, include persons with health problems in the work force, and increase the retirement age for pension, by an agreement of October 2001 (60). Assessments of functional ability in social insurance medicine may be used in decisions on granting sickness benefits, rehabilitation actions and disability pension, or in predicting duration of sickness absence.
Norwegian sickness benefits In Norway, employees can certify themselves sick 12 days per year, but usually not for more than three consecutive days without a sickness certificate issued by a doctor. These regulations have been more flexible since the agreement of October 2001. Employers pay for the first 16 calendar days of certified sickness absence (14 days was the rule until April
15 1998), and the National Sickness Benefit Scheme covers further absence, to a total of maximum 365 calendar days. Sickness benefits can be graded between 100% and 20%. The benefits equal ordinary salary, up to a maximum of NOK 341 166 per year (about € 41 400) (per 01.05.2003). Persons can be granted medical or vocational rehabilitation benefits, or disability pension (within or after the sickness benefit period of maximum one year), depending upon the functional consequences and duration of the health problems (30, chapters 10-12).
Former research In Scandinavia, research about health-related absence from work has mostly dealt with disability pension, but over the last decades also certified sickness absence (4, 18, 28, 34, 43, 46, 49, 59, 69, 91, 102, 109, 117). Much research has examined sociological aspects of disability, looking at work environment and work demands (61, 125). Less has been conducted on reduced function due to health problems. Few studies have looked at the doctors’ issuing of sickness certificates, the doctor-patient relationship in sickness absence, and how doctors assess patients’ work ability, although the practice of doctors is central in sickness absence processes (27, 28, 64, 115, 116).
Musculoskeletal and mental disorders are the diagnostic groups responsible for most certified sickness absence (4, 18, 49, 117). Back disorders and pregnancy related disorders are important diagnostic categories (59, 109, 123, 131).
Lindén’s investigations of persons certified sick in Bergen in 1958 showed that individuals with repeated episodes of sickness absence were responsible for the largest part of the total absence, and that absence concerned a small fraction of employees (69). He found no conclusive association between physical working capacity, expressed as maximal oxygen uptake, and sickness absence.
Age and diagnoses predict duration of absence at the start of episodes of certified sickness absence (112). Tellnes et al have proposed that sickness absence could be used as a health status indicator (113). A recently published study showed that persons with more than five medically certified episodes of sickness absence of more than seven days of duration have a mortality rate 4.8 times greater than those with no such absence, in a 10 year follow-up (62).
Westin and Waddell have contributed with explanatory models of disability (126, 129, 130). Westin has made four models of how disability may develop in relation to a person’s job, based on the person’s level of health, and his/her job requirements (129). In his Type I model of disability, a person’s level of health is lowered relative to the job requirements. In the Type II model, the job requirements are increased relative to the person’s level of health. In the Type III model there is a combination of reduced level of health and increased job requirements, whereas in the Type IV model, disability results due to increased job requirements in spite of improvements in health. Waddell has advocated that disability must be understood and managed in a bio-psycho-social model, because medical, economic or
16 cultural models of disability have shortcomings in explaining the complexity of different factors involved in the concept (126).
Assessments of work ability Some studies have dealt with certified sickness absence and self-assessed work ability (13, 43). Haldorsen et al looked at self-assessed work ability in patients with low back disorders (42). They used five items to make a Graded Reduced Work Ability Scale, constructed for the Norwegian Ministry of Health and Social Affairs (42, 92). The five-point scale elements are about reduced ability to carry out ordinary work, reduced ability to do other work, number of activities and duties that are affected by complaints, the consequences of complaints for well- being and health, and the effect of complaints on continuing working (42). Of the five items, the question about reduced ability to carry out the person’s ordinary work was most predictive of not returning to work.
Ilmarinen et al from Finland have constructed the Work Ability Index which comprises seven items: the person’s subjective assessment of current work ability compared with their lifetime best, work ability in relation to work demands, number of current diseases diagnosed by a doctor, estimated work impairment due to diseases, sick leave during the past year, own prognosis of work ability two years from now, and mental resources (55-57, 120). Among people attending work, it is a strong predictor of future work ability, disability, and mortality (55, 56). Among patients with low back disorders certified sick, work ability perceived as good by the Index question of current work ability versus lifetime best, predicted work resumption (13).
A new procedure for judging medical incapacity was introduced in England in 1995 (9, 58, 110). The procedure includes an “Own Occupation Test” in the Statutory Sick Pay time period, where the work ability of a person is judged by a General Practitioner for up to 6 months, then a Personal Capability Assessment (PCA) (formerly called All Work Test) is performed (25). Trained approved medical services doctors conduct PCA, and consider a person’s work ability against all types of work. The person’s own opinion about function is included in the assessment. The test comprises a physical part that assesses the person’s ability to walk, stand, sit, walk up and down stairs, rise from sitting, bend and kneel, manual dexterity, lift and carry, reach, speak, hear, see, be continent and remain conscious. Other parts of the test assess mental capacity, judge the ability to complete tasks, need of support in daily living, ability to cope with pressure, and ability to interact with other people. Based on these evaluations, a sum score is made. A score of 15 or more is the threshold value for further disablement benefit. In some medical conditions (e.g. malignancies with poor prognosis) PCA is not necessary. The PCA is considered to be a valid and reliable measure of work ability (110). However, few scientific studies using the PCA have been conducted (25).
17 The setting of this study A number of factors may influence sickness absence: the general health in the population, medical, biological, genetic and psychological factors, the compensation level of benefits, legislation, employment rates, economical factors, age, doctors’ and patients’ attitudes to sickness absence, work demands, and more. The macro trends of sickness absence by time are described above. This study focuses on a “micro” situation: how work ability reduction is assessed in consultations by the patients certified sick and their doctors, with follow-up of sickness absence duration. Because one of five consultations in Norwegian general practice involves certified sickness absence, and general practitioners issue more than 80% of all certificates, this study was mainly conducted in general practice, with the exception of Paper VI (38, 64, 99, 113). The setting was patient centred and individually oriented.
More than 800 measures are used for functional assessment of adults in terms of work disability, whereof three have been presented above (9, 42, 58, 79, 92, 110, 120). Most of these consist of several items, are time consuming, and therefore unsuitable for everyday clinical practice. Elements of function and work ability may be measured directly, such as tests of muscle strength, maximum oxygen uptake (e.g. the Åstrand bicycle test), and other physiological tests (138), whereas indicators of function may be self-assessed function (33, 120, 127), or assessments made by observers (110). Measures of such indicators may parallel measures of quality of life (77). To our knowledge there are no tools designed for assessment of work ability in early periods of certified sickness absence in general practice. Given the complexity of possible factors influencing sickness absence, we wanted to use a simple, not time-consuming instrument for assessing function. Questionnaires needed to be quick and easy to complete. Simple questionnaires presumably increase participation and compliance. Global assessment questions were therefore used, instead of several questions constituting indexes. A minimalistic approach, with questionnaires limited to one double-paged A4 format sheet of paper, was chosen.
This study is about assessments of reduced work ability performed after decisions to issue sickness certificates were taken. Data about the processes ahead of that, or information on how persons not certified sick assessed their work ability, have not been collected. The focus is on reduced function, not on a persons’ resources or possibilities despite of health problems.
Because the concepts of function, work ability, and sickness absence are central in this study, they will be presented more detailed in the following.
18 Function
Health problems and reduced function Every society must cope with the consequences of health problems on peoples’ function and work ability. In the development of the industrialised society, different sickness insurance schemes emerged (5). One of the oldest was a disease fund started in 1653 by employees at a printing-house in Antwerp (98). The first social insurance regulations that were formed in Europe, with Germany and the Bismarck government as model in the late 19th century (12), concerned the consequences of accidents and injuries among industrial workers. Later regulations included health-related absence from work, with disease as the central concept (21, 80). The term Arbeitsunfähigkeit (inability to work, “arbeidsuførhet”) was introduced (53).
Disease is used as an umbrella term for disease, injury, and defect (“lyte”) in The Norwegian Social Insurance Scheme. A committee of 1907 (“Sykeforsikringskomiteen av 1907”) defined disease as: “By this law, disease is such a disturbance of the physical and mental health status of a person, that attendance to a doctor is considered necessary”.2 Patients present complaints and symptoms, their subjective feelings – i.e. their illness – to doctors. Doctors use that information in their medical examinations and assessments of the health problems. One social consequence of illness and disease is sickness absence (72, 114). Others have discussed the relations between illness, disease and sickness (4, 18, 87).
The concept of health includes three main components: physical, mental and social (134). They can each have an impact on the function of a person, and are interrelated. Chronic musculoskeletal pain can induce fatigue and depression, and thereby reduced social participation. Depression is associated with pain disorders (23). Treatment of cancer with surgery, cytotoxic drugs or radiation, has physical and mental side effects that may have social implications.
Concepts of reduced function Function is used as an overall term for all activities of human life (133). In physiology it is usually restricted to function at an organ level, e.g. respiratory, muscle, and nerve function.
The first concept used to describe the functional consequences of accidents and injuries was invalidity (47, 63, 90). “Classic” tabular invalidity assesses the physical consequences of an injury, e.g. the loss of a finger, without taking personal or relational aspects into account (47, 63). Work invalidity relates function to the person’s work (47). The concepts of invalidity are now mostly used in insurance medicine for assessing consequences of injuries (medical/tabular invalidity).
Functional ability may reflect what a person can do (performance) in relation to specific demands, or reflect possible function - irrespective of demands (capacity) (25). Capacity is mostly considered as the maximum ability or potential of a person (78). A person’s general
19 function is different from his/her work ability in a remunerative job. The latter has to fulfil certain work demands and work-related social skills. People may have the ability to carry out activities at home or in their leisure time, but may not be able to perform remunerative jobs.
Classification of reduced function The WHO classification system of function has developed from ICIDH – International Classification of Impairments, Disability, and Handicaps (WHO 1980) (132), to the concepts used in ICF – International Classification of Functioning, Disability, and Health (WHO 2002) (133), where focus is on the more positive terms ability, activity, and participation. Disability is used as an umbrella term for impairments, activity limitations, and participation restrictions.
The ICF conceives a person’s function as a dynamic interaction between health conditions and contextual/environmental factors (133). The ICF uses the concepts activity and participation, where activity is the execution of a task or action by an individual, and participation is involvement in a life situation. However, in a work context that distinction is not clear.
Concepts of health-related absence from work Many terms are used to describe absence from work. An overview is given by Tellnes (114). Absence from work caused by a health problem is most frequently labelled sickness absence. Sick leave, sick absence, sickness absenteeism, and medical absence, are synonymous terms. Sickness certification is a declaration issued by a doctor to a person entitled to sickness benefits (114). Other synonymous terms are sick absence certification, certified ill-health, sick leave certification, certification of incapacity or unfitness for work, and sick-listing (28, 114). Sickness certification is about the process of issuing a certificate – to certify sick (“sykmelde”), but also the period of which one is absent from work – an episode of certified sickness absence (“sykmelding”). However, sickness certification and certified sickness absence are used interchangeably to some extent.
Initial certificate is the first document issued by a doctor to declare a persons’ incapacity for work (114). Continuation certificates are issued for prolonging ongoing episodes of certified sickness absence. In Norway, when patients have been certified sick more than eight weeks, the absence is classified as long-term.
Health-related absence from work other than certified sickness absence, such as medical and vocational rehabilitation benefits, and disability pension, is usually not included in the concept of sickness absence. In this study, Papers I – V is about certified sickness absence, whereas Paper VI also includes medical and vocational rehabilitation benefits, and disability pension.
In certified sickness absence, work ability is associated with the job the person is certified sick from. The person is incapacitated for work (“arbeidsufør”). In vocational rehabilitation,
20 work ability is relational to any job, and the term occupational disablement, or disability, may be used (“yrkesufør”/”ervervsufør”). Disability pension is used to describe a persons’ permanent incapacity for work.
In the Norwegian social insurance act, different terms are used to describe health-related absence from work. Certified sickness absence use reduced function (”funksjonsnedsettelse”), medical rehabilitation benefits use work ability (“arbeidsevne”) and functional ability (“funksjonsevne”), vocational rehabilitation benefits use income ability (“inntektsevne”), and disability pension use income ability and work ability (30, chapters 8, 10-12).
Work ability Work ability and work incapacity are contradictory concepts. In accordance with the ICF terminology, the positive term reduced work ability is used here.
According to Nordenfelt, the concept of ability has to be specified (86). First, one has to identify a particular agent (A). Second, one has to specify A’s project or goal: something that A is able to attain, e.g. A is able to perform this action (here remunerative work). Third, one has to specify the circumstances in which A is able to attain this goal, or perform this action, i.e. context (here working life). Thus work ability can be assessed in general, and with respect to the specific job or occupation of the person. Work ability depends on the tasks the individual performs. Nordenfelt notes that all abilities, however simple and basic, are related to a set of circumstances.
According to Ilmarinen (56), factors that affect work ability form a complex relationship between health, competence, values, work environment, and social relations. “Work ability refers to individual and occupational factors that are essential to a person’s ability to cope in worklife. Work ability is the result of the interaction between individual resources and work.” “A person realizes his or her resources at work, … .” A model of how these components influence work ability, from the individual’s point of view, is shown in Figure 2 (56, Juhani Ilmarinen - personal communication 28.01.2003).
21 Figure 2 Components influencing work ability from the individual’s point of view (56, Juhani Ilmarinen - personal communication 28.01.2003. Used by permission of 25.09.2003)
Society Relatives, Family friends WorkWork Abilityability Work Environment Content & demands Community & organisation Management & leadership Values Attitudes Motivation Competence Knowledge Skills Health Functional capacities
Ilmarinen states: “A system of feedback also exists between work ability and its components. A person’s resources receive feedback on how he or she manages at work.” “Work ability is ... a dynamic process that changes through its components throughout life.”
Elements of work ability may be constitutional and genetic - such as one’s physical talents and personality traits, or acquired - such as knowledge, education, training, culture and social factors. People must be able to structure their lives, to get up and go to work, and have energy, endurance and flexibility to cope with their work demands. Work ability is dependent on whether the work is carried out alone, or together with e.g. colleagues, pupils, customers, or patients, and on how things are at home and in the family. What are peoples’ attitudes, goals and motivations for working?
22 Work ability and sickness absence
Models Whether sickness absence will be a consequence of reduced work ability or not, mainly relates to two factors: the consequences health problems have for the person’s work ability, and his/her work demands (61, 130). Given a certain reduced function, a person probably will assess work ability as more reduced in jobs with high work demands, than in jobs with lower demands. Letter A in Figure 3 illustrates this simplified association in certified sickness absence. Work ability is here used as “overall work ability”, or “work ability at large”, not as work ability in an individual context.
Absence may also occur when work demands are increased relative to work ability - letter B in Figure 3. However, situation B is not necessarily associated with a health problem (sickness).
Figure 3 Two simplified models of the associations between work ability and work demands where absence may occur. A – Work ability is lowered relative to work demands. B – Work demands are increased relative to work ability (adapted from Westin, 129)
Level
A Work demands
Work ability
Work demands
Work ability B
Time
The present study focus on work ability reduction associated with health problems (situation A), not on reduced work ability caused by age.
The relations between work ability, work demands and time in an episode of sickness absence may be illustrated as follows (Figure 4).
23 Figure 4 A model of the relations between work ability, work demands and time in an episode of sickness absence
Level
Work ability
Work demands
Time Episode of absence
The health problem is associated with the degree of work ability reduction, and the speed of change – i.e. how fast work ability decreases or increases. Influenza usually reduces work ability substantially and rapidly, whereas in depression or back disorders work ability may be reduced more gradually and to a lesser extent. The natural course and treatment of diseases will influence the speed of work ability changes. Antibiotics may give rapid improvement of work ability in infectious diseases. Depression can be successfully treated over weeks or months, whereas treatment of back disorders may not increase work ability in the short run.
As depicted in situation B in Figure 3, the level of work demands may have an impact on whether absence will occur or not. Despite the increase of sickness absence in recent years, the general health in the population has probably been stable, or has improved (18, 67, 105, 111, 128). Physical work demands have decreased in Norway and Sweden the last decades (36, 89, 111). There is reason to believe that psychological work demands have increased (89, 111, 121). Little is known about how the introduction of computers at work has influenced people’s work ability. The demands for employees’ communication and co-operation skills, and demands for pace, have probably increased. Increased work demands may thus have contributed to the increase of sickness absence (3, 89, 111, 122).
24 Aims This study deals with the relation between work ability and sickness absence. The aims were:
1. To examine how reduced work ability in patients certified sick was assessed by themselves and by their doctors, and to identify factors associated with the assessments.
2. To determine the predictive power of assessed work ability and other factors on the duration of sickness absence.
Material and Methods Definitions Definitions of concepts used in this thesis:
- Work ability is defined as a person’s physical and psychological capacity to perform his/her ordinary, remunerative work.
- Accuracy is defined as the degree to which a measurement, or an estimate based on measurements, represents the true value of the attribute that is being measured (65).
- Predictive accuracy is defined as the difference in percent between model predictions of duration (expected remaining duration) and observed remaining duration of episodes of certified sickness absence (Paper IV).
- Improved predictive accuracy is defined as the improvement of predictive accuracy for the variable of interest (estimated in percent), compared with a variable that had no association with the outcome (a constructed comparison variable) (Paper IV).
Variables Questionnaire I completed by patients and doctors (Papers I – V), and Questionnaire II completed by patients (Paper VI), are presented in the Appendix. The variables are summarised as follows.
Work ability today, Questionnaire I The work ability questions were based on the experiences of others (42, 92). The five ordinal scale answer options were: Very much reduced, Much reduced, Moderately reduced, Not much reduced, and Hardly reduced at all. Very much reduced was used as the first option because all subjects were certified sick.
In the patient questionnaire the question was: To what degree is your ability to perform your ordinary work reduced? (Tick off for the reduction caused by the disorder that you are certified sick because of.)
25 In the doctor questionnaire it was: To what degree is the patient’s ability to perform his/her ordinary work reduced? Ordinary work is the work the patient had at the time of certification.
Work ability today, Questionnaire II (Patient): To what degree does your back disorder reduce your ability to perform your ordinary work today? (Ordinary work is the job from which you are now certified sick)
Medical factors The doctors classified and coded the main sick-listing diagnoses according to ICPC (International Classification of Primary Care) (15), recorded the kind of clinical information sources they used when assessing work ability, and their use of clinical examinations and tests in the consultations.
Work related factors The patients gave information about their work demands, operationalized as number of remunerative working hours per week, daytime or night/shift work, a subjective description of physical work demands, heavy lifting, self-determined breaks, conflicts, and stressful working conditions (95, 104). They also supplied data about satisfaction and influence in their jobs (8).
The patients gave information about their occupations in free text, recoded into groups according to Nordic Classification of Occupations (7).
Certified sickness absence status four weeks ahead Patients and doctors assessed the patients’ certified sickness absence status four weeks ahead by the three options: Probably full-time certification, Probably part-time certification, and Probably returned to work.
Function Function was measured by the six COOP/WONCA charts self-assessed physical fitness, feelings, daily activities, social activities, overall health, and change in health (Questionnaire II, Paper VI) (10, 136). Because of limited space on Questionnaire II, charts without pictures were used.
Other Patients and doctors also evaluated if non-medical factors could have influenced their work ability assessments.
Sickness absence data National registers of certified sickness absence, medical and vocational rehabilitation benefits, and disability pension supplied data about time and duration of sickness absence (16).
26 Collection of data
Papers I - V In the process of constructing Questionnaire I, a pilot study among 16 general practitioners was conducted in 1994 (94). The main study was conducted in January–April 1996, in the county of Aust-Agder, southern Norway, as described in Paper III.
All general practitioners (n=85) and company doctors (n=6) in Aust-Agder County were invited to participate in the project in December 1995. 53 doctors agreed to participate. One did not complete any questionnaires. 52 (57%) doctors actually participated (49 general practitioners and three company doctors). The study had a two-stage design (68). The inclusion of doctors represented a first stage of data collection, and the consultations between doctors and patients represented a second stage. Excluded were certificates issued as documentation for children being sick, to persons having a health problem while on rehabilitation benefits, or certificates only stating return to work.
Early in the sampling process it was noted that some patients did not complete the two questions at the bottom of the instruction side of Questionnaire I. The doctors were informed about this immediately.
Paper VI The inclusion of patients with back disorder certified sick took place at the back disorder outpatient clinic at the Aust-Agder County Hospital September 1997 – December 1998, as presented in Paper VI. Patients were referred to the clinic by their primary health care doctors.
Overview of the sampling An overview of the sampling for Papers I – VI is illustrated in Figure 5. The patients were certified sick by their primary health care doctors.
The main focus of interest for prediction of absence duration was in the early stages of certified sickness absence. Therefore, in the predictive analyses using Questionnaire I, episodes of sickness absence with duration of more than 20 weeks from the start of absence to the time of the response to the questionnaires, were not included.
27 Figure 5 An overview of the sampling for Papers I – VI. GPs = General practitioners. QD = Questionnaires from doctors. QP = Questionnaires from patients. D = Doctors. P = Patients
General practice in Aust-Agder General practice/Aust-Agder back disorder outpatient clinic 91 invited doctors, 52 participated (57%) 220 QP concerning 220 patients (49 GPs, 3 company doctors) 1 051 QD, 724 QP
138 QD exclusions (72 QP) • 57 double recordings 30 exclusions • 30 uncertain certification dates • 17 other diagnoses • 28 other benefits • 10 other benefits • 10 uncertain identities • 3 other B • 13 other A
Primary sample 913 QD concerning 913 patients, 652 QP (71%)
505 C 474 D 364 E 117 F
Work ability Work ability Work ability Prediction Work ability assessed by D & P assessed by D assessed by D & P assessed by D assessed by P New certification Full-time work 52 doctors 52 doctors 190 QP 49 GPs 52 doctors 549 QD & QP 796 QD 408 QD, 268 QP 439 QD & QP
Paper I Paper II Paper III Paper IV Paper V Paper VI
A 13 other exclusions: 5 cases where other consecutive episodes of sickness absence were recorded, 3 cases only stating return to work, 2 cases where all information on the questionnaires were missing, 1 documentation for child being sick, 1 case where the person certified sick did not want a certification, and 1 case with a combination of certified sickness absence and rehabilitation benefits.
B 3 other exclusions: 2 old age pensioned, 1 caused by injury.
C 505 cases had lasted > 7 days at the time of questionnaire completion.
D 474 additional exclusions: 261 no QP, 210 had < 30 working hours per week, 3 no assessment of work ability by doctors.
E 364 additional exclusions: 261 no QP, 85 had lasted > 20 weeks at the time of questionnaire completion, 11 changed to other benefits, 4 no work ability assessed by doctors and by patient in 1 case, and 2 reached maximum benefit time < 156 days after the time of questionnaire completion.
F 117 additional exclusions: 4 no assessment of sickness certification status four weeks ahead by doctors, 113 had lasted > 20 weeks at the time of questionnaire completion.
28 Temporal aspects of sickness absence data
An episode of certified sickness absence, from the date of its onset (Tstart) to the date it ends
(Tstop) is illustrated in Figure 6. An episode means the whole period of certified sickness absence, not just the time interval between one patient-doctor consultation and the next. Tstop can represent return to work, reaching maximum benefit time, or transition to other benefits. In this study, the date of completion of the questionnaires, i.e. when the consultations occurred (the time of inclusion), was used as time zero (T0). The duration from Tstart to T0 determines whether episodes may be classified as new, or prolongations of ongoing episodes of sickness absence. The remaining duration of a sickness absence episode (Dpost) is from T0 to Tstop.
Figure 6 Temporal aspects for an episode of certified sickness absence. See text above for explanation
Tstart T0 Tstop
Dpost
Onset of Remaining End of Consultation absence absence absence
Classification of episodes based on duration from Tstart to T0
The episodes were classified based on the duration from Tstart to T0 , in the different Papers, as illustrated in Figure 7. As an overall classification, the start of episodes of certified sickness absence includes episodes with a duration from Tstart to T0 of until two weeks. Prolongation of ongoing episodes of sickness absence includes episodes with a duration from Tstart to T0 of more than two weeks.
29 Figure 7 The classification of episodes based on the duration from Tstart to T0
(Overall) (Start) (Prolongation)
Tstart T0 1 week Paper I New (50 weeks) Paper II A New Prolongation
5 weeks 20 weeks Paper III New One month Three month
2 weeks Paper IV 1 week 2 week 3 week ••• 20 week
Paper V Short-standing Long- standing (38 weeks) Paper VI B
A The classification of episodes was based on the doctors’ assessments, because all data about time and duration were not available at the time of the analyses. 188 of 211 new episodes had a duration from Tstart to T0 of
< 8 days, and 222 of 227 prolongation of ongoing episodes had a duration of 8 days or more at T0 (maximum 354 days).
B No classification of episodes based on the duration from Tstart to T0 was used.
Follow-up/former sickness absence
The follow-up of the duration of sickness absence was four weeks in Paper V, until one year in Paper III and IV, and until two years in Paper VI.
The duration of all sickness absence one year prior to Tstart was collected for the patients presented in Paper VI.
30 Missing data
Questionnaire I, patients Question 2 about the patients’ occupation was missing on 141 (21.6%) of the 652 questionnaires in the primary sample, due to the lack of completion of the question on the “instruction” page of the questionnaire.
Question 4 about mean remunerative working hours per week had 57 (8.7%) missing answers. Question 5 about physical work demands had an option for “other”. The question had six (0.9%) missing, and 41 cases where “other” was used. Question 8a and 8b, about conflicts between administration and workers, and among workers, had 15 (2.3%) and 68 (10.4%) missing respectively. Question 9 about the patients’ assessments of stress on a visual analogue scale had 17 (2.6%) missing. Question 11 about the patients’ work ability assessments had missing information in one (0.2%) case.
Questionnaire I, doctors Of the 913 questionnaires completed by doctors in the primary sample, seven (0.8%) had missing information about the doctors’ work ability assessments (question 9). Question 15 about the patient’s sickness absence status four weeks ahead had four (0.4%) missing answers.
Questionnaire II Question 9 about the patients’ work ability assessments had missing information in two (1.1%) cases. Question 10 about the patients’ assessments of pain quality (three answer options) had missing information in two (1.1%) cases. Question 11 about assessments of pain intensity on a visual analogue scale had 31 (16.3%) missing. Question 14 about sickness absence status four weeks ahead had seven (3.7%) missing answers. The COOP/WONCA charts (questions 15 – 20) had missing information between 13 (6.8%) (physical fitness) and 17 (8.9%) (social activities) of the cases.
Statistical analyses The patients’ assessments were compared with the doctors’ using percent of agreement (Paper I).
Associations between categorical variables were tested by Chi-square tests (Paper I, II, III), for ordinal variables according to Mantel-Haenszel (Paper II, III), and for continuous variables by t-tests (Paper I, III) (6).
Intercorrelations between variables were checked by Spearman’s rank correlation coefficients (Paper I, III, VI).
Design-based logistic regression analyses were performed (Paper I, II, V). The design-based
31 analyses used the doctors as the primary sampling units. This adjusted for the fact that the assessments of different patients made by a single doctor were not independent.
Cox regression analyses were performed (Paper III, VI).
Positive predictive values (PPVs) for the doctors’ assessments of certified sickness absence status four weeks ahead were calculated (Paper V).
A method for estimating predictive accuracy is presented in Paper IV.
95 percent confidence intervals were estimated. The chosen level of significance was five percent.
Analyses were performed using SPSS£ and STATA£ software.
Full-time/part-time certification
The degree of certification may be looked upon as a consequence of the work ability assessments. The degree of certification was therefore not used as an explanatory variable in analyses of what factors were associated with the work ability assessments. Full-time/part-time certification at inclusion was used to stratify an analysis on work ability and gender in Paper II, and as an explanatory variable concerning return to work in the analyses in Paper VI.
Recoding of diagnoses Injuries are coded in different chapters of ICPC (15). However, injuries are important in sickness absence (117). These ICPC diagnoses were therefore grouped as injuries: eye injuries (ICPC codes F75, F76, F79), fractures, strains and other musculoskeletal injuries (L72 - L81, L96), head injuries (N79 - N81), foreign objects in the respiratory tract (R87), skin injuries (S14, S16, S18), and injuries not coded elsewhere (A80, A81).
Ethical aspects The study using Questionnaire I was approved by the Regional Ethics Committee for Medical Research, the Norwegian Data Inspectorate, and the Legal Affairs Division in the National Insurance Administration. The same recommendations were applied for the study using Questionnaire II.
32 Main results Assessments of reduced work ability Patients and doctors concurred on their assessments (+ 1 answer category) in 81% of the cases in new episodes of certified sickness absence, less so in prolongation of ongoing episodes (Paper I, Paper III). They assessed work ability as very much or much reduced in 2/3, and as moderately reduced in 1/3 of the cases (Paper I).
Full-time/part-time certification Of the 913 patients in the primary sample, 15% (141/910) received part-time certification at the time of questionnaire completion (missing information about degree of certification in three cases). The doctors assessed 77% (107/139) of the patients with part-time sickness absence to have moderately reduced work ability, and 4% to have very much reduced work ability (p < 0.05). Corresponding figures for the patients’ work ability assessments were 50% (55/111), and 15% (p < 0.05).
Diagnoses, medical factors and work demands In new episodes of certified sickness absence, patients assessed their work ability as significantly more reduced in back disorders (n = 38) than in other disorders, and as more reduced the more stressful or physically strenuous their jobs were (Paper I). Doctors assessed work ability as significantly more reduced in patients with depression (n = 27) than in other disorders. The main information sources of the doctors for assessing work ability were statements by the patients in 66% and clinical findings in 34% of the cases. The doctors assessed work ability as more reduced when their assessments were based on clinical findings, and less reduced when non-medical factors had influenced their assessments.
Gender The doctors assessed 60% of female patients, versus 71% of male, to have very much or much reduced work ability (Paper II). The difference was significant in prolongation of ongoing certification only. Women received part-time certification more often than men, 27% versus 11%. Male doctors assessed work ability as less reduced in female than in male patients in univariate analyses, not so in a multiple analysis.
33 Prediction of sickness absence duration Diagnoses predicted remaining duration in new episodes of certified sickness absence, not in prolongation of ongoing episodes of sickness absence (Paper III). Musculoskeletal and mental disorders were associated with longer duration, and respiratory disorders with shorter duration, in new episodes. Diagnoses were most accurate as predictors in the first weeks of absence (Paper IV).
In new episodes, the presence of non-medical factors as judged by the patients was associated with longer duration than the absence of such factors. Non-medical factors mostly were about work related problems and family problems (Paper III).
In prolongation of ongoing episodes of certified sickness absence, very much reduced work ability assessed by patients was associated with longer duration of sickness absence than moderately reduced work ability (Paper III).3 Patients with very much reduced work ability had a mean remaining duration of 97 and 201 calendar days after the time of the work ability assessments in episodes classified as one and three month respectively. Corresponding figures for moderately reduced work ability were 49 and 111 calendar days. Work ability assessed by the patients was most accurate as a predictor of the remaining duration of certified sickness absence, at the time of the work ability assessments, in the interval between eight and 20 weeks from the start of absence, with the maximum at 13 weeks (Paper IV).
The patients’ work demands and the doctors’ work ability assessments were not associated with duration of sickness absence (Paper III).
A longer time until return to work, in patients with back disorders certified sick who attended a back disorder outpatient clinic, was predicted by low self-assessed work ability, a prediction of not returning to work four weeks ahead, and high pain intensity (Paper VI).4, 5 A shorter time until return to work was predicted by back disorders with radiating pain. Ninety-one percent of patients with moderately reduced work ability returned to work in a two-year follow-up versus 63% and 60% of the patients with much reduced and very much reduced work ability respectively. The COOP/WONCA charts physical fitness, daily activities, overall health and change in health were associated with time until return to work in univariate analyses only.
Older patients had a longer duration of sickness absence than younger patients (Paper III, VI).
The doctors’ assessments of who would return to work four weeks ahead had a positive predictive value (PPV) of 84% at the start of episodes of sickness absence, and a PPV of 53% in episodes of three to 20 weeks duration at the time of the assessments (Paper V). The corresponding PPVs for the doctors’ assessments of still certified sick were 72% and 91% respectively.
34 General discussion Methodological considerations
Representativity Information about the 913 patients in the primary study sample was compared with the National Sickness Benefit Register for the same period. Age and diagnoses of participating patients were not significantly different from other patients of participating doctors, as were not age and diagnoses of patients of participating and non-participating doctors. Participating doctors had more female patients certified sick than non-participating doctors, 55% versus 47%. Because the gender of the patients did not affect the main conclusions of this study, that difference is not considered to be of significant importance. The participating doctors had a mean age of 45 years, the non-participating 46 years. The gender distribution between participating and non-participating doctors was not significantly different.
Per 01.01.1996 the county of Aust-Agder had a population of 100,211, which represented 2.3% of the Norwegian population (106). The number of doctors in Aust-Agder per October 5. 1995, registered at the Norwegian Medical Association, was 252, representing 1.6% of the 16,078 doctors in Norway (66). 85 of the 252 were registered as general practitioners (Anders Taraldset, the Norwegian medical association, personal communication 08.04.2003), which equals the number of general practitioners invited to participate in this study. The participating doctors in Aust-Agder are presumed to be fairly representative for doctors in Aust-Agder and in Norway.
The number of certified sickness absence days paid by the National Insurance in 1995 was 8.4 for persons 16-67 years of age in Norway, 9.9 in the County of Aust-Agder (96). Corresponding figures for disability pensions were 8.3, and 10.3. It may be speculated that the higher figures in Aust-Agder reflect higher morbidity than in the rest of the country (82, 118).
Missing data A patient response rate of 71% on Questionnaire I without reminding is considered to be good.
There were few missing answers for the work ability questions and the assessments of certified sickness absence status four weeks ahead.
Occupation was not associated with work ability assessments and prediction of sickness absence duration in any of the analyses. Twenty-two percent missing information about occupation on Questionnaire I is therefore considered to be of minor importance. Occupation on Questionnaire II had no missing data.
Nine percent missing information about mean working hours per week on Questionnaire I may reflect difficulties for some persons to estimate that figure.
The questions of conflicts (Questionnaire I), especially among workers, may be considered a
35 control question. This may explain 10% missing data, despite the assurance that participation in this study would not interfere with the patients’ rights to benefits.
The use of “other” on the question about physical work demands (Questionnaire I) caused difficulties in the interpretation of that option. In the analyses presented in Table 4 in Paper II, 36 “other” were regarded as missing.
Visual analogue scale (VAS) questions may be more difficult to answer than ordinal option answers. On Questionnaire II the VAS pain intensity question had 16% missing, whereas the ordinal pain quality question had 1% missing.
Face validity Do the simple questionnaires used in this study give valid information about such complex phenomena as work ability and sickness absence, or does the simplification give a false picture of reality?
Measures of complex concepts may not be valid even when sophisticated measuring instruments are used. Measurements based on several questions summarized into indexes have an uncertainty of which elements of the indexes are most important. A widely used instrument, the SF-36® (SF for Short Form), was developed from a 102 item basis (127). Now SF-12®, and SF-8TM are also used (33).
The pilot study and the presented study did not bring forward any discussions about interpretations of the work ability question. It seems intuitive that the simple, global, one-item question used here measures work ability. The question was about the patient’s ordinary, remunerative work at the time of certification. Thus the face validity of the question may be considered good (1).
Construct validity The construction of the work ability question was based on the experiences of Haldorsen and co-workers (42, 92). Their Graded Reduced Work Ability scale had an internal consistency by Cronbach’s alpha of 0.71 for patient answers (42). It has been found that self-evaluated work ability correlates well with clinically determined musculoskeletal capacity in healthy women, and in women with back disorders (29). This supports some construct validity of the work ability question (1).
Patients and doctors may exaggerate the reduction of work ability to emphasize the need for sickness benefits. To reduce the possible bias of such “strategic” answers, patients and doctors were assured that participation in the study would not interfere with the patients’ actual rights to benefits. Patients assessed their work ability as less reduced than the doctors did in new episodes of certified sickness absence. The opposite occurred in prolongation of ongoing episodes.
36 Internal validity Diagnoses, medical factors, work related factors, and the presence of non-medical factors were associated with the levels of assessed work ability. This may support some relations between the components of the presented work ability model of Ilmarinen (Figure 2). The present study did not measure competence and values, which are included in his model. The objectives of this study, the methods used, and its findings are considered to correspond, indicating some internal validity of the study (1).
External validity How valid may the presented findings and conclusions be beyond the setting of this study, and what possibilities and limitations are there for generalisations based on this study-design. Self-assessed work ability predicted duration of sickness absence in both datasets, indicating good external validity (1, 75). There are limitations when results are compared between countries. In Norway, certified sickness absence can be given for maximum one year, whereas in Sweden there is no time limit (50). About 80% of all sickness certificates are issued by general practitioners in Norway, as opposed to about 40% in Sweden (28, 64, 113).
It is not likely that the slight differences of wording in the work ability questions on Questionnaire I and II should contribute to biased conclusions about the predictive impact of self-assessed work ability on duration of sickness absence.
Reliability The reliability of the questionnaires used in this study was not subject to a test-retest procedure.
Others have found that the reliability of diagnoses on sickness certificates is good (71, 113). Intra- and inter-observer concordance for diagnostic groups is 97% and 91% respectively (113). In a Swedish study, coding from medical certificates to registers was correct in 98%, whereas diagnoses based on texts in medical records matched diagnostic groups in 72% (71). We found all codes on sickness certificates to be correct in the Sickness Benefit Register in a sample of 25 cases (Paper VI).
There are no references in the literature to the reliability and validity of the duration data of the national registers used. The registers serve as account systems for the payment of sickness benefits, and are therefore revised, audited, and quality controlled. The registers also ensure good follow-up evaluation of the cases. None was lost to follow-up in this study.
Assessments of reduced work ability In new episodes of certified sickness absence, the distribution of work ability assessments was about one third on each of the categories moderately reduced, much reduced, and very much reduced work ability. This applies for both patients and doctors. For the patients, this
37 parallels the assessments of work ability in a study on episodes of certified sickness absence of about six weeks’ duration at the time of the assessments (20).
Most patients were assessed, by themselves and their doctors, to have very much and much reduced work ability. This is to be expected because the study sample only included patients certified sick. In new episodes of certified sickness absence, the doctors assessed work ability as more reduced than the patients did. The patients assessed work ability as more reduced in prolongation of ongoing episodes than in new, whereas the reverse was the case for the doctors’ assessments. This may reflect different clinical situations between issuing a new certificate and prolonging one. It may also indicate increasing discrepancies between patients’ and doctors’ expectations of future sickness absence duration.
An agreement of 81% on work ability assessments made by patients and doctors in new episodes of certified sickness absence could be the result of negotiations that take place in sickness certification consultations (114, 115). Other negotiation situations in general practice also show good agreement between the patients’ and their doctors’ opinions (108). Psychosocial problems seem to be common among work-disabled patients (40). General practitioners are good at identifying their patients’ work related problems (37), and patients’ self-perceived work disability is the socio-economic factor that general practitioners are best at evaluating (39). As gatekeepers for social insurance benefits, doctors frequently have to deal with these problems. It may be easier for doctors to handle work-related problems than more serious psychosocial problems. The issuing of sickness certificates can represent “silent agreement” solutions to unrevealed problems.
Diagnoses and medical factors The most common diagnostic groups were musculoskeletal, mental, and respiratory disorders, as previously found by others (113). In the present study, patients assessed work ability as more reduced in back disorders, and doctors in depression, than in other disorders. These diagnostic groups are large and important in sickness certification (51, 112, 113), and become increasingly important the longer the duration of the certification episodes (17). Back disorders affect all tasks of daily life.
The doctors’ work ability assessments were associated with clinical findings in new episodes of certified sickness absence. It may reflect that doctors use bio-medical conceptual models in their assessments, with more emphasis on objective findings than on patients’ statements (103, 126). However, although depression may present few clinical findings, the doctors assessed it as the disorder associated with the most reduced work ability.
Work demands Patients assessed their work ability as more reduced the more stressful or physically strenuous their jobs were, supporting the assumption that persons assess their work ability as more reduced in jobs with high work demands, than in jobs with lower demands.
38 Other factors The question on non-medical factors was intended to grasp elements which could have influenced the work ability assessments. However, most non-medical factors were work related, as found by others (38).
Prediction of sickness absence duration
Work ability A general question about functional restrictions caused by a health problem probably measures complex phenomena including not only the incapacity posed by ill-health, but also the patients’ life situation and the adaptation to it (76). The patients’ answers to the work ability question most likely include dimensions of their work demands, and probably their work motivation, which may explain why the patients’ self-assessed work ability predicted duration of sickness absence, whereas the doctors’ work ability assessments did not. In long-standing episodes of certified sickness absence doctors may look upon the medical problems as “solved” and reduce their emphasis on the consequences of the problems for the patients’ work ability. With reference to Figure 4, doctors may consider episodes of sickness absence to be finished when their assessments of the patients’ work ability match the doctors’ impression of the patients’ work demands.
It seems likely that there are similarities between reliability measures of self-assessed general health (73, 101), and self-assessed work ability. Reduced self-assessed overall health has a strong predictive impact on disability and death (32, 81, 93).
Age The age of patients predicts duration of certified sickness absence (112). Age has also been known as the most important determinant for sickness rates since actuaries searched for a “law of sickness” in the late 19th century (98).
Sickness absence status four weeks ahead The substantial differences between the PPVs for return to work in short-standing and long-standing episodes of certified sickness absence presented in Paper V suggest that the doctors have been overruled by the course of the illness, or the opinion of the patients, in long-standing episodes.
Prior duration may be a predictor of outcome. With reference to Paper V, recovery (here return to work) can be easy to predict in short-standing respiratory disorders, and conversely, pessimistic predictions (here still certified sick) will be more accurate in prolonged illnesses such as depression. It may thus be argued that the presented findings are obvious due to the high number of respiratory infections in short-standing episodes. Doctors do not necessarily know how self-limiting illnesses are at the time they perform their assessments. Not all
39 respiratory disorders in Paper V were of short duration, and mental disorders were evenly distributed in short-standing and long-standing episodes.
Doctors were good at predicting return to work in short-standing episodes, and continued certification in long-standing episodes in the present study (Paper V). A recent Norwegian study has shown that doctors’ assessments of their patients’ prognoses at eight weeks of certified sickness absence predicts disability pension (35).
Work demands No significant association was found between work demands and duration of sickness absence. Job satisfaction was significantly associated with duration in univariate analyses for new episodes only. No other work related factors were associated with duration in the univariate or multivariate analyses. Few studies have taken diagnoses into account when examining associations between work demands and occurrence, or duration, of sickness absence (74, 88). Some studies have focused on certain diagnostic groups only, such as low back pain (42, 48, 124). Different conclusions from ours about work demands and duration may have been reached for that reason. Another explanatory factor is that work demands may be of importance for becoming certified sick, while when certified sick, one is “out” of one’s work demands. However, the questions of work demands used in this study may not have covered all the different aspects of the term.
Back disorders with radiating pain Patients with back disorder with radiating pain had a shorter time until return to work than patients without radiation. This differs from findings among patients with back disorders in the general population who are certified sick, where patients with radiation have the longest duration (19, 41). When looking at the return to work rates for patients with back disorders certified sick, at more than eight weeks of absence duration, patients with radiation have higher return to work rates than patients without (41). Back disorders with radiation may represent disorders located in the lower back only, whereas other back disorders may represent more complex musculoskeletal disorders (84). In our study we had no information about whether the back disorders were localised or not (Paper VI). Others have found that low back pain existing as part of widespread musculoskeletal pain indicates more reduced functional ability than localised low back pain (83), and predicts long term work disability (85).
An association between aerobic capacity and pain or disability caused by chronic low back pain has not been found (54). Aerobic fitness of patients with chronic low back pain is comparable with the fitness healthy subjects (135).4
40 COOP/WONCA The work ability question presented for the patients at the back disorder outpatient clinic focused specifically on the back disorder. That question may therefore be more related to function in a setting of certified sickness absence, than the more global COOP/WONCA charts. This may explain why the work ability question was predictive of time until return to work in all analyses, whereas some of the COOP/WONCA charts were predictive only in the univariate analyses (Paper VI). The fact that COOP/WONCA charts without pictures were used on Questionnaire II should be considered when comparing the results presented here with other studies.
Temporal aspects Predictors have different predictive values depending on when in the course of sickness absence they are assessed. Diagnoses are predictive of duration at the start of sickness absence episodes, self-assessed work ability becomes predictive later on.
Temporal aspects of the predictors may provide clues to their predictability. The work ability question assessed the patients’ opinion on the day of response, whereas the COOP/WONCA charts retrospectively asked about function the preceding 14 days. The question of sickness absence status in four weeks concerned beliefs about the future. It may be speculated that the important predictive factors of sickness absence duration to assess are the patients’ present and future work ability, not their function in retrospect, because self-assessed work ability and sickness absence status four weeks ahead predicted absence, whereas the COOP/WONCA charts did not. Other functional assessment instruments deal with the situation the last week (20), the last four weeks (33), and the last year or more (33, 120), whereas most of the PCA (Personal Capability Assessment) elements are time neutral (58).
Functional assessment instruments about the situation today will not reflect fluctuations of function. For example, the functional consequences of musculoskeletal disorders may vary on a day-to-day basis. An assessment of today may give a false picture of the situation over time, whereas retrospective assessments may contribute to increased precision of measures. However, recall bias increases by time.
Implications of results At the start of certified sickness absence, diagnoses predict duration of absence. In prolongation of ongoing episodes of absence, the patients’ self-assessed work ability predict duration of sickness absence. The doctors’ assessments of certified sickness absence status four weeks ahead predict duration at the start of, and in prolongation of episodes. Doctors and patients thus supply important prognostic information. This information can be used to identify patients with high probability of short, intermediate and long-term sickness absence.
41 Such differentiation of patients into prognostic groups may be useful in allocating adequate interventions to the right patients at the right time (44, 45).
Depending upon the nature of the health problem, patients with high probability of intermediate and long-term sickness absence may be offered increased follow-up efforts by doctors and social insurance officers. Patients with high probability of short-term sickness absence should not be offered extensive rehabilitation measures.
Conclusions The following conclusions can be drawn from this study that focused on patients’ and doctors’ assessments of work ability in certified sickness absence.
The patients and the doctors concurred on their work ability assessments, mostly at the start of episodes of sickness absence. They assessed work ability as very much or much reduced in 2/3, and as moderately reduced in 1/3 of the cases.
The patients assessed their work ability as more reduced in back disorders than in other sickness certification diagnoses, and as more reduced the more stressful or physically strenuous their jobs were. The doctors assessed the patients’ work ability as more reduced in depression than in other diagnoses, and as more reduced when the doctors’ assessments were based on clinical findings. In prolongation of ongoing episodes the doctors assessed fewer female than male patients to have very much or much reduced work ability.
At the start of episodes of certified sickness absence, the duration of absence was predicted by diagnoses and the doctors’ assessments of who would return to work four weeks ahead. In prolongation of ongoing episodes of sickness absence, the duration was predicted by the patients’ own assessments of reduced work ability, and by the doctors’ assessments of who would still be certified sick four weeks ahead. Back disorders with radiating pain predicted shorter duration of absence than back disorders without radiating pain. Patients with very much reduced self-assessed work ability had twice the duration, and about twice the probability of not returning to work, compared with patients with moderately reduced work ability.
Diagnoses were most accurate as predictors in the first weeks of absence. The patients’ self-assessed work ability was most accurate as a predictor in the interval between eight and 20 weeks from the start of absence, with a maximum at 13 weeks.
The patients’ work demands and the doctors’ work ability assessments were not associated with duration of sickness absence.
42 References 1. Abramson JH. Survey methods in community medicine. London: Churchill Livingstone, 1990.
2. Ainsworth-Vaughn N. Patients claiming power in doctor-patient talk. New York: Oxford University Press, 1998.
3. Ala-Mursula L, Vahtera J, Kivimäki M, Kevin MV, Pentti J. Employee control over working times: associations with subjective health and sickness absences. J Epidemiol Community Health 2000; 56: 272-8.
4. Alexanderson K. Sickness absence in a Swedish county, with reference to gender, occupation, pregnancy and parenthood. Dissertation. Linköping: Linköping University, Department of Community Medicine, 1995.
5. Alexanderson K. Forskning om sjukfrånvaro [Sickness absence research]. Available at: www.fkf.se/aktuellt/kurser/2003/forskarkonferens/kristina_alexandersson.pdf (15.04.2003) (In Swedish)
6. Altman DG. Practical statistics for medical research. Cornwall: TJ Press (Padstow) Ltd, 1996.
7. Arbeidsdirektoratet. Yrkes-klassifisering. Yrkesliste med NYK-nummer (Nordisk yrkesklassifisering) [Nordic classification of occupations]. Oslo: Arbeidsdirektoratet, 1989. (In Norwegian)
8. Arbejdsmiljøinstituttet. Projekt om forskning og intervention i monotont arbejde PRIM [Project about research and intervention in monotonous work]. Available at: http://www.ami.dk/upload/udgivelser/kortspoergeskema2.pdf (07.10.2003) (In Danish)
9. Benefits agency. The all work test. London: Department of Social Security, 1995.
10. Bentsen BG, Natvig B, Winnem M. Questions you didn’t ask? COOP/WONCA Charts in clinical work and research. World Organization of Colleges, Academies and Academic Associations of General Practitioners/Family Physicists. Fam Pract 1999; 16: 190-5.
11. Bjerkedal T, Thune O. Hva koster sykelønnsordningen? [Total cost of sick leave benefits: the case of Norway]. Tidsskr Nor Lægeforen 2003; 123: 662-3.
12. Bjørnson Ø, Haavet IE. Langsomt ble landet et velferdssamfunn. Trygdens historie 1894-1994 [Slowly the country became a welfare state. The history of health insurance 1894-1994]. Oslo: Ad Notam Gyldendal, 1994. (In Norwegian)
13. Bloch FS, Prins R (ed.). Who returns to work & why. A six-country study on work incapacity & reintegration. New Brunswick: Transaction publishers, 2001.
14. Bonner D, Hooker I, White R. Non means tested benefits: The legislation, statutory sick
43 pay (general) regulations. London: Sweet and Maxwell, 1998.
15. Brage S, Bentsen BG, Bjerkedal T, Nygård JF, Tellnes G. ICPC as a standard classification in Norway. Fam Pract 1996; 13: 391-6.
16. Brage S, Nygård JF, Tellnes G. The gender gap in musculoskeletal-related long term sickness absence in Norway. Scand J Soc Med 1998; 26: 34-43.
17. Brage S, Nygård JF, Tellnes GT. Langtids sykefravær i Norge 1989-94 [Long-term sickness certification in Norway 1989-94]. Oslo: University of Oslo, Department of Community Medicine, Section of Social Insurance Medicine, Report 3, 1998. (In Norwegian)
18. Brage S. Musculoskeletal health problems and sickness absence - an epidemiological study of concepts, determinants and consequences. Dissertation. Oslo: University of Oslo, Department of Community Medicine, 1998.
19. Brage S, Lærum E. Rygglidelser i Norge – en epidemiologisk beskrivelse [Spinal disorders in Norway - an epidemiological report]. Tidsskr Nor Lægeforen 1999; 119: 1619-23. (In Norwegian, English summary)
20. Brage S, Fleten N, Knudsrød OG, Reiso H, Ryen A. Egenvurdring av funksjon – utvikling av Norsk funksjonsskjema for bruk i trygdeetaten [Self-assessed function – development of Norwegian functional charts for use by the National Insurance]. Oslo: Rikstrygdeverket, Rapport 5, 2002. (In Norwegian)
21. Bruusgaard D, Nøklebye Heiberg A, Lie K, Malterud K, Noreik K, Westin S. Sykdomsbegrepet i folketrygdloven (§8-3). Konsensusrapport til trygderetten [The concept of disease in the Norwegian Social Insurance Scheme]. Oslo: University of Oslo, Department of Community Medicine, 1994. (In Norwegian)
22. Chew CA, May CR. The benefits of back pain. Fam Pract 1997; 14: 461-65.
23. Crombie IK, Croft PR, Linton SJ, LeResche L, Von Korff M (ed.). Epidemiology of pain. International association for the study of pain. Seattle: IASP Press, 1999.
24. Den kongelige proposisjon om statsbudsjettet medregnet folketrygden for budsjetterminen 1. januar - 31. desember 2003 [The Norwegian State budget for the year 2003] Available at: www.statsbudsjettet.dep.no/2003/dokumenter/html/gulbok/Kap01.htm (08.07.2003) (In Norwegian)
25. Engbers M, Veiersted B. Rapport fra ekspertkonferanse om funksjonsvurdering 27-28 januar 2003 [Report from an expert conference concerning functional assessments 27.-28. January 2003]. Oslo: Statens Arbeidsmiljøinstitutt (STAMI), Rapport 5, 2003. (In Norwegian)
26. Englund L, Svärdsudd K. Sick-listing habits among general practitioners in a Swedish
44 county. Scand J Prim Health Care 2000; 18: 81-6.
27. Englund L, Tibblin G, Svärdsudd K. Effects on physicians’ sick-listing practice of an administrative reform narrowing sick-listing benefits. Scand J Prim Health Care 2000; 18: 215-9.
28. Englund L. Sick-listing – attitudes and doctors’ practice. With special emphasis on sick-listing practice in primary health care. Dissertation. Uppsala: Uppsala University, 2000.
29. Eskelinen L, Kohvakka A, Merisalo T, Hurri H, Wägar G. Relationship between the self-assessment and clinical assessment of health status and work ability. Scand J Work Environ Health 1991; 17 (suppl 1): 40-7.
30. Folketrygdloven [Text of law concerning The Norwegian Social Insurance Scheme]. Available at: http://www.lovdata.no/all/ (07.08.2003) (In Norwegian)
31. Försäkringsmedicin. Allmän försäkring [Insurance medicine. Common insurance]. Stockholm: Riksförsäkringsverket, 1994. (In Swedish)
32. Fylkesnes K, Førde OH. The Tromsø study: predictors of self-evaluated health - has society adopted the expanded health concept? Soc Sci Med 1991; 32: 141-6.
33. Generic health surveys: SF-36®, SF-12®, and SF-8TM. Available at: www.qualitymetric.com/products/SFSurveys.shtml (07.08.2003)
34. Gjesdal S. From long-term sickness absence to disability pension. Studies on predictors of disability pension in Norway. Dissertation. Bergen: University of Bergen, Section for Social Medicine, Department of Public Helath and Primary Health Care, and Program for Health Economics, 2003.
35. Gjesdal S, Rigdal PR, Haug K, Mæland, JG. Predictors of diability pension in long-term sickness absenec. European Journal of Public Health – accepted for publication.
36. Graff-Iversen S, Skurtveit S, Nybø A, Ross GB. Utviklingen i kroppsarbeid hos norske 40 – 42-åringer i tiden 1974 – 94 [Occupational physical activity: trends among Norwegian men and women aged 40–42 years, 1974–94]. Tidsskr Nor Lægeforen 2001; 121: 2584-8. (In Norwegian, English summary)
37. Gulbrandsen P, Hjortdahl P, Fugelli P. General practitioners’ knowledge of their patients’ psychosocial problems: multipractice questionnaire survey. BMJ 1997; 314: 1014-8.
38. Gulbrandsen P, Brage S. Livssituasjon som grunnlag for sykmelding [Life situation as a reason for sick leave]. Tidsskr Nor Lægeforen 1998; 118: 2463-6. (In Norwegian, English summary)
39. Gulbrandsen P, Fugelli P, Hjortdahl P. General practitioners’ knowledge of their 45 patients’ socio-economic data and their ability to identify vulnerable groups. Scand J Prim Health Care 1998; 16: 204-10.
40. Gulbrandsen P, Hjortdahl P, Fugelli P. Work disability and health-affecting psychosocial problems among patients in general practice. Scand J Soc Med 1998; 26: 96-100.
41. Hagen KB, Thune O. Work incapacity from low back pain in the general population. Spine 1998; 23: 2091-5.
42. Haldorsen EMH, Indahl A, Ursin H. Patients with low back pain not returning to work. A 12-month follow-up study. Spine 1998; 23: 1202-8.
43. Haldorsen EMH. Return to work in low back pain patients. Dissertation. Bergen: Bergen University, Department of Biological and Medical Psychology, 1998.
44. Haldorsen EMH, Grasdal AL, Skouen JS, Risa AE, Kronholm K, Ursin H. Is there a right treatment for a particular patient group? Comparison of ordinary treatment, light multidisciplinary treatment, and extensive multidisciplinary treatment for long-term sick-listed employees with musculoskeletal pain. Pain 2002; 95: 49-63.
45. Haldorsen EMH. The right treatment to the right patient at the right time. Occup Environ Med 2003; 60: 235-6.
46. Hansen HT. Kunnskapsstatus for den nordiske trygdeforskningen på 1990-tallet [Knowledge base for the Nordic social insurance research in the 1990’s]. Bergen: Stiftelsen for samfunns- og næringslivsforskning; Rapport 12, 1999. (In Norwegian)
47. Harlem G. Invaliditetsvurdering [Assessment of invalidity]. Tidsskr Nor Lægeforen 1959; 79: 380-3. (In Norwegian)
48. Hemingway H, Shipley MJ, Stansfeld S, Marmot M. Sickness absence from back pain, psychosocial work characteristics and employment grade among office workers. Scand J Work Environ Health 1997; 23: 121-9.
49. Hensing G. Sickness absence and psychiatric disorder – epidemiological findings and methodological considerations. Dissertation. Linköping: Linköping University, Department of Community Medicine, 1997.
50. Hensing G, Alexanderson K, Allebeck P, Bjurulf P. How to measure sickness absence? Literature review and suggestion of five basic measures. Scand J Soc Med 1998; 26: 133-44.
51. Hensing G, Brage S, Nygard JF, Sandanger I, Tellnes G. Sickness absence with psychiatric disorders - an increased risk for marginalisation among men? Soc Psychiatry Psychiatr Epidemiol 2000; 35: 335-40.
52. Hjort P, Damberg G, Holte A, Paulsen K, Riise G. Funksjonsvurdering. Rapport fra prosjekt for inkluderende arbeidsliv [Functional assessments. Report from project for 46 including working-life]. Rikstrygdeverket: Oslo, 2003. Available at: www.trygdeetaten.no/Helsepersonell/pdf/IA81013.1.0_Funksjonsvurdering_rapport.pdf (11.11.03) (In Norwegian)
53. Holzinger H. Die geschichtliche Entwicklung der sozialen Sicherung bei Arbeitsunfähigkeit [Insurances against inability to work]. Freiburg: HochschulVerlag, Hochschulsammlung Philosophie: Sozialwissenschaft, Band 21, 1992. (In German, English summary)
54. Hurri H, Mellin G, Korhonen O, Harjula R, Harkapaa K, Luoma J. Aerobic capacity among chronic low back pain patients. J Spinal Disord 1991; 4: 34-8.
55. Ilmarinen J, Tuomi K, Eskelinen L, Nygård CH, Huhtanen P, Klockars M. Summary and recommendations of a project involving cross-sectional and follow-up studies on the ageing worker in Finnish municipal occupations (1981-1985). Scand J Work Environ Health 1991; 17 (suppl): 135-41.
56. Ilmarinen J. Ageing workers in the European Union – Status and promotion of work ability, employability and employment. Helsinki: Institute of Occupational Health, Ministry of Social Affairs and Health, Ministry of Labour, 1994.
57. Ilmarinen J, Tuomi K, Klockars M. Changes in the work ability of active employees over an 11-year period. Scand J Work Environ Health 1997; 23 suppl 1: 49-57.
58. Incapacity for work: the personal capability assessment. Available at: www.bhas.org.uk/incapacity_intro.htm (07.08.2003)
59. Indahl A. Low back pain – a functional disturbance? Physiology and treatment. Dissertation. Oslo: University of Oslo, Centre for Orthopaedics, National Hospital, 1999.
60. Inkluderende arbeidsliv [Including working-life]. Oslo: Intensjonsavtale av 3. oktober 2001. Available at: http://www.trygdeetaten.no/Arbeidsliv/pdf/intensjonsavtale.pdf (11.11.2003) (In Norwegian)
61. Karasek R, Theorell T. Healthy work: stress, productivity and the reconstruction of working life. New York: Basic Books, 1990.
62. Kivimäki M, Head J, Ferrie JE, Shipley MJ, Vahtera J, Marmot MG. Sickness absence as a global measure of health: evidence from mortality in the Whitehall II prospective cohort study. BMJ 2003; 327: 364-9.
63. Kjønstad A. Fem ulike uførebegreper [Five different concepts of disability]. Tidsskr Nor Lægeforen 1989; 109: 2146-8. (In Norwegian, English summary)
64. Larsen BA, Førde OH, Tellnes G. Legens kontrollfunksjon ved sykmelding [Physician’s role in sickness certification]. Tidsskr Nor Lægeforen 1994; 114: 1442-4. (In Norwegian, English summary)
47 65. Last JM. A Dictionary of Epidemiology. Third edition. New York: Oxford University Press Inc., 1995.
66. Legestatistikk vedrørende medlemmer av Den norske lægeforening [Statistics concerning doctors of The Norwegian medical association]. Available at: www.legeforeningen.no/index.db2?id=8449 (09.10.2003) (In Norwegian and English)
67. Leijon M. Integrating perspectives in Social Medicine - a study using epidemiological and clinical methods with special reference to sickness absence. Dissertation. Linköping: Linköping University, Department of Social Medicine and Public Health, 2002.
68. Lemeshow S, Letenneur L, Dartigues JF, Lafont S, Orgogozo JM, Commenges D. Illustration of analysis taking into account complex survey considerations: the association between wine consumption and dementia in the PAQUID study. Am J Epidemiol 1998; 148: 298-306.
69. Lindén V. Absence from work and work capacity. Dissertation. Bergen: Bergen University, and The National Health Insurance, local office Bergen, 1967.
70. Lipsky M. Street-level bureaucracy: dilemmas of the individual in public services. New York: Russell Sage Foundation, 1980.
71. Ljungdahl LO, Bjurulf P. The accordance of diagnoses in a computerized sick-leave register with doctor’s certificates and medical records. Scand J Soc Med 1991; 19: 148-153.
72. Löfvander M. Illness, disease, sickness. Clinical factors, concepts of pain and sick leave patterns among immigrants in primary health care. Effects of different therapeutic approaches. Dissertation. Stockholm: Karolinska Institutet, Department of Clinical Neuroscience and Family Medicine, and Rinkeby Primary Health Care Centre, 1997.
73. Lundberg O, Manderbacka K. Assessing reliability of a measure of self-rated health. Scand J Soc Med 1996; 24: 218-24.
74. Luz J, Green MS. Sickness absenteeism from work - a critical review of the literature. Public Health Rev 1997; 25: 89-122.
75. Malterud K. Kvalitative metoder i medisinsk forskning: en innføring [Qualitative methods in medical research: an introduction]. Oslo: Universitetsforlaget, 2003. (In Norwegian)
76. Manderbacka K. Methods. How do respondents understand survey questions on ill-health? European Journal of Public Health 1998; 8: 319-24.
77. Mastekaasa A, Moum T, Næss S, Sørensen T. Livskvalitetsforskning [Quality of life research]. Oslo: Institute For Social Research, Report 6, 1988. (In Norwegian)
78. Matheson LN. Functional capacity evaluation. In: Demeter SL, Andersson GBJ, Smith 48 GM (ed.). Disability Evaluation. St. Louis: Mosby, 1996.
79. Matheson LN, Kaskutas V, McGowan S, Shaw H, Webb C. Development of a database of functional assessment measures related to work disability. Occup Rehabil 2001; 3: 177-99.
80. Melvær E. Sykdomsbegrepet i trygden [The concept of disease in the Social Insurance Scheme]. Tidsskr Nor Lægeforen 1992; 112: 3128-31. (In Norwegian)
81. Miilunpalo S, Vuori I, Oja P, Päsänen M, Urponen H. Self-rated health status as a health measure: the predictive value of self-reported health status on the use of physician services and on mortality in the working-age population. J Clin Epidemiol 1997; 50: 517-28.
82. Mouland G, Bratland B, Fagan M, Malde K, Rygh E, Welander F, Wærdahl E, Ytterdahl T. Use of new antidepressants in general practice [Bruk av nye antidepressiver i allmennpraksis]. Tidsskr Nor Lægeforen 1998; 118: 4497-500. (In Norwegian, English summary)
83. Natvig B, Bruusgaard D, Eriksen W. Localized low back pain and low back pain as part of widespread musculoskeletal pain: two different disorders? A cross-sectional population study. J Rehab Med 2001; 33: 21-5.
84. Natvig B. Musculoskeletal pain. From humdrum nuisance to chronic, disabling disease. Dissertation. Olso: University of Oslo, Department of General Practice and Community Medicine, 2001.
85. Natvig B, Bruusgaard D, Eriksen W. Low back pain as a predictor of long-term work disability. Scand J Public Health 2002; 30: 288-92.
86. Nordenfelt L. On the notions of disability and handicap. Scand J Soc Welfare 1993; 2: 17-24.
87. Nordenfelt L. Disease, illness og sickness (translated from English by Erik Falkum). In: Heiberg A, Falkum E, Nylenna M (ed.). Medisinsk mangfold - mangfoldig medisiner. Festskrift til professor Øivind Larsen på hans 60-års dag [Congratulatoria to Øivind Larsen]. Oslo: Universitetsforlaget, 1998. (In Norwegian)
88. North FM, Syme SL, Feeney A, Shipley M, Marmot M. Psychosocial work environment and sickness absence among British civil servants: the Whitehall II study. Am J Public Health 1996; 86: 332-40.
89. Nyman K, Bergendorff S, Palmer E. Den svenska sjukan. Sjukfrånvaro i 8 länder [The Swedish disease. Sickness absence in 8 countries]. Stockholm: Regeringskanselliet, DS; 49, 2002. (In Swedish)
90. Odelstingsproposisjon vedrørende forslag til lov om Folketrygd [Proposals concerning new Norwegian Social Insurance Scheme – White Paper]. OT prp. nr. 22., 1959. (In
49 Norwegian)
91. Olsen KM, Mastekaasa A. Forskning om sykefravær – en oppsummering og vurdering av perioden 1980-96 [Research about sickness absence – a summary and an assessment of the period 1980-96]. Oslo: Institute For Social Research, Report 3, 1997. (In Norwegian)
92. Operasjonalisering av sykdomsbegrepet: En evaluering. Avsluttende rapport (OASE-prosjektet) [Operationalisation of the disease concept: An evaluation. Final report]. Oslo and Bergen: University of Oslo, Department of Community Medicine, Section of Social Insurance Medicine, and Bergen University, Department of Biological and Medical Psychology, Section of Physiological Psychology, 1993.
93. Östlin P. Occupational history, self reported chronic illness, and mortality: a follow up of 25,586 Swedish men and women. J Epidemiol Community Health 1990; 44: 12-6.
94. Reiso H, Nygård J, Tellnes G. Beslutningskriterier ved sykmelding. Rapport fra pilotprosjekt 1994-95 [Decision-making criteria in certified sickness absence. Report from a pilot study 1994-95]. Oslo: University of Oslo, Department of General Practice and Community Medicine, Section of Social Insurance Medicine, and the Aust-Agder County Office of the National Insurance Service, 1995. (In Norwegian)
95. Reiso H, Nygård JF, Brage S, Tellnes G. Arbeidsevne og sykmelding [Work ability and certified sickness absence]. Oslo: University of Oslo, Department of General Practice and Community Medicine, Section of Social Insurance Medicine, Report 2, 1998. (In Norwegian)
96. Rikstrygdeverket. Trygdestatistisk årbok 1996 [National Insurance Statistics for the year 1996]. Oslo: Rikstrygdeverket, 1996. (In Norwegian)
97. Rikstrygdeverket, Utredningsavdelingen. Folketrygden. Nøkkeltall 2002 [Key figures for the Norwegian Social Insurance Administration 2002]. Available at: www.trygdeetaten.no/Trygdeetaten/Statistikker/Nokkeltall/nokkeltall2002.pdf (03.07.2003) (In Norwegian)
98. Riley JC. Disease without death: New sources for a history of sickness. Journal of interdisciplinary history 1987; xvii: 3: 537-63.
99. Rutle O. Pasienten fram i lyset – analyse av legekontaktar i primærhelsetenesta [An analysis of encounters in general practice]. Oslo: National Institute of Public Health, Report 1, 1983. (In Norwegian, English summary)
100. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical diagnostic strategies. In: Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology. A basic science for clinical medicine. Boston: Little, Brown, 1991.
101. Self-rated health in a European Perspective. Stockholm: Forskningsrådsnämnden,
50 2000.
102. Söderberg E, Alexanderson K. Litteraturgenomgång – av publicerade studier om gränssnittet mellan medicinsk praxis och socialförsäkrings-rättslig bedömning [Literature study – published studies of the interface between medical practice and assessments of social insurance legislation]. Linköping: Försäkringsmedicinskt centrum, Rapport 1, 2001. (In Swedish)
103. Solli HM. Medisinsk sakkyndighet, objektivitet og rettferdighet i uførepensjonssaker [Medical expert assessment, objectivity and justice in disability pension cases]. Tidsskr Nor Lægeforen 2003; 123: 2072-5. (In Norwegian, English summary)
104. Statistisk Sentralbyrå. Arbeidslivsundersøkelsen 1993 [Statistics of Norway – Working-life query 1993]. Kongsvinger: Statistisk Sentralbyrå (SSB), Seksjon for Intervjuundersøkelser, 1993. (In Norwegian)
105. Statistisk Sentralbyrå. Helsundersøkelsen 1995 [Statistics of Norway – health survey 1995]. Available at: www.ssb.no/emner/03/01/helseforhold (08.07.2003) (In Norwegian)
106. Statistisk Sentralbyrå. Befolkningsdata [Statistics of Norway – population data]. Available at: www.ssb.no/emner/02/befolkning (08.07.2003) (In Norwegian)
107. Stone DA. The disabled state. Philadelphia: Temple University Press, 1984.
108. Straand J, Sandvik H. Stopping long-term drug therapy in general practice. How well do physicians and patiens agree? Fam Pract 2001; 18: 597-601.
109. Strand K. Pregnancy at work – work conditions, job adjustment and sickness absence in pregnancy. Dissertation. Oslo: University of Oslo, Institute of General Practice and Community Medicine, 2001.
110. Swales K, Craig P. Evaluation of the Incapacity Benefit Medical Test. London: Social Research Branch, Analytical Services Division, Department of Social Security, In-house report 26, 1997.
111. Sykefravær og uførepensjonering. Et inkluderende arbeidsliv [Sickness absence and disability pension, an including working-life]. NOU No. 27; 2000. Available at: www.dep.no/sos/norsk/publ/utredninger/NOU/ (08.07.2003) (In Norwegian)
112. Tellnes G. Duration of episodes of sickness certification. Scand J Prim Health Care 1989; 7: 237-44.
113. Tellnes G, Svendsen K-OB, Bruusgaard D, Bjerkedal T. Incidence of sickness certification. Proposal for use as a health status indicator. Scand J Prim Health Care 1989; 7: 111-7.
114. Tellnes G. Sickness certification in general practice: a review. Fam Pract 1989; 6: 58-65. 51 115. Tellnes G, Bruusgaard D, Sandvik L. Occupational factors in sickness certification. Scand J Prim Health Care 1990; 8: 37-44.
116. Tellnes G, Sandvik L, Moum T. Inter-doctor variation in sickness certification. Scand J Prim Health Care 1990; 8: 45-52.
117. Tellnes G. Sickness certification - an epidemiological study related to community medicine and general practice. Dissertation. Oslo: University of Oslo, Department of Community Medicine, 1990.
118. Tellnes G, Mathisen S, Skau I, Thune O, Ulsberg S, Berg JE. Hvem er langtidssykmeldt i Norge? Fra prosjektet Evaluering av oppfølging av langtidssykmeldte [Who is long-term certified sick in Norway?]. Tidsskr Nor Lægeforen 1992; 112: 2773-8.
119. Timpka T, Hensing G, Alexanderson K. Dilemmas in sickness certification among Swedish physicians. Eur J Public Health 1995; 5: 215-9.
120. Tuomi K, Ilmarinen J, Jahkola A, Katajarinne L, Tulkki A. Work Ability Index. Helsinki: Institute of Occupational Health, 1994.
121. Tveråmo A, Dalgard OS, Claussen B. Økende psykisk stress blant unge voksne i Norge 1990 – 2000 [Increasing psychological stress among young adults in Norway, 1990 - 2000]. Tidsskr Nor Lægeforen 2003; 123: 2011-5.
122. Vahtera J, Kivimäki M, Pentti J, Theorell T. Effect of change in the psychosocial work environment on sickness absence: a seven year follow up of initially healthy employees. J Epidemiol Community Health 2000; 54: 484-93.
123. Välfors B. Acute, subacute and chronic low back pain – Clinical symptoms, absenteeism and working environment. Dissertation. Göteborg: University of Gothenburg, Department of Orthopaedic Surgery, Sahlgren Hospital, 1985.
124. van der Weide WE, Verbeek JH, Salle HJ, van Dijk FJ. Prognostic factors for chronic disability from acute low-back pain in occupational health care. Scand J Work Environ Health 1999; 25: 50-6.
125. Voss M. Work and health. Epidemiological studies of sickness absence and mortality with special reference to work environment, factors outside work and unemployment. Dissertation. Stockholm: Karolinska Institutet, Institute of Environmental Medicine, and National Institute for Working Life, 2002.
126. Waddell G. Models of Disability. Using Low Back Pain as an Example. London: Royal Society of Medicine Press, 2002.
127. Ware JJ, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care 1992; 30: 473-83.
128. Wergeland E. Healthy working conditions in pregnancy. Dissertation. Oslo: University of Oslo, Institute of General Practice and Community Medicine, 2002. 52 129. Westin S. Becoming disabled: a sociomedical analysis of individual adaptations to life after long-term unemployment. Trondheim: Transactions of the Royal Norwegian Society of Sciences and Letters, No 2, 1990.
130. Westin S. Unemployment and health: Medical and social consequences of a factory closure in a ten-year controlled follow-up study. A study from general practice. Dissertation. Trondheim: University of Trondheim, Department of Community Medicine, 1990.
131. Westrin CG. Low back sick-listing. A nosological and medical insurance investigation. Dissertation. Scand J Soc Med 1973; suppl 7.
132. WHO. International classification of impairments, disabilities and handicaps. A manual of classification relating to the consequences of diseases. Geneva: World Health Organization, 1980.
133. WHO. ICF-International Classification of Functioning, Disability, and Health. Geneva: World Health Organization, 2001. Available at: www3.who.int/icf/icftemplate.cfm (03.07.2003)
134. WHO. Definition of Health. Available at: http://www.who.int/about/definition/en/ (07.10.2003)
135. Wittink H, Hoskins Michel T, Wagner A, Sukiennik A, Rogers W. Deconditioning in patients with chronic low back pain: fact or fiction? Spine 2000; 25: 2221-8.
136. WONCA Charts. Available at: www.globalfamilydoctor.com/publications/coop-woncacharts/languages/english.pdf (08.07.2003)
137. Wulff HR. Rationel klinik. Grundlaget for diagnostiske og terapeutiske beslutninger [Rational diagnosis and treatment] København: Villadsen & Christensen, 1987. (In Danish, English version available)
138. Åstrand PO. Experimental studies of physical working capacity in relation to sex and age. Dissertation. Copenhagen: Munksgaard, 1952.
53
Paper I Harald Reiso, Jan F Nygård, Sören Brage, Pål Gulbrandsen, Gunnar Tellnes.
Work ability assessed by patients and their GPs in new episodes of sickness certification. Fam Pract 2000; 17: 139–44.
Reprinted by courtesy of Family Practice and Oxford University Press
Family Practice Vol. 17, No. 2 © Oxford University Press 2000 Printed in Great Britain
Work ability assessed by patients and their GPs in new episodes of sickness certification Harald Reisoa,b, Jan F Nygårda, Sören Bragea, Pål Gulbrandsena and Gunnar Tellnesa
Reiso H, Nygård JF, Brage S, Gulbrandsen P and Tellnes G. Work ability assessed by patients and their GPs in new episodes of sickness certification. Family Practice 2000; 17: 139–144. Background. Sickness certification legislation demands that work ability is reduced due to disease or injury. Most sickness certificates are issued by GPs. Assessment of work ability might introduce conflict in the doctor–patient relationship. Objectives. The aim of this study was to compare the level of work ability assessments by patients and their GPs in new episodes of sickness certification, and to explore how medical conditions and work demands are associated with the assessments. Methods. Forty nine GPs supplied data about 408 patients certified sick Ͻ8 days before ques- tionnaires were filled in. A total of 268 (66%) patients completed corresponding questionnaires. Patients and GPs independently answered the following question using a five-point scale: “To what degree is your (the patient’s) ability to perform your (his or her) ordinary, remunerative work reduced today?” Results. Work ability was assessed by patients as very much or much reduced in 66%, moder- ately reduced in 23% and not much or hardly reduced at all in 11% of the cases. Corresponding assessments made by GPs were 71, 27 and 2%. Patients and GPs agreed well on their assess- ments (± 1 answer category) in 81% (216/266) of the cases. The patients assessed work ability as more reduced the more stressful or physically strenuous their jobs were, and the older their GPs were. The GPs assessed work ability as more reduced the more their assessments were based on clinical findings. Conclusions. The agreement between work ability assessments made by patients and GPs was high, despite patients’ assessments being associated with work demands and GPs’ with medical conditions. Keywords. Functional assessment, sickness absence, sickness certification, sick-listing, work ability.
Introduction work ability” and “incapable of work” are phrases used to describe reduced work capacity.1 Sickness benefit payments account for Ͼ10% of the Doctors have difficult and conflicting roles as social insurance costs in Norway. Persons receiving assessors of function in social insurance medicine. Their sickness benefits must be “. . . incapacitated for work dilemmas concern assessment of work incapacity, duration due to a loss of function that clearly is caused by disease of sick leave, interpreting sickness benefit legislation or injury”, according to Norwegian legislation. Similar and giving social interpretations to medical diagnoses.2 criteria are used in Sweden and the UK, where “reduced Should doctors act as gatekeepers for social insurance benefits, or serve as advocates for their patients when assessing work ability? The roles conflict, and the former Received 23 April 1999; Revised 9 September 1999; Accepted may introduce strain in the doctor–patient relationship. 26 October 1999. This particularly concerns GPs, because they issue Ͼ80% a Institute of General Practice and Community Medicine, of sickness certificates in Norway.3 Department of Social Insurance Medicine, University of Oslo, PO Box 1130 Blindern, N-0318 Oslo and bAust-Agder County Whether sickness certification will be a consequence Office of the National Insurance Service, PO Box 188, N-4802 of reduced work ability depends on the medical con- Arendal, Norway. ditions and work demands of the patients.4,5 There is
139 140 Family Practice—an international journal scarce knowledge about how these factors are associated Ability scale, constructed for the Norwegian Ministry of with each other.2,6,7 Health and Social Affairs.10 In order to measure function, indicators such as self- The GPs classified the main sick-listing diagnoses reported function,8–11 assessments made by observers11 according to ICPC (International Classification of or more specific physiological tests can be used. How- Primary Care), recorded the kind of clinical information ever, most of these methods consist of several items, sources that they used when assessing work ability, and they are time consuming and are therefore unsuitable their use of clinical examinations or tests in the con- for everyday clinical practice. We were not aware of any sultations. Doctors and patients also evaluated whether tool designed for assessment of work ability in early non-medical factors could have influenced their work periods of sickness certification in general practice, and ability assessments. therefore we developed a simple questionnaire for that The patients gave information about their occupations purpose. and work demands. The latter was operationalized as the The present study is aimed at exploring the level of number of remunerative working hours per week, day- work ability assessed by patients and their GPs in new time or night/shift work, a subjective description of physical episodes of sickness certification, the degree to which work demands, heavy lifting, self-determined breaks, these assessments coincided or conflicted, and whether conflicts, stressful working conditions, satisfaction and they were associated with medical conditions or work influence in their jobs. demands. The GPs provided information about their own age and gender and the patients’ age and gender.
Materials and methods Statistical analyses Continuous variables were compared using two-tailed A pilot study was conducted in 1994. Five meetings with t-tests, categorized by chi-square statistics. Diagnoses GPs and social insurance officers were held prior to it, in were recoded into diagnostic groups. order to ensure the relevance of the questionnaires. No Design-based analyses were performed using difficulties were reported in completing 116 question- STATA® software,12 with GPs as the primary sampling naires by 16 GPs and their patients. There was a good units. Design-based multiple logistic regression analyses spread in the use of answer categories and there were included variables with univariate P-values Ͻ0.20 when few missing data. compared with work ability, and age and gender of Work ability in sickness certification was defined as patients and GPs. Assessed work ability was recoded the physical and psychological capacity of a person to into very much/much reduced, and moderately/not much/ perform ordinary, remunerative work. hardly reduced at all in these analyses. Intercorrelation The study was conducted in January–April 1996, in the between independent variables was checked by county of Aust-Agder, southern Norway. The county Spearman’s rank correlation coefficient. had 100 211 inhabitants as per January 1, 1996. The The chosen level of significance was 5%. number of persons with sickness benefit rights, aged 16–66 years, was 47 835 in 1996. The sample All GPs (85) in the county were invited to participate A total of 1001 questionnaires from GPs were collected. in the study. Each doctor could record 25 consecutive 138 questionnaires were excluded due to double record- sickness certification episodes in ordinary daytime prac- ings (57 cases), uncertain certification dates (30), patients tice. All types of contacts were eligible. Questionnaires certified sick while on rehabilitation benefits (28), un- were given to the patients by their GPs, together with a confirmed patient identities (10) or other reasons (13). post-paid envelope addressed to us. The GPs were thus Since assessments of work ability when issuing new blinded to the patients’ responses. The study had a two- sickness certificates, or prolonging them, represent differ- stage design,12 as the selection of GPs represented the ent clinical problems, this study concerns new certificates first stage of data collection and the consultations only. Patients had been certified sick Ͻ8 days before between GPs and patients the second. Patients and GPs the completion of questionnaires in 408 cases. These were assured that participation in the study would not were defined as new episodes of sickness certification. affect the certificates. For episodes registered more than The patient response rate was 66% (268/408). once, only the first record was included in the analysis. Fifty-eight percent (49/85) of the invited GPs partici- pated. The number of patients recorded for each doctor Variables varied between four and 15. The age range of the doctors Patients and GPs answered the following question about was 28–78 years (mean 44 years) and 27% (13/49) were work ability using a five-point scale: “To what degree is women, which was not significantly different from non- your (the patient’s) ability to perform your (his or her) participating doctors. ordinary, remunerative work reduced today?” The The age range of patient respondents was 18–64 question was developed from a Graded Reduced Work years; 57% were women. Respondents were older Work ability assessed by patients and their GPs 141 than non-respondents, mean 40 years versus 36 years Work ability was assessed by the patients as significantly (P Ͻ 0.05). Gender and diagnoses were not significantly more reduced for sinusitis and back disorders than for all different between respondents and non-respondents. other disorders. The patient assessments were not Spearman’s rank correlation coefficients were Ͻ0.50 significantly associated with their own age or gender, nor in all the independent variables that were eligible for the with the gender of the GPs. logistic regression analyses. The GPs assessed work ability as more reduced the The representativeness of the study was checked by more their assessments were based on clinical findings, comparison with the National Sickness Benefit Register. and less reduced when influenced by non-medical factors Participating doctors had more female sick-listed (Table 3). patients than non-participating doctors, 57% versus 44% The information sources of the GPs for assessing work (P Ͻ 0.05). Age and diagnoses of patients in this study ability were based mainly on statements by the patients and of other patients of participating doctors recorded in in 66%, and on clinical findings in 34%. They assessed the register were not significantly different, nor were age work ability as significantly more reduced for depression and diagnoses of patients of participating and non- than for all other disorders. The GPs’ work ability assess- participating doctors recorded in the register. ments were not significantly associated with their own age or gender, or the age or gender of the patients.
Results Discussion The patients assessed work ability as very much or much reduced in 66%, moderately reduced in 23%, and The pilot study did not forward discussions concerning not much or hardly reduced at all in 11% of the cases interpretations of the questionnaire. We assumed that (Table 1). The corresponding assessments made by the the question about work ability had good face validity. GPs were 71, 27 and 2%. Others have found an internal consistency of 0.71 Patients and GPs agreed in 40% (107/266) of their (Cronbach’s alpha) concerning patient answers to the assessments, and had good agreement (± 1 answer cat- Graded Reduced Work Ability scale questions.10 Patients egory) in 81% (216/266) of the cases. Good agreement in this study and in the National Sickness Benefit Register ranged from 64% (16/25) in upper respiratory infections were similar with respect to age, gender and diagnoses. to 100% in depression (16/16). Work ability was assessed We found a distribution of diagnostic groups as in other as more reduced by patients than by GPs in 75 cases, studies.3,6 The reliability of diagnoses on sickness certifi- by more than one answer category in 15 cases, while cates is good; intra- and inter-observer concordance for corresponding figures, where GPs assessed work ability diagnostic groups is 97 and 91%, respectively.3 In a as more reduced than patients, were 84 and 35 (Table 1). Swedish study, coding from medical certificates to regis- Patients assessed work ability as more reduced the ters was correct in 98%, while diagnoses based on texts more stressful their jobs were, and the older their GPs in medical records matched diagnostic groups in 72%.13 were (Table 2). Categorizations and assessments presumably were Patients had a tendency to assess work ability as influenced by the attitudes and experiences of the doc- more reduced when their jobs were physically strenuous. tor. Intra-doctor correlation between the questionnaires
TABLE 1 Work ability assessed by patients and their GPs in new episodes of sickness certification in Aust-Agder County, Norway, 1996
Work ability assessed by GPs
Work ability assessed Very much Much Moderately Not much Hardly reduced Total (%) by patients reduced reduced reduced reduced at all
Very much reduced 39 35 12 1 – 87 (32.7) Much reduced 28 36 22 2 – 88 (33.1) Moderately reduced 8 19 32 3 – 62 (23.3) Not much reduced 7 5 2 – – 14 (5.3) Hardly reduced at all 4 7 4 – – 15 (5.6) Total (%) 86 (32.3) 102 (38.3) 72 (27.1) 6 (2.3) – 266a (100) a Two assessments by patients were missing. 142 Family Practice—an international journal
TABLE 2 Design-based logistic regression analyses of work ability TABLE 3 Design-based logistic regression analyses of work ability assessed by patients in new episodes of sickness certification according assessed by GPs in new episodes of sickness certification according to to work demands, age of GPs and diagnoses in Aust-Agder County, medical conditions in Aust-Agder County, Norway, 1996 Norway, 1996
Work ability assessed by GPs Work ability assessed by patients n OR 95% CI n OR 95% CI Clinical information sources Stress (scale 1–7) Statements by patients 268 1.00 No, little (1–2) 43 1.00 Clinical findings 138 2.75 1.65 4.60 Moderate (3–5) 168 2.00 0.99 4.03 Non-medical factors Much (6–7) 51 4.72 1.67 13.38 No 339 1.00 Physical work demands Yes 67 0.25 0.13 0.47 Mostly seated work 69 1.00 Diagnoses Varied work 175 1.13 0.69 1.88 All other disorders 61 1.00 Physically strenuous work 18 5.23 0.93 29.47 Musculoskeletal disorders Age of GPs (per 10 years) 262 1.50 1.06 2.13 Back 57 1.53 0.63 3.74 Diagnoses Neck/shoulder 27 0.63 0.21 1.83 All other disorders 45 1.00 Other 37 0.86 0.35 2.11 Musculoskeletal disorders Psychological disorders Back 38 5.09 1.60 16.19 Depression 27 3.31 1.03 10.59 Neck/shoulder 15 1.54 0.49 4.80 Neurosis and other 26 1.92 0.53 6.94 Other 28 2.19 0.83 5.78 Respiratory disorders Psychological disorders Influenza 35 1.47 0.58 3.71 Depression 16 1.57 0.44 5.63 Sinusitis 20 1.68 0.44 6.40 Neurosis and other 21 0.91 0.33 2.56 Upper respiratory tract infections 39 0.54 0.22 1.33 psychological disorders Other 28 1.79 0.66 4.83 Respiratory disorders Digestive disorders 18 1.16 0.24 5.66 Influenza 18 2.15 0.69 6.68 Injuries 31 1.70 0.50 5.78 Sinusitis 11 8.30 1.51 45.60 Upper respiratory tract infections 25 1.28 0.44 3.67 Odds ratio (OR) Ͼ1 indicates increased probability that work ability Other 16 1.54 0.46 5.25 was assessed as very much or much reduced. 95% confidence interval Digestive disorders 11 0.79 0.20 3.17 for OR (95% CI). n = 406 [two assessments were missing due to the lack of a report on assessed work ability (1) and diagnosis (1)]. Injuries 18 1.18 0.36 3.85 Clinical information sources, non-medical factors and diagnoses had univariate P-values of Ͻ0.20. Other medical conditions, work demands and occupations had univariate P-values of Ͼ0.20, and were Odds ratio (OR) Ͼ1 indicates increased probability that work ability excluded from the analyses. was assessed as very much or much reduced. 95% confidence interval Age and gender of patients and GPs were excluded in backward for OR (95% CI). n = 262 [six assessments were missing due to the stepwise multiple regression analyses. lack of a report on assessed work ability (2), stress (3) and physical work demands (1)]. Stress, physical work demands and diagnoses had univariate P-values of Ͻ0.20. Other work demands and medical conditions and occupations had univariate P-values of Ͼ0.20, and were excluded Most assessments by doctors and patients were very from the analyses. much or much reduced work ability. This should be ex- Patient age and gender of patients and gender of GPs were excluded pected, because the study sample concerned sick-listed in backward stepwise multiple regression analyses. patients only. GPs assessed work ability as more reduced than the patients did, but there were only a few conflicting assess- is therefore likely. Design-based analyses takes this into ments. This is in contrast to a study that compared patients’ account, as opposed to ‘model-based’ analyses.12 and social insurance medicine doctors’ assessments of Work ability and work demands of persons not sick- function after 26 weeks of disability benefits,11 in which listed were not recorded in the present study, which patients assessed work ability as more reduced than the should be considered when interpreting the results. doctors did. That study also found lower doctor–patient Work ability assessed by patients and their GPs 143 agreement than we did. This might be a reflection of Conclusions the different clinical situations between issuing a new The agreement between work ability assessments made certificate or prolonging one. by patients and GPs was high, despite patients’ assess- Good agreement in work ability assessments by patients ments being associated with work demands and GPs’ and GPs could be the result of negotiations that take with medical conditions. place in consultations concerning sickness certification.6,14 Psychosocial problems seem to be more common among work-disabled than non-work-disabled patients.15 GPs are good at identifying their patients’ work-related prob- Acknowledgements lems,16 and patients’ self-perceived work disability is the The study was approved by the Regional Ethics socio-economic factor that GPs are best at evaluating.17 Committee for Medical Research, the Norwegian Data Doctors frequently have to deal with such problems, as a Inspectorate and the Legal Affairs Division in the gatekeeper of social insurance benefits. Work-related National Insurance Administration. It was supported by problems might be easier to handle than more serious the Norwegian Ministry of Health and Social Affairs, psychosocial problems, and issuing sickness certificates Aust-Agder County Office of the National Insurance can represent ‘silent agreement’ solutions to unrevealed Services, the National Insurance Administration and the problems. University of Oslo. Work ability assessed by patients was associated with stress. The distribution of physical work demands in our study was similar to findings in other studies.14 Work ability assessed by GPs was not associated with the work References demands of the patients. As there were few conflicting work ability assessments by GPs and patients, increasing 1 Bonner D, Hooker I, White R. Non Means Tested Benefits: The the GPs’ knowledge about the work demands of their Legislation, Statutory Sick Pay (General) Regulations. London: Sweet and Maxwell, 1998. patients will not really reduce such conflict. 2 Timpka T, Hensing G, Alexanderson K. Dilemmas in sickness Why was work ability assessed by patients associated certification among Swedish physicians. Eur J Public Health with the age of their GPs? Perhaps older doctors convey 1995; 5: 215–219. more empathy and trust than younger doctors, making 3 Tellnes G, Svendsen KOB, Bruusgaard D, Bjerkedal T. Incidence of sickness certification. Proposal for use as a health status it easier for patients to assess their work ability as more indicator. Scand J Primary Health Care 1989; 7: 111–117. reduced. Is such a ‘Grandpa effect’ the explanation, or 4 Westin S. Becoming disabled: a sociomedical analysis of individual could it be that older doctors have lost their power of adaptations to life after long-term unemployment (dissertation). resistance concerning requests from patients about Trondheim: The Royal Norwegian Society of Sciences and Letters No. 2, 1990: 60–63. sickness certificates, so that patients in need of sick leave 5 Karasek R, Theorell T. Healthy Work: Stress, Productivity and the seek older doctors? One case history inquiry on sickness Reconstruction of Working Life. New York: Basic Books, 1990: certification might support the latter selection hypoth- 31–44. 6 esis, because younger doctors generally were more re- Tellnes G. Sickness certification in general practice: a review. Fam Pract 1989; 6: 58–65. 18 strictive towards sick-listing than their older colleagues. 7 Larsen BA, Førde OH, Tellnes G. Physician’s role in certification Work ability was assessed as most reduced in sinusitis for sick leave. Tidsskr Nor Lægeforen 1994; 114: 1442–1444 and back disorders by patients, and in depression by (in Norwegian, English summary). 8 GPs. These diagnostic groups are large and important Bentsen BG, Natvig B, Winnem M. Assessment of own functional status. COOP-WONCA charts in clinical work and research. 3,19 in sickness certification. Back disorders and depression Tidsskr Nor Lægeforen 1997; 117: 1790–1793 (in Norwegian, become increasingly important the longer the duration English summary). of the certification episodes.20 Can self-assessed work 9 Ilmarinen J, Tuomi K, Klockars M. Changes in the work ability of active employees over an 11-year period. Scand J Work Environ ability by patients with back pain be a good measure of Health 1997; 23 (Suppl 1): 49–57. function compared with other more objective methods? 10 Haaland EMH, Indahl A, Ursin H. Patients with low back pain not Self-evaluated work ability correlates well with clinically returning to work. A 12-month follow-up study. Spine 1998; 23: 1202–1208. determined musculoskeletal capacity in healthy women, 11 21 Swales K, Craig P. Evaluation of the Incapacity Benefit Medical Test. and in women with back disorders. In-house report 26. London: Social Research Branch, Analytical Work ability assessed by GPs was associated with Services Division, Department of Social Security, 1997. clinical findings, as in other studies.7 However, in a study 12 Lemeshow S, Letenneur L, Dartigues J-F, Lafont S, Orgogozo J-M, of low back pain, the proportion of patients without Commenges D. Illustration of analysis taking into account complex survey considerations: the association between wine objective findings increased with the duration of sickness consumption and dementia in the PAQUID study. Am J absence.22 Epidemiol 1998; 148: 298–306. The GPs seemed to focus first on medical conditions, 13 Ljungdahl LO, Bjurulf P. The accordance of diagnoses in a then on the functional consequences for work ability, as computerized sick-leave register with doctor’s certificates and medical records. Scand J Soc Med 1991; 19: 148–153. doctors are supposed to do according to sickness benefit 14 Tellnes G, Bruusgaard D, Sandvik L. Occupational factors in sick- legislation. ness certification. Scand J Primary Health Care 1990; 8: 37–44. 144 Family Practice—an international journal
15 Gulbrandsen P, Hjortdahl P, Fugelli P. Work disability and health- 19 Tellnes G. Duration of episodes of sickness certification. Scand J affecting psychosocial problems among patients in general Primary Health Care 1989; 7: 237–244. practice. Scand J Soc Med 1998; 26: 96–100. 20 Brage S, Nygård JF, Tellnes GT. Long-term Sickness Certification 16 Gulbrandsen P, Hjortdahl P, Fugelli P. General practitioners’ know- in Norway 1989–94. Report 98:3. Oslo: Department of Social ledge of their patients’ psychosocial problems: multipractice Insurance Medicine, University of Oslo, 1998 (in Norwegian). questionnaire survey. Br Med J 1997; 314: 1014–1018. 21 Eskelinen L, Kohvakka A, Merisalo T, Hurri H, Wägar G. 17 Gulbrandsen P, Fugelli P, Hjortdahl P. General practitioners’ Relationship between the self-assessment and clinical assess- knowledge of their patients’ socio-economic data and their ment of health status and work ability. Scand J Work Environ ability to identify vulnerable groups. Scand J Primary Health Health 1991; 17 (Suppl 1): 40–47. Care 1998; 16: 204–210. 22 Vällfors B. Acute, subacute and chronic low back pain: clinical 18 Condren L, Cox J, McCormick JS, Sullivan A. Certification of symptoms, absenteeism and working environment. Scand J unfitness for work. Ir Med J 1984; 77: 159–160. Rehab Med 1985; 17 (Suppl 11): 1–98. Paper II Sören Brage, Harald Reiso.
Arbeidsevne og kjønn – legers vurderinger av sykmeldte. (Work ability and gender – physicians’ assessment of sick-listed patients). Tidsskr Nor Lægeforen 1999; 119: 3737–40. Reprinted by courtesy of Tidsskrift for Den norske lægeforening
Paper III Harald Reiso, Jan F Nygård, Sören Brage, Pål Gulbrandsen, Gunnar Tellnes.
Work ability and duration of certified sickness absence. Scand J Public Health 2001; 29: 218–25.
Reprinted by courtesy of Scandinavian Journal of Public Health and Taylor & Francis
ORIGINAL ARTICLE
Work ability and duration of certi®ed sickness absence
Harald Reiso1,3, Jan F NygaÊrd2 ,So¨renBrage1 ,PaÊl Gulbrandsen1 and Gunnar Tellnes1
Department of General Practice and Community Medicine, 1Section for Occupational Health and Social Insurance Medicine, 2Section for Preventive Medicine and Epidemiology, University of Oslo, Oslo, 3Aust-Agder County Officeof the National Insurance Service, Arendal, Norway
Scand J Public Health 2001; 29: 218±225
Aims: The aim of this study was to examine the association between assessed work ability and the duration of certi®ed sickness absence. Methods: A total of 549 patients and 52 doctors provided questionnaire data about 549 episodes of absence. The episodes were classi®ed as new, one month, or three months according to their duration at the time of questionnaire completion. Their duration after that was used as outcome. Uni-and multivariate Cox regression analyses were performed. Results: In the multivariate analyses, a ``very much reduced’’ work ability assessed by patients was associated with a longer duration than a ``moderately reduced’’ work ability, in both one- and three-month episodes. Musculoskeletal and psycholo- gical disorders were associated with a longer duration, and respiratory disorders with a shorter duration than other disorders in new episodes. Patient age above 50 years was associated with a longer duration than lower age in new and three-month episodes. The doctors’ use of referral and tests in the consultations, and the presence of non-medical factors as judged by the patients, were associated with a longer duration than the absence of those factors in new episodes. The patients’ degree of job satisfaction, and non-medical factors as judged by doctors, were signi®cantly associated with duration only in univariate Cox regression analyses in new episodes. Work demands were not signi®cantly associated with duration in any of the analyses. Conclusions: Work ability assessed by patients may be a useful prognostic indicator of duration in prolonged episodes of certi®ed sickness absence. Further studies using other outcomes, such as disability pensioning, would be of interest to enlighten the concepts of work ability. Key words: duration, epidemiology, functional assessment, sick leave, sick listing, sickness absence, sickness certi®cation, work ability.
Harald Reiso, General Practice and Community Medicine Section for Occupational Health and Insurance Medicine, University of Oslo, PO Box 1130 Blindern, N-0318 Oslo, Norway. Tel: + 47 37 00 43 11, fax: + 47 37 00 43 01, e-mail: [email protected]
BACKGROUND beginning of absence episodes (2, 3). Reduced self- assessed work ability predicts that patients with low Sickness bene®ts account for more than 10% of the back pain, certi®ed sick for 8±12 weeks, do not return social insurance costs in Norway. A person must be to work (4). To our knowledge no other study has ``... incapacitated for work due to a loss of function looked at the predictive importance of assessed work that clearly is caused by disease or injury’’ in order to ability for the duration of sickness absence, or has been obtain such bene®ts. Similar texts of law are used in concerned with the prediction of duration at stages other western European countries, with disease and other than at the start of episodes. function as the key concepts. Maximum sickness bene®t time is 365 calendar days in Norway. Episodes of certi®ed sickness absence AIMS % exceeding 14 days account for 88 of the certi®ed The aim of our study was to examine the association absence days, according to the National Sickness between assessed work ability and the duration of new Bene®t Register. Factors that predict duration may be and prolonged episodes of certi®ed sickness absence. useful for targeting rehabilitation eVorts at groups of patients with increased probability of long-term sick- METHODS ness absence (1). Such predictors can thereby serve as decision-making instruments for doctors and social After a pilot study in 1994, the data collection took insurance oYcers. place between January and April 1996 in the county The age of patients certi®ed sick and diagnostic of Aust-Agder, southern Norway. The county had groups are known predictors of duration at the 100,211 inhabitants in 1996, of whom 47,835 aged
Scand J Public Health 29 Ñ Taylor & Francis 2001. ISSN 0301-7311 Work ability and duration of certi®ed sickness absence 219
16±66 years had sickness bene®t rights. All general (his or her) ordinary, remunerative work reduced practitioners (n=85) and industrial medical oYcers today?’’ (6). The question was developed from a (n=6) in the county were invited to participate in the Graded Reduced Work Ability scale, constructed for study. Each doctor was asked to record up to 25 con- the Norwegian Ministry of Health and Social AVairs secutive, new or prolonged, episodes of certi®ed sick- (4). ness absence in ordinary daytime practice. All types of The doctors classi®ed the main sickness certi®ca- contact were eligible. Questionnaires were given to the tion diagnoses according to ICPC (International patients by their doctors, together with a stamped Classi®cation of Primary Care) (7, 8), and they addressed envelope. The doctors were thus blinded to recorded the clinical information sources they used the patients’ responses. The data collection had a two- when assessing work ability, their use of clinical exam- stage design (5): the selection of doctors representing inations or tests in the consultations, and whether the ®rst stage, and the consultations between doctors patients were referred to specialists, hospitals, or and patients the second. Patients and doctors were physiotherapists, or were not referred. Patients and assured that participation would not aVect the issuing doctors also evaluated whether factors outside the pure of certi®cates. Episodes and patients registered more medical could have in¯uenced their work ability assess- than once were represented only by their ®rst records. ments, by the four-point scale question: ``Do you con- sider that non-medical factors are responsible for your (the patient’s) reduced work ability?’’ The question Time and duration variables, and outcome also had a free text option. An episode of certi®ed sickness absence, from its onset The patients gave information about their occupa-
(Tstart ) to the date it ended (Tstop) is illustrated as the tion (9), job satisfaction, job in¯uence, and work horizontal line in Figure 1. demands. The latter were operationalized as the
Tstop could take place by return to work, or by number of remunerative working hours per week, day- reaching maximum bene®t time. time or night/shift work, subjective description of phys- The time of questionnaire completion, i.e. the time ical work demands, heavy lifting, self-determined of consultation, was used as time zero (T0 ) in the breaks, con¯icts, stress due to working conditions, and analyses. This occurred at a particular time, but personal load by working overtime. because it refers to the onset (Tstart) of the episodes, it The doctors provided information about their own was measured as duration in weeks from Tstart. The and their patients’ age and sex. The type of contact remaining duration (Dpost) was measured in calendar was also recorded. days from T0 to Tstop. Dpost was used as the outcome. We were interested in the prediction of duration in Statistical analyses the early stages of episodes. Episodes whose duration from Tstart to T0 was >20 weeks were therefore not Continuous variables were compared using two-tailed included. The episodes were classi®ed as new, one t-tests, ordinal variables by Mantel±Haenszel tests for month, or three months, according to their duration trend, and categorical variables by chi-squared tests. from Tstart to T0 of 1±7 days, 1±5 weeks, or 6±20 The chosen level of signi®cance was 5%. weeks, respectively. Mean remaining duration (Dpost) for each category Data about time and duration were supplied from of variable was calculated for new, one-month, and The National Sickness Bene®t Register. three-month episodes. Univariate Cox analyses, and backward, stepwise multiple Cox regression analyses were performed for Determinants new, one-month, and three-month episodes. Age and Patients and doctors answered the following question sex of the patients, assessed work ability by about work ability using a ®ve-point scale: ``To what patients and doctors, and diagnoses were included in degree is your (the patient’s) ability to perform your the multivariate Cox analyses. Other variables with
Fig. 1. Time and duration concerning the episodes of certi®ed sickness absence.
Scand J Public Health 29 220 H Reiso et al. p-values <0.20 by change in Õ 2 log likelihood tests in In all, 57% (52/91) of the invited doctors particip- the univariate Cox regression analyses were also ated: 49 general practitioners and 3 industrial medical included. All episodes were monitored to Tstop. None oYcers. The age range of the participating doctors was thus censored. Intercorrelations between variables was 28±78 years (mean 44 years), and 25% (13/52) were checked by Spearman’s rank correlation coeY- were women; ®gures that were similar to that of cients; values <0.50 were considered acceptable. non-participating doctors. Work ability assessed as moderately, not much, or The doctors recorded an average of 11 episodes each hardly reduced at all was recoded as ``moderately (4±23) in the study sample of 549 cases. The age range reduced’’, while much and very much reduced were not of the patients was 17±65 years; 56% were women. The recoded. The ages of the patients and doctors, dia- representativeness of new episodes was checked by gnoses, and work-related variables were recoded into comparison with the National Sickness Bene®t groups. Register. Participating doctors had more female sick- listed patients than non-participating invited doctors: 55% versus 47% ( p<0.05). Age and diagnostic groups The sample of patients with new episodes, and patients of particip- The sampling is illustrated in Figure 2. ating and non-participating doctors recorded in the A total of 277 of the 549 cases in the study sample national register, were not signi®cantly diVerent. concerned new episodes, 124 one month, and 148 three months. The patient respondents in the primary sample of RESULTS 913 cases were signi®cantly older than non-respondents in new episodes (mean 40 years versus 36 years, A total of 77% of the cases concerned ordinary consul- p<0.05), but not in prolonged episodes. Sex and dia- tations, 20% telephone consultations, and 3% short gnostic groups of respondents and non-respondents in over-the-counter contacts. new and prolonged episodes were not signi®cantly The most frequent diagnoses were low back pain diVerent. (76 cases), shoulder/neck disorders (49), depressive
Fig. 2. Illustration of the sampling of cases. (``Maximum bene®t time’’ is the number of days a person has been certi®ed sick within the 6 months previous to the registered episode, and is included in the maximum sickness bene®t time of 365 calendar days.)
Scand J Public Health 29 Work ability and duration of certi®ed sickness absence 221 disorders (42), and upper respiratory/in¯uenza-like than ``moderately reduced’’ work ability in three- disorders (38 cases). month episodes (results not shown). None of the work Work ability was assessed as ``very much reduced’’ demands were signi®cantly associated with duration at by 33% (92/277), 40% (49/124), and 42% (62/148) of any stage of the episodes, not when the following were the cases in new, one-month, and three-month epis- omitted, individually or combined, from the multivari- odes, respectively ( p<0.05). Corresponding ®gures for ate analyses: diagnostic groups, work ability assessed the doctors’ assessments were 34%,27%,and28%. by doctors, or sex of the patients (results not shown). Patients and doctors considered non-medical factors to be present in 127 and 111 cases respectively. Some DISCUSSION 33 of 68 free text comments from the patients con- cerned con¯icts/problems at work or stressful working Method conditions, and 13 concerned problems, caregiving The pilot study did not raise discussions about inter- / tasks, or disease death among family members. The pretations of the questionnaire items. This may indicate corresponding ®gures for 54 comments from the good face validity of the questionnaire. An internal doctors were 19 and 23. consistency of 0.71 by Cronbach’s alpha has been reported for patient answers to the Graded Reduced Univariate Cox regression analyses Work Ability scale questions (4). Intra- and inter-observer concordance for diagnostic Work ability assessed by the patients was signi®cantly groups used on sickness certi®cates exceeds 90% (10). associated with the remaining duration of episodes of Coding from medical certi®cates to registers has been sickness absence (Dpost) by univariate Cox regression found to be correct in 98%, while diagnoses based on analyses in one- and three-month episodes (Table I). texts in medical records match diagnostic groups in The age of the patients, diagnostic groups, the doc- 72% of the cases (11). Responding patients in our tors’ use of referral, non-medical factors as judged by study were similar to non-responding patients, as were patients and doctors, and the patients’ degree of job participating to non-participating doctors. In all, 81% satisfaction, were signi®cantly associated with duration of sickness certi®cates are issued by general practi- by univariate Cox regression analyses in new episodes. tioners in Norway (10). Participating patients were similar to other sick-listed patients, according Multivariate Cox regression analyses to the national register. This indicates the good representativeness of our study sample. Y All Spearman’s rank correlation coe cients concerning Each patient contributed only one registration in the the variables eligible for the multivariate Cox regres- analyses, each doctor up to 23. The attitudes and < sion analyses were 0.50. Work ability assessed experiences of the doctors were probably re¯ected in Y by patients and doctors had correlation coe cients their work ability assessments, making intra-doctor of 0.27, 0.35, and 0.42, in new, one-month, and correlation between questionnaires likely. To our three-month episodes, respectively. knowledge, there are no survival analyses that can take Work ability assessed by patients as ``very much such intra-doctor correlations into account, as cross- reduced’’ was associated with a longer duration of sectional ``model-based’’ analyses can (5). However, sickness absence than work ability assessed as ``moder- another study, which examined the doctors’ attitudes ately reduced’’, in one- and three-month episodes to sickness certi®cation and the duration of absence, (Table II ). showed no association between attitudes and Musculoskeletal and psychological disorders were duration (12). associated with a longer duration, and respiratory dis- orders with a shorter duration, than other disorders in Work ability new episodes. Patient age above 50 years was associ- ated with a longer duration than a lower age in new Self-assessed work ability was associated with the dura- and three-month episodes. The doctors’ use of referral tion of prolonged episodes of sickness absence. In and tests in the consultations and the presence of non- another study, patients with psychiatric disorders and medical factors as judged by the patients were associ- lowered function measured by the Global Assessment ated with a longer duration than the absence of those Functional scale had more and longer sick-leave factors in new episodes. episodes than patients with psychiatric disorders with- Leaving out work ability assessed by the patients in out lowered function (13). Low measurements by other the multivariate analyses made work ability assessed functional assessment tools, such as The Finnish Work by the doctors as ``much reduced’’ and ``very much Ability Index, predict future disability (14, 15). reduced’’ associate signi®cantly with longer duration Can self-assessed work ability be as good a measure
Scand J Public Health 29 222 H Reiso et al.
Table I. Mean remaining duration (Dpost) of episodes of certi®ed sickness absence in new, one-month, and three-month episodesa (Aust-Agder County, Norway, 1996)
Episodes (n=549)
New One month Three months
n Mean Dpost n Mean Dpost n Mean Dpost
Age of patients in years 17±30 74 25 25 73 26 125 31±40 76 28 32 76 33 150 41±50 72 50 40 71 45 153 51±65 55 80 27 90 44 198 Total 277b 43 124 77 148 161 Sex of patients Male 126 40 52 73 65 151 Female 151 46 72 79 83 168 Total 277 43 124 77 148 161 Work ability assessed by patients Moderately reduced 92 36 35 49 30 111 Much reduced 93 43 40 76 56 143 Very much reduced 92 51 49 97 62 201 Total 277 43 124b 77 148b 161 Work ability assessed by doctors Moderately reduced 77 44 38 52 57 134 Much reduced 105 54 52 87 50 167 Very much reduced 95 31 34 88 41 190 Total 277 43 124 77 148 161 Diagnostic groupsc Musculoskeletal disorders 89 68 62 93 80 156 Psychological disorders 38 77 23 86 36 160 Respiratory disorders 74 12 7 45 ± ± Injuries 19 18 6 25 ± ± All other disorders 57 31 26 50 32 173 Total 277b 43 124 77 148 161 Referral No 259 41 112 78 122 161 Yes1877126526160 Total 277b 43 124 77 148 161 Non-medical factors (patients) No 212 35 87 74 112 159 Yes5976338435171 Total (11 missing) 271b 44 120 77 147 162 Non-medical factors (doctors) No 229 41 99 79 110 156 Yes4856256638175 Total 277b 43 124 77 148 161 Degree of job satisfaction (patients) High 190 34 91 69 93 141 Moderate 62 46 22 88 29 178 Low 23 105 10 122 26 189 Total (3 missing) 275b 43 123 77 148 161 aThe episodes had lasted 1±7 days (new), 1±5 weeks (1 month), or 6±20 weeks (3 months) at the time the questionnaires were completed. bp<0.05 by change in Õ 2 log likelihood tests in univariate Cox regression analyses. cRespiratory disorders and injuries were recoded as all other disorders in three-month episodes.
Scand J Public Health 29 Work ability and duration of certi®ed sickness absence 223
Table II. Reduced models from multivariate Cox regression analyses on duration in new, one-month, and three-month episodesa (Aust-Agder County, Norway, 1996)
Episodes
Newb (n=271) One monthc (n=124) Three months (n=148)
n HR 95% CI n HR 95% CI n HR 95% CI Age of patients in years 17±30 69 1.00 25 1.00 26 1.00 31±40 76 0.85 0.61±1.19 32 1.60 0.90±2.86 33 0.74 0.43±1.28 41±50 71 0.68 0.48±0.97 40 1.33 0.77±2.31 45 0.84 0.52±1.37 51±65 55 0.57 0.39±0.83 27 1.15 0.63±2.11 44 0.48 0.29±0.81 Sex of patients Male 124 1.00 52 1.00 65 1.00 Female 147 0.89 0.69±1.14 72 0.88 0.60±1.29 83 1.01 0.72±1.43 Work ability assessed by patients Moderately reduced 92 1.00 35 1.00 30 1.00 Much reduced 89 0.87 0.63±1.20 40 0.65 0.40±1.06 56 0.71 0.43±1.16 Very much reduced 90 0.76 0.54±1.08 49 0.48 0.29±0.80 62 0.44 0.26±0.76 Work ability assessed by doctors Moderately reduced 76 1.00 38 1.00 57 1.00 Much reduced 102 0.92 0.66±1.28 52 0.80 0.50±1.28 50 0.76 0.49±1.19 Very much reduced 93 1.21 0.85±1.70 34 0.75 0.44±1.28 41 0.76 0.48±1.21 Diagnostic groupd All other disorders 56 1.00 26 1.00 32 1.00 Musculoskeletal disorders 87 0.51 0.35±0.73 62 0.71 0.43±1.18 80 1.42 0.91±2.22 Psychological disorders 36 0.60 0.38±0.93 23 0.61 0.33±1.13 36 1.67 0.99±2.81 Respiratory disorders 73 1.66 1.14±2.42 7 1.66 0.68±4.01 Injuries 19 1.01 0.58±1.74 6 2.23 0.87±5.67 Referral No 254 1.00 Yes 17 0.55 0.33±0.93 Use of tests in the consultations No 199 1.00 Yes 72 0.70 0.51±0.93 Non-medical factors (patients) No 212 1.00 Yes 59 0.68 0.49±0.93
Hazard ratio (HR) <1 indicates increased probability of the episodes of having longer remaining duration (Dpost) than the reference category, >1 is shorter; 95% con®dence interval for HR (95% CI ). aThe episodes had lasted 1±7 days (new), 1±5 weeks (1 month), or 6±20 weeks (3 months) at the time the questionnaires were completed. bJob satisfaction, con¯icts between employees, job in¯uence, stress, non-medical factors as judged by doctors, and personal load by working overtime, were excluded. cUse of clinical examinations in the consultations and type of contact were excluded. dRespiratory disorders and injuries were recoded as all other disorders in three-month episodes. of function as other more objective methods? Self- functional restrictions caused by disease probably evaluated work ability correlates well with clinically measures a complex phenomenon, including not only determined musculoskeletal capacity in healthy the incapacity posed by ill-health and the patient’s life women, and in women with back disorders (16). situation, but also adaptation to the situation (21). In our view, similarities are likely between reliability The patients’ assessed work ability was more reduced measurements of self-assessed general health and self- in prolonged episodes than in new, while the doctors’ assessed work ability. Reduced self-assessed overall assessment was the reverse. Other workers have found health has a strong predictive impact on future disabil- that, after 26 weeks of disability bene®ts, function ity and death (17±19). The reliability of self-assessed diVered according to who assessed it: patients’ self- health is good (20). A general question concerning assessments were more reduced than the doctors’
Scand J Public Health 29 224 H Reiso et al. assessments (22). These diVerences may explain why duration in the univariate analyses, nor in the reduced the patients’ assessments in the present study were multivariate analyses. Few studies have taken dia- associated with duration in prolonged episodes, while gnoses into account when examining associations the doctors’ were not. between work demands and occurrence, or duration of sickness absence (25, 26), or have focused only on certain diagnostic groups, such as low back pain (4, Diagnoses and medical factors 27, 31). DiVerent conclusions from ours concerning The diagnostic groups were dominated by musculoske- work demands and duration may have been reached letal, psychological, and respiratory disorders, as found for that reason. previously (10). The ®nding that musculoskeletal and psychological disorders were associated with a longer CONCLUSIONS duration and respiratory disorders with a shorter dura- tion than other disorders in new episodes, parallels Work ability assessed by patients may be a useful pro- earlier ®ndings (3). We have shown in another study, gnostic indicator of duration in prolonged episodes of that patients assessed work ability as more reduced for certi®ed sickness absence. Further studies using other sinusitis and back disorders than for all other disorders, outcomes, such as disability pensioning, would be of while their general practitioners assessed work ability interest to enlighten the concepts of work ability. as more reduced for depression than for all other disorders, in new episodes of absence (6). ACKNOWLEDGEMENTS Referred patients had presumably more serious med- ical conditions than non-referred patients. This may The study was approved by the Regional Ethics explain why referral was associated with longer dura- Committee for Medical Research, the Norwegian Data V tion than non-referral in new episodes. Referral may Inspectorate, and the Legal A airs Division in the also postpone treatment because patients can be put National Insurance Administration. on long waiting lists. The study was supported by the Norwegian Ministry V Y The question about non-medical factors was inten- of Health and Social A airs, Aust-Agder county o ce ded to grasp elements that were not strictly medical, of the National Insurance Services, the National as opposed to the medical factors the text of law Insurance Administration, and the University of Oslo. demands, which could in¯uence the work ability assess- The authors are grateful to the patients and doctors ments. The patients considered such elements to be who participated in the study. working conditions in the majority of cases, while the doctors focused on the patients’ family problems and REFERENCES working conditions. 1. Tellnes G. Sickness certi®cation in general practice: a review. Fam Pract 1989; 6: 58±65. Age 2. Brage S, NygaÊrd JF, Tellnes G. The gender gap in muscu- A patient’s age is known to predict the duration of loskeletal-related long term sickness absence in Norway. Scand J Soc Med 1998; 26: 34±43. new sickness certi®cation episodes (3). In our study, 3. Tellnes G. Duration of episodes of sickness certi®cation. the patient’s age also predicted the duration of Scand J Prim Health Care 1989; 7: 237±44. prolonged episodes. 4. Haaland EMH, Indahl A, Ursin H. Patients with low back pain not returning to work. A 12-month follow-up study. Spine 1998; 23: 1202±8. Work demands 5. Lemeshow S, Letenneur L, Dartigues J-F, Lafont S, Orgogozo J-M, Commenges D. Illustration of analysis Sickness absence is associated with physical and psy- taking into account complex survey considerations: the chosocial work demands and job in¯uence, in addition association between wine consumption and dementia in to medical factors (23±30). We have previously shown the PAQUID study. Am J Epidemiol 1998; 148: 298±306. that patients assessed work ability as more reduced the 6. Reiso H, NygaÊrd JF, Brage S, Gulbrandsen P, Tellnes G. Work ability assessed by patients and their GPs in more stressful or physically strenuous their jobs were, new episodes of sickness certi®cation. Fam Pract 2000; in new episodes of absence (6). However, no signi®cant 17: 138±43. association was found in the present study between the 7. Brage S, Bentsen BG, Bjerkedal T, NygaÊrd JF, Tellnes operationalized elements of work demands that G. ICPC as a standard classi®cation in Norway. Fam we used and the duration of sickness absence. Job Pract 1996; 13: 391±6. 8. ICPC±InternationalClassi®cation of Primary Care, short satisfaction was signi®cantly associated with duration version. RTV form 3.12.c. of October 1992, changed in univariate analyses concerning new episodes only. May 1994. Oslo: National Insurance Administration, No other work-related factor was associated with 1994. (In Norwegian.)
Scand J Public Health 29 Work ability and duration of certi®ed sickness absence 225
9. The Norwegian Directorate of Labour. Classi®cation of 20. Lundberg O, Manderbacka K. Assessing reliability of a occupations. List of occupations with NCO codes measure of self-rated health. Scand J Soc Med 1996; (Nordic Classi®cation of Occupations). Oslo: The 24: 218±24. Norwegian Directorate of Labour, 1989. (In Norwegian.) 21. Manderbacka K. Methods. How do respondents under- 10. Tellnes G, Svendsen KOB, Bruusgaard D, Bjerkedal T. stand survey questions on ill-health? Eur J Public Health Incidence of sickness certi®cation. Proposal for use as a 1998; 8: 319±24. health status indicator. Scand J Prim Health Care 1989; 22. Swales K, Craig P. Evaluation of the Incapacity Bene®t 7: 111±17. Medical Test. In-house report 26. London: Social 11. Ljungdahl LO, Bjurulf P. The accordance of diagnoses Research Branch, Analytical Services Division, in a computerized sick-leave register with doctors’ certi- Department of Social Security, 1997. ®cates and medical records. Scand J Soc Med 1991; 23. Karasek R, Theorell T. Healthy work: stress, productiv- 19: 148±53. ity and the reconstruction of working life. New York: 12. Tellnes G, Sandvik L, Moum T. Inter-doctor variation Basic Books, 1990. in sickness certi®cation. Scand J Prim Health Care 1990; 24. Westin S. Becoming disabled: a sociomedical analysis of 8: 45±52. individual adaptations to life after long-term unemploy- 13. Hensing G, Spak F, Alexanderson K, Allebeck P. Sick- ment [dissertation]. Trondheim: The Royal Norwegian leave among women and the role of psychiatric disorder. Society of Sciences and Letters, No. 2, 1990. Scand J Soc Med 1997; 25: 185±92. 25. Luz J, Green MS. Sickness absenteeism from work ± a 14. Ilmarinen J, Tuomi K, Klockars M. Changes in the work critical review of the literature. Public Health Rev 1997; ability of active employees over an 11-year period. Scand 25: 89±122. J Work Environ Health 1997; 23 (suppl 1): 49±57. 26. North FM, Syme SL, Feeney A, Shipley M, Marmot M. 15. Tuomi K, Ilmarinen J, Setsamo J, Huuhtanen P, Psychosocial work environment and sickness absence Martikainen R, NygaÊrd CH, et al. Summary of the among British civil servants: the Whitehall II study. Am J Public Health 1996; 86: 332±40. Finnish research project (1981±1992) to promote the 27. Hemingway H, Shipley MJ, Stansfeld S, Marmot M. health and working ability of ageing workers. Scand Sickness absence from back pain, psychosocial work J Work Environ Health 1997; 23 (suppl 1): 66±71. characteristics and employment grade among oYce 16. Eskelinen L, Kohvakka A, Merisalo T, Hurri H, workers. Scand J Work Environ Health 1997; 23: 121±9. WaÈgar G. Relationship between the self-assessment and 28. Marmot M, Feeney A, Shipley M, North F, Syme SL. clinical assessment of health status and work ability. Sickness absence as a measure of health status and func- Scand J Work Environ Health 1991; 17 (suppl 1): 40±7. tioning: from the UK Whitehall II Study. J Epidemiol 17. Fylkesnes K, Fùrde OH. The Tromsù study: predictors Community Health 1995; 49: 124±30. of self-evaluated health ± has society adopted the 29. Vahtera J, Pentti J, Uutela A. The eVect of objective job expanded health concept? Soc Sci Med 1991; 32: 141±6. demands on registered sickness absence spells: do per- 18. Ostlin P. Occupational history, self reported chronic ill- sonal, social and job-related resources act as moderators? ness, and mortality: a follow up of 25,586 Swedish men Work and Stress 1996; 10: 286±308. and women. J Epidemiol Community Health 1990; 44: 30. Tellnes G, Bruusgaard D, Sandvik L. Occupational fac- 12±16. tors in sickness certi®cation. Scand J Prim Health Care 19. Miilunpalo S, Vuori I, Oja P, Pasanen M, Urponen H. 1990; 8: 37±44. Self-rated health status as a health measure: the predict- 31. Van der Weide WE, Verbeek JH, Salle HJ, van Dijk ive value of self-reported health status on the use of FJ. Prognostic factors for chronic disability from acute physician services and on mortality in the working-age low-back pain in occupationalhealth care. Scand J Work population. J Clin Epidemiol 1997; 50: 517±28. Environ Health 1999; 25: 50±6.
Scand J Public Health 29 Paper IV Harald Reiso, Jan F Nygård, Pål Gulbrandsen, Sören Brage, Gunnar Tellnes.
To ask the right question at the right time: When is the patient’s self-assessed work ability most accurate as a predictor of the remaining duration of certified sickness absence? Norsk Epidemiologi 2003; 13: 297–302.
Reprinted by courtesy of Norwegian Epidemiology
Norsk Epidemiologi 2003; 13 (2): 297-302 297 To ask the right question at the right time: When is the patient’s self-assessed work ability most accurate as a predictor of the remaining duration of certified sickness absence?
Harald Reiso, MD1,2, Jan F. Nygård, BSc, BA3,4, Pål Gulbrandsen, MD, PhD1, Sören Brage, MD, PhD5 and Gunnar Tellnes, MD, PhD1
1) Department of General Practice and Community Medicine, Section for Occupational Health and Social Insurance Medicine, University of Oslo 2) Aust-Agder County Office of the National Insurance Service, Arendal, Norway 3) Department Group of Basic Medical Sciences, University of Oslo, Norway 4) The Cancer Registry of Norway 5) National Insurance Administration, Oslo, Norway Correspondence: Harald Reiso, Aust-Agder County Office of the National Insurance Service, PO Box 1853 Stoa, N-4858 Arendal, Norway Telephone: + 47 37 00 43 11 Telefax: + 47 37 00 43 01 E-mail: [email protected]
ABSTRACT Aim: To present a model that estimates when predictors most accurately predict the remaining duration of certified sickness absence. Methods: A questionnaire was given to 549 patients who were certified sick, and to their doctors. The questionnaires were completed at consultations that occurred from one to 20 weeks after the start of absence. The duration of the episodes of certified sickness absence was provided by the National Sickness Benefit Register. Four separate analyses were performed for self-assessed work ability, diagnoses, age and gender of the patients. The difference between observed remaining duration of the episodes of absence and their estimated expected duration, in percent, served as a measure of predictive accuracy. The improvement of predictive accuracy, compared with a variable that had no association with the outcome (a constructed comparison variable), was estimated for each of the variables of interest, by weeks of consultations. Results: Self-assessed work ability had no improved predictive accuracy in consultations at one week. It was 10% better at eight weeks, reached a maximum of 12% at 13 weeks, and was 10% better at 20 weeks. The predictive accuracy of diagnoses was at its best of 22% in consultations at one week, and at a mini- mum of 9% better at 11 weeks. The predictive accuracy of age was 7% better at one week, and 10% at 20 weeks. Gender showed no improved predictive accuracy. Conclusions: Self-assessed work ability is most accurate as a predictor in the interval between eight and twenty weeks from the start of absence, diagnoses are most accurate in the first weeks.
Key words: duration (of certified sickness absence), epidemiology, functional assessment, method, predictive accuracy, sick-leave, sick-listing, sickness absence, sickness certification, work ability
BACKGROUND changes over time. A factor with a low or unknown predictive accuracy is of limited use as a prognostic Medicine has to deal with new predictive factors, as tool, and even strong predictors may have low predic- advances in epidemiology, statistics and computing tive accuracy (3). power have made increased use of prognostic models How the predictive accuracy of factors change over possible (1-3). Predictive factors have certain predic- time may provide important information for deciding tive magnitudes, measured by rate or hazard ratios. which interventions to choose, and at what time to use These ratios indicate whether values of independent them. For example, when do measurements of the variables are predictive, relative to reference catego- functional consequences of a disease most accurately ries, or not. Analyses may be performed at different serve as a predictor of the remaining duration of times during e.g. a disease, or an episode of certified certified sickness absence? This information could be sickness absence (4,5). Such an approach does not give useful for doctors to initiate therapy and rehabilitation measures about how useful a prognostic factor per se at the right time (6). Analyses that estimate the predic- is, that is how accurately the factor predicts for tive accuracy of variables, and how it changes during example duration of absence, nor how that accuracy the course of a disease, may thus supply useful prog- 298 H. REISO M.FL. nostic information. Definitions We have shown in a previous study that self- • Accuracy is defined as the degree to which a mea- assessed work ability predicts the remaining duration surement, or an estimate based on measurements, of certified sickness absence in episodes that have represents the true value of the attribute that is lasted more than one week at the time the assessments being measured (8). are made, in a representative group of patients in • Predictive accuracy in this paper is defined as the general practice (7). Diagnostic groups predict remai- difference in percent between model predictions of ning duration at the start of episodes, but not later, duration (expected remaining duration) and obser- while the age of patients predicts remaining duration ved remaining duration of episodes of certified sick- both in new and prolonged episodes. The gender of the ness absence. patients is not predictive of duration. However, • Improved predictive accuracy is defined as the im- although self-assessed work ability, diagnoses and the provement of predictive accuracy for the variable of age of patients are significant predictors of duration of interest (estimated in percent), compared with a absence, their change of predictive accuracy by time variable that had no association with the outcome (a from the start of sickness absence is not known. And constructed comparison variable). thus it is not certain when assessments of these pre- dictors should be made. Variables
AIM The patients answered the following question about work ability: “To what degree is your ability to The aim of this study was to present a model that perform your ordinary, remunerative work reduced estimates when predictors most accurately predict the today?” The answer categories were: “Very much re- remaining duration of certified sickness absence. duced”, “much reduced”, “moderately reduced”, “not much reduced” and “hardly reduced at all” (9). As the METHODS distribution of these categories was strongly skewed, work ability assessed as moderately, not much, or Data collection hardly reduced at all was recoded to moderately re- Data collection took place January-April 1996 in the duced. The other options were left unchanged. county of Aust-Agder, southern Norway (7). The The doctors classified the main sickness certifi- county had 100,211 inhabitants in 1996. All general cation diagnoses according to ICPC (International practitioners and company doctors in the county were Classification of Primary Care) (10). Musculoskeletal, invited to participate in the study. Each doctor recor- psychological, respiratory disorders, and injuries are ded episodes of certified sickness absence made during important groups of diagnoses in sickness absence (11, office hours. Patients who were issued new sickness 12). The diagnoses were therefore recoded into those certificates, or certificates for prolongation of ongoing groups, and all other disorders. The age of patients was episodes of absence, were consecutively included in recoded into four groups; 17-30 years, 31-40, 41-50 the study. No instruction was given that a certain time and 51-65 years. of absence episodes should be given special attention. An episode of certified sickness absence ends by Patients and doctors completed questionnaires inde- return to work, by reaching maximum sickness benefit pendent of one another, and were assured that partici- time – which is 365 calendar days in Norway, or by pation in the study would not interfere with absence transition to other benefits. The observed remaining status. For episodes recorded more than once, only the duration of absence, from the time of the consultation first was included in the analyses. to the end of the episode, was measured in calendar This study focused on prediction of duration early days. The National Sickness Benefit Register supplied in the course of absence. Episodes with duration of duration data (13). No episode was lost to follow-up. more than 20 weeks at the time of consultations were therefore excluded. Analyses For details, see Appendix. In short, linear regressions, The sample stratified by weeks of consultation (one to 20 weeks), From a sample of 567 paired questionnaires from were used to estimate the expected remaining duration patients and doctors, 11 were excluded because they for each episode of certified sickness absence, based concerned patients certified sick while on rehabilita- on observed remaining duration (Stage I analyses). tion benefits, and seven were excluded for other rea- Self-assessed work ability, diagnoses, age and gender sons (7). The study sample consisted of questionnaires of the patients were independent variables in four from 549 patients, 49 general practitioners, and three separate analyses. company doctors. The doctors recorded an average of The difference in percent between expected and 11 episodes each (minimum 4, maximum 23). The observed remaining duration served as a measure of patients’ ages ranged from 17-65 years (mean 41 the accuracy of the prediction. These accuracy years), and 56% were women. measures were used in Stage II linear regressions with PREDICTIVE ACCURACY 299
weeks of consultation. Because the possible range of minimum of 9% better at 11 weeks. The predictive observed remaining duration after consultation is asso- accuracy of age was 7% better at one week, and 10% ciated with the duration of certified sickness absence at 20 weeks. Gender showed no improved predictive before consultation, predictive accuracy is time- accuracy. dependent, which was corrected for by including second-order coefficients (see Appendix for details). Furthermore, estimations of how much better the predictive accuracy for the variables of interest were, compared with a variable that had no association with the outcome (a constructed comparison variable), were performed (see Appendix for details). The improved predictive accuracy was given in percent.
RESULTS Work ability was assessed by the patients as mode- rately reduced in 29% of the cases, much reduced in 34%, and very much reduced in 37% (Table 1). The diagnostic groups were musculoskeletal disorders in 42% of the cases, psychological disorders in 18%, respiratory disorders in 16%, injuries in 5%, and all other disorders in 19%. Self-assessed work ability had no improved predic- Figure 1. The improved predictive accuracy of self-assessed tive accuracy in consultations at one week. It was 10% work ability, diagnoses, age and gender of the patients, by better at eight weeks, reached a maximum of 12% at time of consultations (weeks). Improved predictive accuracy is the improvement of predictive accuracy for the variable of 13 weeks, and was 10% better at 20 weeks (Figure 1). interest (estimated in percent), compared with a variable that The predictive accuracy of diagnoses was at its had no association with the outcome (a constructed compa- best of 22% in consultations at one week, and at a rison variable). N = 549. Aust-Agder County, Norway, 1996.
Table 1. The distribution of the patients’ self-assessed work ability, diagnoses, age and gender of the patients, and the mean observed remaining duration of the episodes of certified sickness absence (with 95% confidence intervals - 95% CI), by weeks of consultations. N = 549 (Aust-Agder County, Norway, 1996).
Consultations (weeks from start of absence) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Work ability (Total) (%) Moderately reduced 92 14 12 9 – 3 – 3 2 1 2 3 1 5 1 1 1 1 2 4 (157) (28.6) Much reduced 93 11 19 6 4 3 3 5 6 5 5 2 6 3 3 6 3 2 3 1 (189) (34.4) Very much reduced 92 14 13 13 9 5 5 7 2 6 3 7 5 3 2 1 5 6 4 1 (203) (37.0) (549) (100.0) Diagnoses Musculoskeletal disorders 89 20 23 12 7 7 5 11 5 6 4 5 9 5 4 4 7 1 4 3 (231) (42.1) Psychological disorders 38 6 10 5 2 1 2 1 3 4 3 1 2 5 1 2 1 6 2 2 (97) (17.7) Respiratory disorders 74 4 1 2 – – 1 – – 1 1 2 – – 1 – – 1 – – (88) (16.0) Injuries 19 1 1 3 1 1 – 1 1 – – – 1 – – – – – 1 – (30) (5.5) All other disorders 57 8 9 6 3 2 – 2 1 1 2 4 – 1 – 2 1 1 2 1 (103) (18.7) (549) (100.0) Age 17-30 years 74 7 6 9 3 4 – 2 1 5 1 3 2 – – 2 2 1 2 1 (125) (22.8) 31-40 years 76 13 12 3 4 2 4 4 3 3 2 2 2 4 1 2 – 2 – 2 (141) (25.7) 41-50 years 72 16 11 7 6 3 3 5 3 – 3 2 4 4 2 3 4 4 3 2 (157) (28.6) 51-65 years 55 3 15 9 – 2 1 4 3 4 4 5 4 3 3 1 3 2 4 1 (126) (22.9) (549) (100.0) Gender Male 126 14 19 14 5 10 6 8 5 4 3 7 4 2 1 5 1 4 4 1 (243) (44.3) Female 151 25 25 14 8 1 2 7 5 8 7 5 8 9 5 3 8 5 5 5 (306) (55.7)
Total number of observations 277 39 44 28 13 11 8 15 10 12 10 12 12 11 6 8 9 9 9 6 (549) (100.0) Mean observed remaining duration (calendar days) 43 67 78 81 91 168 203 156 157 148 94 168 157 147 192 150 198 186 164 150 (95% CI lower) 34 33 49 47 54 101 115 89 95 76 52 107 100 86 113 86 147 126 114 88 (95% CI upper) 53 102 107 114 127 235 291 224 218 221 137 229 213 208 270 214 250 246 215 212 300 H. REISO M.FL.
DISCUSSION Main findings When trying to assess the prognosis of a disease it is Method necessary to take the time of the assessment, relative to Participating patients were similar to other patients the start of the disease, into account. Prognostic tools certified sick regarding age, gender and diagnoses should be predictive and accurate. We consider an im- (7,9). Responding patients were similar to non- proved predictive accuracy of self-assessed work abili- responding patients, as were participating to non- ty of 12%, 13 weeks after the start of certified sickness participating doctors. The National Sickness Benefit absence, to be of possible importance in clinical prac- Register in Norway serves as an account system for tice. Diagnoses had an improved predictive accuracy at the payment of sickness benefits, and is therefore the beginning of absence, age had a gradual increase of revised, audited and quality controlled. The register accuracy by time of consultations, while gender was also ensures good follow-up of the cases. None were neutral. Thus, different information has different lost to follow-up in this study. prognostic value by time. There were few observations in consultations at Other possible applications more than four weeks (Table 1). This should be con- sidered when interpreting the results, because an esti- Methods estimating predictive accuracy may enable mation based on one observation yields high predictive doctors to make more confident statements about prog- accuracy, i.e. estimated equals observed remaining noses, based on measurements performed at different duration. times during the course of diseases. The presented The presented study resembles meta-analysis, be- model may thus be useful in other fields of medicine cause it can be considered as 20 different studies, one that deal with time-dependent data. For example, when for each week, and the regression a meta-analysis that during hepatitis do the levels of liver enzymes most estimates an overall predictive accuracy for the vari- accurately predict the duration of the disease? able of interest. A test for homogeneity would check for differences of accuracy between the studies, but CONCLUSIONS only yield a dichotomous answer (14). Cox models Self-assessed work ability is most accurate as a with time-dependent co-variates, and non-parametric predictor of the remaining duration in the interval be- regression spline models, can take changes of hazard tween eight and twenty weeks from the start of ab- that may occur after the assessments of the predictors sence, diagnoses are most accurate in the first weeks. into account, by using repeated measurements of the predictor (3, 4). Such approaches applied to our data ACKNOWLEDGEMENTS would neither measure how accurately the variable of interest predicts duration, nor how that accuracy chan- The authors thank Håkon K. Gjessing for useful discussions. The Norwegian Ministry of Health and Social Affairs ges by the time of consultation, and would represent supported the study financially. The study was approved by even more complicated analyses than the ones presen- the Regional Ethics Committee for Medical Research, the ted here. Our approach estimates the change in associ- Norwegian Data Inspectorate, and the Legal Affairs Division ation over time. of the National Insurance Administration.
APPENDIX
Stage I analyses
Linear regressions were used to estimate the expected remaining duration (Dexpected) for each episode of absence, for the variables of interest, by when consultations occurred (T0=1,2,…20 weeks), based on the observed remai- ning duration (Dobserved) of the episode. 20 separate, ordinary linear regression analyses were performed with self- assessed work ability, diagnoses and age and gender of the patients as independent variables in four sets of analyses. The ratio of expected remaining duration to observed remaining duration (Dexpected / Dobserved), served as a measure of the accuracy of the prediction. Because a regression yields the mean of the dependent variable as the optimal expected value, approximately half of the observations will be longer, and half will be shorter than the expected value. The direction of such a deviation was not our interest. We wanted a measure of deviation from perfect accuracy. Values of Dexpected / Dobserved of less than 1.0 were kept unchanged, while values larger than 1.0 were inverted. The portions were multiplied by 100, which gives the deviation from perfect accuracy in percent. PREDICTIVE ACCURACY 301
Stage II analyses Correction for time dependency Because there is a maximum possible duration of certified sickness absence of 365 days, there is a non-linear increase in accuracy depending on the time of consultation. At consultation of one week (T0 = 1), observed remaining duration can vary between one and 52 weeks, while at 20 weeks it can vary between one and 32 weeks. Observed remaining duration thus depends on when the consultation occurs, as do expected remaining duration. The probability that expected remaining duration equals observed remaining duration is less likely by coincidence at consultation at one week than at 20 weeks, and the predictive accuracy therefore also depends on when the consultation occur. The difference between predictive accuracy for a variable of interest in consultations at 19 and 20 weeks is larger than the difference at one and two weeks. Therefore the change in predictive accuracy, by when consultations occur, is non-linear. The simplest non-linear model, using a second order factor, was introduced in the Stage II analyses to correct for this. Linear regressions with second order factors were fit to the Stage I accuracy measures and time of consul- tations (one to 20 weeks) for the variables of interest (Stage II analyses). The transformed measures of Dexpected / Dobserved was the dependent variable, T0 with second order factor was the independent variable. The resul- ting second (b2) and first (b1) order coefficients, and the constant term (k), are presented in Table 2.
Table 2. The results of ordinary linear regression analyses, with predictive accuracy in consultations at one to 20 weeks as dependent variable (see Equation 1 and 2), for the variables of interest and the comparison variable (coefficients are multi- plied by 100 to ease their readability). Aust-Agder County, Norway, 1996.
n b2 95% CI of b2 b1 95% CI of b1 k 95% CI of k Work ability assessed by patients 549 -0.14 -0.22 – -0.05 5.18 3.73–6.62 23.4 19.9–26.9 Diagnostic groups 549 -0.07 -0.16 – 0.02 3.60 2.00–5.19 31.6 27.7–35.4 Age-groups of patients 549 -0.11 -0.19 – -0.02 4.44 2.92–5.95 26.6 23.0–30.3 Gender of patients 549 -0.10 -0.18 – -0.01 3.99 2.53–5.46 24.9 21.3–28.4 Comparison variable -0.10 4.06 24.9
These b2s, b1s, and ks were used to calculate the predictive accuracy for the variable of interest (X), by weeks of consultation (T0), as noted in Equation 1. = b 2 + b + (T =1,2,…,20) Equation 1 AccuracyX (T0) 2,X T0 1,X T0 kX 0
Estimation of improved predictive accuracy In order to yield a measure of how much better the predictive accuracy for the variable of interest was, compared with a variable that had no association with the outcome, a comparison variable was constructed. Variables with random values with a normal distribution, a mean of zero and a standard deviation of one, and with a uniform distribution between zero and one, were used in that construction. The means of 1000 Stage I analyses, by weeks of consultation (one to 20), using 1000 different generated variables with random values, were used in Stage II analyses. The means represent the most probable values for the predictive accuracy of any variable, by when the consultations occurred. The use of normally or uniformly distributed random values gave similar means. The uniform distribution was arbitrarily chosen. The resulting b2, b1 and k are presented in Table 2. The estimation of the comparison variable (Z), is expressed in Equation 2, where l is the number of regressions, and j each episode at the time of consultations.
l=1000 l=1000 l=1000 Âbˆ Âbˆ Âkˆ 2,T0 , j 1,T0 , j T0 , j ˆ = l=1 2 + l=1 + l=1 Equation 2 Dexpected (Z) ( )Z ( )Z ( ) T0, j 1000 1000 1000
The predictive accuracy for the variable of interest (X) was compared with the comparison variable (Z), yielding improved accuracy, as noted in Equation 3. b T 2 + b T + k Improved accuracy (T ) = 2,X 0 1, X 0 X (T0=1,2,…,20) Equation 3 0 b 2 + b + 2,Z T0 1,Z T0 kZ
This measure was subtracted 1.00, which means that equal values of the ratio components yielded zero. The deviation from zero was plotted against when the consultations occurred (Figure 1), which visualized the improvement in percent in predictive accuracy for the variables of interest. 302 H. REISO M.FL.
REFERENCES
1. Greenland S. Principles of multilevel modelling. Int J Epidemiol 2000; 29: 158-67. 2. Aalen OO. Medical statistics – no time for complacency. Stat Methods Med Res 2000; 9: 31-40. 3. Schemper M, Henderson R. Predictive accuracy and explained variation in Cox regression. Biometrics 2000; 56: 249-55. 4. Heinzl H, Kaider A, Zlabinger G. Assessing interactions of binary time-dependent covariates with time in Cox proportional hazards regression models using cubic spline functions. Stat Med 1996; 15: 2589-601. 5. Quantin C, Abrahamowicz M, Moreau T, Bartlett G, MacKenzie T, Tazi MA, et al. Variation over time of the effects of prognostic factors in a population-based study of colon cancer: Comparison of statistical models. Am J Epidemiol 1999; 150: 1188-200. 6. http://www.dep.no/sos/norsk/publ/utredninger/NOU/ Governmental report NOU 2000: 27, Sickness absence and disability pension, an including working-life (in Norwegian) (13.02.2003). 7. Reiso H, Nygård JF, Brage S, Gulbrandsen P, Tellnes G. Work ability and duration of certified sickness absence. Scand J Public Health 2001; 29: 218-25. 8. Last JM. A Dictionary of Epidemiology. Third edition. New York: Oxford University Press, 1995. 9. Reiso H, Nygård JF, Brage S, Gulbrandsen P, Tellnes G. Work ability assessed by patients and their GPs in new episodes of sickness certification. Fam Pract 2000; 17: 139-44. 10. Brage S, Bentsen BG, Bjerkedal T, Nygård JF, Tellnes G. ICPC as a standard classification in Norway. Fam Pract 1996; 13: 391-6. 11. Söderberg E, Alexanderson K. Literature study – published studies of the interface between medical practice and assessments of social insurance legislation. Försäkringsmedicinskt centrum, Rapport 2001: 1. Linköping, 2001 (in Swedish). 12. Tellnes G, Svendsen KOB, Bruusgaard D, Bjerkedal T. Incidence of sickness certification. Proposal for use as a health status indicator. Scand J Primary Health Care 1989; 7: 111-7. 13. Brage S, Nygård JF, Tellnes G. The gender gap in musculoskeletal-related long term sickness absence in Norway. Scand J Soc Med 1998; 26: 34-43. 14. Fleiss JL. The statistical basis for meta-analysis. Stat Methods Med Res 1993; 2: 121-45. Paper V Harald Reiso, Pål Gulbrandsen, Sören Brage.
Doctors’ prediction of certified sickness absence. Fam Pract 2004; 21: 192–8.
Reprinted by courtesy of Family Practice and Oxford University Press
Family Practice Vol. 21, No. 2 © Oxford University Press 2004, all rights reserved. Printed in Great Britain Doi: 10.1093/fampra/cmh216, available online at www.fampra.oupjournals.org
Doctors’ prediction of certified sickness absence Harald Reisoa,b, Pål Gulbrandsenc and Sören Braged
Reiso H, Gulbrandsen P and Brage S. Doctors’ prediction of certified sickness absence. Family Practice 2004; 21: 192–198. Background. Doctors’ ability to predict the duration of their patients’ certified sickness absence is important for follow-up efforts aimed at patients with increased probability of long- term absence. Objectives. The aim of this study was to examine the accuracy of doctors’ predictions of their patients’ sickness absence status 4 weeks ahead, and which factors were associated with it. Methods. A questionnaire survey was carried out in primary health care concerning 796 patients certified sick within 140 days after the start of absence. The episodes of absence were labelled short-standing (up to 2 weeks) and long-standing (from 3 to 20 weeks), at the time of consultation. The doctors’ prediction of the patients’ absence status 4 weeks ahead, diagnoses, work ability, clinical information sources used and the presence of non-medical factors that could have influenced the doctors’ work ability assessments were collected. The predictions were compared with the patients’ absence status 4 weeks later by positive predictive values (PPVs) for the statements ‘returned to work’ and ‘still certified sick’. Factors associated with the accuracy of the predictions were analysed by multiple logistic regression analyses. Results. The doctors accurately predicted return to work in 84% [95% confidence interval (CI) 79–87] of the cases in short-standing episodes, and in 53% (43–62) in long-standing episodes. The corresponding PPVs for still certified sick were 72% (62–80) and 91% (85–94). In short- standing episodes, the doctors’ probability of making accurate predictions was higher for respiratory disorders [odds ratio (OR) 2.84; 95% CI 1.36–5.90], than for the reference category ‘all other disorders’, and lower for mental disorders (0.46; 0.24–0.89). In long-standing episodes, the probability was lower for musculoskeletal disorders (0.33; 0.12–0.86) and injuries (0.12; 0.03–0.48). Neither the age nor gender of patients or doctors, nor the degree of work ability reduction, nor other factors were associated with the accuracy of the predictions. Conclusions. The doctors’ predictions were highly accurate for return to work in short-standing episodes, and for still certified sick in long-standing episodes. Diagnoses were associated with the accuracy; other factors, including the doctors’ work ability assessments, were not. Keywords. Prediction, sick listing, sickness absence, sickness certification, work ability
Introduction medical problems and the consequences these problems have for the patients’ work ability. In Norway, employees Doctors act as gatekeepers for sickness benefits, in can certify themselves sick 12 days per year, but usually accordance with sickness benefit legislation.1–3 Twenty not for more than three consecutive days, without a per cent of consultations in Norwegian general practice sickness certificate issued by a doctor. Employers pay for involve certified sickness absence.4–6 In these the first 16 calendar days of certified sickness absence consultations, the doctors assess the importance of (14 days was the rule until April 1998), and the National Sickness Benefit Scheme covers further absence, to a maximum of 365 calendar days. Sickness benefits can be Received 7 April 2003; Revised 15 October 2003; Accepted graded between 100 and 20%. The benefits equal 3 November 2003. aDepartment of General Practice and Community Medicine, ordinary salary, up to a maximum of NOK 341 166 per Section for Occupational Health and Social Insurance Medicine, year (~£28 600, €41 400) (at 1 May 2003). University of Oslo, bAust-Agder County Office of the National Doctors’ ability to predict the length of sickness Insurance Service, Arendal, cHELTEF: Centre for Health absence is important in enabling them to initiate Services Research, PO Box 55 No-1474 Nordbyhagen, Norway d therapy and rehabilitation for patients with increased and National Insurance Administration, Oslo, Norway. 1,7–9 Correspondence to H Reiso at Aust-Agder County Office of the probability of long-term sickness absence. National Insurance Service, PO Box 1853 Stoa, N-4858 Arendal, Norwegian doctors are requested by the National Norway; E-mail: [email protected] Insurance to estimate how many weeks their patients
192 Doctors’ prediction of certified sickness absence 193 will be certified sick. We do not know how accurate to perform his/her ordinary, remunerative work reduced doctors’ predictive assessments are, or how medical today?’ The answer categories were: ‘very much factors are associated with them.10 reduced’, ‘much reduced’, ‘moderately reduced’, ‘not Primary health care doctors have experience in much reduced’ and ‘hardly reduced at all’. Due to few assessing their patients’ need for certified sickness observations in some of the answer categories, work absence.5,11 A recent study in Norway has indicated that ability assessed as moderately, not much or hardly GPs are good at identifying their patients’ work-related reduced at all was recoded as ‘moderately reduced’, problems,12 and at evaluating the patients’ work while much and very much reduced were kept disability.13 Doctors thus probably have a good idea of unaltered. how long their patients will be certified sick, based on The doctors recorded the clinical information sources knowledge about the patients and their diseases. Factors they used when assessing work ability as: ‘the patient’s associated with predictions may be the patients’ age self-reported work ability only’, ‘mainly self-report, and gender, the patients’ work ability, the doctors’ supplemented by clinical findings’, ‘mainly clinical find- diagnostic and medical assessments, and the doctors’ ings, supplemented by self-report’ and ‘clinical findings age and gender.8–10 alone’. The first two were recoded into ‘statements by This study aimed to examine the accuracy of doctors’ patients’ and the last two into ‘clinical findings’. predictions of their patients’ certified sickness absence The doctors also evaluated whether non-medical status 4 weeks ahead, and which factors were associated factors could have influenced their work ability with the accuracy of the predictions. assessments, by the question: ‘Do you consider that non-medical factors are responsible for the patient’s reduced work ability?’ The answer categories were: Materials and methods ‘yes, certainly’, ‘yes, probably’, ‘no, but uncertain’ and ‘absolutely not’, dichotomized into ‘yes’ and ‘no’. Data collection The duration of the episodes of absence was collected Data collection took place from January to April 1996 from the National Sickness Benefit Register.17 No in the county of Aust-Agder, southern Norway.3,8 The episodes were lost to follow-up. county had 100 211 inhabitants in 1996. All GPs and company doctors in the county were invited to The sample participate in the study. The participating doctors In a sample of 913 episodes of certified sickness recorded episodes of certified sickness absence made absence,8 909 concerned episodes for which doctors had during office hours. answered the predictive question. This study focused on The study had a two-stage design,14 the inclusion of prediction of duration early in the course of absence. doctors representing the first stage of data collection, Episodes with duration of Ͼ20 weeks at the time of the and the consultations between doctors and patients assessments were therefore excluded (113 episodes). the second. Patients and doctors were assured that The study sample included 796 questionnaires participation in the study would not interfere with concerning the absence episodes of 796 patients. Fifty- absence status. For episodes recorded more than once, two of 91 invited doctors (57%) participated, 49 GPs only the first was included in the analysis. and three company doctors. The doctors recorded an average of 15 episodes each (minimum nine, maximum Variables 23). The doctors’ age ranged from 28 to 78 years (mean The doctors predicted their patients’ absence status 44 years), and 13 (25%) were female. 4 weeks ahead by the question: ‘What do you believe The episodes of absence were labelled short-standing the patient’s situation concerning certified sickness (up to 2 weeks, n = 486) and long-standing (from 3 to absence will be in 4 weeks?’, with the options: ‘probably 20 weeks, n = 310), depending on their duration at the full-time certification’, ‘probably part-time certification’ time when the consultations occurred, i.e. from the start and ‘probably returned to work’. Probably full-time and of absence to the time of the predictive assessments. part-time certifications were recoded as ‘probably still The study was approved by the Regional Ethics certified sick’. Committee for Medical Research, the Norwegian Data The doctors classified the main sickness certification Inspectorate and the Legal Affairs Division in the diagnoses according to ICPC (International Classi- National Insurance Administration. fication of Primary Care).15 Musculoskeletal, mental, respiratory disorders and injuries are important Analyses diagnostic groups in sickness absence.10,16 The diagnoses Positive predictive values (PPVs), with 95% confidence were therefore recoded into those groups, and ‘all other intervals, for the doctors’ predictive assessments versus disorders’. the patients’ actual absence status 4 weeks ahead were The doctors assessed the patients’ work ability by the estimated. Design-based analyses, with the doctors as 5-point question: ‘To what degree is the patient’s ability the primary sampling units, were performed using 194 Family Practice—an international journal
STATA® software. The analysis adjusts for the fact that Unadjusted results for each variable, and adjusted, the assessments of different patients made by a single reduced models are presented. doctor are not independent. Multiple, design-based, logistic regression analyses Results were performed in a backward, stepwise procedure, with the accuracy of the prediction (right/wrong) as The patients’ age ranged from 16 to 66 years (mean outcome variable. The variables with the highest 40 years), and 55% were women (Table 1). P-values were excluded one by one, keeping the Musculoskeletal (33%) and respiratory disorders patients’ age and gender, and diagnoses, in all steps. (27%) were most common in short-standing episodes,
TABLE 1 Distribution of the determinants, and mean number of patients per doctor, at the time of consultations
Duration of episodes at the time of consultation Total %
Short-standing Long-standing up to 2 weeks from 3 to 20 weeks nn
Age of patients 16–30 years 146 64 210 26.4 31–40 years 127 70 197 24.7 41–50 years 132 87 219 27.5 51–66 years 81 89 170 21.4 796 100.0 Gender of patients Male 219 136 355 44.6 Female 267 174 441 55.4 796 100.0 Diagnoses Musculoskeletal disorders 158 150 308 38.7 Mental disorders 65 69 134 16.9 Respiratory disorders 131 16 147 18.5 Injuries 35 21 56 7.0 All other disorders 96 54 150 18.9 795 100.0 The doctors’ assessments of: the patients’ work ability Moderately reduced 140 112 252 31.9 Much reduced 193 110 303 38.3 Very much reduced 151 85 236 29.8 791 100.0 Clinical information sources Statements by patients 325 237 562 70.6 Clinical findings 161 73 234 29.4 796 100.0 Presence of non-medical factors No 401 237 638 80.2 Yes 85 73 158 19.8 796 100.0 Prediction Probably returned to work 397 99 496 62.3 Probabably still certified sick 89 211 300 37.7 796 100.0 Mean number of patients per doctor (number of D) 9.3 (52D) 6.1 (51D) Minimum/maximum 4/16 1/14 Total number of observations 486 310 796 100.0 Doctors’ prediction of certified sickness absence 195 musculoskeletal (48%) and mental disorders (22%) in Analyses with accurate prediction of returned to long-standing episodes. work and still certified sick as outcomes were also A total of 125 (41%) of the 308 patients with performed. The probability that doctors made accurate musculoskeletal disorders had back problems; 76 (25%) predictions of returned to work was lower for mental had neck/shoulder problems. Eighty-three (62%) of disorders, lower with increasing patient age in short- the 134 patients with mental disorders had depressive standing episodes, and lower for musculoskeletal problems. Thirty-nine (27%) of the 147 patients with disorders in long-standing episodes (not shown in the respiratory disorders had influenza, and 32 (22%) had table). The probability that doctors made accurate upper respiratory tract infections, whereof 38 of 39 and predictions of still certified sick was not significantly 30 of 32 were limited to short-standing episodes. Twelve associated with any variables in short-standing episodes, (75%) of the 16 patients with respiratory disorders in and was lower for injuries in long-standing episodes (not long episodes had pulmonary problems, such as asthma. shown in the table). Twenty-two (39%) of the 56 injuries concerned sprain and bruises, and 13 (23%) fractures, whereof 14 of 22 and four of 13 were limited to short-standing episodes. Discussion The doctor’s work ability assessments were based on clinical findings in 33% of the cases in short-standing Diseases have typical courses. Most respiratory infec- episodes, and in 24% in long-standing episodes. tions are in remission 2 weeks after their onset; The doctors expected that 82% of the patients would depression lasts for several weeks. Although doctors return to work within 4 weeks in short-standing know this, it is not always easy to predict the course of an episodes, and 32% in long-standing episodes. Seventy- illness for individual patients. On the other hand, given four per cent of the patients had actually returned to the purpose of identifying sickness absence of expected work 4 weeks after the predictive assessments in short- long duration, the doctors’ statements were accurate in standing episodes, and 23% in long-standing episodes 91% of the cases in long-standing episodes. Our findings (Table 2). indicate that doctors accurately identify patients with a The PPVs for the doctors’ prediction of returned to poor prognosis of returning to work already after work were 84% in short-standing episodes, and 53% in 2 weeks of sickness absence. However, the ability to long-standing episodes. The PPVs for still certified sick predict the further course of the illness is more difficult were 72 and 91%. Analyses omitting the three company when the episodes have already lasted Ͼ2 weeks. While doctors yielded equal PPVs. the PPVs in short-standing episodes are satisfactory The results of the multiple logistic regression analyses (Table 2), this is not the case for the ability to predict for short- and long-standing episodes are shown in Table 3. return to work in long-standing episodes. In fact, a PPV In short-standing episodes, the probability of the of 53% means that the doctors might as well have tossed doctors making accurate predictions was higher for a coin. Given their knowledge and experience, what can respiratory disorders, and lower for mental disorders, explain this finding? both compared with the category ‘all other disorders’. The overall result is strongly influenced by the large In long-standing episodes, the probability was lower for number of musculoskeletal disorders, which obviously musculoskeletal disorders and injuries. Neither the age cause the doctors problems in predicting the patients’ nor the gender of patients or doctors, nor the degree future work ability. However, the doctors’ problems are of work ability reduction, nor other factors were not confined to this diagnostic group. All specified associated with the accuracy of the predictions. disorders in this study gave much lower estimates of
TABLE 2 Doctors’ predictions of certified sickness absence status 4 weeks ahead, at the time of consultations, versus actual absence status 4 weeks later
Duration of episodes at Doctors’ predictions Actual absence status 4 weeks later Total PPV 95% CI the time of consultation Returned to work Still certified sick nn
Short-standing (up to 2 weeks) Probably returned to work 332 65 397 84% 79–87 Probably still certified sick 25 64 89 72% 62–80 Long-standing (3–20 weeks) Probably returned to work 52 47 99 53% 43–62 Probably still certified sick 20 191 211 91% 85–94
PPV, positive predictive value; 95% CI, 95% confidence interval. 196 Family Practice—an international journal
TABLE 3 Design-based logistic regression analyses of the correctness of the doctors’ prediction of certified sickness absence status 4 weeks ahead, at the time of consultations: unadjusted, and reduced, adjusted modelsa
Duration of episodes at the time of consultation
Short-standing (up to 2 weeks) Long-standing (3–20 weeks)
Unadjusted Reduced, adjusted Unadjusted Reduced, adjusted
ORb 95% CIc OR 95% CI OR 95% CI OR 95% CI
Age of patients (years) 0.98 0.97–1.00 0.99 0.97–1.01 1.00 0.98–1.02 1.00 0.97–1.02 Gender of patients Male 1.00 1.00 1.00 1.00 Female 1.08 0.69–1.69 1.21 0.77–1.89 1.32 0.81–2.16 1.19 0.69–2.05 Diagnoses All other disorders 1.00 1.00 1.00 1.00 Musculoskeletal disorders 0.73 0.42–1.26 0.75 0.43–1.31 0.32 0.12–0.84 0.33 0.12–0.86 Mental disorders 0.45 0.23–0.88 0.46 0.24–0.89 0.54 0.16–1.84 0.54 0.15–1.90 Respiratory disorders 2.79 1.34–5.81 2.84 1.36–5.90 0.22 0.05–0.97 0.23 0.05–1.04 Injuries 3.81 0.86–16.8 3.75 0.88–16.1 0.11 0.03–0.45 0.12 0.03–0.48 Age of doctors (years) 0.99 0.97–1.01 0.96 0.93–1.00 Gender of doctors Male 1.00 1.00 Female 0.90 0.52–1.56 1.54 0.84–2.80 The doctors’ assessments of the patients’ work ability Moderately reduced 1.00 1.00 Much reduced 0.87 0.49–1.56 1.08 0.64–1.82 Very much reduced 1.21 0.63–2.33 1.30 0.61–2.78 Clinical information sources Statements by patients 1.00 1.00 Clinical findings 1.19 0.70–2.03 1.09 0.52–2.25 Presence of non-medical factors No 1.00 1.00 Yes 0.43 0.25–0.75 1.21 0.57–2.58 a The prediction was right/wrong in 395/90 cases in short-standing episodes, in 243/67 cases in long-standing episodes, in the reduced, adjusted models. n ϭ 485 in reduced, adjusted model in short-standing episodes (missing information about diagnoses in 1 case). n ϭ 310 in reduced, adjusted model in long-standing episodes. b Odds ratio (OR) Ͻ 1 indicates reduced probability for the cases to be more accurately predicted than the reference category, Ͼ1 increased. c 95% confidence interval for OR.
predictive probability than the reference diagnosis long-standing episodes of sickness absence. Good in long-standing episodes, even if the estimates were prognoses, here return to work, are hard to identify after not significant for diagnostic groups other than 2 weeks of certified sickness absence. musculoskeletal disorders and injuries (Table 3). In Prior duration may be a predictor of outcome, because mental disorders, the doctors may make positive recovery (here return to work) may be easy to predict in statements about return to work as part of a motivation short-standing respiratory disorders, and conversely that exercise. In predicting the course in injuries, there was a pessimistic predictions (here still certified sick) will be striking difference in the probability for the doctors to more accurate in prolonged illnesses, such as depression. make accurate predictions in short-standing episodes It may thus be argued that our findings are obvious, more [odds ratio (OR) 3.75; 95% confidence interval (CI) like self-fulfilling prophecies, due to the high number 0.88–16.1] and long-standing episodes (OR 0.12; of respiratory infections in short-standing episodes. 0.03–0.48). Patients with injuries returned to work more However, doctors do not necessarily know how self- quickly than the doctors expected in long-standing limiting illnesses are at the time they assess their patients’ episodes, as the probability for the doctors to make future duration of absence. Not all respiratory disorders accurate predictions of still certified sick was also low. in this study were of short duration, and musculoskeletal Our findings indicate that diagnoses alone do not give and mental disorders were evenly distributed in short- the doctors enough information to assess prognoses in and long-standing episodes (Table 1). Doctors’ prediction of certified sickness absence 197 Had the time span for the prediction of the patients’ perceiving themselves as work disabled more frequently absence status been 1 week ahead, instead of 4 weeks, have health-affecting psychosocial problems that are that would of course reduce the probability of the not work related.20 With a mixture of work-related and doctors making accurate predictions of return to work. other problems that the doctors cannot cure, the Being familiar with the presence of non-medical apparent relief given by issuing sickness certificates factors does not seem to help the doctors either. We could be ‘an easy way out’ for both doctor and patient. anticipated that the ability to give accurate predictions would improve if no non-medical factors were present, Method or if the doctors’ work ability assessments were based on Patients in this study were not significantly different clinical findings rather than statements by the patients. from other patients certified sick recorded in the We could not confirm these hypotheses, even if non- sickness benefit register of Aust-Agder in the medical factors apparently could make a difference in corresponding period, with regard to age, gender and short episodes if not adjusted for diagnoses. Another a diagnoses.3,8 Participating doctors had more female priori hypothesis, that the doctors’ assessments of the patients certified sick than non-participating doctors, patients’ work ability would be associated with accurate 55% versus 47%. Because the gender of the patients did predictions, was also not confirmed. Temporal consid- not affect the main conclusions, we do not think that erations made by the doctor could explain this. this difference is of importance. The most common Assessment of the patients’ work ability today is diagnostic groups were musculoskeletal, mental and different from assessing his/her future function. respiratory disorders, as found by others previously.16 The doctors’ answers are not just predictions, but also The National Sickness Benefit Register serves as an plans of action concerning the future absence of the accounting system for the payment of sickness benefits, patients. The doctors had the authority to determine and is therefore revised, audited and quality controlled. whether the predictions would be accurate or not. The The register ensures good follow-up of the patients. substantial differences between PPVs for return to work Episodes with short and long duration at the time of in short- and long-standing episodes suggest that the the predictive assessments differentiate between new doctors have been overruled by the course of the illness, episodes of sickness absence and the prolongation of or the opinion of the patients, and particularly so in ongoing episodes. This says nothing about how long the long-standing episodes. One may question whether this duration of the episodes eventually will be. However, mutual understanding between doctors and patients for one may assume that the doctors have more knowledge longer duration of absence than the doctors had about the patients, and thus the probable duration of expected expresses realistic work ability assessments absence, in prolonged episodes. made by the doctors, motivation exercises towards their Because the PPV for return to work was 84% in patients, or reflects that the doctors’ predictions contain short-standing episodes, increased follow-up efforts elements of the patients’ motivation for work should not be targeted to that group of patients certified resumption. sick. The predictive question may not only reflect an assumed duration of sickness absence. The ability of doctors to predict still certified sick in long-standing episodes (PPV of 91%) could reflect the dialogue and Acknowledgements negotiations that take place between patients and The Norwegian Ministry of Health and Social Affairs doctors in consultations concerning sickness absence. supported the study. Patients have certain expectations when they decide to consult their doctors about sickness absence. It is rare that doctors do not meet the patient’s demands for a References certification. Ninety-five per cent of patients receive certifications by their doctors when the initiative is 1 Bloch FS, Prins R (eds). Who Returns to Work and Why. A Six- taken by the patients.5 Even where GPs do not country Study on Work Incapacity and Reintegration. New Brunswick: Transaction Publishers; 2001. recommend sickness certification, certificates are issued 2 Bonner D, Hooker I, White R. Non Means Tested Benefits: The 11 in 87% of the cases. The patients influence the doctors’ Legislation, Statutory Sick Pay (General) Regulations. London: decisions.18 Sweet and Maxwell; 1998. The last 30 years have gradually brought a shift 3 Reiso H, Nyg°rd JF, Brage S, Gulbrandsen P, Tellnes G. Work ability 19 assessed by patients and their GPs in new episodes of sickness towards patient-centred consultations. This way of certification. Fam Pract 2000; 17: 139–144. thinking probably has influenced gatekeeper functions, 4 Gulbrandsen P, Brage S. Life situation as a reason for sick leave. such as the authorization of sick leaves. We found that Tidsskr Nor Lægeforen 1998; 118: 2463–2466 [in Norwegian, the doctors based their work ability assessments on English summary]. 5 Larsen BA, Førde OH, Tellnes G. Physician’s role in certification for statements by the patients in 71% of cases (Table 1). sick leave. Tidsskr Nor Lægeforen 1994; 114: 1442–1444 [in There is reason to question this practice. Patients Norwegian, English summary]. 198 Family Practice—an international journal
6 Rutle O. An Analysis of Encounters in General Practice. Oslo: ability to identify vulnerable groups. Scand J Primary Health National Institute of Public Health; 1983: Vol 1: 178 [in Care 1998; 16: 204–210. Norwegian, English summary]. 14 Lemeshow S, Letenneur L, Dartigues JF, Lafont S, Orgogozo JM, Com- 7 http://www.dep.no/sos/norsk/publ/utredninger/NOU/ Governmental menges D. Illustration of analysis taking into account complex report NOU 2000: 27, Sickness absence and disability survey considerations: the association between wine pension, an including working-life [in Norwegian] (accessed consumption and dementia in the PAQUID study. Am J 13 February 2003). Epidemiol 1998; 148: 298–306. 8 Reiso H, Nyg°rd JF, Brage S, Gulbrandsen P, Tellnes G. Work ability 15 Brage S, Bentsen BG, Bjerkedal T, Nyg°rd JF, Tellnes G. ICPC as a and duration of certified sickness absence. Scand J Public standard classification in Norway. Fam Pract 1996; 13: Health 2001; 29: 218–225. 391–396. 9 Tellnes G. Sickness certification in general practice: a review. Fam 16 Tellnes G, Svendsen KOB, Bruusgaard D, Bjerkedal T. Incidence Pract 1989; 6: 58–65. of sickness certification. Proposal for use as a health 10 Söderberg E, Alexanderson K. Literature Study—Published Studies status indicator. Scand J Primary Health Care 1989; 7: of the Interface Between Medical Practice and Assessments of 111–117. Social Insurance Legislation. Försäkringsmedicinskt centrum, 17 Brage S, Nyg°rd JF, Tellnes G. The gender gap in musculoskeletal- Rapport 2001: 1. Linköping; 2001 [in Swedish]. related long term sickness absence in Norway. Scand J Soc Med 11 Englund L, Svärdsudd K. Sick-listing habits among general 1998; 26: 34–43. practitioners in a Swedish county. Scand J Primary Health Care 18 Ainsworth-Vaughn N. Patients Claiming Power in Doctor–Patient 2000; 18: 81–86. Talk. New York: Oxford University Press; 1998. 12 Gulbrandsen P, Hjortdahl P, Fugelli P. General practitioners’ 19 Stewart M. Patient-centered Medicine: Transforming the Clinical knowledge of their patients’ psychosocial problems: multipractice Method. Thousand Oaks (CA): Sage Publications; 1995. questionnaire survey. Br Med J 1997; 314: 1014–1018. 20 Gulbrandsen P, Hjortdahl P, Fugelli P. Work disability and health- 13 Gulbrandsen P, Fugelli P, Hjortdahl P. General practitioners’ affecting psychosocial problems among patients in general knowledge of their patients’ socio-economic data and their practice. Scand J Soc Med 1998; 26: 96–100.
Paper VI Harald Reiso, Jan F Nygård, Gudrun S Jørgensen, Rune Holanger, Dag Soldal, Dag Bruusgaard.
Back to Work: Predictors of Return to Work Among Patients With Back Disorders Certified As Sick. A Two-Year Follow-up Study. Spine 2003; 28: 1468–73.
Reprinted by courtesy of Spine and Lippincott Williams & Wilkins
SPINE Volume 28, Number 13, pp 1468–1474 ©2003, Lippincott Williams & Wilkins, Inc. Back to Work: Predictors of Return to Work Among Patients With Back Disorders Certified As Sick A Two-Year Follow-up Study
Harald Reiso, MD,*† Jan F Nygård, BSc, BA,‡ Gudrun S Jørgensen,§ Rune Holanger,§ Dag Soldal, MD,ʈ and Dag Bruusgaard, MD, PhD*
Study Design. A 2-year follow-up study of patients predictors of time until return to work in patients with with back disorders certified as sick. back disorders certified as sick who attend a back disorder Objectives. To identify predictors of return to work. outpatient clinic. [Key words: COOP/WONCA charts, func- Summary of Background Data. Back disorders are tional assessments, pain, return to work, sickness ab- common health problems and the most important disor- sence, sickness certification, work ability] Spine 2003;28: ders associated with absence from work in the welfare 1468–1474 states. Predictors of future absence may be of help in allocating rehabilitation efforts to such patients. Possible predictors include demographic and medical factors, the Back pain is a common health problem, and work inca- patients’ functional status, and former absence. pacity attributed to it probably is increasing.1,2 In 1995, Methods. For this study, 190 patients certified as sick who attended a back disorder outpatient clinic from 13% to 17% of certified sickness absence, rehabilitation September 1997 to December 1998 answered a ques- allowance, and disability pensions in Norway was asso- tionnaire. Demographic data, medical factors, self-as- ciated with back pain.3 Similar figures are found for sessed function, and absence data were recorded. Re- other European countries.4 The prognosis for most pa- turn to work, defined as returning to work for at least 60 tients with low back pain (LBP) certified as sick is consecutive calendar days, was used in Cox regression 5,6 analyses. good. Findings show that 70% of patients with at least Results. According to multiple Cox regression analy- 2 weeks of certified sickness absence have returned to ses, age of 40 to 49 years (HR, 0.52; 95% confidence work within 3 months, and 85% after 6 months.5 interval [95%CI], 0.29–0.94), high pain intensity (HR, 0.30; Predictors of future absence may be of help in allocat- 95%CI, 0.17–0.55), low self-assessed work ability (HR, ing rehabilitation efforts to patients with back disorders. 0.43; 95%CI, 0.25–0.73), and a self-predicted absence sta- 7–9 tus of not returning to work (HR, 0.31; 95%CI, 0.17–0.54) Possible predictors include demographic and medical 3,5,10–12 8,13 predicted longer time until return to work. Back disorders factors, the patient’s functional status, and with radiation predicted shorter time until return to former absence.14,15 work (HR, 2.08; 95%CI, 1.37–3.16). The COOP/WONCA A single question concerning self-assessed work abil- chart’s physical fitness, daily activities, overall health, ity predicted duration of certified sickness absence in pri- and change in health were associated with time until 8 return to work in univariate analyses only, as was the mary health care. Reduced self-assessed work ability in duration of the sickness certification episodes from patients with LBP certified as sick 8 to 12 weeks pre- start to inclusion and the degree of sickness certifica- dicted their chance of not returning to work in 12 tion at inclusion. months of follow-up evaluation.13 Conclusions. Information about the age of the pa- Self-reported pain is a known predictor of LBP dura- tients, diagnoses, pain intensity, self-assessed work abil- 12 ity, and self-predicted absence status may be used as tion. Widespread pain is a strong predictor of persis- tent LBP,11 and pain intensity in LBP patients is strongly associated with disability.10 Function can be measured by self-report,13,16–20 as- 18,19 From the *Department of General Practice and Community Medicine, sessments made by observers such as physicians, or the Section for Occupational Health and Social Insurance Medicine, more specific physiologic tests (e.g., bicycle tests). Some University of Oslo, the †Aust-Agder County Office of the National of these methods are time consuming, and are therefore Insurance Service, Arendal, the ‡Department Group of Basic Medical Sciences University of Oslo, §Havblik Physiotherapy and Training, unsuitable for everyday clinical practice. A work ability Arendal, and ʈthe Back Disorder Outpatient Clinic at Aust-Agder assessment question18 and the COOP/WONCA charts16 County Hospital, Arendal, Norway. are simple instruments for assessing functional status. Supported by the Norwegian Ministry of Health and Social Affairs. Acknowledgment date: May 13, 2002. First revision date: August 16, The aim of this study was to identify predictors of 2002. Second revision date: December 3, 2002. Acceptance date: De- return to work in a 2-year follow-up assessment of pa- cember 10, 2002. tients with back disorders certified as sick. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. No Methods benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. The enrollment of patients in the study took place at the Aust- Address correspondence and reprint requests to Harald Reiso, MD, Aust-Agder County Office of the National Insurance Service, P.O. Box Agder County Hospital back disorder outpatient clinic in 1853, Stoa, N-4858 Arendal, Norway; E-mail: harald.reiso@ southern Norway September 1997 to December 1998. The samfunnsmed.uio.no County had 101,320 inhabitants in 1998. The clinic was estab-
1468 4 9akoWork • Reiso et al 1469Backto lished in August 1997 at the start of a project that focused on Outcome. Return to work, defined as returning to work for early intervention for patients with back disorders certified as at least 60 consecutive calendar days, was used as the outcome. sick.21 The clinic had 625 new patients with back disorder Each patient could have only one return to work event. Na- referred September 1997 to December 1998. Primary health tional registers concerning sickness certification, rehabilitation care doctors were encouraged to refer patients with back dis- allowance, and disability pension supplied absence data. order that they believed would be certified as sick more than 2 months, as soon as possible after the start of the certification. Recoding. Occupations were recoded as administrative/ The patients were asked to complete a simple questionnaire at teaching/office work, trade/service work, care professions, in- the first consultation at the clinic. The questionnaires were dustrial work/construction/primary activity, transportation, forwarded to the National Insurance Administration. Comple- cleaner, and other. The sickness certification diagnoses were tion of the questionnaires was voluntarily, and all information recoded into two groups: back disorders without radiation given by the patients could be withdrawn at any time on their (L02, L03, L84) and back disorders with radiation (L86). request. The patients were assured that their anonymity would Because the number of observations in some of the answer be protected, and that participation in the study would not categories was small, the following variables were recoded. Pain intensity on the VAS scale was recoded into three imply consequences for their certification. 1,3–5,7,13,16,21–23 Patients certified as sick with the International Classification groups. Work ability assessed as hardly re- of Primary Care (ICPC)22 diagnoses L02 (back disorders), L03 duced at all, not much reduced, or moderately reduced was (low back disorders without radiation, excluding L86), L84 recoded as moderately reduced, whereas the much reduced and (back disorders without radiation), and L86 (back disorders very much reduced assessments were kept unaltered. The self- with radiation/nucleus herniation) were included. predicted absence status was dichotomized into returned to work and continued certification. The six COOP/WONCA charts were recoded from 5-point scales to 3-point scales. The Sample. Altogether, 220 referred patients certified as sick sup- two first scale options were combined, as were the two last plied questionnaire data. Subsequently, 30 were excluded: 17 patients because they had other diagnoses, 10 patients because options, whereas the middle option was retained. they were certified as sick while on rehabilitation allowance, 2 patients because they were old age pensioned during the fol- Statistical Analyses. Univariate and backward stepwise mul- low-up period, 1 patient because the back disorder was caused tiple Cox regression analyses were performed. The age, gender, by injury. The final sample included 190 patients. and diagnoses of the patients were included in all the multiple Cox analyses. Variables with P values less than 0.20 by a change of Ϫ2 log likelihood in the univariate analyses were Determinants and Outcome included in the multiple analyses, then excluded backward one by one, guided by which had the highest P value greater than Demographics. The patients’ age and gender were recorded, 0.05 in each step. and their occupations were registered as free text. The proportional hazard assumption was tested numeri- cally by the Schoenfeld residuals, and graphically by plotting Medical Factors. The ICPC sickness certification diagnoses observed Kaplan–Meier survival curves and predicted Cox re- were recorded. Pain today was assessed by a question concern- gression curves. Correlations between variables were measured ing pain quality (whether intermittent or continuous) and a by Spearman’s rank correlation coefficients. The chosen level of question concerning pain intensity using a visual analog scale significance was 5%. (VAS) with choices from 1 (no pain at all) to 10 (intolerable pain). Results Self-Assessed Function. Patients assessed their work ability by answering the following question using a 5-point scale: “To The age range of the patients was 20 to 63 years (mean, what degree does your back disorder reduce your ability to 40 years), and 65% of the patients were men. The three perform your ordinary work today?” The answer categories most common occupational groups were industrial were “very much reduced,” “much reduced,” “moderately re- work/construction/primary activity (43%), administra- duced,” “not much reduced,” and “hardly reduced at all.” The tive/teaching/office work (16%), and care professions patients’ self-predicted absence status was recorded by the (14%). question: “What do you believe your situation concerning cer- The ICPC diagnoses were L86 in 48% of the cases, tified sickness absence will be in 4 weeks?” There were three answer options: “probably full-time certification,” “probably L84 in 32%, L02 in 17%, and L03 in 3%. part-time certification ” and “probably returned to work.” The Findings showed that 130 patients returned to work patients also assessed their function by the six COOP/WONCA for at least 60 calendar days in the 2-year follow-up charts: physical fitness, feelings, daily activities, social activi- period, and that 60 patients reached the maximum pos- ties, overall health, and change in health.11,16 sible sickness absence of, in this study, 2 years (730 cal- endar days) without experiencing any return to work Absence Data. Maximum sickness benefit time in Norway is event. At enrollment in the study, 46 patients (24%) 365 calendar days. Persons not able to work after that can be given a medical or vocational rehabilitation allowance or a were on part-time, 144 patients were on full-time certi- disability pension. All episodes of such sickness absence from fication, and 104 patients (55%) had no episode of ab- work 1 year before inclusion and 2 years after inclusion, were sence in the year before inclusion. The mean duration of recorded, as was the degree of sickness certification at the sickness certification episodes, from their start to inclusion. study inclusion, was 84 calendar days (range, 1–268 1470 Spine • Volume 28 • Number 13 • 2003
Table 1. The Proportion of Patients Returned to Work (for at Least 60 Calendar Days) and the Mean Time until Return to Work in Calendar Days for the Determinants in Aust-Agder County, Norway, 1998
Proportion of Patients Mean Time until Returned to Work Return to Work* P Value†
Age (y) 20–29 73% (30/41) 172 0.45 30–39 73% (40/55) 194 40–49 69% (34/49) 225 50–63 58% (26/45) 174 Gender Male 71% (88/124) 191 0.31 Female 64% (42/66) 197 Diagnoses (ICPC codes) Back disorders without radiation (L02, L03, L84) 61% (60/99) 209 0.01 With radiation (L86) 77% (70/91) 179 Pain (today) Pain quality Pain now and then 77% (57/74) 171 0.01 Continuous pain 62% (71/114) 214 Pain intensity 1–4 (1 ϭ no pain at all) 94% (32/34) 122 0.00 5–6 71% (54/76) 211 7–10 (10 ϭ intolerable pain) 57% (28/49) 262 Work ability (today, self-assessed) Moderately reduced 91% (40/44) 113 0.00 Much reduced 63% (35/56) 186 Very much reduced 60% (53/88) 259 Self-predicted absence status (in four weeks) Returned to work 88% (29/33) 67 0.00 Continued certification 65% (97/150) 232 COOP/WONCA (During the past 2 weeks) Physical fitness (hardest physical activity for at least two minutes) Heavy 87% (20/23) 144 0.02 Moderate 68% (45/66) 200 Light 64% (56/87) 213 Feelings (bothered by emotional problems) No 72% (64/89) 184 0.18 Moderately 68% (30/44) 176 Much 61% (27/44) 255 Daily activities (difficulties doing usual tasks) No 77% (23/30) 139 0.00 Moderately 77% (54/70) 188 Much 57% (44/77) 240 Overall health (health in general the last two weeks) Good 80% (47/59) 152 0.00 Neither good nor bad 64% (46/72) 192 Bad 60% (27/45) 290 Change in health (health now compared with two weeks ago) Better 77% (44/57) 154 0.02 Unchanged 65% (72/110) 219 Worse 50% (5/10) 280 Duration of sickness certification episodes from start to inclusion (d) 1–46 82% (55/67) 143 0.00 47–95 62% (37/60) 227 96–268 60% (38/63) 232 Degree of sickness certification at inclusion Part-time 78% (36/46) 135 0.01 Full-time 65% (94/144) 215
* Estimated among patients who returned to work. † By change in -2 log likelihood in univariate Cox regression analyses. The highest Spearman’s rank correlation coefficients between the variables presented in Table 1 were between degree of sickness certification at inclusion and duration of sickness certification episodes from start to inclusion (correlation 0.45), degree of sickness certification at inclusion and pain quality (0.43), self-predicted absence status and change in health (0.43), and self-predicted absence status and work ability (0.40).
days). In the 2-year follow-up period, 25 patients were certification, reduced physical fitness, emotional prob- granted a disability pension: 17 full-time and 8 part- lems, difficulties with daily activities, bad overall health, time, all attributed to back disorders. worsened health, increasing duration of the sickness cer- Predictors of Return to Work tification episodes from their start to study inclusion, and Continuous pain, high pain intensity, low self-assessed full-time certification at inclusion were associated with work ability, self-predicted absence status of continued longer time until return to work in the univariate Cox 4 1akoWork • Reiso et al 1471Backto
Table 2. Reduced Model From Multiple Cox Regression Analyses of Time until Return to Work in a Two-Year Follow-up in Aust-Agder County, Norway, 1998*
n HR† 95% CI
Age (y) 20–29 38 1.00 30–39 43 0.63 0.36–1.09 40–49 37 0.52 0.29–0.94 50–63 35 0.60 0.33–1.09 Gender Male 104 1.00 Female 49 1.02 0.64–1.62 Diagnoses (ICPC codes) Back disorders without radiation 77 1.00 (L02, L03, L84) With radiation (L86) 76 2.08 1.37–3.16 Pain intensity (Today) 1–4 (1 ϭ no pain at all) 33 1.00 Figure 1. Proportion of patients not returning to work (observed 5–6 75 0.47 0.28–0.78 survival curves) by self-assessed work ability. Aust-Agder County, 7–10 (10 ϭ intolerable pain) 45 0.30 0.17–0.55 Norway, 1998. Work ability (today, self-assessed) Moderately reduced 34 1.00 Much reduced 43 0.63 0.36–1.10 Discussion Very much reduced 76 0.43 0.25–0.73 Self-predicted absence status Method (in 4 weeks) The completion of questionnaires was voluntary. The Return to work 27 1.00 Continued certification 126 0.31 0.17–0.54 number of referred patients certified as sick is not known, nor the number to whom the questionnaires * The following variables were excluded in backward, stepwise analyses (P values in brackets). Pain quality (0.90), degree of sickness certification at were distributed. However, the distribution of the pa- inclusion (0.73), change in health (0.61), feelings (0.56), duration of the sick- tients’ age and gender in this study was similar to that ness certification episodes from start to inclusion (0.44), overall health (0.12), daily activities (0.22), and physical fitness (0.12). N ϭ 153. 37 cases were not among other patients with back disorders certified as included in the analysis because of missing information about pain intensity sick for more than 8 weeks and referred to an outpatient (31 cases), self-predicted absence status (5 cases), and self-assessed work 6,13 ability (1 case). 44 cases censored due to no event. The Spearman’s rank clinic. A selection of patients with a positive attitude correlation coefficient between pain intensity and work ability was 0.35, be- toward the completion of questionnaires may be as- tween pain intensity and self-predicted absence status 0.39, and between work ability and self-predicted absence status 0.40. sumed for this study, and thereby also a possible positive † Hazard ratio (HR) Ͻ1 indicates increased probability for the cases to have attitude toward return to work. But the fact that the longer time to return to work than the reference category, Ͼ1 shorter. 95% confidence interval for HR (95% CI). findings of differences in this group are highly significant, and that the patients’ absence was followed for 2 years weakens arguments that a possible selection should im- ply skewed conclusions. On the other hand, it may be regression analyses (Table 1). Patients who had back argued that such a selection would contribute to weak- disorder with radiation had significantly shorter time un- ened associations. til return to work than patients without radiation in the There are no references in the scientific literature to univariate analysis. the reliability and validity of the duration data of the The COOP/WONCA chart social activities, the pres- national registers used. However, the registers serve as ence of other episodes of absence 1 year before study account systems for the payment of sickness benefits, and inclusion, and occupations had P values exceeding 0.20 are therefore revised, audited, and quality controlled. in the univariate analyses. The registers also ensure good follow-up evaluation of Pain quality, the degree of sickness certification at in- the cases. None were lost to follow-up assessment in this clusion, the COOP/WONCA charts, and the duration of study. the sickness certification episodes from their start to The ICPC diagnoses on the paper copy of 25 certifi- study inclusion were excluded in the backward, stepwise cates (dated between August 3 and December 8 in 1995 multiple Cox regression analyses. Age of 40 to 49 years, and available to author H.R.) were compared with the high pain intensity, low self-assessed work ability, and a register information. The 25 diagnoses all are coded cor- self-predicted absence status of continued certification rectly in the register. Intra- and interobserver concor- predicted longer time until return to work in the multiple dance for diagnostic groups used on sickness certificates analyses, whereas diagnoses with radiation predicted a exceeds 90%.24 Coding from medical certificates to reg- shorter time until return to work (Table 2). isters was found correct in 98% of the cases, whereas The test of the proportional hazard assumption by the diagnoses based on texts in medical records match diag- Schoenfeld residuals yielded a P value of 0.19. Observed nostic groups in 72% of the cases.25 and predicted survival curves were close together. The Cox regression analyses solved the problem of censor- observed survival curves are presented in Figure 1. ing (i.e., not all patients returned to work within the 1472 Spine • Volume 28 • Number 13 • 2003 follow-up period) and made adjustments between vari- present and future work ability, not their function in ables possible. The prerequisites for using Cox regression retrospect, because self-assessed work ability and self- were fulfilled because the proportional hazard assump- predicted absence status were predictive in this study, tion was not rejected. Return to work for at least 60 whereas the COOP/WONCA charts were not. calendar days was considered to be a robust measure of a return-to-work event. The COOP/WONCA charts were Conclusions used as functional measures because of their simplicity, Information about the age of the patients, diagnoses, and because there is some tradition for their use in pain intensity, self-assessed work ability, and self- 26 Norway. predicted absence status may be used to predict time Predictors until return to work in patients with back disorders cer- Age is a known predictor of duration of certified sickness tified as sick who attend a back disorder outpatient absence.7–9 clinic. Patients with back disorder with radiation had shorter time until return to work than patients without radia- Key Points tion. This is different from findings among patients with back disorders in the general population who are certi- ● High pain intensity, low self-assessed work abil- fied as sick, in which patients with radiation have the ity, and a self-predicted absence status of not re- longest duration.3,5 Patients referred to the outpatient turning to work in 4 weeks predicted longer time clinic should, according to the aim of the clinic, have until return to work in a 2-year follow-up evalua- more serious back disorders than other patients with tion of patients with back disorders certified as sick back disorders, which may explain this difference, al- who attended a back disorder outpatient clinic, though that argument ought to concern both patients whereas back disorders with radiation predicted with and those without radiation. It may be that back shorter time until return to work. disorders with radiation represent disorders located in ● Some COOP/WONCA charts assessing func- the lower back only, whereas other back disorder diag- tional status were associated with time until return noses represent more complex musculoskeletal disor- to work in univariate analyses only, as were dura- ders. This study had no information about whether the tion of the sickness certification episodes from start back disorders were localized or not. Others have found to study inclusion and the degree of sickness certi- that LBP existing as a part of widespread musculoskele- fication at inclusion. tal pain indicates more reduced functional ability than localized LBP,27 and predicts long-term work Acknowledgments disability.28 The authors express gratitude to the patients who par- The work ability question focused on the back disor- ticipated in the study. der specifically, and may therefore be more related to function in a setting of sickness certification than the References more global COOP/WONCA charts. This may explain 1. Croft P. Is life becoming more of a pain? BMJ 2000;320:1553–4. why the work ability question was predictive of time 2. Palmer KT, Walsh K, Bendall H, et al. Back pain in brain: comparison of two until return to work in all the analyses, whereas some of prevalence surveys at an interval of 10 years. BMJ 2000;320:1577–8. the COOP/WONCA charts were predictive only in the 3. Brage S, Lærum E. Spinal disorders in Norway: an epidemiological report (in Norwegian, English summary). Tidsskr Nor Lægeforen 1999;119:1619–23. univariate analyses. 4. Andersson GBJ. Epidemiology of low back pain. Acta Orthop Scand 1998; Self-predicted status of absence in 4 weeks may resem- 69(Suppl 281):28–31. ble self-assessed health as a predictor of morbidi- 5. Hagen KB, Thune O. Work incapacity from low back pain in the general 23,29,30 population. Spine 1998;23:2091–5. ty. Self-assessed health may be considered a self- 6. Indahl A, Velund L, Reikerås O. Good prognosis for low back pain when left fulfilling prophesy, but the fact that it also predicts untampered: a randomized clinical trial. Spine 1995;20:473–7. mortality serves as an argument against such a 7. Brage S, Nygård JF, Tellnes G. The gender gap in musculoskeletal-related 29,30 long-term sickness absence in Norway. Scand J Soc Med 1998;26:34–43. conclusion. 8. Reiso H, Nygård JF, Brage S, et al. Work ability and duration of certified sickness absence. Scand J Public Health 2001;29:218–25. Temporal Aspects of the Predictors 9. Tellnes G. Duration of episodes of sickness certification. Scand J Prim Health Because return to work is an event in the future, the Care 1989;7:237–44. temporal aspects of the predictors may provide clues to 10. McGorry RW, Webster, BS, Snook SH, et al. The relation between pain intensity, disability, and the episodic nature of chronic and recurrent low their predictability. The work ability question assessed back pain. Spine 2000;25:834–41. the patients’ opinion the day of the response, whereas the 11. Thomas E, Silman AJ, Croft PR, et al. Predicting who develops chronic low COOP/WONCA charts were retrospective, concerning back pain in primary care: a prospective study. BMJ 1999;318:1662–7. 12. van den Hoogen HJ, Koes BW, Deville W, et al. The prognosis of low back function during the preceding 14 days. The self-predicted pain in general practice. Spine 1997;22:1515–21. absence status concerned beliefs about the future. It may 13. Haaland EMH, Indahl A, Ursin H. Patients with low back pain not returning be speculated that the important predictive factors to to work: a 12-month follow-up study. Spine 1998;23:1202–8. 14. Kolstø G, Holm S, Tuhus A. Rapport om Sykepengeutviklingen i 1996: Hva assess in patients with back disorder certified as sick who Skyldes Økningen i Sykefraværet (Report about the change in the payments attend a back disorder outpatient clinic concern their for certified sickness absence in 1996: what causes the increase of absence) 4 3akoWork • Reiso et al 1473Backto
(in Norwegian). Report 1997;7. Oslo: the National Insurance Administra- health: has society adopted the expanded health concept? Soc Sci Med 1991; tion, 1997. 32:141–6. 15. Linde´n V. Absence From Work and Work Capacity [dissertation]. Bergen: 24. Tellnes G, Svendsen KOB, Bruusgaard D, et al. Incidence of sickness certifi- University of Bergen, 1967. cation: proposal for use as a health status indicator. Scand J Prim Health 16. Bentsen BG, Natvig B, Winnem M. Questions you didn’t ask? COOP/ Care 1989;7:111–7. WONCA Charts in clinical work and research. World Organization of Col- 25. Ljungdahl LO, Bjurulf P. The accordance of diagnoses in a computerized leges, Academies and Academic Associations of General Practitioners/Family sick-leave register with doctors’ certificates and medical records. Scand J Soc Physicists. Fam Pract 1999;16:190–5. Med 1991;19:148–53. 17. Ilmarinen J, Tuomi K, Klockars M. Changes in the work ability of active 26. Natvig B, Rutle O, Bruusgaard D, et al. The association between functional employees over an 11-year period. Scand J Work Environ Health 1997; status and the number of areas in the body with musculoskeletal symptoms. 23(Suppl 1):49–57. J Rehab Res 2000;23:49–53. 18. Reiso H, Nygård JF, Brage S, et al. Work ability assessed by patients and their 27. Natvig B, Bruusgaard D, Eriksen W. Localized low back pain and low back GPs in new episodes of sickness certification. Fam Pract 2000;17:139–44. 19. Swales K, Craig P. Evaluation of the Incapacity Benefit Medical Test: In- pain as part of widespread musculoskeletal pain: two different disorders? A house report 26. London: Social Research Branch, Analytical Services Divi- cross-sectional population study. J Rehab Med 2001;33:21–5. sion, Department of Social Security, 1997. 28. Natvig B, Bruusgaard D, Eriksen W. Low back pain as a predictor of long- 20. WONCA. Available at: http. //www.globalfamilydoctor.com/publications/ term work disability. (Accepted for publication in Scand J Public Health coop-woncacharts/languages/english.pdf. Accessed: February 19, 2002. January 2002). 21. Hagen TP, Østtveiten HS. Social Security Benefits Under the Knife: An anal- 29. Miilunpalo S, Vuori I, Oja P, et al. Self-rated health status as a health mea- ysis of the Purchase of the National Insurance Service’s Pilot Scheme “Treat- sure: the predictive value of self-reported health status on the use of physician ment of the sick.” (in Norwegian, English summary) NIBR report 16 1999. services and on mortality in the working-age population. J Clin Epidemiol Oslo: Norwegian Institute for Urban and Regional Research, 1999. 1997;50:517–28. 22. Brage S, Bentsen BG, Bjerkedal T, et al. ICPC as a standard classification in 30. Ostlin P. Occupational history, self-reported chronic illness, and mortality: a Norway. Fam Pract 1996;13:391–6. follow-up of 25,586 Swedish men and women. J Epidemiol Community 23. Fylkesnes K, Førde OH. The Tromsø study: predictors of self-evaluated Health 1990;44:12–6.
Point of View
Steven J. Atlas, MD, MPH
Low back pain is a common condition and a frequent understood. A large number of physical, psychosocial, cause of disability claims. As a medical condition, low employment, and economic factors have been shown to back pain is most often caused by a musculoligamentous affect disability outcomes.5 The reasons for uncertainty process that is hard to identify reproducibly, but has a about which prognostic factors are associated with dis- well-defined natural history associated with gradual im- abling back pain are complex,6 and may relate in part to provement but frequent recurrence. The history and how disabling low back pain is thought to occur.7,8 In physical examination usually are sufficient to determine the United States, the workers’ compensation model initial management in which pain and physical function views low back pain as an injury resulting from an acci- are the primary outcomes. There is no reason to believe dent. In other countries where the disability system does that low back pain that prevents an individual from not mandate proof of work-relatedness, low back pain working represents a different clinical entity. However, may be viewed more as a disease resulting from multiple work-related or disabling low back pain is increasingly risk factor exposures. Despite the differences in disability recognized as a complex disorder in which psychoso- evaluation systems throughout the developed world or in cial, economic, and ergonomic factors are more im- the hypotheses for the development of disabling low portant in predicting disability outcomes than physio- back pain, the magnitude of the problem and the lack of logic or anatomic measures of severity. Psychosocial effective treatments appear to be remarkably similar factors, which refer to aspects of a patient’s psycho- wherever the problem is studied. logical state or interactions with his or her social envi- In this context, the current study by Reiso and col- ronment (e.g., the workplace, disability system, or a health leagues adds to the evidence that psychosocial factors are care provider) that can influence the patient’s symptoms, predictors of long-term work absence in disabling low have been shown to play a role in the development and back pain.9 This study evaluated patients from one Nor- persistence of low back pain.1,2 Indeed, the pathophysio- wegian back pain clinic. A convenience sample of pa- logic model of low back pain and its medical treatments has tients with disabling low back pain completed a ques- done an inadequate job of improving work-related low tionnaire at their initial visit, and outcomes were back pain outcomes or predicting those at highest risk for long-term disability.3,4 available through retrospective review of a national dis- Despite an extensive literature, the factors associated ability registry. As the authors note, the major limitation with disabling low back pain and its outcomes are poorly of their study is the validity and generalizability of their findings because of a seemingly low participation rate From the General Medicine Division, Massachusetts General Hospital, and uncertainty about the eligibility of those who did not Harvard Medical School, Boston, Massachusetts. answer the questionnaire. 1474 Spine • Volume 28 • Number 13 • 2003
In addition, the study population included individuals pensation Insurance Corporation of New Zealand and the National Health Committee, 1997. with widely divergent duration of disability at the initial 2. Linton SJ. A review of psychological risk factors in back and neck pain. Spine visit. Although duration of disability was a univariate 2000;25:1148–56. but not a multivariate predictor, it would have been in- 3. Hadler NM. Back pain in the workplace: what you lift or how you lift matters far less than whether you lift or when. Spine 1997;22:935–40. teresting to know how patient self assessment variables 4. Nachemson A. Work for all: for those with low back pain as well. Clin were influenced by the duration of disability. Orthop 1983;179:77–85. Finally, identifying predictors of disability outcome 5. Margoshes BG, Webster BS. Why do occupational injuries have different health outcomes? In: Mayer TG, Gatchel RJ, Polatin PB, eds. Occupational does not necessarily clarify how to improve these out- Musculoskeletal Disorders. Philadelphia, PA: Lippincott Williams & comes. Medical and surgical treatments can improve Wilkins, 2000:47–61. symptoms and function for those who have low back 6. Krause N, Frank JW, Dasinger LK, et al. Determinants of duration of dis- 10 ability and return to work after work-related injury and illness: challenges disorders without improving disability outcomes. It is for future research. Am J Ind Med 2001;40:464–84. likely that patient-centered efforts that promote return to 7. Andersson GBJ. Factors important in the genesis and prevention of occupa- work and address these psychosocial factors may hold as tional back pain and disability. J Manipulative Physiol Therap 1992;15: 43–6. much or more promise for improving disability out- 8. Dempsey PG, Burdorf A, Webster BS. The influence of personal variables on comes as new medical or surgical therapies directed at work-related low back disorders and implications for future research. JOc- the underlying back disorder. cup Environ Med 1997;39:748–59. 9. Reiso H, Nygard JF, Jorgensen GS, et al. Back to work: predictors of return References to work among patients with back disorders certified sick: a two-year follow- up. Spine 2003;28:1415–20. 1. Kendall NAS, Linton SJ, Main CJ. Guide to Assessing Psychological Yellow 10. Atlas SJ, Chang Y, Kammann E, et al. Long-term disability and return to Flags in Acute Low Back Pain: Risk factors or Long-Term Disability and work among patients who have a herniated lumbar disc: the effect of disabil- Work Loss. Wellington, New Zealand: Accident Rehabilitation and Com- ity compensation. J Bone Joint Surg Am 2000;82:4–15.
Appendix The appendix includes explanatory notes, a list of participating doctors, and the questionnaires used - in Norwegian, and one way translated to English.
Explanatory notes 1 “Sykepenger ytes til den som er arbeidsufør på grunn av en funksjonsnedsettelse som klart skyldes sykdom eller skade. Arbeidsuførhet som skyldes sosiale eller økonomiske problemer o.l., gir ikke rett til sykepenger” (30, §8-4). 2 “Ved sykdom forstås etter denne lov en sådan forstyrrelse av den legemlige og åndelige sunnhetstilstand at legehjelp anses påkrevet.” 3 Two other questions about assessments of work ability on Questionnaire I, the patients’ question 12 and 13, and the doctors’ question 10 and 11, are not presented in any of the papers, because they were not predictive of sickness absence duration in any analyses. 4 An Åstrand bicycle test performed among 51 of the patients presented in Paper VI was not associated with time until return to work. 5 After the publication of Paper VI, a question of interaction between self-assessed work ability and age has been raised. The use of such an interaction term in the analyses presented in Table 2 in Paper VI had a p value of 0.28, and did not alter the predictive importance of the other factors presented in the table.
Participating doctors Torkil P. Andersen, Bjørn Bratland, Jørn Brodwall, Per Gunnar Eng, Mark Fagan, Ole Christian Fiane, Dag Findreng, Michael Golding, Ole Kjell Gundersen, Nils Petter Hagen, Rein Holanger, Olav Mikkel Holen, Nina Holm, Magn-Bjørn Hornnes, Brit I. Høgetveit, Randi Jakobsen, Marie I. Wilson Jeppesen, Gunnar Johannesen, Finn Tysland Johnsen, Harald Jomaas, Kari Hilde Juvkam, Jan Eigill Larsen, Kjeld Malde, Gunnar Mouland, Johannes Netland, Dag Nordhaug, Nils Nærdal, Torgrim Næverdal, Bodil Aasvang Olsen, Jan G. Ribe, Peter Selås, Robert Sillerud, Dag Steinar Skagestad, Trygve Skonnord, Inger-Tori Solberg, Gunnar Spikkeland, Odd K. Starheim, Jan Arthur Staurset, Anne-Marit Storaker, Audhild Egeland Torp, Per Øyvind Torp, Kristin Namtvedt Tuv, Knut Tuv, Bernt Erik Tylden, Lars Bjarne Tøsse, Brita Slettebø Ugland, Sveinung Vatnedalen, Per O. Walle, Elisabeth Wærdahl, Anne Berit Aagedal, Trygve Aanjesen, Kåre J. Aasen.
Questionnaire I – patients/doctors Answered by patients and doctors. Presented in Norwegian, and one way translated to English.
Lege Pasient Til legen
Takk for at du vil være med på prosjektet Beslutningsinstrument ved sykmelding!
Bakgrunn På baksiden av dette arket er spørsmålene du som lege bes svare på. Spørreskjemaet fylles ut ved kontorkonsultasjon, ved telefonkonsultasjon eller ved enkel kontakt (frammøte eller bud), hvor det utstedes sykmeldingsblankett. Vennligst ikke registrer tilfeller hvor sykmeldingsblanketten er dokumentasjon på barns sykdom, eller sykdom under attføring. Sykmeldingstilfeller ved legevakt registreres heller ikke. Skjema fylles ut en gang for en og samme sykmeldingsepisode. Alle svarene som gis er basert på din egen vurdering.
Personvern Alle opplysningene som gis behandles strengt fortrolig, og de vil ikke få konsekvenser for sykmeldingene. Datatilsynet har gitt prosjektet konsesjon for opprettelse av personregister. Prosjektledelsen vil følge de anbefalingene om bruk og oppbevaring av data som Datatilsynet, Den regionale komité for medisinsk forskningsetikk i helseregion II, og Rikstrygdeverkets juridiske kontor har gitt i sine vurderinger av prosjektet. Undersøkelsen foregår blant sykmeldte i Aust-Agder i januar 1996.
Levering av spørreskjema til de sykmeldte I de tilfeller hvor sykmeldingsblanketten utstedes ved telefonkonsultasjon eller ved frammøte/bud, ber vi om at spørreskjemaet til den sykmeldte enten overleveres vedkommende idet han/hun henvender seg i legekontorets skranke for å hente den utfylte sykmeldingsblanketten, eller at skjemaet med svarkonvolutt vedlegges sykmeldingsblanketten som oversendes vedkommende per brev.
Hvem foretar undersøkelsen Ansvarlig for undersøkelsen er Fylkestrygdekontoret i Aust-Agder i samarbeid med Rikstrygdeverket og Seksjon for trygdemedisin ved Universitetet i Oslo. Er det noe du lurer på i forbindelse med undersøkelsen, kan du ringe rådgivende overlege Harald Reiso ved Fylkestrygdekontoret i Aust-Agder, tlf. 37 02 40 90.
Om utfylling av spørreskjemaet Bruk bare ett kryss for å markere svaret på hvert spørsmål, med unntak av spørsmål 13. Spørsmål 10 og 11 oppfattes kanskje som mest aktuelle ved langvarige sykmeldinger, men vi ber deg svare på de spørsmålene også ved sykmeldinger av kortere varighet. I spørsmål 14 skal legen angi på hvilket grunnlag vurderingen av den sykmeldtes arbeidsevne ble gjort. Også legens historiske kjennskap til aktuell klinikk eller funn hos den sykmeldte kan tas med ved legens valg av avkryssing. Særlig vil det gjelde ved muntlig kontakt mellom legen og den sykmeldte, per telefon eller ved frammøte, i forbindelse med utstedelsen av sykmeldingsblanketten.
Utfylte skjema sendes samlet til Fylkestrygdekontoret i Aust-Agder, ved Harald Reiso. Bruk vedlagte forhåndsfrankerte svarkonvolutt. Med hilsen Gunnar Tellnes Harald Reiso
Overlege, Professor dr. med. Rådgivende overlege Rikstrygdeverket og Fylkestrygdekontoret i Aust-Agder Seksjon for trygdemedisin Institutt for allmennmedisin og samfunnsmedisinske fag Universitetet i Oslo
Snu arket! 1. Sykmeldingsblanketten er utfylt ved: 10. I hvilken grad tror du pasienten evner å ta annet eller tilpasset arbeid, på nåværende eller 1 Kontorkonsultasjon 2 Telefonkonsultasjon annen arbeidsplass? 3 Enkel kontakt (frammøte/bud) 1 I svært stor grad 2 I stor grad (dag, måned, år): 2. Dato for utfylling 3 I middels stor grad ______/ ______- ______4 I ikke særlig stor grad 5 I ubetydelig grad 3.1 Pasientens fødselsdato (dag, måned, år): ______/ ______- ______11. Tror du pasienten har reelle muligheter, i 3.2 Pasientens initialer: ______dagens arbeidsmarked, til å ta annet eller tilpasset arbeid, på nåværende eller annen arbeidsplass?
4. Pasientens kjønn 1 Ja, helt sikkert 2 Ja, trolig 1 Mann 3 Nei, usikker 2 Kvinne 4 Nei, helt sikkert ikke 5. Tilfellet gjelder 1 Ny sykmelding 12. Mener du at det er forhold utenfor de rent 2 Forlengelse av sykmelding medisinske som gjør at pasientens arbeidsevne er nedsatt? 6. I hvilken grad er pasienten sykmeldt i dag? 1 Ja, helt sikkert 1 100% 2 Ja, trolig 2 Delvis sykmeldt. Angi prosentgrad: ______% 3 Nei, usikker 4 Nei, helt sikkert ikke
7. Første fraværsdag for denne sykmeldingsperioden: Dersom du har krysset av på svaralternativ 1 eller 2, kan du Dato (dag, måned, år): skrive her eller i margen hvilke forhold det gjelder? ______/ ______- ______
8. Diagnose 13. Hvilke metoder eller undersøkelser/tester brukte du i denne konsultasjonen? 8.1 Hoved-diagnosen som er årsak til sykmeldingen: (Sett ett eller flere kryss) ______Anamnese Klinisk undersøkelse av pasienten Hoveddiagnose, ICPC nr.: ______Supplerende undersøkelser (f. eks. blodprøver, spirometri, EKG m.m.) 8.2 Eventuelt viktigste bi - sykmeldingsdiagnose: Henvisning til samarbeidspartnere (f. eks. sykehus, spesialist, fysioterapi, m.m.) ______
Bidiagnose, ICPC nr.: ______14. På hvilket grunnlag vurderte du den sykmeldtes arbeidsevne i dag? Vurderingen baserte seg hovedsakelig på: Arbeidsevne per dags dato 1 Pasientens utsagn om egen arbeidsevne alene, eller 9. I hvilken grad er pasientens evne til å utføre sitt 2 Mest utsagn, noe klinikk/funn, eller vanlige arbeid nedsatt? 3 Noe utsagn, mest klinikk/funn, eller (Her angis den nedsettelse som skyldes de plagene pasienten 4 Klinikk/funn alene sykmeldes for. Med vanlig arbeid menes det arbeidet pasienten hadde på sykmeldingstidspunktet) 15. Hvordan tror du den sykmeldtes situasjon i 1 Svært mye nedsatt forhold til sykmelding vil være om 4 uker? 2 Mye nedsatt 3 Middels nedsatt 1 Vedkommende vil trolig være helt sykmeldt 4 Ikke særlig nedsatt 2 Vedkommende vil trolig være delvis sykmeldt 5 Ubetydelig nedsatt 3 Vedkommende vil trolig være helt friskmeldt
Takk for hjelpen! Doc. number: Patient number: Dear doctor Thank you for participating in this project concerning certified sickness absence. On the back of this sheet you find the questions we request you to answer. Use the questionnaire at ordinary consultations, telephone consultations, or at short over-the-counter contacts where sickness certificates are issued. Please do not record certificates that serve as documentation of children’s disease, or certificates issued for patients on rehabilitation benefits. Certificates issued at house-calls should not be recorded. Complete the questionnaire once during an episode of certified sickness absence. All answers should be based on your own assessments. Personal rights All information is handled with discretion, and will not influence the absence episodes. The study was approved by the Regional Ethics Committee for Medical Research, the Norwegian Data Inspectorate, and the Legal Affairs Division in the National Insurance Administration. The study is conducted among patients certified sick in the county of Aust-Agder in January 1996. Distribution of questionnaires to the patients When certificates are issued by telephone consultations, or by short over-the-counter contacts, we ask you to distribute the questionnaires to the sick-listed persons when they take contact to get the certificates. Otherwise, enclose the questionnaires and the accompanying post-paid envelopes together with the certificates that are mailed to your sick-listed patients. Who is in charge of the investigation? The Aust-Agder County Office of the National Insurance Services is in charge of the investigation, in co- operation with the National Insurance Administration and the Department of Social Insurance Medicine, the Institute of General Practice and Community Medicine, at the University of Oslo. If you have any questions concerning the study, do not hesitate to contact the senior medical advisor Harald Reiso, telephone 37 02 40 90. About completing the questionnaire Use only one tic-mark to indicate the answer category for each question, except question 13. Question 10 and 11 may be considered most appropriate in prolonged episodes of certified sickness absence, but we ask you to answer them when absence of shorter duration is concerned also. In question 14 you shall indicate your main information sources for assessing the sick-listed’s work ability. Also your historical knowledge of clinical findings of the actual patient can be taken into account. This is especially the case in telephone consultations, or for short over-the counter contacts. Completed Questionnaires are forwarded to The Aust-Agder County Office of the National Insurance Services, attention Harald Reiso. Use the enclosed post-paid envelope. Best regards Gunnar Tellnes Harald Reiso
Senior doctor, Professor PhD Senior medical advisor The National Insurance Administration and The Aust-Agder County Office of the National Insurance Services The Institute of General Practice and Community Medicine, The Department of Social Insurance Medicine, The University of Oslo
Turn! 1. This sickness certificate is issued by: 10. To what degree is the patient capable of 1 Office consultation performing other form of work than his regular 2 Telephone consultation job, at his present or other workplace? 3 Short over-the-counter contact 1 Extremely capable 2. Date of consultation (day, month, year): 2 Very capable 3 Moderately capable ______/ ______- ______4 Minimally capable 5 Incapable 3.1 Patient’s date of birth (day, month, year): ______/ ______- ______11. Do you believe that the patient has a realistic chance, in the present job market, to get other 3.2 Patient’s initials: work that takes his reduced function into 4. Patient’s gender consideration, at his present or other workplace? 1 Man 2 Woman 1 Yes, certainly 2 Yes, probably 5. The certification concerns 3 No, but uncertain 1 New certification 4 Absolutely not 2 Prolongation of previous certification 12. Do you consider that non-medical factors 6. Percent certification today? could have influenced the patient’s reduced work 1 Full-time (100%) ability? 2 Part-time, percentage: ______% 1 Yes, certainly 2 Yes, probably 7. The first day of absence concerning this episode of 3 No, but uncertain certified sickness absence: 4 Absolutely not Date (day, month, year): If you ticked off alternative 1 or 2, you may write here, or ______/ ______- ______in the margin, what factors this concerns:
8. Diagnoses 13. What methods or tests did you use in this 8.1 The primary diagnosis responsible for the sick- consultation? (Tick off one or more options) leave: Patient history ______Physical findings Supplementary testing Primary diagnosis, ICPC number: ______(e.g. laboratory tests, spirometry, ECG etc.) Referral to other healthcare workers 8.2 Other contributing diagnosis: (e.g. consultants, hospitals, physical therapists etc.) ______14. What was the main clinical information Secondary diagnosis, ICPC number:______source for your assessment of the patient’s work ability today?:
Work ability today 1 The patient’s self-reported work ability only, or 2 Mainly self-report, supplemented by clinical 9. To what degree is the patient’s ability to findings, or perform his/her ordinary work reduced? 3 Mainly clinical findings, suppl. by self-report, or (Tick off for the reduction caused by the disorder for which 4 Clinical findings alone the patient is sick-listed . Ordinary work is the work the patient had at the time of certification) 15. What do you believe the patient’s situation 1 Very much reduced concerning certified sickness absence will be in 2 Much reduced four weeks? 3 Moderately reduced 4 Not much reduced 1 Probably full-time certification 5 Hardly reduced at all 2 Probably part-time certification 3 Probably returned to work
Thank you! Lege Pasient
Til deg som er sykmeldt Bakgrunn I samarbeid med din vanlige lege, utfører vi et forskningsprosjekt om sykmeldinger, og vi håper du kan hjelpe oss i den forbindelse. Nederst og på baksiden av dette arket finner du noen spørsmål som vi ønsker å få svar på. Personvern Alle opplysningene som gis av deg behandles strengt fortrolig, og de vil ikke få konsekvenser for din sykmelding. Din lege får ikke vite hva du svarer. I undersøkelsen vil du få et kodenummer, og alle opplysningene blir på den måten anonymisert. Det er kun noen få personer i ledelsen av prosjektet som har adgang til kodesystemet, og disse personene har taushetsplikt. Når undersøkelsen er ferdig, blir alle data tilintetgjort. Prosjektledelsen vil følge anbefalingene om bruk og oppbevaring av data som Datatilsynet, Den regionale komité for medisinsk forskningsetikk i helseregion II og Rikstrygdeverkets juridiske kontor har gitt i sine vurderinger av undersøkelsen. Prosjektet foregår blant sykmeldte i Aust-Agder i januar 1996. Frivillighet Deltakelse i undersøkelsen er frivillig. Opplysningene du gir vil bli brukt til å øke våre kunnskaper om hvordan sykdom virker inn på din og andre sykes arbeidsevne, i forhold til hvilke belastninger som er på arbeidsplassene. Vi ønsker derfor å vite mer om hvilke arbeidsbelastninger du har i ditt arbeid, og hvordan du selv vurderer din arbeidsevne. Dersom du på et senere tidspunkt ønsker å trekke tilbake de opplysningene du har gitt, kan du når som helst gjøre det ved å kontakte Fylkestrygdekontoret i Aust- Agder. Hvem foretar undersøkelsen Ansvarlig for undersøkelsen er Fylkestrygdekontoret i Aust-Agder i samarbeid med Rikstrygdeverket og Seksjon for trygdemedisin ved Universitetet i Oslo. Er det noe du lurer på i forbindelse med undersøkelsen, kan du ringe rådgivende overlege Harald Reiso ved Fylkestrygdekontoret i Aust-Agder, tlf. 37 02 40 90. Om utfylling av spørreskjemaet Sett bare ett kryss for å markere ditt svar på hvert av spørsmålene. Spørsmål 12 og 13 oppfatter du kanskje som mest aktuelle hvis du har vært langvarig sykmeldt, men vi ber deg svare på de spørsmålene selv ved sykmelding av kortere varighet. Opplysninger om hvor lang din sykmelding blir, vil bli innhentet fra trygdekontoret. Legg utfylt skjema i vedlagte forhåndsfrankerte konvolutt og lim igjen. Konvolutten leveres i skranken på kontoret til legen din, eller legges i nærmeste postkasse. Med hilsen Gunnar Tellnes Harald Reiso
Overlege, Professor dr. med. Rådgivende overlege Rikstrygdeverket og Fylkestrygdekontoret i Aust-Agder Seksjon for trygdemedisin Institutt for allmennmedisin og samfunnsmedisinske fag Universitetet i Oslo
1. Skriv her hvilken kommune du bor i: 2. Skriv her hva slags yrke eller stilling du har: Snu arket! Arbeidsbelastninger Arbeidsevne 3. Hvilken arbeidstidsordning har du vanligvis? Hvordan du vurderer din arbeidsevne akkurat i dag
1 Dagarbeid 11. I hvilken grad er din evne til å utføre ditt vanlige 2 Fast kveldsarbeid arbeid nedsatt? (Her angis den nedsettelsen som skyldes de 3 Fast nattarbeid plagene du er sykmeldt for) 4 Skiftarbeid (gjelder også sjøfolk/offshore-arbeid) 1 Svært mye nedsatt 5 Annet: ______2 Mye nedsatt 3 Middels nedsatt 4. Hvor mange timer er du vanligvis i lønnet arbeid 4 Ikke særlig nedsatt (i gjennomsnitt) hver uke? 5 Ubetydelig nedsatt Antall hele timer: ______12. I hvilken grad tror du at du evner å ta annet eller 5. Hvilken beskrivelse passer best på det arbeidet du tilpasset arbeid, på nåværende eller annen arbeidsplass? utfører? 1 I svært stor grad 2 I stor grad 1 Mest stillesittende arbeid 3 I middels grad 2 Arbeid som krever at en står eller går mye 4 I ikke særlig grad 3 Arbeid hvor man går og løfter mye 5 I ubetydelig grad 4 Tungt kroppsarbeid 5 Annet: ______13. Tror du at du har reelle muligheter, i dagens arbeidsmarked, til å ta annet eller tilpasset arbeid, 6. Må du i ditt arbeid daglig løfte noe som veier mer på nåværende eller annen arbeidsplass? enn 20 kilo, og i tilfelle hvor mange ganger pr. dag? 1 Ja, helt sikkert 1 Nei, ikke aktuelt for meg 2 Ja, trolig 2 Ja, mellom 1 og 4 ganger daglig 3 Nei, usikker 3 Ja, mellom 5 og 20 ganger daglig 4 Nei, helt sikkert ikke 4 Ja, mer enn 20 ganger daglig 14. Mener du at det er forhold utenfor de rent 7. Kan du selv bestemme når du vil ta pauser fra medisinske som gjør at din arbeidsevne er nedsatt? arbeidet? 1 Ja, helt sikkert 2 Ja, trolig 1 Ja, som regel 3 Nei, usikker 2 Ja, i noen grad 4 Nei, helt sikkert ikke 3 Sjelden Dersom du har krysset av på svaralternativ 1 eller 2, kan du 4 Nei skrive i kanten på dette arket hvilke forhold det gjelder? o 8. Opplever du konflikter eller dårlige arbeidsforhold på din hovedarbeidsplass? 15. Føler du at det arbeidet du nå er sykmeldt fra, er Mellom ledelsen og ansatte Mellom ansatte det rette arbeidet for deg? 1 Aldri 1 Aldri 2 Sjelden 2 Sjelden 1 Ja, i svært stor grad 3 Av og til 3 Av og til 2 Ja, i stor grad 4 Ofte 4 Ofte 3 I middels stor grad 5 Daglig 5 Daglig 4 Nei, ikke i særlig grad 5 Nei, i svært liten grad 9. Hvor stressende eller psykisk belastende synes du ditt arbeid er? Sett en ring rundt det tallet som passer best 16. Føler du at du selv har innflytelse på hvordan etter din oppfatning: du utfører det arbeidet du nå er sykmeldt fra?
(1 = Ikke stressende Svært stressende = 7) 1 Ja, i svært stor grad 2 Ja, i stor grad 1 2 3 4 5 6 7 3 I middels stor grad 4 Nei, ikke i særlig grad 10. Dersom du arbeider overtid, hvordan synes du at 5 Nei, i svært liten grad antall overtidstimer er med hensyn til ditt personlige velvære? 17. Hvordan tror du at din situasjon i forhold til sykmelding vil være om 4 uker? 1 Passe 2 For mye 1 Jeg vil trolig være helt sykmeldt 3 For lite 2 Jeg vil trolig være delvis sykmeldt 4 Ikke aktuelt, jobber ikke overtid 3 Jeg vil trolig være helt friskmeldt Takk for hjelpen! Doc. number: Patient number: Dear patient Background In co-operation with your doctor, we are conducting a research project concerning certified sickness absence, and hope that you will participate in it. We are interested in your answers to the questions in this questionnaire. Personal rights All information that you give will be handled with discretion, and will not influence your absence. Your doctor will not be informed of your answers. You will be assigned a code number, which will make your answers anonymous. Few persons in the project have access to the code system, and they have professional secrecy. When the project is finished, the questionnaire will be destroyed. The project will follow the recommendations about use and storage of data given by the Regional Ethics Committee for Medical Research, the Norwegian Data Inspectorate, and the Legal Affairs Division in the National Insurance Administration. The study is conducted among patients certified sick in the county of Aust-Agder in January 1996. Participation Your participation is voluntary. The information you provide will be used to increase our knowledge about how health problems may affect your, and other sick-listed person’s work ability. We therefore want to know more about the work demands you experience in your job, and how you assess your work ability. If You want to withdraw the information You have given later, be free to do so by contacting the Aust-Agder County Office of the National Insurance Services. Who is in charge of the investigation? The Aust-Agder County Office of the National Insurance Services is in charge of the investigation, in co- operation with the National Insurance Administration and the Department of Social Insurance Medicine, the Institute of General Practice and Community Medicine, at the University of Oslo. If you have any questions concerning the study, do not hesitate to contact the senior medical advisor Harald Reiso, telephone 37 02 40 90. About completing the questionnaire Use only one tic-mark to indicate the answer category of each question. Question 12 and 13 you may consider not to be actual unless you are long-term sick-listed, but please try to answer them even though your absence will be of short duration. The National Insurance Administration will supply information about the duration of your absence. Put the completed questionnaire into the accompanying post-paid envelope. Deliver the envelope to personnel at the doctor’s office, or at the nearest mailbox. Best regards Gunnar Tellnes Harald Reiso
Senior doctor, Professor PhD Senior medical advisor The National Insurance Administration and The Aust-Agder County Office of the National Insurance Services The Institute of General Practice and Community Medicine, The Department of Social Insurance Medicine, The University of Oslo
1. Write where you live here: 2. Write your occupation here: Turn! Work demands Work ability
3. Which shift do you work? How do you assess your work ability today? 11. To what degree is your ability to perform your 1 Day 2 Evening ordinary work reduced? 3 Night (Tick off for the reduction caused by the disorder that you are 4 Rotating shift (including seamen/offshore-work) certified sick because of) 5 Other: ______1 Very much reduced 2 Much reduced 4. How many remunerative hours per week (mean) do 3 Moderately reduced you usually work? 4 Not much reduced 5 Hardly reduced at all Number of hours: ______12. Are you capable of performing other work, or a job 5. Which of the following describes your job best? that takes your health problem into account, at your present or other workplace? 1 Mostly sitting 2 Mostly standing/walking 1 Extremely capable 3 Walking with a great deal of lifting 2 Very capable 4 Heavy physical labour 3 Moderately capable 5 Other: ______4 Minimally capable 5 Incapable 6. Do you have to lift items weighing more than 20 kg at 13. Do you believe that you have a realistic chance, in work? If so, how often? the present job market, to get other work that takes your reduced function into consideration, at your present or 1 No other workplace? 2 Yes, 1-4 times daily 3 Yes, 5-20 times daily 1 Yes, certainly 4 Yes, more than 20 times daily 2 Yes, probably 3 No, but uncertain 7. Can you take a break when you wish? 4 Absolutely not 14. Do you consider that non-medical factors could have 1 Yes, generally 2 Yes, at times influenced your reduced work ability? 3 Seldom 1 Yes, certainly 4 Never 2 Yes, probably 3 No, but uncertain 8. Are there conflicts or poor working conditions where 4 Absolutely not you work? If you ticked off alternative 1 or 2, you may write here in the margin which factors this concerns? o Between administration & workers Among workers 1 Never 1 Never 15. Do you think your present job is the right job for you? 2 Seldom 2 Seldom 1 Yes, absolutely 3 Now and then 3 Now and then 2 Yes, probably 4 Often 4 Often 3 Maybe 5 Daily 5 Daily 4 No, not especially 5 No, not at all 16. Do you have the possibility to influence the tempo 9. How stressful is your work? Circle the number which and content of your present job? is most appropriate in your opinion: 1 Yes, nearly completely (1 = No stress Extremely stressful = 7) 2 Yes, significantly 3 Somewhat 1 2 3 4 5 6 7 4 No, not much 5 No, not at all 10. If you work overtime, is the amount of overtime in 17. What do you believe your situation concerning accordance with your wishes? certified sickness absence will be in four weeks? 1 Yes 1 Probably full-time certification 2 No, too much overtime 2 Probably part-time certification 3 No, too little overtime 3 Probably returned to work 4 I don’t work overtime Thank you! Questionnaire II – patients Answered by patients. Presented in Norwegian, and one way translated to English.
Aktiv rygg er bedre rygg i Aust-Agder Til deg som er sykmeldt for ryggplager Bakgrunn I samarbeid med din vanlige lege, utfører vi et prosjekt som prøver å bedre behandlingstilbudet for ryggpasienter i Aust-Agder. Som ledd i å vurdere virkningen av ulike behandlingstiltak, har vi utarbeidet noen spørsmål som vi håper du kan svare på. Spørsmålene er på begge sider av dette arket! Personvern Alle opplysningene som gis av deg behandles strengt fortrolig, og de vil ikke få konsekvenser for din sykmelding. I undersøkelsen vil du få et kodenummer, og alle opplysningene blir på den måten anonymisert. Det er kun noen få personer i ledelsen av prosjektet som har adgang til kodesystemet, og disse personene har taushetsplikt. Prosjektledelsen vil følge Datatilsynets anbefalinger om bruk og oppbevaring av data. Frivillighet Deltakelse i prosjektet og besvarelse av dette spørreskjemaet, er frivillig. Dersom du på et senere tidspunkt ønsker å trekke tilbake de opplysningene du har gitt, kan du når som helst gjøre det ved å kontakte Fylkestrygdekontoret i Aust-Agder, ved prosjektleder Jørn Lindalen, telefon 37 00 43 00. Om utfylling av spørreskjemaet Sett bare ett kryss for å markere ditt svar på hvert av spørsmålene. Utfylt skjema legges i vedlagte blanke konvolutt, og leveres til den som du fikk skjemaet av i dag. Med hilsen Ryggpoliklinikken ASA Fylkestrygdekontoret i Aust-Agder
1. Din fødselsdato (dag/måned/år): _____ / _____ / _____ og ditt personnummer (5 siffer): ______
2. Ditt fornavn og etternavn: ______, ______
3. Dagens dato (dag/måned/år): ______/ ______/ ______4. Din bokommune: ______
5. Skriv her hva slags yrke eller stilling du har: ______
6. Skriv antall år med utdannelse du har utover niårig (eventuelt sjuårig) grunnskole: ______
7. Ditt kjønn:1 Mann 8. Har du i det hele tatt vært sykmeldt tidligere? 1 Ja 2 Kvinne 2 Nei
9. Din arbeidsevne akkurat i dag. I hvilken grad 1 Nei medfører dine ryggplager at din evne til å utføre ditt 2 Ja vanlige arbeid er nedsatt? (Med vanlig arbeid menes det 13. Om røyking akkurat i dag arbeidet du nå er sykmeldt fra) 1 Jeg røyker ikke 1 Svært mye nedsatt 2 Jeg røyker, men ikke daglig 2 Mye nedsatt 3 Jeg røyker daglig mindre enn 10 sigaretter 3 Middels nedsatt 4 Jeg røyker daglig, fra 10 til 19 sigaretter 4 Ikke særlig nedsatt 5 Jeg røyker daglig 20 sigaretter eller mer 5 Ubetydelig nedsatt 14. Hvordan tror du at din situasjon i forhold til 10. To spørsmål om smerteplager. Hvor mye sykmelding vil være om 4 uker? smerter føler du at ryggplagene gir deg akkurat i dag? 1 Ingen smerter 1 Jeg vil trolig være helt sykmeldt 2 Smerter av og til 2 Jeg vil trolig være delvis sykmeldt 3 Kontinuerlige smerter 3 Jeg vil trolig være helt friskmeldt
11. Sett også ett kryss på streken nedenfor i forhold til det som passer best etter din oppfatning: (1 = Ikke smerter i det hele tatt Uutholdelige smerter = 10) Snu arket! 1 ______10
12. Om røyking. Har du noensinne vært dagligrøyker? Bruk bare ett kryss for å markere ditt svar på hvert av spørsmålene
15. Fysisk form / Kondisjon 18. Sosiale aktiviteter I løpet av de siste 2 ukene: Hva var den største I løpet av de siste 2 ukene: Har din fysiske eller fysiske belastningen du greide eller kunne greid i psykiske helse begrenset dine sosiale aktiviteter og minst to minutter? kontakt med familie, venner, naboer eller andre?
1 Meget stor (f. eks. løpe fort) 1 Ikke i det hele tatt 2 Stor (f. eks. jogge i rolig tempo) 2 Bare litt 3 Moderat (f. eks. gå i raskt tempo) 3 Til en viss grad 4 Lett (f. eks. gå i vanlig tempo) 4 Ganske mye 5 Meget lett (f. eks. gå sakte, eller kan ikke gå) 5 I svært stor grad
16. Følelsesmessig problem 19. Bedre eller dårligere helse I løpet av de siste 2 ukene: Hvor mye har du vært Hvorledes vil du bedømme helsen din akkurat i dag, plaget av psykiske problemer, slik som indre uro, fysisk og psykisk, sammenlignet med for 2 uker angst, nedforhet eller irritabilitet? siden?
1 Ikke i det hele tatt 1 Mye bedre 2 Bare litt 2 Litt bedre 3 Til en viss grad 3 Omtrent uforandret 4 En god del 4 Litt verre 5 Svært mye 5 Mye verre
17. Daglige aktiviteter 20. Samlet helsetilstand I løpet av de siste 2 ukene: Har du hatt vansker med I løpet av de siste 2 ukene: å utføre vanlige gjøremål eller oppgaver enten Hvorledes vil du vurdere din egen fysiske og innendørs eller utendørs, på grunn av din fysiske psykiske helse i allminnelighet? eller psykiske helse?
1 Ikke vansker i det hele tatt 1 Svært god 2 Bare lette vansker 2 God 3 Til en viss grad 3 Verken god eller dårlig 4 En god del vansker 4 Dårlig 5 Har ikke greid noe 5 Meget dårlig
Utfylt skjema legges i vedlagte blanke konvolutt, og leveres til den som du fikk skjemaet av i dag. Takk for hjelpen! Active back Dear patient with back disorder Background In co-operation with your doctor, we are conducting a research project that will try to improve therapy for patients with back disorders in Aust-Agder. The project concerns patients certified sick. In order to evaluate the effect of different therapeutic efforts, we hope you can answer the following questions. Questions are on both sides of this form! Personal rights All information you provide will be handled with discretion, and will not influence your sick leave benefits. In the project you will receive a code-number, making your answers anonymous. Only a few persons in charge of the project will have access to the code-system, guaranteeing your privacy. The project will follow the recommendations about use and storage of data given by the Norwegian Data Inspectorate. Participation Your participation is voluntarily. If you want to withdraw the information you have given later, be free to do so by contacting the Aust-Agder County Office of the National Insurance Services, at 37 00 43 00. About completing the questionnaire Use only one tic-mark to indicate the answer category of each question. Put the completed questionnaire into the enclosed envelope and deliver it to the person who gave it you. Best regards The low back pain outpatient clinic at Aust-Agder County Office of Aust-Agder County Hospital The National Insurance Services
1. Your date of birth (day/month/year): _____ / _____ / _____ your citizen number (5 digits): ______
2. Your first and last name: ______, ______
3. Today’s date (day/month/year): ______/ ______/ ______4. Your address: ______
5. Your occupation : ______
6. Number of years of education beyond compulsory school: ______
7. Your gender:1 Man 8. Have you ever been certified sick? 1 Yes 2 Woman 2 No
9. Your work ability today. To what degree does 12. Smoking habits. Have you ever smoked on a daily your back disorder reduce your ability to perform basis? your ordinary work today? (Ordinary work is the job 1 No from which you are now certified sick) 2 Yes 1 Very much reduced 2 Much reduced 13. About smoking today 3 Moderately reduced 1 I do not smoke 4 Not much reduced 2 I smoke, but not daily 5 Hardly reduced at all 3 I smoke, but less than 10 cigarettes daily 4 I smoke between 10 and 19 cigarettes daily 10. Regarding your pain. How much pain do you feel 5 I smoke 20 cigarettes or more daily that the low back pain disorder gives you today? 1 No pain 14. What do you believe your situation 2 Intermittent pain concerning certified sickness absence will be in 3 Continuous pain four weeks?
11. Mark with an x on the following line what best 1 Probably returned to work describes your pain : 2 Probably part-time certification 3 Probably full-time certification (1 = No pain at all Intolerable pain = 10) 1 ______10 Turn! Use only one tic mark to answer the following questions
15. Physical fitness 18. Social activities During the past 2 weeks… What was the hardest During the past 2 weeks… physical activity you could do for at least 2 Has your physical or emotional health limited your minutes? social activities with family, friends, neighbours or groups? 1 Very heavy (ex. run, at fast pace) 2 Heavy (ex. jog, at slow pace) 1 Not at all 3 Moderate (ex. walk, at a fast pace) 2 Slightly 4 Light (ex. walk, at a medium pace) 3 Moderately 5 Very light (ex. walk, at a slow pace or not able 4 Quite a bit to walk) 5 Extremely
16. Feelings During the past 2 weeks… How much have you 19. Changes in health been bothered by emotional problems such as How would you rate your overall health now feeling anxious, depressed, irritable or downhearted compared with 2 weeks ago? and sad? 1 Much better 1 Not at all 2 A little better 2 Slightly 3 About the same 3 Moderately 4 A little worse 4 Quite a bit 5 Much worse 5 Extremely
17. Daily activities 20. Overall health During the past 2 weeks… How much difficulty During the past 2 weeks… have you had doing your usual activities or tasks, How would you rate your health in general? both inside and outside the house because of your physical and emotional health? 1 Excellent 1 No difficulty at all 2 Very good 2 A little bit of difficulty 3 Good 3 Some difficulty 4 Fair 4 Much difficulty 5 Poor 5 Could not do
Put the completed questionnaire into the enclosed envelope and deliver it to the person who gave it you. Thank you!