1

A SIMPLIFIED, QUALITATIVE STRATEGY FOR THE ASSESSMENT OF OCCUPATIONAL RISKS AND SELECTION OF SOLUTIONS: CONTROL BANDING

EEN VEREENVOUDIGDE KWALITATIEVE METHODE OM BEROEPSGEBONDEN RISICO’S TE BEPALEN EN OPLOSSINGEN TE SELECTEREN: CONTROL BANDING

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft, op gezag van de Rector Magnificus Prof.ir. K.C.A.M. Luyben voorzitter van het College voor Promoties, in het openbaar te verdedigen op woensdag 22 december om 10.00 uur

door David Mark ZALK Masters of Public Health, University of California, Berkeley geboren te Boston, Massachusetts, Verenigde Staten

2

Dit proefschrift is goedgekeurd door de promotor: Prof.dr. A. R. Hale

Copromotor Dr. P.H.J.J. Swuste

Samenstelling promotiecommissie:

Rector Magnificus, voorzitter Prof.dr. A.R. Hale, Delft University of Technology, promoter Dr. P.H.J.J. Swuste, Delft University of Technology, copromotor Prof.dr P.Vink, Delft University of Technology Prof.dr. F.J,H, van Dijk, University of Amsterdam, Academic Medical Center Prof.dr.ir. A. Burdorf, Erasmus University, Rotterdam Prof. M.P. Guillemin, Lausanne University, Switzerland Dr. P.W. Johnson, University of Washington, USA

ISBN 978-1-4507-4664-6

This work performed, in part, under the auspices of the US Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-BOOK-461828

3

A simplified, qualitative strategy for the assessment of occupational risks and selection of solutions:

Control Banding

by

David M. Zalk

4 Het volmaakte is de vijand van het goede.

The perfect is the enemy of the good.

 2010, D.M. Zalk  Cover photograph by D.M. Zalk  Cover photo: Shipbreaking in South Africa

All rights reserved. No part of this publication may be reproduced in any form by print, photoprint, microfilm or any other means without written permission of the rightful claimant(s). This restriction concerns the entire publication or any part of it.

5 FOREWORD

L‘chayim! In Hebrew it is a toast that means ―to life‖ and is said when drinking to a person‘s health or well-being. As the completion of this thesis is indeed a celebration, it is important to give thanks to all those who have given assistance to get me to this milestone. Of course, I am grateful to Andrew Hale my promoter and Paul Swuste, my good friend and instigator, for his guidance. I would especially like to thank my parents and brother. To my mother, Sophia, for paving the way for this opportunity by being an example of excellence and having full faith in my potential. To my father, Bertram, for his inspiring me to make a difference in the world. To my brother, Reuven, for his bringing smiles to those around him as well as his religious inspiration to the family. I would also like to thank my father-in-law, Dr. Marvin Kirsh, for his strong and continual support and my late mother-in-law, Helen Kirsh, for her eternal optimism for my potential to succeed. This effort could not have been accomplished without my friends and colleagues, both at LLNL and around the world, who have helped throughout this journey and whom I always feel blessed to have in my life. I would especially like to thank Paul Oldershaw for his groundbreaking Control Banding vision and Berenice Goelzer for her longstanding support and encouragement for keeping things simple on behalf of the world‘s workers. Thanks to the International Occupational Hygiene Association for their support of Control Banding globally. Much of the US Control Banding work could not have been done without the help of the Control Bandits, with a special thanks to my longstanding partner in crime Deborah Imel Nelson. Then there‘s my best work ever, my three boys. To Joshua, whose strength to overcome obstacles is an inspiration; Jacob, whose humor and compassion are endless; and Jesse, my loving baby bear and dancing machine. Of course, no great work can be built without a solid foundation. My wife Janice is indeed that foundation, the love of my life. She has given me undying support, through highs and lows, and encompassing all things important to me. She is the laugh that breeds smiles and an amazing mother to my boys. However, the real inspiration for this thesis is all the workers who have taught me how to be a better professional and human being. This is especially true for all the workers I have not met, and will not meet, whose daily efforts put them at unnecessary risk. They do this so they can go home to their families and put bread on their table. In the Talmud it is said ―saving one life is like saving the entire world‖; this is the essence of preventing worker exposures to unnecessary risks and the cornerstone of my dedication to completing this thesis. To the over 2.5 billion workers who have never seen a health and safety professional, I offer the efforts of my work and a heartfelt L‘chayim!

6 CONTENTS LIST CHAPTERS

Chapter 1 Introduction 8 Chapter 2 History and evolution of control banding: a review. Published: Journal of Occupational and Environmental Hygiene 16 Chapter 3 Examination and evaluation of custodial ergonomics to modify training procedures. Published: Professional Safety 39 Chapter 4 Evaluating the control banding nanotool: a qualitative risk assessment method for controlling nanoparticle exposures. Published: Journal of Nanoparticle Research 49 Chapter 5 Review of qualitative approaches for the construction industry; designing a risk management toolbox. Submitted: Annals of Occupational Hygiene 69 Chapter 6 Risk level based management system: a control banding model for occupational health and safety risk management in a highly regulated environment. Published: Industrial Health 88 Chapter 7 Barrier Banding: a concept for safety solutions utilizing control banding principles. Submitted: Safety Science 102 Chapter 8 Discussion and Concluding Remarks 117

APPENDICES

Appendix A Application of a pilot control banding tool for risk level assessment and control of nanoparticle exposures. Published: Annals of Occupational Hygiene 123 Appendix B Grassroots ergonomics: initiating an ergonomics program utilizing participatory techniques. Published: Annals of Occupational Hygiene 137 Appendix C Participatory Occupational Hygiene; a path to practical solutions. Published: Asian-Pacific Newsletter 145 Appendix D IOHA/ICOH Declaration on Occupational Hygiene. Published: Asian-Pacific Newsletter 147 Appendix E – Carpenter shop wood dust control - practical experience to reduce hardwood dust exposures below the ACGIH TLV. Published: Applied Occupational and Environmental Hygiene 151

ACRONYM GLOSSARY 168

REFERENCES 170

SUMMARY 183

CURRICULUM VITAE 184

7

CHAPTER 1 INTRODUCTION

Control Banding and Solutions Around the world there are 2.8 billion workers. Due to occupational risk annually there are 2.3 million deaths, 270 million non-fatal injuries, and 160 million work-related diseases for the global workforce (Takala 2010; ILO 2005a; Fingerhut et al., 2005). These injuries and illnesses cost more then 5% of the worldwide Gross Domestic Product (GDP) (ILO 2005b). To assist in identifying solutions toward preventing these injuries and illnesses for the 2.8 billion workers globally, World Health Organization (WHO) estimates that, at best, 10 – 15% of these workers have access to a basic standard of occupational health services (WHO 2003). Therefore, approximately 2.5 billion of the world‘s workers do not have access to occupational safety, health, and hygiene (OSHH) professionals and their ability to address and reduce occupational risks that lead to the occurrence of work-related illnesses, diseases, and safety-related accidents and injuries is very limited. The multidisciplinary risks that workers face, as well as the adverse outcomes that often result, are well understood in the OSHH scientific literature. What has been lacking in this literature is a simplified strategy for delivering solutions to the 2.5 billion workers facing unnecessary occupational risks on a daily basis.

In the not too distant past international experts were in agreement that the scarcity of solutions within the OSHH scientific literature was an issue requiring improvement (Swuste 1996, Roelofs et al. 2003). In recent years a strategy for qualitative occupational risk management (ORM) has gained international attention and has provided a robust framework for improvement (Nelson and Zalk 2010). This strategy, known as Control Banding, offers a simplified approach for reducing work-related risks. The objective of Control Banding is to develop a qualitative risk assessment that stratifies a given workplace hazard across -usually- four levels or ‗bands‘ of risk that are placed in a risk matrix, using four risk levels of severity on one axis and the probability of outcome on the other axis, with the outcome (intersecting point from the two axes) leading directly to control solutions. This means that there is an opportunity to deliver preventive methods to reduce occupational risks for workers lacking access to OSHH experts. This success with Control Banding and its solutions-based focus has also presented a quandary in the world of the occupational hygiene sciences, where quantification is historically king. Control Banding uses a less rigid, qualitative approach, since even some of the long considered state-of-the-art, literature-based scientific findings OSHH professionals rely on change on a regular basis. In this respect the European Union (EU) is learning cutting edge Control Banding lessons from India and the United States (US) is being educated by recent Control Banding developments in Chile (Fingerhut 2008). It is Control Banding‘s simplification in reducing work-related exposures, and changing the science of quantitative risk assessment to the art and science of qualitative risk assessment, that is redefining ORM. However, within this simplification is the fundamental question of whether Control Banding can achieve risk reduction comparable to quantitative methods.

The success of Control Banding did not occur in isolation. Oliver (1902) not only predicted increased risk in the ―dangerous trades‖ ―as international competition becomes keener, and our manufacturers endeavour to produce more cheaply by increasing the speed of their machinery,‖ but also the outcome that ―workpeople incur risks without knowledge of the danger or the means of preventing it.‖ This prescience found traction in the 1980s, addressing these issues in an evolutionary process, which was an historical progression of a series of national and international 8 solutions-based initiatives. Without these initiatives, and the experts who succeeded in developing methods to spread solutions internationally in the absence of the web-based capabilities of modern day, Control Banding as it is currently known would not be possible. This historical progression of research-based programs promoting solutions began with the International Labor Organization (ILO) Work Improvement in Small Enterprises (WISE) and Work Improvement in Neighborhood Development (WIND), the WHO and the International Occupational Hygiene Association (IOHA) Prevention And Control Exchange (PACE), and national solution initiatives in Australia, Denmark, and the US, as well as the EU Solbase databank of solutions (Kogi 2010; Kawakami and Kogi 2001; Swuste et al., 2003). In developed countries, the establishment of occupational health and safety laws and regulations drove the search and dissemination of efficient and practical models to prevent work-related diseases. Early initiatives to compile solutions include the noise control solutions from the United Kingdom (UK) Health and Safety Executive (HSE 1983) and in mining from Australia (Mitchell and Else 1993), and chemical substitution strategies from Denmark, the UK, the US, and The Netherlands to reduce health hazards (Swuste and Hale 1994). This process has evolved into efforts such as the European Solbase, with many nations teaming together to develop a database of effective controls for workplace hazards and reduction of occupational risks (Swuste et al 2003).

Achieving a Global Initiative The eventual combining of paths for Control Banding and the historical solutions initiatives came about as a matter of necessity. Efforts by the WHO and ILO to compile the global burden of work-related risks have been essential for crystallizing the need for action in assisting the 2.5 billion workers without access to OSHH professionals (ILO 2005a; ILO 2005b, Fingerhut et al., 2005). However, it is important to note that the inability of developing countries to address workplace illnesses, injuries, and accidents is not a direct outcome of an aggressive and productive global economy as the world‘s most competitive, economically developed countries are also the safest (ILO 2005b). The best of our OSHH prevention sciences, as implemented in many Established Market Economies (EMEs), have regulatory enforcement assisting large enterprises and governmental organizations and the regulatory development and the growth of the OSHH professions. However, these opportunities have neither been effectively transferred to small- and medium-sized enterprises (SMEs) in EME countries nor to all industries in non-EME countries (ILO 2005b). The opportunity for Control Banding grew out of the need to address this massive gap as the current context and construct of our quantitatively focused OSHH sciences are not in a position to be applied in a manner that would significantly affect the 85 - 90% of the world‘s workers that remain unprotected.

The historical solutions initiative to assist non-EME countries in this manner has its roots in Southeast Asia and participatory methods (Kawakami and Kogi 2001). This region of the world has succeeded in numerous grassroots initiatives across the OSHH professions over the years and is a prime example of participatory training and interventions to achieve practical solutions (see Chapter 3). The WISE program began in 1993 as a three-year project in the Philippines and is based on ILO-developed methodologies from as early as 1976 (ILO 2005b). The approach of WISE is to provide simple tools for identification and prioritization of safety and health measures that seek to prevent injury and illness in the smaller industries (Kawakami and Kogi 2001). The WIND training program has a focus on agriculture and was developed in 1995 for application in Vietnam within local farming villages (ILO 2005b). The training manual for WIND focuses on improving safety, health, and working conditions in family-based agricultural settings with a primary emphasis on easy and practical solutions. It is within this framework of sharing successes and assisting SMEs with solutions that a common need grew to develop an ORM 9 model that could implement controls without relying on quantitative methods and the OSHH professionals. The solutions initiatives developed the framework to share information, whilst Control Banding would develop the necessary model.

The standard Control Banding model groups hazards into stratified risk ‗bands‘ within a risk matrix, identifying commensurate controls within ‗toolkits‘ to reduce the level of risk and promote worker health and safety. Shared amongst Control Banding solution approaches are also task-to-control initiatives that are not banded in a risk matrix, rather the level of controls increase as the risk of the task increase. Distinct from the Control Banding approaches are the historic progression of solution initiatives that present a more comprehensive ORM approach to achieve prevention. These initiatives emphasize participatory approaches together with simplified solutions to better manage the reduction of occupational risks over time. These simplified solutions, collected through research in the literature and in practice, can be found in many of the controls that are the outcome of Control Banding strategies. An optimal approach harnesses the strengths and utility of both Control Banding and solutions initiatives to achieve primary prevention of work-related risks in a simplified and practical manner. Although the primary focus of this thesis is Control Banding, an example of this differentiation presented in this thesis can be found in Chapter 3 and Appendix B on ergonomics. Chapter 3 focuses on the development of two qualitative Hazard Banding approaches for ergonomics. Hazard Banding is a preliminary step for establishing a relative level, or band, of risk for a given hazard that does not have controls as a direct outcome. Preferably, these ergonomics Hazard Banding results lead to changes of procedures, acquisition of ergonomic tools, and development of trainings on all these outcomes that collectively serve as controls to reduce high-risk postures. Appendix B focuses more holistically on developing ergonomics programs for SMEs in developing countries using Hazard Banding, participatory methods, and solutions initiatives that collectively can be used in leading toward a Control Banding approach. The combination of Hazard Banding, Control Banding, and solutions initiatives to achieve a comprehensive and simplified ORM model can be seen in the multidisciplinary programs in chapters 5 and 6.

Qualitative Model In generic terms, Control Banding is a qualitative approach to risk assessment and ORM that groups occupational risk control strategies into bands based on their level of rigor. In the initial Control Banding theory there is a 4-level strategy for controlling exposures to chemicals, in ascending order of rigor, covering: (1) good IH practices, including personal protective equipment (PPE), (2) engineering controls, including local exhaust ventilation (LEV), (3) containment, and (4) the need to seek specialist advice depending on the particular CB strategy. The approach for chemical substances is to assign them to control bands based on their potential hazard. These hazards are captured by risk phrases, indicators of toxicity, or other hazardous properties and in some cases the potential for exposure based on quantity in use, volatility or dustiness. Later Control Banding approaches seek to expand this theory beyond bulk chemicals. Other Control Banding strategies focus on the task performed to directly assign PPE and control options without banding risk and without an interim step of assessing potential exposure.

Within the initial Control Banding concept the only established program, or toolkit, available to the public at the turn of the century was the UK‘s Control of Substances Harmful to Health (COSHH) Essentials (see Chapter 2). These chemical exposure reduction models, intentionally simplified for use by SMEs, are based on a historical progression of Control Banding models used in the biological and pharmaceutical industries in the 1980s and 1990s. A fundamental difference in these Control Banding approaches, although they both serve to develop a qualitative 10 risk-based control matrix often in the absence of toxicological data, is that older models were developed for OSHH expert use and the newer models are primarily intended for use by non- experts within SMEs. In as much as Control Banding strategies have been applauded as a tool for SMEs, they have also received their share of criticism. The COSHH Essentials is a focus of much of this criticism as it was released as an SME toolkit without any validation or evaluation of its utility in practice (see Chapter 2). The NIOSH Control Banding document (2009) summarizes many of these criticisms, issues, and weaknesses including: estimation of control approaches compared to solid scientific protocols, reduction of the need for traditional exposure assessment, qualitative outputs not being as good as quantitative scientific protocols, and its narrow utility for bulk chemical processes. These criticisms will be addressed in the subsequent chapters.

For the majority of the toolkits currently available, the art of qualitative risk assessment is quite basic (NIOSH 2009). A hazard is defined and considered, the level of risk is stratified to a minimal number of components, and the commensurate controls necessary to reduce each risk level to an acceptable outcome becomes the output. Determining the number of risk levels, which fits the concept of bands, is a result of balancing the intricacy/complexity of the hazard with the needs of the worker. Throughout this process, keeping in mind that the worker is the end-user of the method is often the most difficult thing to remember. Theoretically, for the worker, there is functional understanding that there are two risk levels relating to their tasks when performing work; one that is unsafe and should not be done (red light) and one that is safe and should be done (green light). Having three risk levels, or red-yellow-green lights if you will, have been found in practice to have its limitations with the middle, yellow light option leading to inappropriate judgment over a wide range of risk as it can cover a wide range of often frequent adverse outcomes. To cope with this ambiguous middle band, the classic Control Banding delineation of bands, or risk levels (RLs) tends toward four as it essentially divides the yellow into a better decision matrix to ensure the commensurate controls are in place (Figure 1).

In the end it is all about scaling prevention to a given situation in a toolkit designed as a complement to the OSHH professions, rather than a replacement. The more difficult a toolkit is perceived to be in its conceptual phase, the less likely it is to make it to the development phase.

Figure 1. A basic Control Banding 4 x 4 model for delineating Risk Levels. A Risk Level (RL) is an outcome of probability and severity inputs, with RL1 as the lowest risk and RL4 the highest.

11

When the COSHH Essentials, the first online available toolkit, began being used within countries throughout the world - in part through the efforts of IOHA and the WHO Collaborating Centre Network for Occupational Health (WHOCC) – at some level it was felt that it would become the qualitative risk assessment of choice. To the surprise of many, although it was often the first toolkit to be used within different countries, it was rarely put into practice without first being modified to match the requirements and/or environment of the adopting country. Rather, national organizations often found that regulatory requirements would not support a program developed for the UK as a national approach and would take the fundamental approach to the COSHH Essentials as the basis for development of their own toolkit to fit their own national regulations. The driver for resource expenditure for the development of national toolkits was quite often updates in national regulatory requirements for risk assessments. These updates often left the SMEs, typically 80 – 90% of their working population, without the ability to achieve regulatory compliance in the same manner as was seen in the UK. This situation is best exemplified in the EU Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) regulations and the outcome of national toolkits developed in the Netherlands and Germany. It can also be contrasted with the US, where regulations remain painfully static with no modern updates for risk assessment, which is reflected in the fact that the national development of toolkits is currently non-existent. It is within this environment that the WHOCC has played an important role of matching, or twinning, EME and non-EME entities to share information and assist in toolkit development. The International Control Banding Workshops (ICBWs) have also played an important role in uniting the history of the solutions-based initiatives with the development of Control Banding strategies. Supported throughout by IOHA in a manner comparable to the PACE initiative, the ICBWs afforded a globally supported opportunity for establishing needs, sharing of knowledge, and offering a forum for the continued validation and evaluation of Control Banding toolkits available. Most recently, the 3rd (Pilanesburg), 4th (Seoul), 5th (Cape Town), and 6th (Rome) ICBWs have in many ways reversed the role of EMEs sharing their knowledge with non-EMEs as the developing countries are quite often the ones sharing their lessons learned in a manner that significantly assists developments in the EU and US (Nelson and Zalk 2010).

12 Qualitative Better Than Quantitative? Today we look at a buffet of Control Banding opportunity that has the broad flavors of international cuisine and now a deep support of the OSHH professionals, the chefs if you will, that are continuing to add to the menu. Control Banding has found broad support in scientific literature when there is an absence of knowledge, whether in the absence of occupational exposure limits (OELs) or when new hazards emerge, such as with pharmaceuticals or nanotechnologies, where the toxicology and epidemiology bases lag far behind the use of the chemicals. Significant advances were made with the development of the Control Banding Nanotool, a toolkit for controlling nanomaterial exposures (see Appendix A). The Nanotool has played an important role in the professional acceptance of Control Banding as its qualitative risk assessment Risk Level matrix (as seen in Figure 1) is now considered by many national entities to be a more powerful ORM model then its quantitative counterpart (see Chapter 4). By having Control Banding grace the pages of prestigious research journals and becoming an integral approach for national exposure reduction programs, it now has an opportunity to become an integral part of the future of OSHH sciences. Control Banding is also, quite surprisingly in many instances, serving as an invaluable risk communication tool within and between the OSHH professions, even serving to help Occupational Physicians in the EU better understand the role and value of Occupational Hygiene, a focus on prevention assisting in growing this valuable profession with an outcome that can only serve to better protect workers (Zalk and Heussen 2010). Currently in continued development are the Barrier Banding concepts (see Chapter 7) and Ergonomics approaches (see Chapter 3 and Appendix B), both harnessing Control Banding‘s risk communication potential and opening the door to invaluable future multidisciplinary collaborations to achieve prevention. Together, both through the solutions-based initiatives and the growth of qualitative risk assessment, these collaborations lead toward more simplified occupational health and safety management systems based on Control Banding that afford utility to both non-experts and experts alike. The success of the Control Banding ORM model for comprehensive OSHH programs, like the Risk Level Based Management System (RLBMS - see Chapter 6), present an excellent example of combining resources to maximize the most beneficial outcome for workers. The future is in a multidisciplinary union of both qualitative and quantitative approaches across the OSHH sciences, as this unified approach is proving successful for SMEs as well as the largest of industrial and research enterprises. This multidisciplinary union assists independent professional experts to maintain their independence with improvement in risk communication (Zalk and Heussen 2010). However, for non-experts, this multidisciplinary approach intentionally does not keep each discipline separate, but rather combines the professions into a singular toolbox approach by necessity (see Chapter 5, NIOSH 2009). Thackrah (1832) saw this multidisciplinary need two centuries earlier, taking to examining the working professions to understand ―the atmosphere they breathe, -the muscular exercise they take, - the postures of body they maintain.‖ Perhaps Control Banding in the 21st century presents the missing link, taking the best of the quantitative and qualitative OSHH sciences to achieve the simplification of solutions necessary to achieve a substantive reduction in the global burden of work-related disease, illness and injury.

Thesis questions and structure Against this background of the development of Control Banding, the main objective of this thesis is to determine Control Banding‘s viability across the OSHH professions. Therefore a central question must be asked to address this objective. Can Control Banding‘s simplified, qualitative models and solutions initiatives be integrated within the multidisciplinary requirements of ORM to address appropriate control needs are implemented commensurate to risk, to assist in reducing

13 work-related injury and illness? Achieving this would require answering the following sub- questions: 1) Can the application of Control Banding and Hazard Banding strategies expand beyond bulk and liquid chemicals to establish a foundation for a unified ORM model across the OSHH professions?  See Chapters 2, 3, 4 and 7. 2) Can qualitative risk assessment models select the appropriate controls at a level comparable with traditional quantitative models?  See Chapters 2, 4 and 6. 3) Can multidisciplinary ORM models that integrate a qualitative risk assessment approaches be developed within an occupational trade and within a regulatory framework?  See Chapters 5 and 6.

The structure of this thesis is based on the development of qualitative strategies toward addressing the central question. It begins with an historical review of the Control Banding topic (Chapter 2) and follows with its successful expansion beyond bulk chemicals with simplified, qualitative models and solutions for ergonomics (Chapter 3) and nanomaterials (Chapter 4). Building on this foundation, qualitative ORM models for a multidisciplinary assessment of occupational risk and selection of solutions is presented for the construction industry (Chapter 5) and as an occupational health and safety management system in a regulatory framework (Chapter 6). Ending with an eye toward the future, a strategy for delivering safety solutions within a Control Banding model is proposed (Chapter 7) and thesis findings are discussed (Chapter 8).

The concept of thesis began with OSHH experts in agreement that the need for achieving prevention for the global workforce was of utmost importance. However, the scope of the Control Banding strategies was a far too narrow in application at the time, covering only bulk chemicals in a manufacturing industry. However, there was also a significant split amongst OSHH professionals on the value and importance of Control Banding. Therefore, aspects of the first two sub-questions above can be found in the next chapter addressing the history and evolution of Control Banding. Chapter 2 addresses both of these issues by distinguishing the modern incarnation of Control Banding from its historical predecessors. It presents the available scientific literature on the validation and evaluation of toolkits to establish the foundation to clearly identify strengths and weaknesses and the path forward for the future research. This includes identifying the expansion necessary to include a broader base of chemical exposures and the inclusion of point source exposures (e.g. silica) that are common in the construction and mining trades. Chapter 3 also addresses this expansion with a Hazard Banding approach for identifying high-risk postures to focus in on the commensurate controls necessary to reduce musculoskeletal disorder risks. This approach is expounded on in Appendix B with a collection of simplified solutions initiative approaches for ergonomics. This control strategy emphasizes the use of a participatory approach beneficial for a practical approach to the primary prevention work-related injuries and illnesses. Collectively this practical primary prevention approach affords an ORM program that addresses risk reduction methods for the prevention of musculoskeletal disease common to the world‘s workplaces. Control Banding‘s strength and versatility is presented in Chapter 4, with an evaluation of the Nanotool as the first established method to control occupational exposure to nanomaterials. This chapter also addresses the first two thesis sub-questions and many Control Banding criticisms by directly and adequately

14 addressing the validation of a qualitative risk assessment against its traditional quantitative counterpart.

With the scope of Control Banding strategies and solutions initiatives further developed, Chapter 5 and 6 explore the development of multidisciplinary ORM models in addressing the third thesis sub-question. Chapter 5 focuses on Control Banding approaches for the construction trades, uniting individual toolkits and multiple solutions-oriented approaches within a singular, trade- specific toolbox. When extending the multidisciplinary potential of the Control Banding strategies, there is the need to provide a singular approach to achieve practical primary prevention of work-related illness and injury for a global workforce. Chapter 6 presents a multidisciplinary ORM model as an occupational health and safety management system that integrates both qualitative and quantitative risk assessment approaches within a regulatory environment that is effective for the largest of industries in any country. This Risk Level Based Management System (RLBMS) was developed for use at the Lawrence Livermore National Laboratory (LLNL) as an occupational health and safety management system that combines both Control Banding strategies and the solutions initiatives approach with an emphasis on participatory methods that maximize worker involvement in its processes. LLNL is a US national research laboratory that has a high degree of regulatory oversight at both national and local levels. In managing its own processes and workplace risks, LLNL has developed the RLBMS to manage the prioritization of the OSHH experts‘ work, the integration of Control Banding strategies, the sharing of solutions using quantitative and qualitative assessment techniques, and the establishment of a common risk communication language to benefit workers and enhance their participation in procedure development. In Chapter 7 the topic of banding in occupational safety is investigated in a manner that addresses both the strengths and weaknesses of utilizing a Control Banding process to reduce occupational accidents and injuries. Finally, a discussion of the findings in this thesis and a summary of the main conclusions are found in Chapter 8.

The appendices contain articles that focus on practical developments in occupational hygiene, the profession that developed the Control Banding strategies. Both chemical-specific examples as well as international initiatives that focus on exchanging solutions-based information are presented. Appendix A shows the logic behind the development of the Control Banding Nanotool to address a topic that had stifled the occupational hygiene profession‘s quantitative foundation, and shows its versatility in leading with a qualitative-based solutions approach. Appendix B, as discussed above, shows a holistic approach for creating ergonomics programs for SMEs and in developing countries by combining the outcomes of Hazard Banding approaches with participatory methods and solutions initiatives to control musculoskeletal disorders. Appendix C identifies how occupational hygienists can learn from the other OSHH professions to increase practicality, here borrowing from the participatory ergonomics approach to better assist workers worldwide. The author‘s efforts to ratify the IOHA/ICOH Declaration to Strengthen the Position of Occupational Hygiene, an international effort to bolster the focus on practical prevention in occupational health, can be found in Appendix D. Participatory occupational hygiene in practice can be found in Appendix E, where the difficulty of controlling for wood dust in a carpenter‘s shop led to early solutions-based efforts to better protect workers with control methods they assisted in developing.

As Control Banding is the major focus of this thesis, some repetition on the topic can be seen in the introduction to many chapters. Although this was reduced as much as feasible in Chapters 1 and 8, it is hoped that the readers understand the need for this redundancy in what are reprints of published papers. 15

CHAPTER 2 HISTORY AND EVOLUTION OF CONTROL BANDING: A REVIEW Journal of Occupational and Environmental Hygiene, 5(5):330 – 346 (2008) David. M. Zalk1 and Deborah Imel Nelson2 1. Lawrence Livermore National Laboratory, P.O. Box 808 L-871, Livermore, CA 94551-0808 [email protected] 2. Geological Society of America, P.O. Box 9140, Boulder, CO, 80301-9140, [email protected]

Control Banding (CB) strategies offer simplified solutions for controlling worker exposures to constituents often encountered in the workplace. The original CB model was developed within the pharmaceutical industry; however, the modern movement involves models developed for non-experts to input hazard and exposure potential information for bulk chemical processes, receiving control advice as a result. The CB approach utilizes these models for the dissemination of qualitative and semi-quantitative risk assessment tools being developed to complement the traditional industrial hygiene model of air sampling and analysis. It is being applied and tested in small and medium size enterprises (SMEs) within developed countries and industrially developing countries; however, large enterprises (LEs) have also incorporated these strategies within chemical safety programs. Existing research of the components of the most available CB model, the Control of Substances Hazardous to Health (COSHH) Essentials, has shown that exposure bands do not always provide adequate margins of safety, that there is a high rate of under-control errors, that it works better with dusts than with vapors, that there is an inherent inaccuracy in estimating variability, and that when taken together the outcomes of this model may lead to potentially inappropriate workplace confidence in chemical exposure reduction in some operations. Alternatively, large-scale comparisons of industry exposure data to this CB model‘s outcomes have indicated more promising results with a high correlation seen internationally. With the accuracy of the toxicological ratings and hazard band classification currently in question, their proper reevaluation will be of great benefit to the reliability of existing and future CB models. The need for a more complete analysis of CB model components and, most importantly, a more comprehensive prospective research process remains and will be important in understanding implications of the model‘s overall effectiveness. Since the CB approach is now being used worldwide with an even broader implementation in progress, further research toward understanding its strengths and weaknesses will assist in its further refinement and confidence in its ongoing utility.

INTRODUCTION A foundation of the modern movement for Control Banding (CB) strategies is derived from programs initiated in the United Kingdom (UK) by the Health and Safety Executive (HSE). The need to provide guidance and assistance to small and medium size enterprises (SMEs, which employ about 90% of the UK workforce (Garrod 2003)) in meeting requirements to conduct risk assessments of chemical exposures led to the HSE development of a program known as the Control of Substances Hazardous to Health (COSHH) Essentials. In 1998 the HSE published a series of papers outlining a CB strategy of creating a model in which the hazard was combined with the potential exposure to determine a recommended level of control approach. European Union (EU) risk phrases were used to rank the hazard of a chemical, and potential for exposure was estimated by the quantity in use, and the volatility of liquids or dustiness of solids. The scheme uses information associated with hazardous chemicals to develop hazard groups. These 16

hazard groups are derived for a variety of chemicals and are designated by experienced toxicologists. When a hazard group associated with a chemical is selected by the manager of a Small- and Medium-Sized Enterprise (SME), toxicological expertise is utilized without the need for an on-site expert. This is an important foundation for the eventual consideration of the exposure potential to the chemical. The remainder of the decision making process includes the volume of chemical used, and likelihood of the chemical becoming airborne, estimated by the dustiness or volatility of the source compound. When these parameters are entered into a work sheet, the suggested control approach is identified. The end product is the selection of a control guidance sheet with both general and specific advice for common tasks (Jackson 2002).

In the development of the CB model, Maidment stressed the importance of limiting the number of factors in the model to reduce its complexity and increase its applicability for non-experts (Maidment 1998). Although in theory there can be a stratification of risk across many levels, each additional level leads to a more intricate tool for the SME manager, which as an end product may hamper its overall intended utility. To achieve this balance of simplicity and effectiveness Maidment suggested four categories, or ―bands‖, to assist in preventing exposure to chemicals. These four control strategies are a grouping of three levels of engineering containment based on sound industrial hygiene (IH) principles, with professional IH expertise as a fourth category. Within this model, these generic control strategies have also been adapted to address chemical exposure potential where the control guidance sheet (CGS) approaches may not be appropriate or practical. These other CB strategies utilize the banding approach to assist in directly assigning personal protection equipment (PPE) such as an appropriate level of respiratory protection and addressing dermal exposure potential (Garrod 2003).

In a historical context, the banding of risk began in the 1970s and 1980s relating to explosive events, radiation, lasers and biological agents. The pharmaceutical industry should be credited with the initiation of exposure control categorization utilizing an industrial hygiene basis with its work in the late 1980s and early 1990s (Sargent 1998; Naumann, et al. 1996). During this period approaches to protect workers handling products with limited pharmacological and toxicological data led to efforts to stratify toxicological hazards and link them directly to simplified, commensurate control strategies during the production phases of product development (Farris, et al. 2006; Tait 2004). These control approaches for pharmacological agent exposures were divided into five hazard categories (Naumann, et al. 1996). This effort to address the growing potency of newly developed compounds followed the path of the microbiological and biomedical industries controlling exposures to increasingly toxic microorganisms within the four categories of the Biosafety Level approach (CDC 1993). Formally, the establishment of in-house Occupational Exposure Bands (OEBs) by the Association of the British Pharmaceutical Industry (ABPI (1995)) assisted the product development phase of the industry to achieve a method for compliance with the COSHH regulations in a manner later adapted to the COSHH Essentials to address chemical exposures. There were several forces beyond the regulatory realm that also led to the CB model‘s adaptation and expansion into the chemical arena. Perhaps the most significant was the recognition that the traditional process of establishing occupational exposure limits (OELs), against which measurements of airborne concentrations of chemicals could be compared to ensure that exposures are controlled, was quickly losing ground by orders of magnitude to the increasing number of chemicals posing a threat to worker health1. Forces that drive the evolution of the CB model continue to this day. The nanotechnology industry is seeing itself akin to pharmaceutical and microbiological industries in that they are facing similar limitations in toxicological data. A CB model that addresses exposure to nanoparticlulate has recently been 17

presented in concept as a practical approach to achieve exposure control in the absence of this data (Maynard 2007).

REVIEW OF CB LITERATURE

The peer-reviewed literature on CB approaches (mostly relative to COSHH Essentials) can be summarized according to the development of the models, the use of databases to support the models, and the models‘ validation. CB has its roots in a number of qualitative (Swuste et al., 2003) and semi-quantitative (Balsat et al, 2003) risk assessment approaches which began to appear in the 1970s and evolving in the 1980s relating the assessment of catastrophic failure probabilities at large chemical facilities (Money 2003). An example of this is a risk matrix describing the likelihood and probable severity of an event, e.g. an explosion or release of toxic material, developed for use by a large chemical enterprise. As Money (2003) presented, there are a number of relevant strategies that were borrowed from and built upon during previous efforts and it is not always possible to trace the steps by relying on chronological appearance in the peer-reviewed literature. What is evident is that there was much exchange of information and ideas amongst occupational health practitioners and scientists in the chemical, biological, and pharmaceutical industries during that period of time (Garrod 2003; Naumann et al., 1996; Money 1992).

Model development; linking toxicology to control In an early, perhaps the first, published report in which toxicological data were linked directly to an appropriate level of control. Money (1992) presented a structured approach to design and operation of a chemical plant that handles aromatic amines, nitro compounds, and equivalent agents with carcinogenic potential based on a carcinogenic ranking system. This was a broad approach for ensuring that appropriate measures would be in place to control risks from these chemicals from both routine and abnormal operations. It was truly simple in that it utilized a basic exposure scenario where the only determinant of exposure was the veracity of the toxicological data. Money suggested that this approach, which covered both inhalation and skin contact, should be applicable to similar approaches ranking relative hazards of chemicals (Henry and Schaper 1990; Gardner and Oldershaw 1991; Woodward et al., 1991).

This toxicology-to-control approach described by Money (1992) began by using four categories of toxicological outcome relating to carcinogenic potential, collapsed from a system utilizing six that considers both carcinogenic potency and weight of evidence (Crabtree et al., 1991). Money argued that while it is important to distinguish the potencies of different substances, in reality such a separation is artificial and impractical. Linearly matched with these four levels of carcinogenic potency were four levels of controls, progressing in complexity and stringency. Putting them together, these toxicology-to-control levels are then summarized as: (1) for all chemicals, use good basic IH; (2) for suspected animal carcinogens, increase to isolation of moderate exposure potential; (3) for suspected human carcinogens with moderate exposure potential increase, to containment and regular audits; and (4) for proven carcinogens with low exposure potential, increase to automated bulk transfers and process control.

The toxicology-to-control model was also applied by Nauman et al. (1996) to exposures to pharmaceutical active ingredients in laboratory and manufacturing operations. The pharmaceutical industry had traditionally used risk assessment methods to establish OELs for active ingredients; however, the increasing potency of these agents led to a new approach based on the Biosafety Level concepts used in laboratories. Substantiated by a large database of air 18

monitoring data for various operations they were able to distinguish five hazard categories (or performance-based exposure control limits, PB-ECL), based on toxicological and pharmacological properties of these agents and on the engineering controls and administrative procedures known to be effective in controlling exposure levels.

The Chemical Industries Association (CIA) further addressed toxicological information for chemical agents in their guidelines for safe handling of colorants (second version) (CIA 1993). In this document, inputs of hazard categorization (1-4), hazard classification (e.g., toxic, corrosive), associated risk phrase, and guideline control level (8-hour TWA) were linked to control recommendations for each hazard category. As both the CIA guidelines and the COSHH regulations were created in the UK, an ongoing discussion of chemical agent models began to develop. According to Guest (1998) the advice of the COSHH Approved Code of Practice, i.e., to set a self-imposed working standard for chemicals which did not have an official OEL, could not be followed by industry or government, due to the technical complexity of establishing OELs, the lack of adequate toxicological databases and experts, and the sheer volume of substances covered in the European Inventory of Existing Substances (EINECS (1987)). These factors led the CIA to develop chemical categorization guidelines for their member organizations.

Building on the earlier CIA guidance (1993) and the work of Gardner and Oldershaw (1991), the later CIA guidelines (1997) incorporated the Chemical Hazardous Information and Packaging (CHIP) Risk Phrases and guideline control levels, in addition to data on adverse effects in humans. The purpose of these guidelines was to provide a simple, broad-based, integrated approach for use by CIA members in classifying hazards. The categories were to be called OEBs and would only be developed when there were no other in-house, national, or international OELs. They would define the upper limit of acceptable exposure. As the number of control strategies is usually limited to approximately four levels, this approach was designed to cover 6 orders of magnitude, plus a special category. The upper limits (OEB C for dusts, OEB D for gases / vapors) were designed to ―reflect good occupational hygiene practice‖ and the maximum dust concentration in the COSHH regulations (10 mg/m3).

Model development; the exposure prediction step At this juncture, no one had yet factored the probability of exposure into the risk assessment and risk management aspects of a CB model. Although it had not yet been incorporated into the equation, much work was being conducted during the 1990s on predicting exposures. For example, Burstyn and Teschke‘s (1999) review on the methods of studying the determinants of exposure included work tasks, equipment used, environmental conditions, and existing controls. In evaluating the risk, a dedicated exposure model was used that is based on Cherrie and Schneider (1999) by providing subjective exposure assessment using a structured approach based on descriptive workplace activities and environment. Using this model, subjective exposure assessment showed significant correlation with exposure measurements across 63 jobs and four different agents (asbestos, toluene, mixed respirable dust, and man-made mineral fibers). This serves as an excellent example of how dissecting existing models can lead to criteria to be used in developing other exposure control models and future toolkits.

In studies of determinants of exposure reviewed by Burstyn and Teschke (1999), there was little attention devoted to volume of product used, and less to the physical characteristics of chemicals in use. The HSE played a pivotal role in developing a regulatory approach based on these concepts used to date (Russell et al., 1998; Brooke 1998). While the work of the HSE was 19

based in large part on that of the UK CIA (EINICS 1987; CIA 1997), which categorized substances into OEBs, it is apparent from the preceding discussion that many other groups have contributed to the development of COSHH Essentials. The challenge facing the HSE was to develop guidance which was practical for SMEs, used available hazard information, was easy to use and understand, and which relied upon readily available information (see Table I). These goals can be realized by using European risk phrases (R-phrases) and simple predictors of exposure to conduct a generic risk assessment, which leads to straightforward recommendations on risk management, i.e., control approaches.

Table I. Factors used in HSE‘s core model (Russell et al., 1998). HEALTH +  GENERIC  HAZARD EXPOSURE RISK CONTROL POTENTIAL ASSESSMENT APPROACH Substances Substances Combination of Type of allocated allocated to a health hazard approach to a hazard dustiness or and exposure needed to band using volatility potential factors achieve R-phrases band and a determine adequate band for the desired level of control scale of use control

The COSHH Essentials approach, as it later came to be known, builds on earlier approaches (Naumann et al., 1996; Henry and Schaper 1990; Gardner and Oldershaw 1991; CIA 1997; RCS 1996; CIA 1992). It also offers two other significant advances: it is specifically developed for SMEs and it includes control advice. The key components of the model include the hazard banding, exposure potential, and control approaches. Hazard banding is described more fully below (Brooke 1998). It is important to point out, however, that from a British perspective, COSHH Essentials is limited to substances classified under CHIP, thereby excluding, e.g., pesticides and pharmaceuticals, which are outside the scope of those regulations, and also process-generated hazards such as wood dust, silica dust, and welding fumes. Exposure banding is a function of physical properties leading to likeliness for the material to become airborne (volatility of liquids or dustiness of solids, and the quantity in use) (Maidment 1998). These elements are combined to determine the appropriate control approach (see Table II). Therefore, there is perhaps a stronger link in the modern evolution of the CB model to the work of Burstyn (1999) and Cherrie (1999) than to the earlier toxicology-to-control approaches. Later versions of COSHH include PPE Essentials, offering advice for gloves and respirators, and for addressing dermal risks. Another feature of the COSHH Essentials web site is the newer Direct Advice topics for accessing hazard guidance by specific tasks, services, and processes (e.g., foundries, woodworking, beauty treatments, pubs, clubs and restaurants).

Table II. Control approaches used in COSHH Essentials (Russell et al., 1998). Control approach 1 – General ventilation. Good standard of general ventilation and good working practices. Control approach 2 – Engineering control. Ranging from local exhaust ventilation to ventilated partial enclosure. Control approach 3 – Containment. Containment or enclosure, allowing for limited, small scale breaches of containments. Control approach 4 – Special. Seek expert advice. 20

The developers felt that operation-based control guidance sheets (CGS) would provide the best format for advising SMEs. The approximately 100 CGS now available are structured according to a standard format (Oldershaw 2001). This format contains sections on: design and equipment, maintenance, examination and testing, cleaning and housekeeping, PPE, training, supervision, a short list of references, a sample schematic of an engineering control, and an employee checklist for proper utilization of controls. Russell et al. (1998) states that use of the scheme will not in itself constitute a suitable and sufficient workplace risk assessment; it must therefore be considered as guidance and not a replacement for traditional IH. Employers should still consider other factors in their risk assessments, such as the need for health surveillance and the need to monitor exposure to ensure adequacy and suitability of controls. Similarly, it was pointed out that an over-protective approach would lack credibility, and deter promotion efforts and implementation, whereas an under-protective approach would not protect workers. Weighing these factors, it was generally agreed in the model development that a conservative approach would be the most responsible.

Brooke (1998) outlined three criteria for the toxicological basis of the UK approach: (1) simple and transparent, (2) make best use of available hazard information, and (3) recommend control strategies that vary according to degree of health hazard. The R-phrases that are agreed to throughout the EU facilitated these criteria, as they address all relevant toxicological endpoints. This idea had been proposed previously be Gardner and Oldershaw (1991) and had formed the basis of similar strategies (ABPI 1995; CIA 1987; RSC 1996). Brooke noted differences between these approaches and that of the HSE. COSHH Essential includes alignment between dust and vapor target exposure ranges and dose level cut-off values and is based on achievement of exposure levels anywhere in the target range, whereas the CIA recommends that exposures should be maintained ―as low as reasonably practicable‖ (ABPI 1995; Guest 1998; CIA 1997). Brooke‘s article achieved two goals: first, it explained the assignment of R-phrases to the Hazard Bands A-E utilized in the COSHH Essentials; and second, it compared these assignments to health-based OELs. The hazard bands, which are based on toxicological considerations, each divided by an order of magnitude in concentration range. As the relationship between the part per million (ppm) concentration of a vapor and the mg/m3 concentration is a function of its molecular weight (and also temperature and pressure, though not discussed in this article), the working group which oversaw development of this approach decided to adopt a pragmatic approach and to align the exposure bands as seen in Table III below. Due to this alignment, ―in mg/m3 terms, the concentration range for substances in vapor form is substantially higher than that for the substance in particulate form, for the same toxicological hazard band.‖

21

Table III. Allocating R-phrases to hazard bands (Brooke 1998). Hazard band Target R-phrases airborne concentration range (Note 1) A >1–10 mg/m3 R36, R38, all dusts and dust; >50-500 vapors not allocated to ppm vapor another band (Note 2) B >0.1-1 mg/m3 R20/21/22, R40/20/21/22 dust; >5-50 ppm vapor C >0.01-0.1 R48/20/21/22, R23/24/25, mg/m3 dust; R34, R35, R37, >0.5-5 ppm R39/23/24/25, R41, R43 vapor D <0.01 mg/m3 R48/23/24/25, R26/27/28, dust; <0.5 ppm R39/26/27/28, R40 Carc. vapor Cat. 3, R60, R61, R62, R63 E See specialist R40 Muta. Cat. 3, R42, advice R45, R46, R49 S: skin and eye Prevention or R34, R35, R36, R38, R41, contact reduction of R43, Sk (Note 3) skin and/or eye exposure

In writing about the development of the model, Maidment (1998) stressed the importance of limiting the number of factors in the model to control its complexity and applicability. This simplicity was to be balanced with the hazard and exposure potential parameters necessary to predict an adequate control strategy. Toward this end, control strategies were collapsed into four main categories (Table II). Since characteristics of exposure potential can be summarized as those related to physical properties and those related to substance handling, Maidment focused on the dustiness of solids, and the volatility of liquids. The study indicated that three dustiness bands would adequately describe the properties of dusts and maintain the simplicity of the model: low, medium, and high. For liquids, the volatility of a liquid would be captured by consulting a graph of boiling point versus operating temperature, separated into three regions: low, medium, and high volatility. As a subsequent characteristic of operational factors the scale of the operation was classified as small-scale, medium-scale, and large-scale.

With these three articles (Maidment 1998; Russell et al., 1998; Brooke 1998) the wider occupational safety and health community was thus introduced to the basics of the COSHH Essentials approach. While this strategy leans heavily on the work of historical models and approaches, it has a number of unique features, including an electronic version accessible via the internet. It meets all six of Money‘s (2003) core principles (understandability, availability, practicality, user-friendliness, confidence on the part of users, and transparent, consistent output). While welcoming the move by HSE to provide guidance in the form of CGS, Hudspith and Hay (1998) pointed out an additional obstacle to worker protection: communications barriers within companies. They recommended that HSE continue to stress the value of 22

workforce involvement in health and safety issues. Despite its attributes, however, the COSHH Essentials model is subject to a number of limitations relative to the development of the model, development of databases, use of the model, and its validation and verification.

Validation and Verification For purposes of this paper, validation focuses on the establishment of the soundness of a given model, whereas verification requires the evidence necessary to confirm its effectiveness. While it would be useful to validate a variety of the CB strategies proposed, only COSHH Essentials has been developed and implemented to the point that it has been the subject of almost all validation efforts. Also receiving attention is the International Labor Organization (ILO) Chemical Control Toolkit (ILO Toolkit), produced in collaboration with the HSE and the International Occupational Hygiene Association (IOHA). The ILO Toolkit is based on the HSE COSHH Essentials and is adapted for use worldwide (Jones and Nicas 2004). For validation purposes, three aspects of model evaluation were applied by Tischer et al. (2003) to COSHH Essentials. These aspects to validate the model include: internal (conceptual) validation of the model‘s assumptions and structure, external (performance) validation of the model predictions corresponding to professional IH monitoring data, and operational analysis of the understanding and implementation of the model‘s outcomes respective to its target group.

However, before presenting these model aspects there are still many questions to be answered in all three categories. Kromhout (2002a) took strong exception to the lack of exposure monitoring in ―generic risk assessment tools like COSHH essentials and expert systems like the Estimation and Assessment of Substances Exposure (EASE)…‖ as they ―…are known to be inaccurate and they do not take into account the various components of variability in exposure levels…‖. Kromhout built a strong case, estimating the variability in an eight-hour shift to be between 3 and 4000 fold, and delineating the sources of variability as spatial, between workers, and between groups. He argued that while providing exposure controls without having measured exposure concentrations would save money in the short term, in the long run it would be ―penny wise but pound foolish‖.

Topping (2002a) responded that these arguments ignored the range of competencies in the workplace, and the number of firms handling chemicals however, he concurred that the use of ―quality exposure data is extremely valuable for assessing the effectiveness of control measures‖. He did not directly address Kromhout‘s variability concern, but instead relied on the premise that COSHH Essentials is not intended to replace monitoring, but rather to provide needed help to SMEs. Topping pointed out that the cost of conducting the extensive monitoring suggested by Kromhout would be ―astronomical‖ and that the capacity to do so does not exist. He allowed that the COSHH Essentials were designed to ―err on the side of caution,‖ that the strategy had been peer reviewed by the British Occupational Hygiene Society (BOHS) experts, and that there had been no complaints about the recommended controls being too stringent without addressing the lack of research to show that the controls have even been put into place, let alone that they have been verified to achieve the intended exposure control. Kromhout (2002a) replied that he and the editor of the Annals of Occupational Hygiene questioned the role of tools like COSHH Essentials in the ―collapse of full time training of occupational hygiene professionals in Britain through lack of demand for expertise.‖ Kromhout‘s strongest criticism was that EASE and COSHH Essentials had not been properly evaluated prior to release, and that peer review by BOHS experts could not replace the rigorous evaluation of testing for reproducibility, validity, and peer review of results in the scientific literature. It was recommended that COSHH and EASE be used in the initial screening process. According to 23

Maidment (1998), the core model was validated by predicted dust and vapor exposure ranges, and their corresponding three-tiered hierarchy of engineering controls with measured data, and by extensive peer review of the logic and content by experts. He noted that it was extremely difficult to find quality data for comparisons, and further, that the information describing control strategies often seemed to indicate that several control strategies were in use. Limited comparisons were described in his manuscript; heavy reliance was placed on peer review during the model‘s development and validation and specifically involved the HSE Advisory Committee on Toxic Substances (including Guest, Brooke and Money) and experts of the BOHS (Maidment 1998; Topping 2002a). When taken as a whole, Topping did not address Kromhout‘s concerns of this unpublished peer review process. Therefore, not addressed are the potential weaknesses that one might find in the scientific literature when internal and external validation of the model is performed.

Brooke‘s (1998) work in comparing the R-phrases and resulting target airborne concentrations to the relevant health-based OELs on national lists (UK and German Maximum Allowable Concentrations (MAK) began to address the first category on internal validation for the COSHH Essentials. The work of Jones and Nicas (2004) reported below looked at both internal validation of the ILO Toolkit as compared to the UK HSE model and the external validation of the COSHH Essentials. The work of Tischer et al. (2003) and Maidment (1998) focused on the external validation and began to answer some of the questions relating to performance validation. A glaring weakness in the research at this time is present regarding the operational analysis of the given CB models.

Brooke was the first to identify the inherent difficulty in assigning dusts and vapors to equivalent bands designated elegantly by orders of magnitude (Table III). Resulting from this alignment of the bands, dusts have a higher margin of safety than vapors, especially for repeated exposure toxicity based R-phrases. Emphasizing the generic nature of this CB model and its provision for ―adequate control‖ Brooke concluded that the margins offer ―considerable reassurance‖ for vapors and ―even greater reassurance‖ when used for dusts. Much of the model‘s weakness in this regard was balanced against the intended non-expert SME end-user with no risk assessment background. With this in mind Brooke explicitly noted that the model used in practice would require ―continued evaluation of the allocation of the R-phrases to the hazard bands, such that the scheme may be revised and improved in the light of practical experience‖ (Brooke 1998). Brooke also reported that some categories of materials were arbitrarily assigned to a higher hazard category based on their toxicity characteristics, and this would provide an extra factor of 10. It must be pointed out that the Hazard Band values are generally in the same order of magnitude as OELs (see Table III) and also that it is not uncommon for acceptable risk levels of OELs, which are based on a 40-hour work week that accounts for worker recovery periods, to be in the range of 10-4 to 10-3. In contrast, acceptable risk values in environmental settings, which are based on continuous, involuntary exposure (168 hours per week) of all members of the population with no recovery period (Jayjock et al., 2000), are in the range of 10-6 to 10-5. Without understanding the basis of these underlying risk parameters, the problem then lies more in the lack of overall acceptance of higher-risk levels for occupational settings as compared to environmental settings. Solving this issue will require an improved communication of the reasons behind this risk differential and, therefore, a greater understanding of risk acceptability in occupational settings.

Jones and Nicas (2004; 2006a) reported less positive results in their evaluation of the ILO Toolkit. The ILO Toolkit, as discussed above, was based on the COSHH Essentials strategy, 24

but may not have been subject to the same periodic updates and revisions. They concluded that the calculation of safety margins (No Observed Adverse Effect Level (NOAEL), or the Lowest Observed Adverse Effect Level (LOAEL), divided by the high air concentration of the hazard band) resulted in values of <100 for Hazard Groups B and C, and <250 for Hazard D for vapors. They noted that these values should be in the range of 1000 to 10,000 for R48/20 (Danger of serious damage to health by prolonged (inhalation) exposure), depending on whether the NOAEL or LOAEL was utilized as the basis of calculation. That study made these calculations based on the generic COSHH criteria, to avoid any errors caused by incorrect assignments of hazard bands. A comparison of the R-phrases (taken from the HSE ―Approved Supply List‖ (National Chemical Emergency Centre at http://www.the-ncec/cselite)) assigned to commonly- used solvents indicated that the hazard group ratings assigned by the ILO Toolkit were lower than seen in the COSHH Essentials, for 12 of 16 solvents. In 5 cases, the ILO Toolkit included an S notation (skin hazard) which was not on the R-phrases. Jones and Nicas (2004; 2006a) suggested that the authors of the ILO Toolkit should reconsider the hazard classification plan as the variations among CB strategies reduce trust on the part of users. Based on the small safety margins between doses that cause significant effect in animals and the exposure bands in the toolkits being evaluated, they also suggested target exposure levels be made available to end users. Without offering these to the user to evaluate whether exposures are in line with the minimal margin, a false sense of health protection in the workplace is permitted (Jones and Nicas 2006a).

Tischer (2002; 2003) and colleagues at the German Federal Institution for Industrial Safety and Medicine (BAuA) conducted the first and most complete external validation of the COSHH Essentials to date, based on independent measurement data. The primary empirical basis for their analysis was measurement data collected within the preceding decade during several BAuA field studies. Some data were also provided by the chemical industry. Tischer‘s team also set out to address the external validation of the COSHH Essentials exposure model. While stating that the accuracy of the model was represented by agreement between predicted and observed, they believed that statistical tests are not useful due to the uncertainties in empirical data such as variability, errors in measurements, or false or incomplete information. Due to a lack of available data for some professions, only those with more complete data sets were used in this study. There were apparently 958 data points available for evaluation: 732 for liquids, and 226 for solids.

The BAuA data were all obtained from their own laboratories, and all workplace measurements were conducted as per the German Technical Rules. Sampling durations were usually 1-4 hours, and were task-based, i.e., they corresponded to a specific scenario. Over 95% were personal samples. Sources of uncertainty considered were volatility / dustiness, scale of use, and control strategy. For example, the uncertainty associated with volatility (of pure substances) was judged to be low, but quite complicated when mixtures were considered. Dustiness was considered to be a problem requiring additional attention. Scale of use was judged to be straightforward. (Most of the available data corresponded to the medium scale of use, with very little in the milliliter or tonne ranges.) Because of the limited quantity of data available, these researchers limited their analyses to scenarios in which the control strategy could be determined from the historical reports, generally matching one of the four control strategies. Comparisons of the predicted and actual data were conducted using frequency polygons overlaid with the range of predicted values and by calculating the percentage of the cases which were correctly or incorrectly predicted. Most of the data points fell within the predicted ranges. Per Balsat et al. (2003), Tischer (2001) found that the 95th percentile of data from different operations fit within 25

the ranges predicted by the COSHH Essentials model. Exceptions were scenarios where some of the limited data points for solvent exposures were above the predicted range, such as in carpentry workshops and with adhesives applications where the chemical product are spread over a large surface area reflecting small-scale, dispersive operations. Exceedances also occurred in the handling of powdery substances in kilogram quantities under local exhaust ventilation.

Jones and Nicas (2006) also performed external validation by evaluating the ability of the COSHH Essentials to select an appropriate control approach and whether these controls achieved reduction of exposure concentrations. They compared reported air monitoring data and related use of ventilation systems, taken from the National Institute for Occupational Safety and Health (NIOSH) Health Hazard Evaluations (HHEs) for 31 vapor degreasing operations with 7 different solvents and 20 bag filling operations with 17 particulates (#42). R-phrases for these liquids and dusts were obtained from the HSE National Chemical Emergency website (8 substances), the Australian ―Approved Criteria for Classifying Hazardous Substances (2002) and the Hazardous Substances Data Base (HSDB) of the United States (US) National Library of Medicine (6 substances), and the Internet (9 substances). Volatility information was obtained from the HSDB, and dustiness and scale-of-use were obtained from the NIOSH HHEs. Using this information, Jones and Nicas determined the appropriate control approach, and compared the actual measured exposures to the maximum value of the exposure band of the recommended exposure band. This comparison resulted in two types of control errors: situations in which insufficient exposure control occurred in the presence of local exhaust ventilation (LEV) (under- control errors), and situations in which sufficient exposure control occurred in the absence of LEV (over-control errors). They found under-control errors in 96% of the 163 cases where LEV was present in vapor degreasing operations, and in 55% of the 49 cases where LEV was present in bag filling operations (Jones and Nicas 2006b).

Their findings led Jones and Nicas (2006a; 2006b) to multiple conclusions. They found that the exposure bands do not provide consistent, or adequate, margins of safety and the high rate of under-control errors highlighted the need to evaluate the effectiveness of installed LEV systems using capture efficiency and/or air monitoring techniques. The limited assignment of ―dustiness‖ ratings to dusts complicates the model‘s process and indicates that specific guidance must be provided in cases where there is insufficient or inappropriate hazard information and that guidance on contacting professional assistance for engineering controls should be included on Task Guidance Sheets. Additionally, the R-phrase procedures, which include concentration ―cut-off‖ values (e.g., the hazard classification would not be for a mixture with

Ruden and Hansson (2003) investigated the accuracy of the EU classifications for acute oral toxicity for 992 substances by comparing their acute toxicity categorization (―very toxic‖, ―toxic‖, and ―harmful‖) to the acute oral toxicity data available in RTECS (Registry of Toxic Effects of Chemical Substances). Acute oral toxicity in rats is used because, although of minor importance for the complete toxicological profile, it offers a gauge of immediate toxicity with many substances lacking long-term data. They found that of the 992 substances that had enough data to undergo this evaluation, 15% (152) were assigned too low a danger class, and 8% (79) too high. Of those too low, or under-classified, 26 should be classified as ―very toxic,‖ 49

26

should be ―toxic,‖ and 77 should be ―harmful‖. According to Ruden and Hansson, the EU classifications rules indicate that once a substance is placed into a category based on specific toxicological data, it cannot be downgraded to a lower category based on additional information. It is when different studies of appropriate scientific quality would lead to different categorizations that the rules are less clear. In this instance, the authors indicate that there is an ―informal policy‖ in the EU to base its final classification on the most adverse outcome. A number of possibilities for this under-classification issue were noted, including variations in toxicity data from different laboratories; however, more issues arise relating to the EU informal policy. If the EU Commission has access to data not in RTECS, and these data support higher classifications, then the policy should default a substance categorization to a higher hazard class and not a lower one. Other possibilities exist such as the frequency of updating classifications, insufficient toxicity data searches, or problems with the RTECS database. Regardless, it is difficult to accurately pinpoint a causal relationship as there is a ―lack of transparency‖ in the Commission‘s classifications. For future classifications of substances the authors recommended the scientific basis be published to afford this transparency so when similar issues arise they can be addressed and rectified (Ruden and Hansson 2003).

Although not always recognized as a validation parameter, the COSHH Essentials‘ CB model had ease-of-use and simplicity as intended design parameters for the non-expert end user (Maidment 1998; Brooke 1998; Oldershaw 2001). Therefore, the results of an HSE survey to determine the utility of the COSHH Essentials should also be considered. A telephone survey was performed with 500 chemical purchasers who have used the older, paper version of COSHH Essentials (Evans 2005a). The survey indicated that it had been utilized by 80%, with only 5% finding it difficult to use and 95% willing to recommend it to other companies. In addition, 75% of those surveyed had taken action to control chemical exposures. Actions taken when utilizing the COSHH Essentials model included: chemical substitution (18%), changing work procedures (25%), changing the control measure used (36%), providing information or training to workers (48%), and checking existing control measures to ensure they are working (67%).

Variations of the chemical model Users of CB strategies quickly realized that one strategy would not fit all needs. Variations of the model and its use in practice have been developed by several nations including France, Germany, Belgium, The Netherlands, and Singapore and also by corporations. Interest in CB strategies on the part of the European occupational hygiene community was spurred by the introduction of the Chemical Agents Directive in 1998 (1998; 2003). Several approaches have resulted. The French approach evaluates the probable effectiveness of risk management in protecting workers at the company level (UIC 1999). It suggests appropriate references to provide guidance based on the type of substance and handling procedures. In June 2007, a new European law on chemicals, REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals), entered into force; at the same time, the European Chemicals Agency (ECHA) began operations. This law shifts greater responsibility to industry to manage the risks from chemicals and to provide safety information on the substances (ECA 2007). The European Chemical Industry Council (CEFIC) exposure management system (CEMAS) (Money 2001; Money 2003) intends to provide a guidance tool for SME, to collect workplace exposure data which can be coupled with hazard information and deliver advice on risks and risk management, recommending whether exposure monitoring should be conducted.

27

The European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC (2002)) approach is a tiered and targeted risk assessment that could aid in the registration of large number of chemicals under REACH. This is a streamlined approach which applies CB concepts in a tiered manner with Tier 0 screening out chemicals not presenting an immediate risk to humans or the environment; Tier 1 identifying uses of a chemical which may present further risks, to be investigated in greater depth in Tier 2. In Tier 1, margins of exposure (MoE) are compared with generic OELs for the chemical‘s hazard category, while Tier 2 assessments are conducted in accordance with EU risk assessment principles. Toward that end, a database known as Solbase shows potential as a source from which CGSs could be developed. With partners from throughout Europe, Swuste et al. (2003), have tested Solbase, both for usability of the software, and suitability of the recommendations yielded by Solbase, using 535 new and existing solutions. Although most of these solutions currently relate to manual or material handling, noise and vibration, machine guarding, and other safety issues, few address air contaminants. The databank can be queried either by production process, or by hazard.

Much of the literature for the evaluation and validation of CB models has been related to a concerted effort to create and drive a research agenda through workshops. This approach has been proven useful for developing earlier solutions-based programs beyond their national origin. Early solution-based initiatives include the noise control solutions from the UK HSE (1993), exposure reduction in mining from Australia (Mitchel and Else 1993), and chemical substitution strategies from Denmark, the UK, the US, and The Netherlands to reduce health hazards (Swuste and Hale 1994). A model for communication and evaluation of these programs began at the first IOHA Scientific Conference in 1992 with a workshop on sharing knowledge of preventive measures. This culminated in a 1994 WHO experts meeting to stimulate the interchange of solutions toward the reduction of occupational risk and the formation of the Prevention and Control Exchange (PACE) working group (Swuste et al. 1995). This process has evolved into efforts such as the European Solbase, with many nations teaming together to develop a database of effective controls for workplace hazards and reduction of occupational risks (Swuste et al., 2003). International CB workshops have been held in London (2002), Cincinnati (2004), and South Africa (2005). The workshops have led to an international agreement for coordinating the work of international agencies and their partners and a global implementation strategy for CB models. An example of this collaboration is the appropriate international forum that the workshops have provided for the Globally Harmonized System of Classification and Labeling of Chemicals (GHS). The GHS is a uniform, internationally developed, and standardized protocol for the toxicological basis for assigning chemicals to standardized hazard statements on labels and safety data sheets in manner that builds on the EU R-phrase process. From the beginning of the UK model‘s development it was made clear that when more data became available, chemical substances would need to be reclassified (Guest 1998). Should there be future reclassification efforts, it has been recommended that the scientific basis and decision matrix for these hazard classifications be standardized and readily available to achieve transparency for subsequent evaluations (Ruden and Hansonn 2003). GHS is presented as a proper approach to build transparency to the process by including a core set of label elements to work towards harmonized hazard statements for each category and class of chemicals covered. It also has a harmonized approach to classifying mixtures of these chemicals. The GHS has also adopted the concept of the ILO Toolkit as part of its overall process to include exposure control approaches in parallel with its efforts for a chemical standardization process. While this may take some time to accomplish, it will eventually provide

28

consistency of information on over 1500 commonly used chemicals and include a centralized procedure for frequent updates of information (Balsat et al., 2002).

Through these CB workshops, a process emphasizing the utility of available CB models has also led many countries to adapt and use them within their existing occupational risk management approaches. A two-stage risk assessment strategy (Regetox) was developed and tested in Belgium (Balsat et al., 2002; Balsat et al., 2003; Basat et al., 2004) in response to the European Chemical Agents Directive 98/24/EC (EC 1998), which requires companies to assess and manage chemical risks in the workplace. In order to minimize the number of chemicals (and resulting costs) for which risk assessment must be conducted, the first stage of the strategy utilizes the French ―ranking of potential risk‖ based on R-phrase, annual quantity in use, and frequency of use, as described above (Vincent and Bonthoux 2000). Only products receiving a rating of medium or high are carried forward to the second stage, which utilizes the COSHH Essentials. When mixtures are being handled, the risks are evaluated for each harmful component according to the composition by weight of the mixture12. For cases in which contaminants are generated during the process, e.g. aerosols generated during spray painting, the EASE model is used. Feasibility studies conducted in two firms revealed lacking or inadequate MSDS. There was only one case in the two companies in which the strategy failed to reveal need for improvement in the work situation. The authors felt that simple examination of the work situation would have indicated the need for semi-quantitative risk assessment. Further lessons drawn from the trial are that most companies are not prepared to comply with the European Chemical Agents Directive, and that the use of the Regetox approach can be helpful to companies, but requires training of ―prevention advisors‖ and a strategy to involve employers, staff members, and workers to assist in collecting basic information for the risk assessment (Vincent and Bonthoux 2000).

The Semi-Quantitative Risk Assessment (SQRA) developed in Singapore is intended to facilitate identification of chemical hazards, evaluation and potential for exposure, risk level determination, and prioritization of appropriate controls to address the identified risks. With the SQRA there are three methods for performing exposure evaluation that include personal exposure monitoring, using exposure factors and parameters, and/or applying empirical and theoretical formulas to estimate exposures at the plant or process design stage. The ILO Toolkit, which was renamed as the International Chemical Control Toolkit (ICCT) during the SQRA‘s development, was tested in parallel with applications of the SQRA to evaluate their utility and to perform comparisons based on theoretical and empirical aspects (Yap 2004). Direct comparison of the two approaches was stratified by their respective control approaches as compared to the SQRA risk level approach. The empirical comparison of the models uses actual personal air monitoring data used to derive the SQRA method‘s risk level to assess against the Toolkit‘s control approach. This comparison was performed on 27 selected SME processes including metal-working, paint manufacturing, chemical processing, printing, dry cleaning, and electronics industries. The results of the theoretical comparison indicate that the Toolkit and the SQRA method are somewhat consistent with any differences between the control approach and risk level being at most one to two bands. In the majority of cases using the empirical comparison it was determined that the Toolkit over-evaluates the risk relative to SQRA, leading to more conservative approaches relating to controls (Yap 2004).

Germany is the third largest chemical producer in the world, and the largest chemical exporter in the world (Adelmann 2001). As such, it has taken its responsibility to assist in sound management of chemicals in developing countries (Tischer and Scholaen 2003). Under its 29

Convention Project on Chemical Safety, the technical arm of the German Development Agency‘s Society for Technical Cooperation (GTZ) has developed a Chemical Management Program Guide as part of its Pilot Project on Chemical Safety. The Chemical Management Guide is a method to demonstrate and document how chemical safety in emerging countries and small businesses can be improved and sustainability implemented in line with international standards. In more than 130 partner countries, GTZ is supporting 2,700 development projects and programs with the aim to improve the living conditions and perspectives of people in developing and transition countries. It has been implemented at international sites in Argentina, Indonesia, and EU countries. The GTZ chemical management guide and pilot project on chemical safety is a unique program developed specifically to meet the needs of small businesses and developing countries for addressing chemical hazards. A participatory training process is utilized to work to the selection of control technologies. The GTZ program acknowledges that CB models may be too sophisticated for many small enterprises in developing countries; medium- and larger enterprises often have more MSDS on site and therefore they have a greater potential for conducting risk assessments using the ILO Toolkit (Tischer and Scholaen 2003).

Building upon the COSHH Essentials approach, countries have also begun to develop their own CB models to address national regulatory requirements and professional approaches. Stoffenmanager (www.Stoffenmanager.nl - accessed 14 October 2007) is a web-based software tool built for SMEs to assist in working safely with chemical substances. Their CB model factors in an exposure potential through the use of an interactive chemical risk management approach. This model approach was developed in The Netherlands to assist SMEs in assessing, prioritizing, and controlling risks associated with hazardous substances. The tool is based on the COSHH Essentials and a modified version of the Cherrie and Schneider (1999) inhalation exposure model (Tielemans 2007). Stoffenmanager is currently a generic tool that supports the inventory of the hazardous substances, assessing and controlling risks in a risk inventory, obtaining a plan for control measures, making instruction sheets for the workplace, and helping in storage according to guidelines. For the risk inventory, the employer uses R-phrases categorized according to COSHH Essentials. Then the employer completes an exposure assessment, involving response to 7 questions to determine the chemical‘s exposure class. The tool automatically calculates a risk score, a relative risk ranking. Thus an initial assessment of the health risk is completed. Using the tool‘s risk score, the employer can then calculate the efficacy of various control measures and choose the most effective ones (Tijssen et al., 2004). The Stoffenmanager model has been recently evaluated utilizing targeted field surveys for many dust (i.e. animal feed, flour processing, textile, and construction) and liquid (i.e. solvents for metal, car body repair, and printing) industry exposures in comparison with existing exposure data (Tielemans 2007). This comprehensive validation study has initially found relatively good initial correlation of the non-expert Stoffenmanager score with expert evaluation overall for inhalable dusts (rs = 0.83) and liquids (rs = 0.81). This validation process for the Stoffenmanager model remains an ongoing process and is intended to remain a dynamic process with continual updating (Tielemans 2007).

Developed through the cooperation of corporations within the Norwegian oil industry, KjemiRisk is an assessment of chemical health risk based on experience and practice in these industries (Smedbold 2004). The tool takes the following into account: physical properties of the chemical, the handling of the chemical, and the appropriateness of the technical, organizational and personal barriers established to control the chemical exposure, and the duration and frequency of the work task using R- and safety phrases (S- phrases) as its basis. 30

Similar to R-phrases, S-phrases are also required by the EU to appear on each label and safety data sheet for hazardous chemicals as part of the classification, packaging, and labeling of dangerous substances provision (Council Directive 67/548/EEC). Chemicals are grouped into one of five health hazard categories based on R- and S-phrases. As part of the KjemiRisk application, 15 common tasks are defined and the handling of the chemical, its physical state, duration and frequency of use, potential for exposure, and the appropriateness of controls in place are used in the conceptual model. The risk assessment is divided into two phases which include the potential risk and the final risk. These are adjusted for risk based on a judgment of the reliability and appropriateness of the established barriers and/or controls. The risk assessment provides a full risk evaluation of task-based work procedures based on an evaluation of risk for illness related to lungs, internal organs, and skin. KjemiRisk can be considered both a rough risk assessment tool when used by line managers or health and safety generalists and an expert tool when used by industrial hygienists. It is currently available in Norwegian and English as an individual or a network application when integrated with an appropriate server. Expansion of web applications, improvement of reporting functionalities, and substitution of capabilities are currently being considered for development (Smedbold 2004).

Developing and implementing CB risk assessment / management programs is critically important to many industries which process and market hazardous chemical substances. CB is an invaluable universal tool for assessing and managing these chemical risks. There is an important difference between industries which employ commodity industrial chemicals (e.g., bulk petro-chemical, health care, etc.) and those with unique proprietary chemicals (e.g., pharmaceutical, many industrial / commercial products, etc.). Many commodity industrial chemicals are well characterized, therefore, they can be assessed and controlled using existing CB models (COSHH Essentials, etc.). (Note that there is no need to use CB models to manage chemicals for which OELs exist.) However, industries which process and market unique proprietary chemicals must customize their risk assessment / management CB models for their operations using three essential steps: (1) performing appropriate hazard assessments to classify and communicate hazards; (2) assessing worker exposures in the workplace during specific operations; and (3) communicating, implementing, and verifying the proper control measures. Exposure Control Practices (ECPs) – specific guidance on the control measures – are a valuable tool for managing chemical risks. The ECPs provide a "feedback loop" to ensure that workers are protected and exposures are controlled to the desired levels. ECPs should be based on the Hierarchy of Control principles. Also, they must be verified as part of the exposure assessment program. However, they enable much more robust risk assessment / management than do traditional IH approaches.

Further Model Evolution Both the UK and ILO CB models focus on the use of bulk chemicals. In addition to chemical agents, which are covered by other UK regulations (i.e. asbestos, lead, and pesticides), they also are not intended to address process-generated emissions. These are chemical agent exposures created by the task, or not purchased in bulk, and include construction-related hazards such as silica dust, welding fume, and wood dust exposures as examples. Silica exposures in mining or construction have an excellent track record for existing interventions and practical solutions- based outcomes. These include standardized recommendations and subsequent reduction of exposures relating to the implementation of specified control solutions (Goldsmith 1997; Flynn and Susi 2003; Rappaport et al., 2003). The UK HSE has already begun to adapt the CB model for broader chemical agents and expansion of the COSHH Essentials approach toward direct control advice. Exposures generated by these processes do not have Risk Phrases and require a 31

different practical approach. The UK HSE has developed a CB process for some of these exposures by directing the user to job-specific control advice sheets relating to initial selected professions such as dry cleaning, hairdressing, and paint spraying (Tischer 2002; Evans 2005b). Taking this a step further, the Silica Essentials is also directing users to control advice sheets that are industry and task-based and do not require the additional step of inputting data (Evans 2005c). Instead the user selects the control advice directly by activity, avoiding the interim exposure prediction step of the COSHH Essentials model. The Silica Essentials is another CB model that is currently being evaluated in implementation and validation efforts internationally, including in Africa and Latin America.

The stratification of risk that began in the 1970s is now being considered for application in a variety of occupational health, hygiene, and safety professions as well in major industries. The international CB workshops have been an essential element in establishing uniform research agendas for evaluating CB strategies. They have also served to initiate the expansion of chemical-oriented models to best address practical prevention of a broader spectrum of work- related illness, disease, and injury. Topics discussed at these workshops that are beginning to be addressed include the provision of national-level guidance and coordination, pilot projects at the state level, and creation of an Occupational Risk Management (ORM) Toolbox. The ORM Toolbox approach is intended to broaden the CB model to include a more comprehensive exposure control basis for globally common industries such as construction and agriculture that require a multidisciplinary approach for chemical and ergonomic, safety, and environmental concerns. Current efforts have begun for the development of a CB model for a Construction Toolbox, addressing these composite, potential exposures by trade and task (Spee 2005). To achieve the ORM Toolbox approach a broader, multidisciplinary approach for trade-related exposures is needed.

Applying the CB model in a multidisciplinary fashion requires some brief consideration of differences between the fundamental approach to IH, ergonomics, and occupational safety. Concepts on exposure and variability of exposure are well developed in the IH profession. These concepts are hardly present in occupational safety. Ergonomics and occupational safety both have a strong focus on design and redesign, which is much less developed in IH. Therefore, as CB models are being developed to address musculoskeletal disorders and occupational injuries, they may find professionals in these specialties well conditioned to this simplified adaptation. While CB strategies like the Silica Essentials are being developed to address locally generated exposures, as in the construction industry, the exposure factors relating to ergonomics are also being evaluated. Another IH to ergonomic comparison is that chemical production involves the development of new products which may never be fully researched and can logarithmically expand the variety of exposure routes and sources for a given worker. In contrast, ergonomics has a finite group of well-researched and defined risk factors and effective programs (Stewart et al., 2005; Zalk 2001). The ILO Ergonomics Checkpoints document is an example of well-researched and internationally validated models that is being developed as the basis of Ergonomics Toolkits (Zalk 2003; Zalk 2006a). Efforts in The Netherlands have begun to consider the incorporation of occupational safety requirements with a focus on traumatic injury (Swuste 2005). Occupational safety is not restricted to chemical safety, but a more general approach is considered, focussing on causes of both major and minor occupational accidents. It has been presented that classifications already exist for various variables of accident causation, which can be viewed as an analogous ‗banding principle‘ where safety phrases can be applied in a manner similar to risk phrases. In IH practice control of exposure takes place after the central event occurred, the emission of the hazardous substance. In 32

occupational safety, barriers are active both before and after the central event. Therefore these barriers, including management factors, have a strong relation with the quality of safety management systems, and these factors are important parameters for risk prevention73. The endpoint of this CB model would not necessarily lead to control advice as much as an identification and implementation of barriers. This barrier banding model would apply these phrases to provide information on the type of hazard of accident scenarios or related situations and will guide the type of precautions needed deal with these scenarios or situations (Zalk 2006b).

Moving back to the roots of the modern CB movement, nanotechnology industries are also finding a limitation in toxicological data in a manner similar to their biological and pharmaceutical counterparts. They also have to achieve a risk management program with an insufficient basis for traditional IH quantitative risk assessment approaches. An important distinction is that they have a longer track record of CB models to work with in developing a control approach. To develop the concept, Maynard (2007) has combined the proven effectiveness of CB in controlling exposures in an intensive research and development industry, such as in pharmaceuticals, with the utility of COSHH Essentials model. A conceptual CB model is presented which offers the same four control approaches of the UK model as stratified by corresponding ‗impact‘ and exposure indices. This model proposes combining engineered nanomaterial composition parameters such as shape, size, surface area, and surface activity with their exposure availability in terms of dustiness and amount in use and linking these indices to bands with corresponding control approaches. This nanomaterial CB model, although not developed in practice, is presented similarly to COSHH Essentials in that it is a useful concept that affords a pragmatic approach to exposure control and is considered to be an alternative rather than a substitute to traditional IH risk assessment and control (Maynard 2007).

DISCUSSION

Underpinning the toxicological basis of the UK approach is the importance of an accurate toxicological rating and hazard band classification by suppliers of chemical substances. Given this critical need for CB models, there is a need to reevaluate the assignment of R-phrases to chemical substances (Brooke 1998; Ruden and Hanson 2003). This process should go beyond work with the COSHH Essentials model and become a central focus for the different CB models available and in development. Significant concerns have been raised about the accuracy of EU classifications of chemical substances (Ruden and Hanson 2003). If COSHH Essentials has been designed to be slightly over-protective (Russell et al., 1998), then the 15% of evaluated EU classifications that were assigned too low a danger class should be considered a substantial issue to be addressed (Ruden and Hanson 2003). Other confounding issues for the model also require further evaluation. The margins of safety are possibly inadequate for many vapors and some may need to be classified into higher hazard bands (Brooke 1998; Jones and Nicas 2004; Jones and Nicas 2006b). There is also a variation in hazard band assignment between the COSHH Essentials model and ILO Toolkit (Jones and Nicas 2004; Jones and Nicas 2006a). In addition to model validation efforts, experts who have written on the CB topic confirm its potential value as a risk assessment and risk management tool in the workplace. They also express caution about the need for systematic, critical evaluation of the approach before widespread adoption.

According to Money (2003), ―no systematic evaluation of the actual impact and effectiveness of the schemes has been undertaken … no systematic assessment has been undertaken of the impact that CB approaches have had on the management of risk at the workplace or other levels. 33

Thus, in terms of future developments in the area, it would appear that before further refinements are considered, there needs to be an extensive and systematic evaluation of the uptake and impact of a number of the key approaches.‖ Swuste et al. (2003) referenced Kromhout (2002b), stating that ―The COSHH Essentials has met some criticism in the literature, focusing on the lack of a proper evaluation before its introduction into the occupational arena, as well as the generic nature of the tool, which will lack precision and accuracy in situations where these are required.‖ Tischer (2003) and his colleagues have said that in the German occupational hygiene community, ―…there was consensus that the scheme [COSHH Essentials] had great potential for further development. On the other hand, with respect to the exposure predictive model it has been argued that, due to its generic character, reliability and accuracy (safety) may have been sacrificed for the sake of simplicity and transparency. However, this assumption is not based on real measurement data and instead reflects the low degree of confidence generally enjoyed by generic models.‖

The impetus for the modern movement of CB was the regulatory driven need to address chemical exposures for the majority of the UK workforce. COSHH Essentials, the UK model, was created by experts who, with much thought, chose a simplified model to achieve maximum utility in addressing this need. The work of Cherrie and Schneider (1999) served to strengthen this decision by showing that a structured approach based on descriptive workplace activities provided significant correlation with exposure measurements. The dissection and examination of this CB model remains an ongoing endeavor. However, its effectiveness in achieving its intended utility is often overlooked as a prime component. Results from HSE survey on the use and application of the COSHH Essentials model infer utility in the UK (Evans 2005a). This CB model, however, is also being considered for use in many other countries around the world, including the US. For this reason the COSHH Essentials model has received additional attention and should receive more to ensure an ongoing critical evaluation to determine whether the model delivers target exposure ranges, offers controls commensurate to exposure potential, is used appropriately, and has improved control of exposure.

In consideration for its implementation, Oldershaw (2003) cautioned that the COSHH Essentials approach cannot be adopted uncritically by other countries; further, the approach must be seen in the context of personal protection, training, health surveillance as appropriate, etc. A key point is that the approach is not meant to replace exposure measurement, interpretation, substance, and chemical control. These studies and expert comments presented in the literature are not ―project stoppers,‖ but rather emphasize the need for collection of data under controlled scenarios to validate the predictions of the model. Under-prescription of control could lead to serious injury, while over-prescription could lead to significant unnecessary expense, especially for SMEs. Of the two types of error, i.e., under-control (recommendation of inadequate level of control) is potentially more serious than over-control. In this model‘s development the general approach was to be conservative or slightly over-protective (Russell 1998).

Internal evaluations of the UK model have shown under-control error for small-scale, dispersed use of solvents and some powder handling operations (ABPI 1995; Tischer et al., 2003) as well as for vapor degreasing and bag filling operations (Jones and Nicas 2006b). These results seem to confirm Kromhout‘s argument (Kromhout 2002a) that potential misclassification of exposure bands can consequently affect assignment to control bands. Brooke‘s work predicted this potential; however, this concern was essentially addressed with expectations that the model‘s scheme and the allocation of hazard bands with R-phrases would be consistently evaluated and improved (Brooke 1998), however the research has not shown this to date. However, with 34

external evaluation the COSHH Essentials model has also been found to deliver a significant level of confidence in the target exposure ranges (Tischer et al., 2003). German BAuA comparisons of the model‘s outcomes compared to personal exposure monitoring data, in a number of different industries, were well within range for work with solids and medium scale liquids (Tischer 2002; Tischer et al., 2003) although some under-control error with liquids was found in their work as well as Brooke‘s (1998). The ILO Chemical Toolkit, based on COSHH Essentials, has also been shown to indicate more conservative control solutions based on comparisons with the Singapore‘s SQRA method utilizing personal exposure monitoring data for deriving risk level approaches (Yap 2005). Comparisons indicate that, for the majority of the 27 processes selected, the Toolkit equally- or over-evaluated the risk relative to the SQRA (Yap 2005). Within these validation efforts there has been an acknowledged paucity of data with which to validate CB models (Maidment 1998; Swuste et al., 2003; Money 2003; Tischer et al., 2003; Kromhout 2002a; Jones and Nicas 2006b). There is also a limited range of exposure situations with which to compare predictions (Tischer et al., 2003). There has also been difficulty in ascertaining reported control classification (Maidment 1998) proper characterization of specific work parameters, and materials in use (Jones and Nicas 2006b) for comparison of predicted and actual exposures. The need for health surveillance data / environmental monitoring must be evaluated (Russell et al., 1998); particularly when toxicological data are limited (Guest 1998). The ongoing need for personal monitoring (air and wipe tests) must be strongly emphasized. The use of the CB models is to complement, not replace the traditional IH approach to risk and exposure assessment. Therefore, personal monitoring is needed to bolster a system that evaluates the effectiveness of controls initially and over time. It will continue to be an essential requirement that ongoing monitoring is needed to detect breaches in containment systems and effectiveness of LEV, even if previously verified5.

The work of Jones and Nicas (2006a; 2006b) has received much attention as its critique of the COSHH Essentials and ILO Toolkit have indicated a high prevalence of control errors and the potential for an inappropriate confidence in the workplace chemical exposure reduction. HSE members responded to their COSHH Essentials evaluation (Jones and Nicas 2006b) clarifying that their CB model is not intended to predict exposure, but rather to identify adequate control approaches (Evans and Garrod 2006). This is a difficult statement to justify in that the exposure prediction step is what separates the COSHH Essentials model from the earlier toxicology-to- control pharmaceutical CB model. It was also indicated that the article (Jones and Nicas 2006b) did not actually evaluate the COSHH Essentials as, of the workplace exposures utilized, none of the controls in place were recommended by their CB model. Non-HSE members also responded (Money et al., 2006) to these Jones and Nicas articles, noting that the intent of COSHH Essentials is its utility in obtaining and implementing appropriate risk control advice and that user evaluation trials have indicated a higher likelihood of achieving this than if presented in a less accessible or understandable format. Jones and Nicas replied to this commentary (Jones and Nicas 2006c) indicating that without a recommended prospective study of COSHH Essentials, evaluation of its components is necessary. While confirming their approach and remaining skepticism of the model‘s outcomes, they do address their study‘s limitations in that the variability of engineering control efficiency may also be seen in the high rate of under-control findings. Their margin of safety applications in their assessment of the ILO Toolkit (Jones and Nicas 2006a) also requires evaluation. Their reliance on safety margins may not be appropriate for validation studies in that their conclusions are heavily dependent on the critical effect‘s relative toxicity. Higher consequence toxicological outcomes such as cancer require a much larger safety margin than for lower outcomes such as irritation, and may therefore affect the probability of an under-control finding with more adverse toxicological outcomes. 35

An important distinction in the development of the UK model is that the current objective of the COSHH Essentials is to achieve exposure levels anywhere in the exposure band, whereas the CIA recommends that exposures should be maintained ―as low as reasonably practicable‖ (CIA 1993; Guest 1998). This disconnect with the trade association should be further investigated. Although comparisons to solid chemical exposures have been promising, model validation efforts have shown that it is difficult for researchers to retrospectively evaluate the dustiness of particulates and it may therefore be difficult for SME managers to do the same41. For liquid chemical exposures, under-control error in small-scale solvent applications, although consistent with Brooke‘s (1998) reservations on vapor‘s equivalency with dusts as in Table III, can in part be attributed to industrial tasks that spread relatively minute quantities over a large surface area, increasing exposure potential. An adjustment or acknowledgement within the control guidance sheets can be made for tasks with these processes, however the model‘s weakness with vapors must be further evaluated. In addition, the Regetox approach (Balsat et al., 2003) presented composition by weight for both liquid and solid mixtures in evaluating risks for each harmful component in a workshop that prepares plasticizing mix. Composition by weight is appropriate for solids, but this may skew the estimation of potential risk for liquids as composition should be by molar fraction due to the difference in volatility of various liquid components.

Promising information is just beginning to be put forth in the evaluation of The Netherlands‘ Stoffenmanager CB model. Their approach has benefited from the ongoing critique of the COSHH Essentials and ILO Chemical Control Toolkit which has assisted in their decision to utilize exposure assessment prioritization in its banding strategy which has been derived from the international validation process which include the international CB workshops (Tielmans et al., 2007; Tijssen et al., 2004). Stoffenmanager serves as an excellent example of how dissecting existing models can lead to criteria to be used in developing other exposure control models. Its initial validation study remains an ongoing process, but preliminary information shows that the current generic version of Stoffenmanager indicates its utility as an exposure assessment tool for SME managers and may be an appropriate CB model for use in Tier 1 scenarios relating to REACH. Future efforts include an English version of the generic model in late 2007, creating opportunities for wider international use and further validation of the model and verification of the effectiveness of its control outcomes. Also in progress is an expansion of this CB model into branch specific versions that is expected to become a standard in The Netherlands, and the development of a dynamic web-based data exchange module called STEAMBASE (SToffenmanager Exposure And Modeling dataBASE) (Tielemans et al., 2007) which may be an important foundation for the prospective studies that are a consensus in CB literature.

Regulatory requirements in the UK were a driver to develop the COSHH Essentials CB model for non-experts to address exposure to chemicals. The model was simplified by design due to the many SME managers under this regulation who do not have easy or affordable access to professional judgment. The pharmaceutical and biological agent exposure control models, the evolutionary predecessors of the modern CB movement, were and are intended for use in Large Enterprises (LE). Due to their size, these industries typically have adequate access to professional expertise and funding for engineering controls and their maintenance. Models relating to pharmaceutical agents, as an example, can therefore be more intricate and achieve greater accuracy as they are implemented and maintained by trained professionals. The lack of toxicological data and availability of established OELs are the common bases for the creation of both the pharmaceutical agent and chemical exposure control models.

36

The developers of the COSHH Essentials and the related ILO Chemical Control Toolkit both deliberately chose a less complex model in order to achieve simplicity. The ease-of-use of the UK model has been for the most part achieved for the intent of its development in the UK -- use and application by SME managers. However, a key distinction between the models is that the ILO developed its international version for use by non-experts everywhere in the world. This expertise may not available due to limited funds, such as in the EU or the US, or due to the relative absence of the IH profession in most industrially developing nations worldwide, affecting LEs as well as SMEs (Eijkemans and Fedetov 2005). It was understood in the development of the modern CB models that a practical exposure control tool for non-experts may in practice compromise a level of accuracy when compared to the advice of experts. As important as this is to achieve utility for the intended audience -- whether for SMEs, developing countries, or for experts and non-experts alike in the absence of OELs -- validation of these models has indeed pointed out areas where this accuracy has been compromised. The focal point then becomes one of perceived risk and the variable levels of acceptability of risk, a perception that varies from country to country, from culture to culture.

The historical basis for the modern CB models was that they were to be used by experts within a research and development environment. The need for this approach was primarily related to the absence of OELs, such as in the biological, pharmaceutical, and now the nanotechnology industries. Validation of these models is complicated in that traditional exposure assessment may not be possible at this time without a proven toxicological basis, as is especially apparent with nanoparticulate (Maynard 2007). What all these CB models have in common is achieving a level of approachability to what otherwise may remain only in the hands of those with access to expert judgment. They also share a certain acceptance of risk and inaccuracy. Adaptation of the existing models beyond bulk chemical use has been assisted by this cumulative CB discussion in that developers can learn from still ongoing evaluations and benefit from a growing acceptability of simplicity in achieving exposure reduction. The practical nature of the silica, ergonomics, and injury prevention CB model approaches indicates that they are likely to succeed; however, not without the same rigor of validation and evaluation that should be given to all CB models. In developing multidisciplinary CB strategies it has become apparent that involvement of stakeholders is helpful in defining minimum performance standards, whether required by regulation or by circumstance.

CONCLUSION

Further research remains a requirement for all CB models. This includes further internal validation of CB model components, broader external validation of the model predictions when compared to expert interventions, and especially the need for operational analysis of the model as implemented to achieve intended outcomes. A prospective research process therefore remains essential to achieve an understanding of the implications of the model as applied and how this correlates to its overall effectiveness for its target group. This will assist in addressing the remaining questions as to how control recommendations are being implemented and maintained and whether they are achieving the intended exposure reduction. The lack of this information has led many to question the overall effectiveness of CB models in that they have knowingly chosen simplicity at the expense of accuracy and, therefore, protection of the worker. This research needs to be performed and the results folded into an improvement process for CB models, which must include continual reevaluation of R-phrases and GHS Hazard Statements, in order to scientifically address these questions. In addition, further field studies are also vital to this research as they are necessary for providing essential validation and verification data 37

which in turn will improve our practical understanding of the strengths and weaknesses of each of the models. In the absence of this information, the CB models as currently available are best used when OELs do not exist or as initial risk assessment screening tools that at some level include expert input and traditional IH monitoring.

It seems that lost in these scientific validation discussions are the billions of workers who do not have access to expert advice. When further research is performed it must not stop short at the dissection of models. It must use the lessons learned from the process to build a better model that does have a place in the hands of non-experts. CB models are therefore, in essence, an opportunity to simplify the best of scientific information into a format that is accessible to the multitudes. Expert IH advice in practice is expensive and is non-existent in many countries, rendering it inaccessible to so many. This fact should not be used as an excuse to apply unvalidated control models blindly, but rather to serve as an impetus to expand the reach of this expertise and to develop it where it does not exist. With this in mind, the modern CB movement should continue to seek the finest technical expertise to make the models as good as possible. Seeking perfection will only ensure that the prevention of work-related disorders will not be achieved for the majority of the world‘s workforce.

ACKNOWLEDGEMENTS

We would like to thank all those who have assisted and the organizations that have supported all aspects of the International Control Banding Workshops, whose efforts have assisted in developing and driving the CB research agenda. Special thanks to Paul Swuste for his keen insight and direction in the development of this manuscript. Thanks also to Henri Heussen for his timely update on the Stoffenmanager validation process. Note; *Work performed in part under the auspices of the US Department of Energy by the Lawrence Livermore National Laboratory under contract W-7405-Eng-48. UCRL- JRNL-223247.

38

CHAPTER 3 GRASSROOTS ERGONOMICS: AN EFFORT TO MODIFY CUSTODIAL TRAINING Professional Safety, 3:21 – 25 (1997) David M. Zalk, John C. Tolley, and Yong Kim Hazards Control Department Team 4 LLNL, Livermore, Ca

Abstract

By working with custodians to identify high risk procedures, ergonomics is utilized to enhance health and safety behavior. The purpose of this ergonomic analysis is to modify the custodian education and training program to reduce the number of ergonomic related incidents utilizing posture based video analyses. The video analyses of Lawrence Livermore National Laboratory custodians have identified some awkward postures and the need to modify custodial training by teaching back care, the balancing of right and left hand use, neutral body postures, and the awareness of high risk custodial procedures.

Introduction

The Lawrence Livermore National Laboratory (LLNL) Plant Engineering Department (PE) has approximately 950 employees. The Maintenance/Operations (M/O) department of PE has approximately 650 employees and contains all of the traditional crafts such as electricians, carpenters, and pipefitters as well as the Gardener and Custodian groups. The Custodian Department has approximately 150 employees that are equally divided between males and females. Each of the custodians is responsible for cleaning some 30,000 square feet each day for 40 hours per week. In total, they are responsible for the cleaning needs of some 5.6 million square feet at LLNL. In addition to normal custodian functions such as cleaning restrooms, vacuuming, dumping trash, carpet cleaning, and floor care, LLNL custodians respond to emergency situations such as rainwater leaks into office areas.

Safety and CQI Plant Engineering has been very involved in Continuous Quality Improvement (CQI) in every aspect of its operations, including safety. In 1992, M/O recognized the need for a different approach to safety than the traditional ones. With the guidance of an outside consultant, they instigated a process titled Grassroots Safety Leadership which is based on behavioral and CQI principles.

Key elements of this process are empowering employees to run their own safety program and safety culture change. The employee culture has far more impact on safety than the physical problems that exist. This safety culture has empowered the custodians at LLNL to address issues, like ergonomics, that concern their health and safety. In fact, this ergonomic study is a result of the grass roots safety culture at Plant Engineering

The custodians approached health and safety professionals to assist in solving some long- standing problems associated with their training procedures. The main goal of this project was to identify high risk custodial activities using the teamwork of a custodial working group. The custodial working group consists of custodians, industrial hygienists, and safety professionals

39

organized to use existing ergonomic and custodial expertise to address the concerns of the custodians.

Workers Compensation Since the process of changing the safety culture began, Workers Compensation costs for PE&M/O are only 20 percent of what they were in the year 1991. An informal study was recently made of the actual cost of a lost workday injury in the Custodian group. An existing procedure for estimating true Workers Compensation related costs was used for this informal study (Simmonds 1992). It was determined that the minimum cost per hour was $171, and depending on circumstances, could be as high as $450 per hour for a custodian out with a lost work time injury. This certainly underlines the important role ergonomics can play in an organization's overall efforts to reduce employee injuries. Reducing work related injuries not only helps to minimize human suffering, it also offers economic benefits.

JOB ANALYSIS

Identifying Custodial Concerns

Education. A primary goal of the health and safety professionals on the custodial working group was to educate the other members about general ergonomic theories. Included in this education process is the need for using correct postures, emphasizing correct shoulder and hand positions, the mechanics of manipulating heavy objects, and avoiding extreme body positions and postures that could result in injury. After this initial education process the custodial working group soon became a wealth of information. Tasks that the working group considered of primary concern due to extreme body positions were: dusting high shelves, dumping 55 gallon containers, and the wiping of sinks, shower walls, doors, and table tops. Tasks the custodial working group considered to be of secondary concern due to repetitive motion potential were: picking up garbage cans, using spray bottles, and vacuuming. Lower priority tasks consisted of: sweeping with a straw or push broom, cleaning toilets, and mopping.

Video Taping. The custodial working group decided to video tape these custodial procedures for a variety of reasons. Video offers a continuous record of the custodial work and the possibility for frame by frame analysis. Studying a video taped process is an effective and relatively inexpensive way of capturing and observing the necessary musculoskeletal information (Ghosh 1993). Finally, it was agreed that the health and safety professionals would do more than just watch the raw footage. Video opened the opportunity to develop postural analyses. This gave an added dimension to this ergonomic study that should provide useful information to the health and safety professionals and the working group. Once the procedures were identified and prioritized, the custodians were assigned the task of organizing the personnel to be involved with the video taping sessions. The health and safety professionals requested that the custodians being taped should be representative of their general population.

The males and females selected were to reflect the ranges of height in the custodial group. Many factors made the video taping of custodians difficult to coordinate. For security reasons, the use of a video camera is restricted in many of the facilities in which the custodians work. Many custodians were also wary of being video taped because they perceived this as an opportunity for their supervisors to be overly critical. This video taping became an opportunity to build trust with the custodians. It was agreed that tapes would be viewed only by health and safety 40

professionals for ergonomic considerations. Other parties would be allowed to see these tapes only with the custodian's permission. It was also emphasized that these video sessions would serve as an opportunity for the custodians to help many other people in their profession.

Test Participants. Six male and three female custodians were selected for the video taping of the tasks. From this group, two were classified as taller than average, two were classified as shorter than average, and the remaining five were of average height. This mix of heights is important for correct representation of LLNL custodians and is based on anthropometric data (Pheasant 1986). Many custodial procedures, as well as the equipment, are standardized. The adaptation to their tools and procedures were important items to look for during the video's initial viewing. Each custodian was video taped for a minimum of 15 minutes. One video camera was used and an attempt was made to keep the camera orthogonal to the custodians to best capture their posture for video analysis. Traditionally, two cameras are used to capture stationary work activities from different angles (Karlqvist 1994). However, the variety of cramped locations and the randomness of custodial activities made the use of two cameras infeasible. The goal was to record the same nine custodians performing the studied tasks in order to compare and contrast their work techniques and ergonomic problems. A total of 190 minutes of custodial work was recorded.

Video Analysis

Initial Viewing

Body Positions. Many awkward body positions and postures were observed when the video was viewed for the first time. Back flexion and twisting were common occurrences. Some of the bending and twisting observations were a result of improper technique and body mechanics. Trash can emptying and relining is a good example. The small trash cans are at floor level, and their weights vary from five to twenty pounds. Since these cans are never in uniform locations, back flexion, twisting, and awkward lifting were often viewed with this procedure.

Body Adaptation. Most of the back flexion and twisting was a result of custodian adaptation to the limited space of the work area. Cleaning bathrooms and showers, and working in cramped offices, require the custodians to adapt their body positions so they can perform the required cleaning procedures. This inevitably led to increased flexion and twisting of the back. Some custodians were also observed adapting to their tools. The inappropriate length of handles of many wet mops, brooms, and dust mops resulted in an increase in back flexion and twisting. This effect was not viewed when the handles were an appropriate length for the custodian. Some awkward positions in the wrist, shoulder, upper back, and neck were also a result of tool adaptation.

One above average height custodian using an average length mop handle was observed stooping at the neck and shoulders during the entire mopping procedure. In addition, this custodian was observed adapting his wrists to non neutral positions in order to manipulate the mop. Again, these extreme positions were not as prevalent when the tool handle length was appropriate. A procedure that resulted in similar awkward postures was the cleaning of the showers. Since heavy hand wiping and hand scrubbing is required, having a wiping/scrubbing utensil at the end of an appropriate length handle should help avoid extreme body posture problems. The correct

41

tool length or handle shape can create a mechanical advantage to reduce the force the custodian has to exert as well as promoting less extreme postures.

High Risk Postures. The initial viewing of the video taped tasks identified areas of concern for further analysis. The lower back, wrist, and shoulder postures were seen to vary dramatically between custodians performing similar procedures. Whether it was adaptation to the work space or the tools one was using, or even improper technique, these three body posture categories were consistently viewed as having significant potential for developing both short and long term ergonomic related injuries.

Video Analysis Methods

Analyses

Two postural analysis methods were used by this study. The two analyses were performed by stopping the video tape every minute and recording all the necessary information at this interval (Corlett 1979; Ghosh 1993). This allowed both a random approach to identifying high risk activities and also an opportunity to maximize the number of data entry points. The first video analysis is a posture targeting method that is used to create a three dimensional picture of each of the prioritized custodial procedures. The posture targeting will help the custodial working group to identify which custodial procedures require modification of training for awareness of high risk postures. The second is a postural analysis of custodians focusing on increased risk postures in the five joint categories of right wrist, left wrist, right shoulder, left shoulder, and lower back. Using these video analysis methods the information recorded at each interval was used to both identify high risk custodial procedures and also give detailed data of which targeted body parts were performing custodial procedures with increased risk postures. An Industrial Hygienist and a Safety Engineer performed these video analyses together. This allowed for professional interpretation and discussion of each data entry point for both analyses.

Posture Targeting

By Procedure. Posture is an important aspect of interpreting work load and the related possibility of acquiring musculoskeletal injuries (Grandjean 1977). Corlett's posture targeting was the best method for the entire custodial working group to get an ergonomic related perception of high risk body positions associated with different custodial procedures. Corlett's posture targeting diagram is designed to capture three dimensional postures at specific points in time such as those encountered during a freeze frame analysis. Video taping affords the ability to pause the video and examine the postures at set intervals (Corlett 1979). At each video tape analysis interval (every minute), the video tape was paused. The appropriate job procedure was identified and the three dimensional posture targeting position was recorded as a dot on the Corlett chart.

Left and right shoulder positions and back positioning were separately recorded for each procedure. Shoulder positioning was utilized as an indicator of neck stress and pain symptoms due to their interrelationship with the trapezius muscle load {Hagberg 1981, Jensen 1993}. The three dimensional wrist positioning was not recorded because wrist views are limited with only one video camera.

42

Eleven custodial procedures were recorded using this method of analysis. The random nature of this analysis technique controlled how many custodians were viewed for each procedure. A good three dimensional visual picture for the custodial working group was obtained for identifying high risk positions while cleaning bathrooms, cleaning showers, emptying trash, mopping, sweeping, vacuuming, and dusting. Two examples of the posture targeting three dimensional picture are given in Figures 1 and 2. The high risk motions of the right shoulder during mopping are reflected in Figure 1. The twisting and flexion of the back and the many high risk shoulder positions associated with bathroom cleaning are reflected in Figure 2.

Figure 1. Mopping

Figure 2. Cleaning Bathroom

43

Postural Analysis Techniques At each video tape analysis interval a separate postural analysis was also performed. Postural information regarding all nine of the custodian test subjects was obtained for the five target categories of the left and right wrist, left and right shoulder, and lower back. Data points were recorded for each of the five target categories at every minute interval. A data point was determined for each category based on that joint's most extreme position. Total data points were accumulated for every joint category, per each custodian. The percentage of time spent in a joint posture classification was determined by dividing the number of data points per classification by the total number of data points collected per joint category. If a target point was not visible or determinable from the video tape, then a data point was not entered for that minute interval. The categories of recordable wrist, shoulder, and lower back joint postures for this postural analysis were based on existing experimental data (Baluyutt 1995; Genaidy 1995). Analyses utilizing postural classification have been useful in interpreting postural stresses (Genaidy 1994; Punnett 1991; Stetson 1991).

Wrist. The classifications for both the right and left wrist are: Neutral flexion/extension (0 - 15°), Moderate flexion/extension (16 - 45°), Severe flexion/extension (>45°), Radial deviation (>10°), and Ulnar deviation (>10°). The more extreme position of either flexion, extension, or deviation was recorded for each wrist at every minute interval.

Shoulder. The classifications for both the left and right shoulder are: Neutral elevation (0 - 15°), Light elevation (16 - 45°), Moderate elevation (46 - 90°), and Severe elevation (>90°). The more extreme position of either flexion or extension was recorded for each shoulder at every minute interval.

Lower Back. The classifications for the lower back are: Neutral flexion (0 - 15°), Moderate flexion (16 - 45°), Severe flexion (>45°), Lateral bending (>15°), and Rotation (>15°). The more extreme position of either flexion, bending or rotation was recorded at every minute interval.

Increased Risk Postures The postural analysis performed as part of this ergonomic study was designed to identify awkward posture trends for LLNL's custodians. In order to study these trends, increased risk categories were derived from the original postural classifications.

Wrist. Increased risk classifications for the wrist correspond with severe extension/flexion, radial deviation, and ulnar deviation. Previous studies have indicated that severe extension and flexion of the wrist was ranked as having a high perceived stressfulness (Genaidy 1995). Ulnar and radial deviation were categorized as an increased risk wrist posture due to the extreme positions viewed during the statistical analysis. The origin of nerve compression related to cumulative trauma syndrome (CTS) and hand/wrist activity is especially high when the wrist is deviated from positions other than neutral (Moore 1995).

Shoulder. Increased risk classifications for the shoulder correspond with moderate and severe shoulder elevations. Moderate and severe shoulder elevations had the highest relative posture stressfulness ranking (Genaidy 1995). Shoulder pain syndromes and cumulative trauma disorder (CTD) of the shoulder region have also been associated with awkward postures and hands working above shoulder height (Sommerich 1993). 44

Lower back. Increased risk classifications for the lower back correspond with moderate and severe extension, lateral bending, and rotation. Genaidy's studies have shown that the lower back receives its highest relative postural ranking for stressfulness while in flexion >15° and during lateral bending and rotation (Genaidy 1995). Pain and fatigue have been linked to lower back postures such as twisting, lateral bending, and flexion (Boussenna 1982).

Results

Wrist and Shoulder The postural analysis has revealed that LLNL custodians have an increased risk of repetitive motion problems occurring in the right wrist and right shoulder. Both the right wrist and right shoulder are in the increased risk classification approximately 100% more than the left wrist and left shoulder (Table 1). The right wrist is in the increased risk classification 33% of the time as compared to 16% with the left wrist. Similarly, the right shoulder is in the increased risk classification 44% of the time as compared to 22% with the left shoulder.

Lower Back The postural analysis has revealed that LLNL custodians have an increased risk of potential back injuries. The lower back is in the increased risk classification 68% of the time (Table 1). LLNL custodians were found rotating or twisting their lower back 24% of the time. Often these increased risk activities are performed while lifting substantial loads (emptying trash) or while applying a significant force (scrubbing bathrooms and shower walls).

Table 1. Postural Analysis

% = The percentage of time spent in each joint‘s classification determined by the number of data points per classification divided by the total number of data points collected per joint category.

45

Discussion

Training Modification

Ambidexterity. Postural analysis results indicate the need to modify custodial procedures by distributing their work load more evenly across both hands. Reducing the repetitive nature in work procedures reduces both the wear and tear of tendons and their sheaths and the worker's level of fatigue (Sommerich 1993). The postural analysis has indicated that the LLNL custodians are more likely to develop cumulative trauma in their right shoulder and right wrist. The custodians need to be educated on using both the left and right arms equally while performing their tasks. A 100% increase in right wrist and right shoulder use in the increased risk classification indicates an increased possibility for repetitive motion related injuries in those joints. The nature of custodial work does not allow the shoulders and wrists to remain in near neutral positions all the time. Educating the custodians on balancing the use of the right and left arms equally should help to balance the repetitive motions in those joints. There is an expected period of transition for some custodians to adapt their muscles on their left hand side to performing custodial tasks. Balancing the use of the right and left arms can serve as an administrative control to be used in conjunction with neutral posture awareness.

Neutral Postures. Both the video analyses indicated a need to educate custodians about body mechanics. Special emphasis will be placed on making the custodians aware of the neutral positions of the wrist, shoulder, and back. This instruction will be aimed at making LLNL custodians conscious of when they are in the extreme positions that make up the increased risk classifications. The need to educate the custodians about an increased risk for CTDs is vital to both the employees and their supervisors. Work-related musculoskeletal disorders, such as neck, shoulder, upper limbs, and wrist related CTDs, have increased dramatically in recent years and now account for over 60% of reported occupational illnesses in the United States (BLS 1994).

Back Care. Special emphasis will be placed on the high incidence of awkward back postures that were reflected by both these analyses. At the end of this study two custodians suffered lost workday lower back injuries while performing custodial procedures. LLNL has an intensive effort in place to educate employees on ergonomically related back injuries. An excellent back care class is already in existence at the Plant Engineering department. This short class will be modified for the custodians so that it utilizes the information from this ergonomic study.

Modified Procedures. The information from the Corlett posture targeting will be presented to the custodial working group. The working group will interpret the three dimensional information to determine where to adapt the existing custodial training procedures. Each custodial procedure will address the high risk body postures that are frequently encountered while performing the task. This method of training will be used to reinforce the neutral posture and ambidexterity awareness training previously discussed.

Conclusion

An increased risk of musculoskeletal related disorders is associated with awkward body postures (Armstrong 1993; Fine 1987; Genaidy 1995). Video analyses of LLNL custodians has identified some of these awkward postures and the need to modify custodial training by teaching 46

back care, the balancing of right and left hand use, neutral body postures, and the awareness of high risk custodial procedures. The nature of custodial procedures makes it impossible to eliminate high risk body postures. However, teaching the balancing of right and left hand use should help to reduce repetitive motions in the right wrist and right shoulder. Combining this with neutral posture education and the awareness of increased risk procedures should help to reduce the overall incidence of repetitive motions in increased risk categories. A follow up video analysis will be performed to confirm the effectiveness of the modified training procedures.

Acceptance Preliminary feedback shows a high level of acceptance to the modified training procedures. Custodians, supervisors, and high level management have all commented favorably on the training's ease of understanding, simplicity of instruction, and most importantly, its high retention potential. However, for this method of training modification to succeed, it must come from the custodians. The grass roots safety culture that began this ergonomic study process must also be instrumental in the associated changes and implementation. Custodians have noted that if the modification of training comes from the supervisor level, it is less likely to be accepted. It is agreed that a separate supervisor training be created to discuss this ergonomic study, increased risk behaviors, and give a general education about ergonomics. The custodial working group will also be involved in the implementation of the modified training. Selected custodians will be trained to spot high risk behaviors and to effectively record them. Recorded high risk behaviors will then be anonymously addressed in another custodial working group to track the success of the modified training. Since the custodians implemented this study, they are also interested in keeping an eye on each other to reduce these high risk behaviors from within.

Practical Implications The ergonomics awareness level of many custodians has increased sharply since our initial training of the custodian working group to identify high risk procedures. An example of this awareness level, and the effectiveness of the modified training, comes from an ongoing case study. One custodian that was videotaped suffered work related tendinitis in her right wrist after the videotaping. This was also identified in the postural analysis. Her right wrist was in high risk postures 40% of the time while her left wrist was not recorded in any high risk postures. Although this was not a lost workday case, she was required to wear a wrist and forearm brace. Her situation was noted, and her progress was followed during this ergonomic study. The above training procedures were presented to her informally. By focusing on the neutral postures and positions, she became more aware of awkward wrist and arm positions. Within two weeks she was able to perform her custodial duties without the use of the brace. Later, after the postural analysis was performed, the need for ambidexterity was introduced to her. Once she was informed that custodians used their right side in a much higher proportion to the left she agreed that she also uses her right side most of the time. She realized this right away because she is actually left handed. Currently she is no longer experiencing the symptoms that initially required her to wear the brace.

Creating Your Own Program By participating in this ergonomic investigation and being part of the solution, the custodians share in the success of this study. They also bear the responsibility of implementing the modified training procedures. The integral role of custodians in this study should not keep an interested group or individual from building on the foundation created at LLNL and Plant Engineering. This information can serve as an initiation into designing your own custodial 47

ergonomic program. The structure of this study is also adaptable to many other non-traditional ergonomic evaluations. This is a form of analysis that is easy to learn and inexpensive to implement because it does not require excessive time and effort. More importantly, it has the potential to save both workers and managers significant pain and suffering.

48

CHAPTER 4 EVALUATING THE CONTROL BANDING NANOTOOL: A QUALITATIVE RISK ASSESSMENT METHOD FOR CONTROLLING NANOPARTICLE EXPOSURES Journal of Nanoparticle Research, 11(7):1685 – 1704 (2009) David M. Zalk1, Samuel Y. Paik1, Paul Swuste2 1: Lawrence Livermore National Laboratory, P.O.Box 808, L-871, Livermore, CA, 94551 USA email: [email protected] phone: +1 925 422 8904 2: Safety Science Group, Delft University of Technology, P.O.Box 5015, NL-2600 GA, Delft, The Netherlands, email: [email protected] phone: +31 015 27 83820

Abstract

Control Banding (CB) strategies offer simplified processes for controlling worker exposures in the absence of firm toxicological and exposure information. The nanotechnology industry is an excellent candidate for applying such strategies with overwhelming uncertainties of work-related health risks posed by nanomaterials. A recent survey shows that a majority of nanomaterial producers are not performing a basic risk assessment of their product in use. The CB Nanotool, used internationally, was developed to conduct qualitative risk assessments to control nanoparticle exposures. Nanotoxicology experts have requested standardization of toxicological parameters to ensure better utility and consistency of research. Such standardization would fit well in the CB Nanotool‘s severity and probability risk matrix, therefore enhancing the protection of nanotechnology industry workers. This paper further evaluates the CB Nanotool for structure, weighting of risk factors, and utility for exposure mitigation, and suggests improvements for the CB Nanotool and the research needed to bolster its effectiveness.

Key words: nanoparticle, nanomaterial, control banding, risk assessment, risk level, qualitative, CB Nanotool.

Introduction

A fascinating case study for an industrial hygienist (IH) is presented in nanotechnologies due to the properties of some nanomaterials that include a high degree of reactivity, ability to deposit in various regions of the respiratory tract, ability to cross normally impenetrable barriers (e.g., blood-brain barrier, skin), and the lack of a clear toxicological basis for setting nanomaterial- specific occupational exposure limits. The applications for engineered nanoparticles seem endless, and substantial efforts are being put forth by both government and private industries into the research and development of nanotechnologies. However, it is becoming increasingly clear that the very properties that make nanoparticles technologically beneficial may also make them hazardous to humans and the environment and news on nanoparticle health effects are often major news items at popular newspapers, like the Dutch NRC (NRC 2008) and San Francisco Chronicle (Fernholm, 2008). A recent Dutch NRC article refers to the similarity between carbon nanotubes and asbestos, both in their dimensions as well as their pathogenicity (Poland et al. 2008), and a recent San Francisco Chronicle article refers to the potential adverse effects of silver nanoparticles on the environment.

While there is an increasing public demand for evaluating the risks associated with nanomaterials, any attempt to quantify the risks of nanoparticles is fraught with uncertainties. Just to name a few: 1) the contribution of a nanoparticle‘s physical structure to its overall toxicity 49

is not fully understood; 2) lung deposition and alveolar clearance appears to be significantly different between nanoparticles and their larger counterparts; 3) there is no consensus on the relevant indices of exposure, as particle size and surface area are likely to be much more important than mass; and 4) there is a lack of clear information on exposure scenarios and populations at risk. Control Banding (CB), which is described in greater detail below, has proven to be an effective strategy for controlling worker exposures in the absence of toxicological and exposure information, and has been mentioned as a potentially useful concept for managing nanomaterial exposures in the workplace (Warheit et al. 2007a; Thomas et al. 2006; Maynard 2007; Schulte et al. 2008). The CB strategy facilitates decisions on appropriate levels of control, based upon product and process information, without complete information on hazards and exposure scenarios. The CB Nanotool was recently developed and has been implemented in many countries utilizing a qualitative decision matrix for a risk assessment that leads to commensurate controls (Paik et al. 2008). Now that the CB Nanotool is beyond its pilot stage, it is appropriate to evaluate its construct, applications, and efficacy and to invite continued dialogue, use, and improvement of the tool within the IH community.

Control Banding CB is not intuitively understood by name alone and can therefore be considered to be a notation for experts, which may come as a surprise to many people. The term originates from the field of IH and represents a qualitative instrument to assess risks for chemical substances in order to generate solutions and control measures (Russel et al. 1998). The instrument uses categories, or ´bands´, of health hazards, which are combined with exposure potentials, or exposure scenarios, to determine desired levels of control. The bands of health hazards for some control banding approaches are based upon the European Union risk phrases, while exposure potentials include the volume of the chemical used and the likelihood of the chemical becoming airborne, estimated by the dustiness or volatility of the source compound (Maidment 1998).

Originally, the concept of banding risks of chemical substances and their exposure controls started in the pharmaceutical industry around 20 years ago. Here, the limited availability of pharmacological and toxicological data of products handled by workers was the main motive to develop control strategies as part of a risk management approach. The foundation of the present movement for control banding is derived from a program of the British Health and Safety Executive (HSE) to assist small and medium enterprises in their risk management approach so they could have a simplified method to comply with regulations requiring all users of chemicals to assess their risks and implement appropriate controls to protect their workers. In a series of papers in 1998, the instrument was published (Annals 1998). International CB workshops and activities further refined the instrument, and explored possibilities to apply the control banding approach to other domains, like ergonomics, occupational safety, and recently to control nanoparticle exposure (Zalk 2001; Annals 2003; Swuste 2007; AIHA 2007; Zalk and Nelson 2008; Paik et al. 2008). Although CB has received criticism (see for instance Kromhout 2002a; Swuste et al. 2003; Jones and Nicas 2006a; ACGIH 2008), the focus on controls is a strong point of the instrument and makes it applicable for operations with many uncertainties in hazard, exposure, and consequence data (ACGIH 2008). Much of the criticism has focused on issues relating to the simplicity of the CB approach and the mistaken belief that CB has forsaken the experts and their traditional, quantitative methods. With nanoparticulate exposure and its many toxicological and quantitative measurement uncertainties, however, it can be argued at this time that the CB approach may in fact be superior to the traditional methods.

50

Risk Assessment of Nanomaterials with Control Banding

A recent survey indicates that 65% of companies working with engineered nanomaterials (NM) do not perform any kind of risk assessment relating to their product use (Helland et al. 2008). Therefore, the development of a standardized risk decision framework is necessary and has been called for in many of the latest investigative studies (Schulte et al. 2008; Warheit et al. 2008; Hallock et al. 2008). A conceptual CB model was presented by Maynard (2007) using ´impact´ and ´exposure´ indices. This conceptual model combines engineered nanomaterial composition parameters (shape, size, surface area and surface activity) with their exposure availability (dustiness and amount in use). These indices are linked to bands with four corresponding control approaches. The control approaches are a grouping of three levels of engineering containment, based on sound IH principles; (i) general ventilation, (ii) fume hoods or local exhaust ventilation, and (iii) containment. The fourth level is ´seek specialist advice´, referring to specialist IH expertise. In the recently published paper on the pilot ´CB Nanotool´, the feasibility of using CB principles is further developed and put into practice (Paik et al. 2008). Here the control band for a particular operation is based on the overall level of risk determined for that operation. This risk level (RL) is the result of a combination of a severity score and a probability score for that operation (Figure 1), analogous to the impact and exposure index described by Maynard.

Figure 1. Risk level (RL) matrix as a function of severity and probability scores. Control bands are based on overall risk levels1.

Probability Score

Extremely Unlikely Less Likely Likely Probable (0-25) (26-50) (51-75) (76-100)

Very High

(76-100) RL 3 RL 3 RL 4 RL 4

High (51-75) RL 2 RL 2 RL 3 RL 4 Medium (26-50) RL 1 RL 1 RL 2 RL 3 Low (0-25) RL 1 RL 1 RL 1 RL 2 Severity score

Control bands by risk level: RL 1: General Ventilation RL 2: Fume hoods or local exhaust ventilation RL 3: Containment RL 4: Seek specialist advice 1(Paik et al. 2008). Reprinted by permission of the British Occupational Hygiene Society, License #2114970359639.

51

The biggest challenge in developing the CB Nanotool was in the determination of the weightings for the different risk factors. To accomplish this, a group of experts at Lawrence Livermore National Laboratory (LLNL) was convened in more than 20 meetings over a 6-month period to address health, safety, and environmental control of NM to protect the health of both workers and the public while balancing the needs and requirements of researchers to continue their operations in a safe manner. These experts formed an ―Institutional Project Team‖ (IPT) that was internally established at LLNL for the purpose of developing LLNL‘s first institutional NM safety program. Members of the NM IPT included: 1) LLNL‘s medical director, with over 30 years of experience, who oversaw the development of LLNL‘s medical surveillance program for Nanoparticle Workers and was to use the CB Nanotool‘s RL outcome as a basis for determining levels of medical surveillance; 2) The director of LLNL‘s Nanoscale Synthesis and Characterization Laboratory, with over 20 years of experience, who provided input into the physical characteristics that could affect the severity aspects of NM; 3) LLNL‘s Nanotechnology Safety Subject Matter Expert (SME), with over 10 years of experience, who played a key role in developing the tool and is a co-author of this paper; 4) LLNL‘s medical programs division lead, with over 25 years experience, who developed a medical history questionnaire and medical monitoring scheme for higher risk workers; 4) A field environmental analyst, with over 20 years experience, who provided input on the proper waste characterization of NM ; 6) An instructor from the safety training division, with over 15 years of experience, who developed an institutional web-based course on NM safety; 7) and an IH, with over 15 years experience, who provided a critique of the CB Nanotool and its practicality for use in the field. The NM IPT met on a weekly basis over these six months. These meetings were approximately one hour in duration and discussed not only the CB Nanotool itself, but how it fit into the broader context of a comprehensive nanotechnology safety program.

Once the CB Nanotool was developed, integrated into the LLNL nanotechnology safety program as the required risk assessment approach for all work with NM, and implemented as part of the LLNL pilot program, further expert review and input was sought. One mechanism was through expert solicitation via email and phone correspondence prior to submission of the original ‗CB Nanotool‘ manuscript. Dr. Remko Houba, who is currently affiliated with the ArboUnie Expert Centre for Chemical Risk Management in The Netherlands, provided valuable input. Dr. Houba, who has over 15 years experience as an IH and investigated the population at risk in the Dutch NM research and manufacturing industries, offered his insight in the CB Nanotool including issues regarding the validity of the tool versus IH professional judgment and the importance of including mistiness (e.g., from spraying applications) as part of the dustiness index. Dr. Houba also concurred with the weighting for ―unknown‖ factors and believed it adequately addressed the uncertainty that is prevalent in this relatively new field of NM safety. Another mechanism was through peer reviews of the original ‗CB Nanotool‘ manuscript that was submitted to the Annals of Occupational Hygiene Journal. One reviewer commented on whether or not dermal considerations were adequately addressed in the design of the CB Nanotool. Another reviewer commented on the role of total surface area of the nanoparticles and how this factor is addressed. Both of these issues are discussed in greater detail in the current study. While the actual identities of the peer reviewers are not known to the authors of this paper, they are presumed to be experts in this field as they were chosen by the editor of the Annals of Occupational Hygiene Journal as peer reviewers of the manuscript. Yet another mechanism was through presentation of the CB Nanotool as part of two international conferences. One of them was on 20 October 2008, during the Organisation for Economic Co-operation and Development (OECD) Working Party on Manufactured NM Workshop on Exposure Assessment and Exposure Mitigation. Another was the International Commission on Occupational Health (ICOH) Congress in Cape Town, 52

South Africa at the 5th International CB Workshop on 25 March 2009. At these workshops, the value of the CB Nanotool became quite apparent; however, there were many excellent questions asked that focused on the expert judgment behind the CB Nanotool, the weighting values to determine RLs, and whether additional risk assessments were available to further evaluate its outcomes as compared to expert IH judgment. This paper is presented as a direct outcome of the OECD NM Workshop and the information presented is bolstered by professional consultation at the ICOH Congress. In addressing these cumulative questions, and in developing a further transparency of the CB Nanotool, this paper was developed to offer an expert review of the most recent research in evaluating the initial judgment behind the CB Nanotool and revisits the tool‘s scoring parameters based on this cumulative information available to date.

Pilot CB Nanotool Scoring Parameters

As described in Paik et al. (2008), an important consideration in developing the CB Nanotool was that information on many of the factors related to severity would be unknown or uncertain. While it was recognized that traditionally, an unknown hazard would be treated as a high hazard, it was also that defaulting to the worst-case assumption is overly conservative for most hazards and would place an unnecessary burden on those managing the risk and limit the tool‘s usefulness. For that reason, it seemed appropriate to assign a given factor with ‗unknown information‘ 75% of the point value of ‗high‘. The implication, seen in Figure 1, is that for a nanotechnology operation where nothing is known, RL 3 (containment) is required. In this scenario, if just one rating of any of the factors were to be ‗high‘, the tool would require an RL 4 assignment for the activity, the maximum control. The breadth and depth of the scoring factors is provided in greater detail within the original Paik et al. (2008) article and therefore will not be included to the same degree within this paper. The information presented below reflects a summary of the severity factors, probability factors, and the maximum scores attributed to each of these factors.

Severity Factors Based on the literature available prior to publication of the pilot CB Nanotool, the list of factors below were considered to determine the overall severity of exposure to nanoscale materials. These factors influence the ability of particles to reach the respiratory tract, their ability to deposit in various regions of the respiratory tract, their ability to penetrate or to be absorbed through skin and their ability to elicit biological responses systemically. The division of severity factor points taken cumulatively is 70% for the NM and 30% for the parent material (PM). Research to date does not contraindicate the potential for engineered NM to be more toxic than its PM. The individual factors that make up the NM severity factors are as follows:

Surface chemistry of NM: surface chemistry is known to be a key factor influencing the toxicity of inhaled particles. Points are given based on a judgment of whether the surface activity of the nanoparticle is high, medium or low. High: 10 Medium: 5 Low: 0 Unknown: 7.5

Particle shape of NM: the highest severity score is given to fibrous or tubular-shaped particles. Particles with irregular shapes (anisotropic) have higher surface areas than isotropic or spherical particles. Tubular, fibrous: 10 Anisotropic: 5 Compact/ spherical: 0 Unknown: 7.5

53

Particle diameter of NM: the severity score was based on the particles‘ deposition in the respiratory tract, regardless of the region in the respiratory tract. Additional research on the toxicological significance of particle size is needed to improve the basis for these weighting factors. 1 – 10 nm: 10 11 – 40 nm: 5 < 41 – 100 nm: 0 Unknown: 7.5

Solubility of NM: poorly soluble, inhaled nanoparticles can cause oxidative stress, leading to inflammation, fibrosis, or cancer. Since soluble NM can also cause adverse effects through dissolution in the blood, severity points are assigned to soluble NM as well, but to a lesser degree. Insoluble: 10 Soluble: 5 Unknown: 7.5

Carcinogenicity of NM: points are assigned based on whether the NM is carcinogenic or not, regardless of whether the material is a human or animal carcinogen. Little information is available. Yes: 7.5 No: 0 Unknown: 5.625

Reproductive toxicity of NM: points are assigned based on whether the NM is a reproductive hazard or not. Little information is available of this factor. Yes: 7.5 No: 0 Unknown: 5.625

Mutagenicity of NM: points are assigned based on whether the NM is a mutagen or not. Little information is available of this factor. Yes: 7.5 No: 0 Unknown: 5.625

Dermal toxicity of NM: points are assigned based on whether the NM is a dermal hazard or not. Little information is available of this factor. Yes: 7.5 No: 0 Unknown: 5.625

Toxicity of PM: although research agrees that NM can be more toxic than PM, it is a good starting point for understanding the NM toxicity. Points are assigned according to the OEL of the bulk material. 0 – 1 μgm-3: 10 2 – 10 μgm-3: 5 < 41 – 100 μgm-3: 2.5 > 100 μgm-3: 0 Unknown: 7.5

Carcinogenicity of PM: Points are assigned based on whether the PM is carcinogenic or not. Yes: 5 No: 0 Unknown: 3.75

Reproductive toxicity of PM: points are assigned on whether the PM is a reproductive hazard or not. Yes: 5 No: 0 Unknown: 3.75

Mutagenicity of PM: points are assigned on whether the PM is a mutagen or not. Yes: 5 No: 0 Unknown: 3.75

Dermal hazard potential of PM: points are assigned on whether the PM is a dermal hazard or not Yes: 5 No: 0 Unknown: 3.75

54

The overall severity score is determined based on the sum of all the points from the severity factors. The maximum score is 100. Since nanoparticles usually behave much differently than their PM due to their small scale, greater consideration was given to the NM characteristics (70 possible points out of 100) than to the parent material characteristics (30 possible points out of 100). An overall severity score of 0-25 was considered low severity; an overall severity score of 26-50 was considered medium severity; an overall severity score of 51-75 was considered high severity; and an overall severity score of 76-100 was considered very high severity.

Probability Factors

The probability scores are based on factors determining the extent to which employees may be potentially exposed to nanoscale materials.

Estimated amount of NM used during operation: for NM embedded on substrates or suspended in liquid, the amount is based on the NM compound itself and not the substrate or liquid portion. > 100 mg: 25 11 – 100mg: 12.5 0 – 10 mg: 6.25 Unknown: 18.75

Dustiness/mistiness: since employees are potentially exposed to nanoparticles in either dry or wet form, this factor encompasses both dustiness and/or mistiness of the NM. Knowledge of the operation (e.g., handling dry powders versus liquid suspensions of nanoparticles) would be a means to estimate dustiness/mistiness. Due to the size of NM, visibility may not be a reliable means to estimate overall dustiness/mistiness. A CB Nanotool design feature is that a rating of ´none´ for dustiness/mistiness level (and only for this factor) automatically causes the overall probability score to be ―Extremely Unlikely‖, regardless of the other probability factors, since the other factors will not be relevant if no dust or mist is being generated. Examples of operations resulting in a ‗none‘ rating are handling of carbon nanotubes embedded on fixed substrates and working with non-agitated liquid suspensions. High: 30 Medium: 15 Low: 7.5 Unknown: 22.5

Number of employees with similar exposure: points are assigned by the number of employees assigned to this activity. More employees means a higher probability of an employee being exposed. > 15: 15 11 - 15: 10 6 - 10: 5 Unknown: 11.25

Frequency of operation: points are assigned based on the frequency of the operation, as more frequent operations are more likely to result in employee exposures. Daily: 15 Weekly: 10 Monthly: 5 Less than monthly: 0 Unknown: 11.25

Duration of operation: points are assigned based on the duration of the operation, as longer operations are more likely to result in employee exposures. > 4h: 15 1 – 4 h: 10 30 – 60 min: 5 < 30 min: 0 Unknown: 11.25

The overall probability score is based on the sum of all the points from the probability factors. The maximum score is 100. An overall probability score of 0-25 was considered extremely unlikely; an overall probability score of 26-50 was considered less likely; an overall probability score of 51-75 was considered likely; and an overall probability score of 76-100 was considered probable. Based on the severity score and probability score for an operation, the overall level of risk and corresponding control band is determined by the matrix shown previously in Figure 1.

55

Judgment Behind the Score Weighting

There was a great deal of research and consideration of the collective information available during the development of the CB Nanotool. In concept, as described above, it was easiest to begin with the realization that traditional IH did not provide a comprehensive and accurate quantitative risk assessment of NM. The use of available quantitative instruments, such as condensation particle counters and nanoparticle surface area monitors, needs to be balanced with their potential biases (summarized in ISO, 2007). Condensation particle counters can have both positive and negative biases depending on the particle size distribution of the particles they are measuring. The CPC Model 3007 (TSI Inc., Shoreview, MN), for example, measures particles in the 10 nm to >1000 nm range and provides a number concentration. This instrument would significantly underestimate the actual number concentration of nanoparticles if the particle count median diameter were less than 10 nm or significantly overestimate the actual number concentration of nanoparticles if the particle count median diameter were greater than 100 nm. This would be problematic in most environments where naturally occurring nanoparticles would tend to co-exist with the engineered nanoparticles in question. Similarly, nanoparticle surface area monitors that rely on diffusion charging of the sampled particles can have significant biases when particles greater than 100 nm exist in the sampled air stream. Below 100 nm, active surface (inferred from the attachment rate of positive unipolar ions to particles) is a function of the square of the particle diameter and is considered a good indicator of external surface area. However, this relationship breaks down for larger particle size; therefore, the measurement would no longer correlate well the particles‘ external surface area. In addition, the expense of the available and more accurate exposure monitoring tools could be seen as cost-prohibitive, especially in the face of so much uncertainty. Once a decision was made to build a qualitative approach, it was also easy to decide on using the 4X4 risk model that is utilized in many of the CB strategies. The 4X4 risk matrix has been found over time to balance ease-of-use with an appropriate level of rigor to develop a binning of established and graded control approaches in a historically acceptable manner (Maidment 1998; ANSI 2000; Zalk and Nelson 2008). The research also presented a relatively consistent set of factors that should used in the model, which each needed to be included in the scheme for scoring input; however, the weighting of each factor relative to the others was a bit more involved and required a relative risk approach in line with the available research (Robichaud et al. 2005).

Severity

Physicochemical characteristics NM (40 points): Research showed a strong agreement that the physicochemical aspects of NM structure have a predominant effect on their potential toxicity (Maynard 2007; Warheit et al. 2007a; Thomas et al 2006). Therefore, both the physical parameters (particle shape and diameter) and chemical parameters (surface chemistry and solubility) were weighted equally with 20 points attributed to each parameter as research did not indicate that one parameter or the other led to a more elevated risk. This decision was also based on the fact that appropriate standardization of testing did not appear available in the literature, only that both of these considerations were necessary when evaluating the potential toxicity of a given NM (Powers et al. 2006).

Toxicological characteristics NM (30 points): Having taken into account the more generic health hazard parameters of NM, it was also necessary to account for the toxicological concerns that might be related to research on specific NM effects. As the research on NM as a whole had not delved into these specific toxicological aspects to date, agreement by experts invariably 56

noted that the more classic toxicological outcomes for an individual NM product should also be considered (Maynard 2007; Powers et al. 2006). Therefore, the toxicological adverse outcomes that would lower any prospective occupational exposure limit were included and these were: carcinogenicity, reproductive toxicity, mutagenicity, and dermal. From an IH perspective, it is difficult to consider weighting these adverse outcomes as anything other than equally as any one of these toxic effects will lead to an appropriate lowering of its OEL to avoid a health hazard.

Toxicological characteristics of PM (30 points): As stated earlier, the properties that make NM unique in their utility also have the potential to create unique toxicological considerations. Without more specificity of this issue presented in research publications, it is necessary to start with the likelihood that much more of this toxicological information would be available for the bulk PM. Therefore, equal weight was given for the research-derived toxicological characteristics for both the NM and PM, with both at 30 points. This also gave an appropriate greater weighting to the physicochemical aspects of NM (40%), which are being heavily researched, than for the specific toxicological outcomes of both the NM (30%) and its PM (30%). A decision was made to use the same toxicological characteristics for PM and NM, dividing each of their points equally, although greater weighting was given to the NM (30%) then to the PM (20%) to reflect concerns expressed in the research. To make up the additional 10 points to equalize the PM toxicity with NM toxicity, the PM‘s OEL was included in the PM toxicological outcome determination as this is more holistic in offering a relative weight to a more broad classification of epidemiology and toxicology issues. Thus the PM OEL (10%) was given twice the value of any of the individual PM toxicological characteristics (5%).

Probability

Dustiness/Mistiness (30 points): In determining the factors that would lead to potential exposure to employees, the primary consideration would be based on the opportunity for the NM in question to become airborne. Experts are in agreement that the most important factor for determining the potential for exposure, and therefore the potential for bioavailability and translocation systemically, is in regards to inhalation (Warheit et al. 2007a; Maynard 2007; Thomas et al. 2006; Powers et al. 2006; Tsuji et al. 2006; Holsapple et al. 2005). The consideration was therefore a balance between its ability to become airborne, to disperse easily, and the amount of material used. It was determined to give Dustiness/Mistiness the greatest weight of the probability factors (30%). It was also given consideration that many of the CB Nanotool users performing an initial screening of NM activities could default here to ‗unknown‘ if no other parameters for airborne potential were readily available (Donaldson et al. 2006). Then, if the RL outcome were too restrictive with the weighting of an ‗unknown‘ score, a decision could to made to use quantitative measurements to assist in scoring this category. This focused use of quantitative monitoring tools is considered a more appropriate and cost-efficient application and is not confounded by the biases of using multiple monitoring devices simultaneously. In addition, although dustiness and mistiness are characterized together, mistiness in isolation would likely have a lower score then dustiness as the nanoparticles would be in the form of wet suspensions. This score for mistiness would therefore be more analogous to a lower score for dry, agglomerated particulates than when compared to non-agglomerated, highly dispersed particulate in a similar operation.

Estimated amount of chemical used (25 points): The more material that is used, the better chance that it will become available as a potential source term for employee exposure. The weighting of the amount of chemical used in a given task was considered to be a slightly lower 57

relative risk (25%) than the consideration for the airborne potential (30%). The authors also considered the combination of dustiness and amount used as being the primary exposure probability factors, in deference to Maynard‘s (2007) use of this as the only exposure factors, and therefore wanted this combination to be greater (55%) than the remaining factors that are task-specific (45%). This overall weighting is not entirely based on the relative risks presented in research for these factors due to the fact that this information is acknowledged as not being available in sufficient depth to make such a determination (Nasterlack et al. 2008; Tsuji et al. 2006; Holsapple et al. 2005). Therefore IH expertise was utilized to make this relative risk delineation based on the decades of combined field practitioner experience for the factors culminating in exposure.

Opportunity for exposure (45 points): For all of the discussion on the toxicological aspects of working with nanoparticles, the focus can now be given to the more classical nature of the traditional IH profession. Exposures and the potential for employee uptake are typically seen as a function the length of the task at hand and the periodicity of which that task is performed. Taking on aspects of epidemiology and a statistical view of the potential for variance from the mean, the more workers there are that are performing a given task the higher the probability of exposure. Therefore, these three aspects relating to exposure opportunity were given an equal weighting with frequency of operation, duration of operation, and the number of employees performing each given 15% of the probability factors scoring.

Addressing expert opinion

Surface area: There was some professional consideration given as to whether total surface area should be considered an exposure characteristic or a severity characteristic. Total surface area was not included as a severity characteristic because all the other severity characteristics pertained to properties inherent to a given NM or PM and did not consider dosage or exposure. However, since particle size and particle shape are characteristics inherent to NM that would result in a greater total surface area, at the same mass concentration, these were included as severity parameters. Surface area relating to exposure characteristics is captured in the dustiness/mistiness scoring factor and is accounted for in its greater weighting for probability of exposure. Elevated dustiness/mistiness levels for a given activity will have a higher concentration of airborne nanoparticulate and a much higher surface area concentration than lower dustiness/mistiness levels. With dustiness/mistiness as such a heavily weighted scoring factor, especially with the potential for a lack of visual evidence to appropriately characterize this aspect, it is recommended for the evaluator to create their own decision matrix to decide whether they should make this a qualitative judgment or perhaps investing in a quantitative assessment specific to this input factor.

Dermal exposure: There were a few experts that questioned how dermal considerations were addressed in the design of the CB Nanotool. One issue was that the dustiness/mistiness input factor includes a design feature that defaults to ‗Extremely Unlikely‘ if there is no potential for airborne NM during a given process. It was mentioned during a third-party review of the CB Nanotool that this default appears to discount the potential for the dermal exposure route and therefore its relevancy (Ryman-Rasmussen et al. 2006). In actuality, the potential for dermal exposure and uptake through various external uptake routes (eg. ocular, hair follicle) can be considered entirely influenced by highly dispersible nanoparticulate, affecting dermal exposure through both airborne routes as well as its deposition on working surfaces. If there is no airborne exposure, then dermal exposure is isolated to the source term, which can be controlled with 58

gloves while handling the product. Another point of discussion was the weighting of the dermal toxicity parameters overall. As the research is indeterminate for the potential of dermal penetration of NM through intact skin, in question was the consideration of this route as an equivalent severity consideration. The equal weight of NM dermal toxicity was given to not only address this one aspect, but also in consideration of the other factors that encompass cutaneous toxicity in a manner that also includes the potential for absorption as well as penetration.

Frequency and duration: Some analysis was given toward the inclusion and weighting of the duration and frequency of a given task in determining the potential for exposure. As a primary reference in support of this CB approach for NM, Maynard (2007) considered dustiness and amount as the only factors to be considered within the exposure index. The weight to these two factors is given in protecting the employee first, regardless of the frequency and duration of a given task. In the CB Nanotool, the greatest weighting in the probability scoring is given to the dynamics of the source term – dustiness and amount – as these are the focus of the controls that are derived from the toolkits‘ application. However, the consideration of frequency and duration, in addition to number of employees potentially exposed, give a practicality counterweight to the probability of exposure. Consideration of these additional factors were not seen as conflicting with the two primary factors, but rather supplementing them. That is, if a task takes a few minutes and is performed a couple times a year, this must also be given consideration in affecting the overall potential for exposure.

OEL of PM: Giving only 10% of the severity weight to a well researched, professionally derived, and science-based OEL for the PM was considered by some to be insufficient. In consideration of the relative value of the PM OEL, the authors of the CB Nanotool felt that its 10 points did not stand in isolation. The toxicological and epidemiological aspects that drive a PM‘s OEL to lower and more conservative values are often the same as the identified critical effects (e.g. carcinogenicity, reproductive toxicity, mutagenicity, and dermal) which would each add an additional 5% to the severity weighting up to a theoretical maximum of 30%.

Number of Employees: Experts at the ICOH Congress‘ 5th International CB Workshop questioned the 15% weighting given to the number of employees as part of the probability of exposure. The value of this weighting was agreed upon by the expert working group at LLNL as there is a large working population at this national research laboratory. At LLNL, there can be a significant number of researchers working with NM as part of numerous projects, phases, and tasks at any given time that it deserved a comparable weighting to frequency and duration. It was decided that, even with engineering controls potentially in place, the variability of individual working habits and approaches to NM research applications with a large research population supported this weighting. It should be noted that no risk assessment approach, especially those with a qualitative basis, should be adopted prima facie. The CB Nanotool was developed for the NM working parameters at LLNL and, although adopted by many organizations and even as a best practice (IRSST 2009), it should not be put directly into practice without consideration and evaluation given to the weighting of factors that may be pertinent to individual facilities. Therefore, this weighting factor may not be appropriate for research organization with only a few workers and this weighting value may be distributed into other probability factors as deemed appropriate.

Uncertainty: One of the experts commented on the fact that through his numerous discussions with companies using NM in their processes, it appeared that the most common approach that is 59

currently applied in The Netherlands is that no exposure to NM is accepted because of the uncertainty associated with NM. Similar sentiments were expressed during presentations of the CB Nanotool during the two aforementioned professional conferences. While the CB Nanotool, by assigning different levels of engineering controls based on RL, potentially allows some level of exposure for certain types of activities, the assignment of 75% of the rating score of high for ―unknown‖ factors appeared to satisfy most experts in terms of erring on the safe side for relatively unknown materials and operations.

Validation: Appropriately, many experts have questioned the ability to develop the parameters to truly validate the pilot CB Nanotool. The problem is that there is a lack of a gold standard to accomplish this for NM. In practice, this question remains a major topic of discussion for chemical control CB strategies; however, publications have begun to fill this research need that is building confidence in the approach in the face of uncertainties (Zalk and Nelson 2008; Marquart et al. 2008; Tielemans et al. 2008). This question is more appropriately compared to the scarcity of validation publications for CB schemes utilized in the pharmacological industries. CB has been an accepted practice for the risk assessment and control of new and more potent pharmaceutical components and has been successfully in place within the industry for over a decade though very little validation data has been presented in research publications (Farris et al. 2006; Naumann et al. 1996). Often in this industry, it is the recommended control that has been put in place that is monitored for its containment effectiveness using standardized, mock particulate (e.g., lactose) that have established analytical detection methods. In a similar manner, quantitative particle counters have been used in selected screening opportunities to compare rogue NM particle counts as compared to background levels. During the implementation and evaluation of the pilot CB Nanotool, this approach was used to facilitate the assignment of the appropriate dustiness/mistiness level to specific operations. The scenarios presented as case studies in Paik et al. (2008) focused on a sampling of representative and existing research and development (R&D) activities within the LLNL institutional safety document database. Prior to the existence of the CB Nanotool, expert IH advice was used to select the most appropriate controls for a given activity with NM. The IH would also utilize best practices such as the NIOSH ―Approaches to Safe Nanotechnology‖ publication. Therefore outcomes were directly compared with existing IH expertise which as close as we can come to a validating method without the existence of a gold standard. We provided this validation within the Paik et al. (2008) article and good agreement was found at the time between the IH and the CB Nanotool. Since that time, many more applications were reviewed and a much larger database for comparison has been developed and is presented below in Table 1. In addition, many more research papers in publication have been consulted, to make a formalized decision on our ‗pilot‘ determination, the validity of the findings, and parameters for refinement for the CB Nanotool in practice.

Further applications of the CB Nanotool

Despite the limitations presented, the CB Nanotool is a transparent and logical method. Although much research has been performed within the sciences relating to NM since the first publication of the tool (Paik et al. 2008), data on nanomaterial health effects is still limited and it is expected that this stream of information will continue to expand rapidly (Yang et al. 2008; Warheit et al. 2008; Hallock et al. 2008). Therefore, as specific studies are published, severity parameter scores that where once ‗unknown‘ can now be more accurately portrayed and users of the tool can adjust their input and affect the severity score. More importantly, as one cannot 60

control the pace of science, users of the CB Nanotool can immediately seek to address some of the parameters relating to the probability of exposure to reduce the final overall RL. For experts in the IH field, this is a common activity; however, for CB Nanotool users new to the exposure sciences, this is an essential learning opportunity in a simple and practical format. An additional 27 risk assessments with the CB Nanotool have been performed since the initial publication of the CB Nanotool and the results and discussion are briefly discussed below.

Table 1. Additional activities assessed by the CB Nanotool

Activity Scenario Description Name or Current Severity Probability Overall Recommended Upgrade Number description of Engineering band band Risk Engineering Engineering nanomaterial Control Level Control Based Control? Recommended Without on CB Nanotool by IH Expert Controls Risk Level 1 Synthesis of metal ZnO, SnO2, TiO2, Containment High Extremely RL2 Fume hood or No oxide nanowires on PBZrTiO3, Unlikely local exhaust substrates wthin a tube BaTiO3, and ventilation furnace SrTiO3 nanowires 2 Synthesis of silver and Ag oxide Fume hood or High Less RL2 Fume hood or No copper oxide nanoparticles, Cu local exhaust Likely local exhaust nanoparticles oxide ventilation ventilation nanoparticles 3 Activities related to Ag, Cu, Ni, brass, Containment High Likely RL3 Containment No operating and Au and Pt maintaining a vertical nanoparticles tube quench furnace and horizontal tube furnace 4 Deposition of liquid- Polymer latex, General Medium Extremely RL1 General No suspended gold, platinum, ventilation Unlikely ventilation nanoparticles onto palladium surfaces using low nanoparticles voltage electric fields 5 Preparation of samples. Carbon black, Al Fume hood or Medium Extremely RL1 General No Activities include oxide, Mg oxide, local exhaust Unlikely ventilation cutting, slicing, polycrystalline ventilation grinding, lapping, diamond polishing, chemical suspension, etching, colloidal silica, Pd electrochemical powder, carbon polishing and ion nanotubes etching. 6 Water is poured into Carbon nanotubes Fume hood or High Extremely RL2 Fume hood or No container with liquid- local exhaust Unlikely local exhaust suspended carbon ventilation ventilation nanotubes 7 Gold nanoparticles Gold nanoparticles General Medium Extremely RL1 General No used to test carbon ventilation Unlikely ventilation nanotube filter 8 Mixing polystyrene Polystyrene General Medium Extremely RL1 General No spheres with buffer, spheres, ventilation Unlikely ventilation etching nanostructures nanostructures onto semiconductors 9 Addition of quantum Cadmium Fume hood or High Extremely RL2 Fume hood or No dots onto porous glass selenide, lead local exhaust Unlikely local exhaust sulfide ventilation ventilation 10 Growth of palladium Palladium Fume hood or High Less RL2 Fume hood or No nanocatalyst nanocatalyst local exhaust Likely local exhaust ventilation ventilation 11 Sample preparation and Gold, silver General Medium Less RL1 General No characterization nanoparticles ventilation Likely ventilation

61

12 Sample preparation and Iron oxide, silicon Fume hood or Medium Less RL1 General No characterization dioxide, aluminum local exhaust Likely ventilation oxide, carbon, ventilation ceramic aerogels and nanopowders 13 Synthesis of aerogel Zinc, titanium General Medium Extremely RL1 General No nanoparticles ventilation Unlikely ventilation 14 Synthesis of aerogel Silica, iron, General Medium Extremely RL1 General No chromium, copper, ventilation Unlikely ventilation zinc nanoparticles

Table 1. Additional activities assessed by the CB Nanotool (continued)

Activity Scenario Description Name or Current Severity Probability Overall Recommended Upgrade Number description of Engineering band band Risk Engineering Engineering nanomaterial Control Level Control Based Control? Recommended Without on CB Nanotool by IH Expert Controls Risk Level 15 Synthesis and optical CdSe quantum General High Less RL2 Fume hood or Yes characterization of dots, germanium ventilation Likely local exhaust nanoparticles quantum dots, iron ventilation oxide, gold, lead sulfide nanoparticles 16 Sample preparation and CdSe quantum Fume hood or High Less RL2 Fume hood or No charcaterization of dots local exhaust Likely local exhaust CdSe nanodots ventilation ventilation 17 Sample preparation and Carbon Fume hood or High Less RL2 Fume hood or No charcaterization of diamonoids local exhaust Likely local exhaust carbon diamonoids ventilation ventilation 18 Sample prepration and Gold, silver General High Less RL2 Fume hood or Yes characterization using nanoparticles ventilation Likely local exhaust laser microscopy ventilation 19 Preparation of Gold, copper, General High Extremely RL2 Fume hood or Yes nanofoam sample for aluminum, nickel ventilation Unlikely local exhaust microscopy nanoparticles ventilation 20 Preparation of carbon Carbon nanotubes General High Extremely RL2 Fume hood or Yes nanotubes sample for ventilation Unlikely local exhaust microscopy ventilation 21 Machining (e.g., Silica aerogels, Fume hood or High Less RL2 Fume hood or No turning, milling) of tantulum aerogels, local exhaust Likely local exhaust aerogels and nanofoams metal nanofoams ventilation ventilation for target assembly (copper, gold), carbon nanofoams 22 Site wide waste Various General High Less RL2 Fume hood or Yes sampling activities ventilation Likely local exhaust ventilation 23 Waste Accumulation Various General High Less RL2 Fume hood or Yes Area activities, ventilation Likely local exhaust including waste ventilation management, waste packaging, etc. 24 Analysis of Various General High Less RL2 Fume hood or Yes nanomaterial waste ventilation Likely local exhaust samples in the ventilation laboratory 25 RHWM field tech Various General High Less RL2 Fume hood or Yes activities, including ventilation Likely local exhaust waste management, ventilation waste packaging, waste sampling, etc. 26 Purification and Carbon nanotubes Fume hood or High Less RL2 Fume hood or No functionalization of local exhaust Likely local exhaust carbon nanotubes ventilation ventilation

62

27 Purification and Carbon nanotubes Fume hood or High Extremely RL2 Fume hood or No functionalization of local exhaust Unlikely local exhaust carbon nanotubes ventilation ventilation

Out of the 27 additional activities that were characterized, the CB Nanotool recommendation was equivalent to the existing controls for 16 of them, a higher level of control for 8 of them, and a lower level of control for 3 of them. These data suggest that the CB Nanotool produced control recommendations that were generally equal to or in some cases more conservative than the existing controls that were implemented through expert IH judgment. The results were consistent with what the authors hoped to achieve through the tool, which was to develop a consistent approach that would generally err on the safe side, in light of the uncertainty associated with the health effects related to NMs.

Due to the novelty of the CB Nanotool, all the risk assessments with the CB Nanotool that are presented in Table 1 were completed by field IHs in conjunction with the Nanotechnology Safety SME. In reviewing an operation for the first time to collect information required for the risk assessment, the Nanotechnology SME accompanied the field IH in interviewing the workers and touring the work location. A ―Nanomaterial Information Form‖ was developed soon after the CB Nanotool was developed to facilitate consistency in collecting the required information for entry into the CB Nanotool. The data inputs for the specific activities were resolved through face-to-face discussions between the field IH and Nanotechnology Safety SME. It was apparent from this process that while some questions did come up (e.g., is the duration of the activity based on the duration of direct worker interaction with the exposed NM or is it based on the duration of the activity itself, contained or otherwise?), the CB Nanotool was fairly easy to use and thorough instructions were included in LLNL‘s institutional NM safety document. Therefore, taken holistically, the entire process was considered valuable, not just in the utility of the CB Nanotool, but in the logical, cooperative, and educational process behind the creation of the qualitative risk assessment.

The process of implementing the CB Nanotool has been an excellent educational opportunity for the IH as well as the user. Both groups can benefit in that the individual parameters to be considered when scoring a task or procedure during its risk assessment are effective bases for risk communication by the IH and user alike. One application of the CB Nanotool was the evaluation of the grinding and shaping of a NM product where the RL outcome presented a control that was too expensive to implement given the facility‘s limitations. The employee assisting the IH in filling out the CB Nanotool‘s score parameters immediately began considering which of the probability factors he could adjust for the task to lower the RL to the existing controls in place. Although it was possible to consider reducing the amount of NM used during the task, as well as the frequency of the operation, the discussion of risk pointed the IH to offer an alternative reduction factor. The weighting of the dustiness/mistiness was seen as an opportunity to offer a more quantitative evaluation rather then the qualitative one at the time. A condensation nuclei counter (P-trak, TSI, Inc) was used for this operation to determine if any particles in the 20 – 1000 nm range were created in excess of background levels. This selective use of a quantitative measuring device was the most appropriate investment in this task‘s evaluation from the IH‘s point of view and the results determined that the existing controls were indeed appropriate in limiting employee exposures to background levels. More importantly, this standardized language for the discussion of risk between experts and non-experts opened the door for a greater understanding of the potential hazards during this activity and the employees were very grateful to receive this information.

63

Discussion

The CB Nanotool was not intended to be a static tool for a given task or procedure. This can be seen both in the valuation and utility of working with unknown aspects of risk factors as well as the relative values of each component within the CB Nanotool. This begins with an individual, task-specific risk assessment that was designed in a way that allows the user to have the opportunity to revisit their evaluation once more knowledge is obtained on any or all of the components deemed ‗unknown‘ in the initial qualitative evaluation. In the same manner, it is explicitly noted in the original manuscript that the tool itself can be updated in terms of any and all of its individual components as research and process knowledge is further developed. It was for this reason that the CB Nanotool was considered a ‗pilot‘ toolkit in its initial form as the relative importance of each of the implicit factors and their relative values may change as more research on the adverse effects of NM becomes more standardized in publications. Within this discussion, an effort was made to take the latest information available into account and offer opinions on what the next version of the CB Nanotool might look like as part of an overall risk management approach.

Risk management of operations involving nanoparticles primarily is managing exposure scenarios to these particles. Managing means that these scenarios are known and effective barriers are installed to control exposures. Slowly, as information becomes available on exposure scenarios, the quality of barriers and the population at risk will vary, both in production and research facilities (see for instance Bałazy et.al. 2006; Borm et.al. 2008; Paik et.al. 2008; Schulte et.al. 2008). The importance of a sound risk management approach is obvious. It is not only the public‘s attention serving as a driver, but also the necessity to provide adequate protection to workers in laboratories and production environments. The CB Nanotool is an instrument that can facilitate risk management, but there are obvious limitations to the tool. One of them is the relevant factors and the scores of these factors, determining the severity and probability, and hence, overall risk level. These factors and scores refer to the present state-of- the-art in characterizing risks from nanoparticle exposure. However, the relative importance of one factor compared to another may change as more knowledge on the adverse effects of nanoparticles becomes available. Another limitation of CB in general is the inability to address process changes, like automation, elimination of transport routes, etc. These process changes are dominant factors in the observed decline of exposure levels throughout industry (Kromhout and Vermeulen 2000).

Severity Factors: In consideration of the health effects potentially related to NM and the environmental safety and health (ES&H) protocol necessary to perform appropriate risk assessments, the majority of the physicochemical aspects appear to have received a further confirmation (Warheit et al. 2008; Yang et al. 2008). There remains a strong emphasis on the particle surface chemistry, surface area, solubility, particle number, shape, and its biological availability for translocation (Yang et al. 2008; Warheit et al. 2007a). Extrapulmonary translocation varies in degree of toxicological consequence due to differences in chemical composition, particle size, and surface characteristics including surface electrostatic charge on inhalation leading to higher deposition rates (Yang et al. 2008). Additional research has shown when the exposed mass of selected NM of the same size are held constant, it is the structure (e.g. anatase > rutile) that is the toxic differential and that surface area alone is not enough to address pulmonary exposure (Liao et al. 2008; Warheit et al. 2007b). It can be noted that there is no evidence on ingestion as a route of exposure and dermal penetration remains in study 64

(Warheit et al. 2007a). A toxicological characteristic for NM that has been noticed in the recent research, in addition to those already mentioned in the original CB Nanotool publication, is the potential for NM be considered asthmagenic (Hallock et al. 2008; Orthen 2008). As the asthmagenic potential of the PM is also well researched, it was decided that ‗asthmagen‘ should be added to the toxicity scoring of the CB Nanotool for both NM and PM; however, the scoring would be divided amongst these toxicity factors equally and no additional weighting would be given (see Table 2) to the overall weighting for toxicity. In addition, as research on NM as a whole appears to create more questions than it answers, it seemed appropriate that the weighting for PM OELs should be adjusted as follows with two orders of magnitude given between the scoring factors:

< 10 μgm-3 = 10 points 10 μgm-3 – 100 μgm-3 = 5 points 101 μgm-3 - 1 mgm-3 = 2.5 points > 1 mgm-3 = 0 points Unknown = 7.5 points

Other than these moderate changes, the current research studies have served to confirm not only the CB Nanotool‘s risk assessment approach, but also its intent to provide a broad characterization of potential NM toxicological considerations. Warheit et al. (2008) have emphasized that as more and more studies are being conducted, there is further confirmation that enhanced toxicity is found for NM as had been earlier postulated and an increasing variation of physiochemical effects from the growing number of nanoparticulate materials is being observed. This points to an even greater need for standardization for toxicological characterization studies. The CB Nanotool would fit well within the framework of a database that would result from such standardization.

Probability Factors: All of the research presented above confirms the importance given to the CB Nanotool‘s weighting of both dustiness/mistiness and estimated amount of chemical used. The same logic for offering a higher score relating to the NM‘s ability to become airborne has been given even greater emphasis in the more recent publications. The physicochemical focus remains on the biologically available surface area and its ability to translocate systemically. The unique properties of a given NM, inherent in its design and aiding its intended utility, also seem to afford an elevated, persistent, and comprehensive ES&H risk potential. Therefore, the CB Nanotool‘s conservative approach to capture and weight the factors that reflect the probability for a NM to become airborne and persist in the work environment relative to a given task‘s exposure potential appear to remain consistent with the pervasive expert call for a precautionary approach in implementing controls and worker protection (Yang et al. 2008; Warheit et al. 2008; Hallock et al. 2008; Stern and McNeil 2008, Orthen 2008).

65

Table 2. Severity and Probability Factors and Maximum Point Per Factor (NM: Nanomateral; PM: Parent Material)

* Bolded: Pilot Nanotool revisions resulting from this article‘s evaluation.

Validation: A high level of consistency has been found when comparing the CB Nanotool RL outcomes to expert IH recommendations. It can be seen that there is a tendency for the CB Nanotool‘s qualitative risk assessment approach to err towards the conservative at times; however, IH experts also agree that it is better to err toward over-control rather than under- control (Zalk and Nelson 2008). An important question in testing the validity of the toolkit is to ascertain the most appropriate comparison to the tool‘s RL outcomes. For each of the examples given, there is an existing control in place as recommended by an expert IH. In all of the examples given, quantitative measurements were not taken to determine the initial control in place, just professional evaluation and expert judgment. It can be argued that as the CB Nanotool gives a research derived scoring parameter in a comprehensive and structured manner, the broad-based qualitative input is more valuable then an expert‘s judgmental opinion in the absence of an OEL for comparison. If for no other reason, this non-quantitative expert opinion is at best subjective and therefore highly dependent on the IH‘s NM expertise in particular. Perhaps the best comparison is for an IH review of tasks using chemicals that have no OEL, which experts tend to agree is the most appropriate application of CB toolkits (ACGIH 2008; Zalk and Nelson 2008).

Controls: Although the control outcomes themselves do not require adjustment, further research has shown an optimum face velocity for work with dry NM within hoods. The use of a local enclosure within a hood can minimize powder dispersion during handling processes. It should be recommended to avoid higher face velocities when working in hoods with dry powder forms of NM, with an optimum face velocity range of 100 fpm, as some light density NM during transfer operations has been seen to escape at low or high face velocities (Hallock et al. 2008).

66

NM in Industry: Consideration is now being given for a CB Nanotool approach for NM within industry as opposed to R&D. In good part this scaling production volume can also find a counterpart in the pharmaceutical industry which also utilizes an analogous CB approach. For larger scale processes with NM in manufacturing, with such a relatively uncertain toxicological footing, the input factors would require a more appropriate relative value and the control options potentially more robust and that may lead to additional expense. First and foremost, the mass utilized will more likely be orders of magnitude greater than the mass typically used in R&D applications and therefore the primary factor affecting variations in the probability of exposure among the different activities will be dustiness/mistiness. To aid in consistency for the scoring inputs of an industrial CB Nanotool strategy, there should be process specific information that is uniform to manufacturing. As proposed in research, there should be task-based ‗airborne‘ factors derived by industry for standardization (Schneider 2008). The utility of ‗dustiness‘ within a set range is already a uniform application in many CB strategies and exposure models (Tielemans et al. 2008; Zalk and Nelson 2008). In addition, quantitative evaluations of control effectiveness should be considered an essential part of the validation effort. However, perhaps in a manufacturing process there should also be the expectation of Material Safety Data Sheets (MSDSs) for the product used and that the MSDSs would be designed to communicate both NM and PM parameters that could be directly transferred into an industrial CB Nanotool.

MSDS Improvement: The majority of MSDSs for NM, if they are available, provide most of their ES&H information based on the bulk PM. The opportunity for MSDSs to become an integral part of NM risk assessment, exposure prevention, and risk management needs to be addressed. The majority of chemical control CB strategies utilize R-phrases as inputs to the toolkit in order to derive appropriate controls and reduction of work-related exposures (Zalk and Nelson 2008). The majority of NM experts agree that research parameters affording comparisons and sharing of findings is a primary requirement for controlling exposures (Warheit et al. 2008; Yang et al. 2008, Liao et al. 2008). In practice, the toxicological information available on nanoparticulates is minimal and will require deference toward ‗unknown‘ for an individual NM property until this standardization occurs. The real question is not when will this information be put forward, but whether it will be put forth in a consistent manner that will be useful and interpretable for future users of the CB Nanotool. Currently, the research publications that are in circulation seem to be more appropriate for expert dissemination and not necessarily for health and safety professionals in general, let alone managers and technicians. The request for uniformity of descriptive information about NM, captured in a database of set research parameters, should also be listed on MSDSs, which would afford users of the CB Nanotool the latest and most accurate input factors for product appropriate hazard information that would lead to a process specific risk assessment.

Conclusion

The fact that the majority of NM users in industry are not performing even the most basic risk assessment of their product in use, as indicated in the aforementioned recent survey, is unconscionable and must come to an end. CB strategies are known over decades to offer a simplified control of worker exposures when there is an absence of firm toxicological and exposure information and the nanotechnology industry fits this classification perfectly. The overwhelming uncertainties of work-related health risks posed by NM have appropriately led many experts to suggest CB as a solution for these issues. The CB Nanotool was created to fulfill this request. To do this, an expert group of IH professionals, NM and aerosol specialists, 67

and an excellent depth of CB experience have sought to bring this request into the hands of practitioners. As presented, the CB Nanotool has been developed, implemented, and been proven to afford a qualitative risk assessment toward the control of nanoparticle exposures. The international use of the CB Nanotool reflects on its need and its possibilities, but the expansion of its use will assist in ensuring that risk assessments by NM users are both accessible and affordable. While minor changes considered necessary to adapt the CB Nanotool to the growing research publications have been presented, these changes are not expected to result in risk levels that are significantly different from those produced by the CB Nanotool in its original form.

The need for standardization of toxicological parameters has also been emphasized by nanotoxicological researchers. This is to afford better utility and consistency of research with NM as their use and exponential growth in application continue. A standardized database of toxicological research findings should be harnessed and presented in a format, preferably captured in MSDSs, feeding directly into the CB Nanotool severity and probability risk matrix. Making the latest research available for experts and practitioners alike will play an important role in the protection of workers in the nanotechnology industries. This study‘s evaluation of the CB Nanotool, its structure, weighting of risks, utility for exposure mitigation, and improvements place the CB Nanotool in the middle of directing the research still to come, maximizing its effectiveness for all those involved in the nanotechnology industries. It should be recognized that CB toolkits must always be used with some degree of caution. The different factors considered, weighted, and influencing the overall risk levels and control bands are determined as educated ‗guesses‘ as to factor importance and range delineation. Any CB toolkit requires frequent use, validation, and evaluation of recommended control effectiveness. The authors, therefore, strongly encourage an active dialogue within the IH community and further utilization of this or other similar tools for a wide range of applications as these efforts will undoubtedly improve and refine the tool.

Acknowledgement

This manuscript is in part based on a presentation at the OECD Working Party on Manufactured Nanomaterials (WPMN) on exposure assessment and exposure mitigation in Frankfurt, Germany, 21 October 2008 titled ‗Manufactured nanomaterials Control Banding Nanotool, a qualitative risk assessment method: it might be hazardous at the bottom‘. Funding, in part, was provided by US DOE by LLNL Contract (DE-AC52-07NA27344); Lawrence Livermore National Security, LLC. LLNL-JRNL-413240.

68

CHAPTER 5 REVIEW OF QUALITATIVE APPROACHES FOR THE CONSTRUCTION INDUSTRY; DESIGNING A RISK MANAGEMENT TOOLBOX Submitted to The Annals of Occupational Hygiene Zalk DM1, Spee T2, Gillen M3, Lentz TJ3, Garrod A4, Evans P4, Swuste, P5

1) Lawrence Livermore National Laboratory, PO Box 808 L-871, Livermore, CA 94551 USA [email protected] 2) Arbouw, PO Box 213, 3840 AE Harderwijk, the Netherlands [email protected] 3) National Institute for Occupational Safety and Health [email protected], [email protected] 4) Health and Safety Executive, [email protected], [email protected] 5) Delft University of Technology, Safety Science Group, P.O.Box 5015, NL-2600 GA Delft, the Netherlands [email protected]

Abstract

This paper presents the framework and protocol design for a construction industry risk management toolbox. The construction industry needs a comprehensive, systematic approach to assess and control occupational risks. These risks span several professional health and safety disciplines, emphasized by the United States (US) National Institute for Occupational Safety and Health National Occupational Research Agenda mentioning projects for not less than seven topics: falls, electrocution, struck-by hazards, noise, silica, welding fumes, musculoskeletal disorders. Yet, according to the International Social Security Association, ―whereas progress has been made in safety and health, the construction industry is still a high risk sector.‖ Small- and medium-sized enterprises (SMEs) employ about 80% of the world‘s construction workers. In recent years a strategy for qualitative occupational risk assessment and risk management, known as Control Banding (CB) has gained international attention as a simplified approach for reducing work-related risks. CB groups hazards into stratified risk ‗bands‘, identifying commensurate controls to reduce the level of risk and promote worker health and safety. We review these qualitative solutions-based approaches and identify strengths and weaknesses toward the design of a simplified Control Banding systematic ‗toolbox‘ approach for use by SMEs in construction trades. This toolbox design proposal includes international input on multidisciplinary approaches for performing a qualitative risk assessment determining a risk ‗band‘ for a given project. The risk band is used to identify the appropriate level of training to oversee the work, leading to commensurate and appropriate control methods to perform the work safely.

Key words: control banding, construction toolbox, barrier banding, risk level based management system, universal precautions, qualitative risk management, risk assessment

Introduction

The construction industry is serviced by a collection of trades, many of which have attendant hazards, a high risk of injury or illness, and involve working in a changing environment. Despite the existence of recognized and effective solutions and guidance for reducing risks from these hazards, too frequently they are poorly implemented. The consequences of construction hazards can be severe in terms of morbidity and mortality. Analysis of these incidence data calculated an average cost of US$27,000 per incident in construction, almost double the US$15,000 cost per case for all industry (Waehrer et al., 2007). 69

An estimated 7% to 10% of the global workforce works in the construction industry. But the sector accounts for 30% to 40% of occupational fatal accidents worldwide: at least 60,000 per year (Murie, 2007; ILO, 2005a). The risks are similar worldwide, and are in many cases safety- related (Holmes et al., 1999). Falls from heights can cause significant injuries, are often fatal, and the fundamental approach necessary to prevent this accident outcome was described over 3300 years ago [Deuteronomy 22:8]. Even so, the numbers for construction fatalities, injuries, and related costs are generally flat or continue to rise in the US, New Zealand, Taiwan, and The Netherlands (NL) (Ale et al., 2008; BLS, 2007; Bentley, et al. 2006; Chi et al., 2005).

In addition to injury risks, construction workers are also exposed to a variety of health hazards. Potential hazardous substance exposures include:  solvent vapours from glues and paints  acids and alkalis used for cleaning  reactive compounds such as epoxy resins  insulation materials, e.g. mineral wool  ‗natural‘ products e.g. quartz from a stone, concrete or brick cutting, wood dust  fume from heating or burning, e.g. torch cutting, welding, diesel exhaust, bitumen

Many construction tasks also present physical hazards, e.g. noise, vibration, and handling loads. Occupational hearing loss in the construction sector remains significant, even in nations with strong regulations (Nelson et al., 2005). An estimated 30% of construction workers have musculoskeletal disorders (MSD) and back pain, even though basic solutions have been available for 100 years (ILO, 2005a; Weinstein et al., 2007). Industry recognition of health hazards is lower than that for injury hazards. The casual nature of employment in construction is likely to conceal disorders and diseases. For example, MSDs could cause the worker to leave construction, or respiratory disease might not develop until later in life. Complaint data from NL indicate construction workers generally do not complain about hazardous substances, with the exception of a few specific jobs. However, over 50% of all construction laborers complain about dust, apparently without being aware that virtually all construction dust contains hazardous substances like silica and wood (van Thienen and Spee, 2008). Even when implemented nationally, control solutions are rarely known or used widely, despite near- identical hazards.

Construction Industry Needs The construction industry is dominated by SMEs who lack full time safety and health staff. In the US and NL, about 80% of the construction companies are SMEs with fewer than 10 employees (USCB 2004; Waeireer 2007; Geraedts and Wamelink 2008) and in Great Britain (GB) they account for 73% of the industry‘s fatalities (HSE 2008). As is true with general industry, the accident and occupational disease rates are often twice as high among SMEs, compared with large enterprises (Malchaire 2004). The construction industry is highly competitive and work is typically awarded to the lowest bidder. Construction worksites are by their nature temporary, and typically involve multiple contractors and subcontractors each present for only a portion of the project. These and other attributes contribute to hazards and complicate national safety and health enforcement efforts. The end result is that poor occupational safety, health and hygiene (OSHH) protection and enforcement shift disproportionate human and economic costs to the construction worker, their families and communities, (Watterson 2007). Globally, construction employers commonly utilize immigrant 70

workers. These employees typically speak non-native languages, may have low literacy skills resulting in many languages being spoken on a worksite, and may have general communication issues between employers and employees that go beyond language and literacy (Rath, 2002). Small employers often do the same job as their employees and are without time to search for prevention, risk assessment, and control information. As construction industry management is often output-oriented, as long as quality, time, and cost criteria are met, little thought is given to ensure protective measures are used and followed. Often employees decide how the job is done. Therefore, ‗solutions initiatives‘ are best aimed at employee and employer (Watterson, 2007).

It is unrealistic to expect most SME employers to distinguish among separate OSHH fields. Small construction employers have been shown to view OSHH risks as the responsibility of employees instead of something integrated into their company management systems (Holmes, 1995). Few understand accident prevention or detailed hazard awareness, often with controls unavailable or opting for the cheapest control measure (Lingard and Holmes, 2001; Haslem, 2005). Regulatory enforcement of control measure use is weakest with construction SMEs, and nearly non-existent in most countries for accident and MSD prevention (Haslem, 2005; Vedder and Carey, 2005; Watterson, 2007). Effective enforcement as a means of promoting control solutions use requires intense and sustained efforts that is unlikely to occur given limited resources and expertise. Consequently, more effective approaches will involve better mechanisms for reaching SMEs with holistic solutions to industry challenges, rather than a reliance on enforcement and punitive strategies.

Health and Safety Perspectives Construction hazards have received considerable attention over the last two decades. Researchers internationally have examined hazards, consequences, and costs and developed numerous interventions and tailored controls (Thienen and Spee, 2008; LI, 2008; BLS, 2007; CDM, 2007; Waterson, 2007; Flanagan et al., 2006; Haslem et al., 2005). Construction has also grown as a specialty practice area for OSHH professionals. Although injury and fatality rates have become relatively flat over the last two decades for larger firms, injury rates remain high for SME construction firms. Numerous research needs remain and transfer of research from peer-reviewed journals to construction practice, and between countries, has been slow. There is increasing recognition that implementation of evidence-based public health interventions of all types is hampered by the near total absence of systems and infrastructure for marketing and distributing information to end users (Kreuter and Bernhardt, 2009). Many countries are getting more involved with transfer, ranging from developing ―Research to Practice‖ programs in the US, to improved packaging of technical guidance in GB, to industry in NL. Increasingly, research programs are attempting to develop more comprehensive approaches to transfer research findings to practice.

Respirable quartz dust (silica) provides an excellent example of these gaps and challenges. Silicosis in construction has been an issue for decades and preventative methods well known to OSHH professionals (Wagner, 1995). Awareness of silica hazards among SME contractors has lagged behind awareness of injury hazards. Silica as a recognized hazard in NL began getting attention in the early 1990s (Lumens and Spee, 2001). Despite several initiatives taken to reduce silica exposure in construction, inconsistent application of control measures has led to early signs of silicosis in newer construction workers (Onos et al., 2003; Tjoe Nij and Heederik, 2005). Since then, the Dutch government and the sector jointly invested approximately 16 million euro for developing and implementing measures to reduce silica exposure (Staatscourant, 2001). In 2007, however, Dutch Labor Inspectorate inspections showed only 71

30% of the construction companies take measures against dust at high exposure levels (LI, 2008). In NL, therefore, one can conclude that the investment has not led to implementation of more control measures, so far.

Increasingly important is focusing on preventive and control methods for common work-related hazards (Kristenson, 2005). In shifting the focus to 'prevention', it is vital to transfer information comprehensibly, so workers and employers can understand the hazards and risks, how they apply, and how to use the control measures properly (Fingerhut, 2008; NIOSH, 2009). To significantly affect injury and illness rates in the construction industry, a consistent and coordinated message must present a simplified method for ensuring risk assessment, risk prioritization, and workable solutions readily available to workers. Given the similarity of construction hazards and control implementation problems across different countries, a strong case can be made for increased global collaboration and better utilization of limited resources.

Objective

Several initiatives have been presented to overcome the variety of hazards in the construction industry that are multidisciplinary and multinational. The National Institute for Occupational Safety and Health (NIOSH) National Occupational Research Agenda identifies priority construction challenges and related goals for seven specific hazards: falls, electrocution, struck- by hazards, noise, silica, welding fumes, and MSDs (NIOSH 2008). The International Social Security Association has created a construction-based declaration stating ―whereas progress has been made in safety and health, the construction industry is still a high risk sector with respect to accidents and occupational diseases, often resulting in premature death or disability retirement‖ (ISSA, 2009). The declaration resolves to ―all nations‖ that ―massive action must be taken‖ to address this situation and ―the main focus should be risk prevention.‖ NIOSH resource also notes ―construction work is an important example for showing how an application moving directly to exposure controls based on the task performed is the best use of the Control Banding strategy‖ (NIOSH, 2009). The purpose here is exploring possibilities for a multidisciplinary approach addressing these initiatives utilizing international input. The goal of this paper is investigating the feasibility of utilizing CB strategies to develop a toolbox model that addresses risk prevention for the hazards that threat the construction worker.

Methods

This analysis is divided into two parts: (i) overview of solutions-based models to derive multidisciplinary elements necessary for the construction industry and (ii) designing a toolbox framework to develop risk ‗bands‘ for construction projects and identify commensurate control methods to perform comparable work safely. The overview provides a presentation of available solutions-based models are discussed and analyzing them according to their strengths and limitations. Emphasis is given to research findings that can be standardized into usable information products for contractors and workers ensuring solutions generated can be transferred between countries. An original toolbox framework is then designed from both existing and new elements that are necessary and practical for multidisciplinary application based on research for organizing and delivering solution options for construction SMEs.

72

Solutions-Based Models

Solutions-based OSHH research for controlling risk or exposure is currently divided by hazard:  Chemical risks - silica, solvents, welding fumes, asbestos, and lead.  Physical risks - ergonomics risks, noise, vibration, heat, and cold.  Safety risks - working at heights, with energized equipment/machinery, unstable structures.

Various governmental, professional, and industry groups have developed guidance, recommendations, and solutions for use by construction employers. There are a number of ways to organize solutions-based information. We have divided up solution approaches into three models that reflect relative complexity of evaluation and control. This provides a helpful perspective given that the most straightforward low complexity solutions are most likely to be successfully communicated to, and implemented by, SMEs. These three risk approaches are:  Low complexity - Universal Precautions  Medium complexity - CBs, subdivided as - Task-to-Control - Risk–to–Control, and  High complexity - Expert Driven

Universal Precautions Model We describe the least complex approach as ‗universal precautions‘ because it involves using a basic control in every situation involving a particular hazard. The term is commonly used in occupational health to describe the use of hand and face personal protective equipment (PPE) in health care with all patients to avoid contacting bloodborne pathogens and other bodily fluids. An example of a common universal precaution in construction is the use of hardhats and steel- toed shoes by all site workers. These precautions are universal because they are implemented on every worksite regardless of work scope or tasks performed. Universal precautions also include other types of measures higher in the control hierarchy, such as machinery guarding, requirements for ladders, prohibited activities, safe traffic routes, equipment maintenance, residual current devices, and requirements for extracted or emission-suppressed tools. The approach is widely used for safety hazards, both in guidance and regulations.

The US Occupational Safety and Health Administration (OSHA) appear to concur with universal precaution terminology in a ―Quick Card‖ for construction PPE. The single-page Quick Card also covers very basic information on selection, use, and care for prescribed PPE. Construction universal precautions listed are PPE for the: eye, face, foot, hand, head, and hearing protection. Simple guidance from the GB Health and Safety Executive (HSE) expands on this model: ―The absolutely essential health and safety toolkit for the smaller construction contractor‖. This sets out checklists for many of construction hazards, presenting short, simplified, and standardized controls by industry sector.

A universal precautions model represents a simple binary approach: when or wherever the hazard is present the same control approach is used. The simplicity of this approach is a major strength. It is the easiest to communicate to employers and employees. It serves to create a baseline of holistic construction rules that every employee should be trained on and aware of, before entering a construction site. It works well with easily recognized hazards. Disadvantages of universal precautions may arise when this approach is applied to complex hazards because the approach may result in overprotection and perceived burden in some cases and under- 73

protection in others. While this may not be a significant issue for low cost controls such as hard hats, it is more challenging for more costly controls. While universal precaution approaches should always be considered, they may not be suitable for complex hazards ranging in severity based on site-specific factors.

Control Banding Model CB is a medium-complexity approach to evaluation and control of hazards. It involves a structured evaluation of tasks, operations, or work settings. It does not involve quantitative exposure assessment, utilizing operation-specific objective information such as quantities of materials used, exposure properties of substances handled, and nature or duration of tasks. CB originated for use as an alternative approach for controlling chemical exposures. It relies on decision rules derived from prior quantitative studies of various exposure factors. CB allows users to make meaningful inferences about likely exposures and controls needed to reduce them. Thus it represents qualitative risk assessment and risk management approaches. CB groups workplace hazards into four or five stratified risk ‗bands‘, identifying commensurate control measures. Such risk assessment is necessarily generic, so the bands used should apply precautionary assumptions. While OSHH professionals have viewed CB and simplification as a lesser option to quantitative methods, recent application of CB to nanomaterial exposure control has altered that view significantly (Paik et al., 2008; Zalk et al., 2009; NIOSH, 2009). An important application for CB is where uncertainty is high, such as when exposure limits do not exist but substances can be reliably grouped based on similarity to better studied substances. This is also true when tasks vary in nature, setting and duration, as is often the case in the construction industry.

Development and validation of CB has accelerated internationally, resulting in occupational risk management models being built upon CB principles (Zalk and Nelson, 2008; NIOSH 2009; Zalk et al., 2010a). Most CB ‗toolkits‘ are national initiatives to control SME employees' exposures to chemicals, especially substances without exposure limits. CB strategies have also expanded recently to ergonomics and injury prevention (NIOSH, 2009). CB publications have helped increase control use by providing an evidence-based alternative to quantitative exposure assessment. CB efforts have helped emphasize the need to increase control implementation for SMEs. Although researchers and OSHH practitioners have long worked hard to communicate solutions, even through the late 1990s literature was inconsistent in reporting intervention effectiveness and good practice (Roelofs et al., 2003).

Increasing global need has expanded the popular definition of CB as a risk assessment-banding- control model to include ―task-to-control‖ approaches focused on controls needed for specific tasks. Since 2000, national and international collaborations facilitated through the World Health Organisation and the International Labour Organization have been a major driver. These global organisations have provided significant impetus to CB, backing an array of initiatives to prevent work-related injury and illness wherever OSHH expertise is lacking (Fingerhut 2008). US NIOSH highlights construction sector attributes, such as pre-job planning for preventing and managing construction hazards, suggesting that CB approaches could fit well (NIOSH, 2009). Researchers commonly use construction task-based exposure models to evaluate and understand the episodic, highly variable exposures associated with construction activities, and develop work practice and control solutions. While construction is a definitive multidisciplinary activity, such research and solutions tend to not be easily accessible for the worker. OSHH expertise available to SMEs is lacking: meanwhile, regulatory standards and guidance call specific practices and control measures for defined tasks. To date, no set of tools has been developed for construction 74

contractors; although many ―task-based‖ control measures exist and align with CB concepts. To address variability needs, the CB model divides into two sub-categories: task-to-control (T2C) and risk assessment.

Task-to-Control Model The T2C approach organizes control recommendations by task, rather than by level of risk. PPE may also be recommended for those controls not sufficiently reducing exposures. Standardizing tasks afford multidisciplinary perspectives across the OSHH professions. Although standardizing construction industry tasks is a substantial challenge, there are both common tasks and task components that can meet this expectation. In GB, the Control of Substances Hazardous to Health (COSHH) Regulations were not well understood by SME employers who lacked OSHH expertise, prompting requests for simplified approaches to compliance. This led to ‗COSHH essentials‘, a CB toolkit combining hazards of chemicals or products, and potential exposure, to identify appropriate control measures via 'Control Guidance Sheets' (CGS). Strengths of the T2C model include addressing risks, for a variety of hazards, by identifying established standardized control solutions for individual tasks. Limitations include potential for excessive or insufficient control that standard approaches present. This is particularly true for construction tasks with substantial variability and a heavy reliance on PPE rather than the well- established hierarchy of controls.

Silica and Asbestos. In 2005, new evidence for risk to health at the existing GB silica exposure limit, (0.3 mg/m3 as an 8-hour Time Weighted Average) emerged (HSE, 2005). This created needs for task-related guidance describing a precautionary degree of good control practice the new limit at 0.1 mg/m3. This guidance, ‗Silica essentials‘, is identified by industry: here 'COSHH essentials for construction'. ‗Silica essentials‘ is a good example of using 33 construction-related CGS for point-source hazards generated at worksites, with good exposure assessment and controls established. Also, a series of HSE CGS for 'non-licensed' asbestos work applies to lower risk asbestos work in the construction sector. These present the T2C approach in a similar manner as 'Silica essentials', with commensurate approaches to reducing exposures.

This T2C approach to the CB model is effective for SMEs. It gives imperative advice on control measures for defined tasks. It utilizes research findings on exposures associated with tasks along with research on effectiveness of controls. The advice reflects expert consensus, tested by SMEs for usability, acceptability and comprehensibility. Advice is accessed by selecting the appropriate task, such as rock drilling, tile pressing, or abrasive blasting, and downloading the CGS (HSE, 2008b), with ancillary guidance (e.g. appropriate respirator selection; health surveillance for silicosis). Strengths include field application and evaluation, with the ‗Silica essentials‘ CGS being implemented in Southern Africa and Latin America, often adapting the CGS to local conditions and resources. Many have been translated to Spanish and Portuguese (Muianga et al. 2009). In addition, the US Center for Construction Research and Training (CPWR) targets silica control advice directly to workers, in a manner comparable with HSE CGSs. These are in two formats: Construction Solutions Work Practices, and Hazard Alerts (CPWR 2004). Other national studies also identify limitations of relying on respirators alone for silica exposure control in Worksafe Victoria (Australia) and Worksafe BC (Canada). NIOSH has developed a number of task-based ―Workplace Solutions‖ for grinding concrete, tuckpointing, breaking concrete with a jackhammer, and rock drilling.

Ergonomics. Currently the limitation is that there are no true CB toolkits for ergonomics; however, recognizing that a number of approaches and supportive research fit ‗T2C' can be seen 75

as a strength. For example, tiling and plastering involve a significant amount of removing and installing materials, overhead and at floor level, with extensive heavy lifting, twisting and carrying. Construction Solutions Work Practices (developed by CPWR) is a 'Construction Solutions' database and an excellent source of information for controlling MSD risks. CPWR offers control solutions in a quasi-CGS format for carrying heavy materials, stooped postures for floor level work, and stressed hand and wrist activities. CPWR solutions for lifting and carrying include using lightweight concrete materials, tool extension devices for work at floor level, and ergonomically designed hand tools to mitigate wrist and arm MSDs 'NIOSH Simple Solutions' for construction ergonomics offers a solutions-based approach for floor level or overhead lifting, and for handling, and hand-intensive work in the form of Tip Sheets (NIOSH, 2007). Acknowledging that sharing worker-developed ergonomic solutions remains limited, the CPWR Construction Solutions database affords an opportunity for posting to share trade-based ergonomic risk reduction solutions. Simplified solutions to reduce MSDs for work at height and floor level can also be found in Ergonomics Checkpoints (ILO, 1996). These might form an appropriate basis for the development of CB ergonomic toolkits, because such practical approaches aim specifically at SMEs internationally and have been updated and improved (Kogi, 2006; Kogi and Caple, 2008). Task-based ergonomics solutions are effective, but ‗participatory ergonomics‘ has the most sustainable effect. ‗Participatory methods‘ means developing solutions, in collaboration with workers and SME managers, that include construction SME strategies (Zalk, 2001; Hignett et al., 2005; Kogi and Caple, 2008). These methods led to a measured risk reduction of up to 34% for lifting and carrying materials (de Looze, et al. 2001; de Jong and Vink, 2002).

Noise. Noise-induced hearing loss remains common and exposure to high noise levels remains a major issue for construction workers globally (Nelson et al., 2005). Noise and hearing protection belong both to T2C and universal precautions. Although the potential for application of noise reduction and barrier solutions in construction have been around for decades, they are viewed as cost-prohibitive even though there are few data to substantiate this perception (HSE, 1983; Thinksafe, 2007). Both US and GB research indicates that the least expensive and most beneficial noise control practice is to ensure that construction equipment is working and is maintained appropriately (Suter, 2002). However, the construction industry will always have to use hearing protection and its use in US construction worksites is common. But a study of construction workers in their first three years of apprenticeship found measurable hearing loss, even though average noise exposures were measured below 90 dB(A) (Seixas, et al., 2005a). Research has produced substantial information, but correlating elevated noise levels to either trade-based or task-based construction activities is difficult for prioritizing intervention resources (Neitzel, et al., 1999; Seixas, et al., 2003; Seixas, et al., 2005b).

Despite the investment to develop a quantitative task-based solution, University of Washington found that a qualitative evaluation provided a better exposure prediction (Neitzel et al., 2009). Recent research by Neitzel focused on methods to increase the actual use of abundantly available hearing protection by deploying a semi-quantitative noise level indicator with an alert. The device, worn by a construction worker, lit up at elevated noise levels as a reminder to use hearing protection. This approach created an effective, low-cost individual worker‘s T2C that fits noise variability and construction SME needs. Wearing the device for just two months ensured significant increases in workers' hearing protection use another two months later. Combining this approach with HSE control solutions can lead to a CB toolkit.

76

Safety Risks. There is a limitation in that there are no existing CB strategies for 'Safety'. Research is working in this direction, however, and T2C resources to prevent construction injuries continue to appear. A CB-like approach for safety risks, known as Barrier Banding, aims to simplify injury prevention (Zalk, 2006; Zalk, et al., 2010b). Rather than 'control measures', safety identifies 'barriers' such as machinery guarding or fall protection to prevent injury. Such barriers are related strongly to safety management systems. The technique involves check-phrases that describe potential accident scenarios, which guide the user towards appropriate precautions against injury (Swuste, 2007, Zalk, 2006). This comprises:  an innovative look at construction site safety risks  identifying preventative measures necessary to reduce these risks, and  having control measures and barriers in place before work starts.

There are a number of solutions-based initiatives that follow this approach, with downloadable information for both construction workers and employers. For example, NIOSH funded CPWR development of an electronic Library of Construction Occupational Safety and Health (eLCOSH) identifying appropriate risk reduction methods for common construction injuries. In 1997 four stakeholder focus groups began an evaluation of eLCOSH to ensure organization of existing available construction OSHH materials. The website has since grown to 836 documents, with 150 in Spanish. The Construction Solutions database is another internet-based resource from CPWR and NIOSH, referred-to under 'noise', above. It was developed in response to needs of contractors, supervisors and workers in the field who want immediate and accurate information when confronted with problems. The database provides control solutions to construction industry hazards by identifying interventions and control measures demonstrated as effective. Content is created from peer-reviewed literature and other available information. The Construction Solutions database describes T2C measures in practical, end-user terms, organized by hazard and task. The success and value of the database for users will depend on having sufficient coverage for hazards and controls for a given task or trade. It will be launched in phases as various sections are completed. It can support, and encourages, construction user feedback and suggestions for other solutions. This T2C approach for safety solutions is growing with a NIOSH group developing a general Solutions Database for SMEs and other countries working on similar databases (Australia, GB, NL). A task-based prototype for hazards and control measures for "Masonry, Cement, and Plaster" is being tested on focus groups.

Excellent standard implementation guides are offered in Australia through Worksafe Victoria in booklet format. The T2C measures presented for concrete cutting and drilling include a step-by- step approach to address electrical safety, working at heights, barriers necessary for safe work with equipment, and PPE necessary for specific tasks - an expansion of universal precautions. The booklet also provides a checklist for site and equipment safety, and site-specific Job Safety Analysis worksheets. These ensure that prior to starting, comprehensive safety measures are in place, commensurate control measures are implemented, and responsible person identified.

Risk-to-Control Model A limited selection of CB toolkits are available that could help the construction sector. The Risk-to-Control model utilizes a Risk Level (RL) matrix approach, utilizing qualitative risk assessment techniques. Parameters for defined tasks determine the RL. This RL approach assists in ensuring control measures are sufficient to control risks. Strengths of this approach are that controls are stratified commensurate to risk, it follows the hierarchy of control, and it promotes

77

the potential for substitution. The strengths and limitations of CB toolkits have been well described (Zalk and Nelson, 2008; ACGIH, 2008; NIOSH, 2009).

While 'COSHH essentials' could be utilized as a construction RL approach, its applications are limited because the focus is the fixed workplace. 'COSHH essentials' has been evaluated for adequacy of its recommended control for tasks requiring solvent-based components (Tischer et al., 2003; Tischer et al., 2009) and may also be useful for tasks using large quantities of dry materials, such as cement. Recent evaluation of CB against quantitative data sets with solvents and in carpentry-related production show excellent control of exposures below established limits, however a probabilistic model shows CB does not guarantee compliance except for volatile liquids in closed systems and solids with local exhaust ventilation (Tischer et al., 2009).

The premier RL approach is 'Stoffenmanager Construction', developed as a pilot to control silica dust exposures for plasterers and tilers (Zalk and Spee, 2008). Its concept is that risk assessment and advice should be based on quantitative exposure assessments wherever possible and modelled calculation should only be applied when there were no exposure data. Stoffenmanager for control of exposure to silica dust has three risk assessment routes, depending on the data available. The program determines which route to use:  Route 1 uses quantitative exposure data as well as quantitative data about the effectiveness of control measures. It calculates an exposure factor (EF), a reduction factor (RF) and a remaining exposure factor for each task  Route 2 uses a database of about 4000 building materials and products (developed for the Product Group Information System Arbouw) to locate workplace instruction sheets  Route 3 deals with products that are not in the database. As in chemical control version of Stoffenmanager, the assessor must enter data.

Factors. 'Stoffenmanager Construction' uses exposure and reduction factors. This is a useful technique to assess the control requirement for a task, through the degree to which an exposure limit is exceeded. According to the European Standard 689, a situation is adequately controlled if the probability of exceeding the limit is 5% or less. Therefore, the 95th percentile (95th %) value of exposure data sets is used as the measure of exposure for the task.  EF is calculated by dividing the measured exposure by the applicable limit value, whether this limit is health-based or performance-based.  RF shows to what extent a defined control measure is capable of reducing an exposure. It is calculated by dividing the exposure concentration with no control measure by that with a defined control measure. The RF is also calculated as the 95th % with and without control. Where there is more than one control measure, the RFs for each separate measure are multiplied, eg RF LEV x RF water suppression  Divide EF by RF. The remaining value is the EF after control measures have been applied. If the remaining EF is greater than 1, additional control measures are necessary.

In summary, CB utilizes existing researcher data on exposures and control effectiveness. It packages this information into decision rules that allow the selection of controls to be tailored to either specific tasks (T2C) or the most common scenarios and control options (Risk-to-Control or RL Approach). The advantage of CB is that it offers flexibility to accommodate workplace variation and increases likelihood that appropriate controls are used. CB can be considered more complex for employers to use than T2C, therefore reliability can increase when employers have had some basic CB training (Zalk et al., 2010a). It lends itself to a ―competent person‖ 78

approach, where a trained individual employed by the contractor has the knowledge and authority to implement controls. Disadvantages of CB approaches are the lack of quantitative exposure assessment and direct involvement of an OSHH professional. Therefore, overexposures could occur if controls are not implemented correctly or maintained over time.

Expert-Driven Model The expert model approach represents the traditional approach whereby an employer utilizes an OSHH professional to evaluate site specific hazards, recommend controls, and manage a safety and health program or system to insure effective implementation. This conventional approach is commonly seen as the 'gold standard'. Strengths are obvious as expert derived prescriptive solutions effectively follow the established hierarchy of controls and are the standard of OSHH professions. This approach provides the highest confidence that hazards are properly evaluated and controlled, especially for highly complex hazards. OSHH professionals are expected to know about research describing exposures and control effectiveness from professional journals and learning about developments via professional conferences. OSHH professional employment by construction employers also helps to institutionalize and integrate OSHH at the organization level and into other key business areas such as design and procurement. The major limitation of this approach is that most SMEs do not employ OSHH professionals as they are expensive and in many parts of the world are rare or non-existent. OSHH professionals also tend to specialize, thus requiring multiple consultants. While the expert-driven model will always be critical for large employers and complex hazards, limitations of this approach have fuelled interest by governments and international organizations in developing CB approaches that provide meaningful alternatives for SMEs.

Framework for a Construction Toolbox Model

Banding of Risk Over the last few years CB has expanded into wider OSHH fields, beyond substances and chemical products. These are termed 'toolkits': a 'toolbox' may contain several toolkits (NIOSH, 2009). Toolkits use the CB approach of ‗banding‘ for assigning the risk of a given hazard to one of several - typically four - levels. Expert models use quantitative risk assessment methods to stratify the risk and define the boundaries between strata. They use data to assign the efficacy of control measures into bands. International collaborations are developing toolkits as 'preventive' measures for OSHH experts and non-experts alike. Toolkits should guide non-professionals, and inform professionals. International CB workshops have promoted CB expansion, research, and publications. They have been a focal point for original applications of simplified risk banding and control approaches for silica exposure, ergonomics, and safety (Zalk and Nelson 2008, NIOSH 2009) and have led to this international collaborative opportunity. Construction industry risk banding requires identifying the right safety solutions across OSHH disciplines to achieve injury and illness prevention. Assigning a band to a project, just as for tasks, is a practical approach to identify and reduce risks relating to work-related accidents and disease.

Construction stakeholders under NIOSH assembled a National Construction Agenda. The agenda provides a framework upon which to develop and promote a construction toolbox, including intermediate goals that build upon task-based control measures, and moves towards pre-job planning, awareness training, competent person training, and CB use. However, interest of US researchers and practitioners in CB construction applications has not led to the development of a toolbox for use by ‗duty holders‘ – employers, designers and contractors. A variety of applications, generally falling under the rubric of ―task-based‖ approaches or toolkits, 79

align with CB risk banding concepts. For example, construction researchers recognized early on that ―task-based sampling‖ was most appropriate for understanding and controlling exposures with construction activities (Flanagan, et al., 2006; Flynn and Susi, 2003; Croteau, et al., 2002). Multiple opportunities to utilize risk-based CB strategies in the construction industry exist.

Examples. Two major OSHA health standards for construction –for lead (1926.62) and asbestos (1926.1101) - include task-based alternatives that reinforce regulatory applicability.  The lead standard required that specific precautions be used for identified tasks, ranging from work involving lead based paints (e.g., manual scraping, manual sanding, heat gun applications) to welding, cutting, and torch burning (1926.62(d)(2)).  The asbestos standard created four classes of work, delineated by risk, with tailored specific precautions for each. It incorporates provisions to treat suspect material as ―presumed asbestos-containing materials‖ as an alternative to bulk analysis testing.

Preliminary versions of an OSHA proposal for silica in construction also included an option for employers to follow specified controls for 8 listed tasks as an alternative to exposure assessment and competent person provisions. Each standard utilizes the concept of a ―competent person‖ capable of identifying hazards, selecting the appropriate control strategy, and with authority to take prompt remedial measures. A GB asbestos decision tool leads to comparable banding decisions – whether or not the job is licensed. If licensed, the full provisions of the Control of Asbestos Regulations 2006 apply. If 'not licensed', there are CGS for common, low risk tasks.

Level of Risk Safety science has qualitative and semi-quantitative tools to assess risks (Zwaard and Passchier, 1995). Risk is seen as a numeric variable. Simplistically, it combines an adverse effect (or its severity) with a probability of that consequence. Probabilities and consequences are classified into groups and provided with a value. Simple multiplication of values for probability and consequence produce a risk score, used to compare one risk with another for prioritization. An exposure/frequency estimate of hazard and probability of specified consequences for scenarios are compared (Zwaard and Goossens, 1997). Other variations of these tools can include the numbers exposed or the degree to which risk can be controlled via ‗relative ranking‘. The RL model most often used is divided into four categories, with determination by frequency and severity of hazard. This RL approach has been used recently for an occupational hygiene qualitative risk assessment to control nanoparticle material exposure (Zalk et al. 2009) and as part of a Risk Level Based Management System (RLBMS). The RLBMS is a risk-based occupational risk management model, designed to focus OSHH resources on the highest workplace risks. This model is also auditable, so it fits OSHH management systems and national regulatory oversight (Zalk, et al., 2010a).

The RLBMS model is an appropriate framework for bringing together the control-focused research and solutions-based approaches (Zalk et al., 2010a). RLBMS has shown consistency in integrating multiple solutions across the OSHH professions utilizing an RL delineation of hazards and commensurate controls. Strengths include the use of qualitative risk assessment to achieve regulatory compliance and standardizing a wide variety of tasks. Trades and tasks of the construction industry are more limited, especially for SMEs. RLBMS and solutions-based toolkits offer simple identification and control measures that can fit into a single toolbox. It requires identifying OSHH risks and risk reduction steps for each project phase. The RLBMS approach provides a step-by-step mechanism for creating this toolbox. It takes control solutions

80

from research and good practice, organized within simple project-based and task-based formats. This simple format requires a toolkit structure layered by level of complexity of control solutions. Also necessary is addressing the National Construction Agenda (stakeholders' consensus) to secure pre-job planning to ensure the appropriate level of worksite supervisor training to identify controls matching task-related OSHH risks. This step requires a project- specific RL. Therefore, within this framework design proposal we first provide a method to obtain a project specific RL and then present the layered toolkit structure by control levels (CL).

Pre-Job Hazard Analysis (PJHA) In addressing the need for a project specific RL, Table 1 presents a concept for a PJHA checklist establishing a project-specific band by RL. The table is based on the research-derived hazards inherent to construction trades. Its purpose is to score the hazards within a given construction project. The 'severity checklist' contains common chemical, physical and safety hazards determined independently from specific tasks. The ‗probability checklist‘ determines 'exposure': the project scale, duration, number of workers, overall dustiness, and overall ‗potential energy‘ estimate for project completion. This estimate is based on the concept: higher inherent potential energy yield higher adverse outcome risk from its release. Silica exposures are higher from grinding versus manual breaking; metal fume exposures are higher from torch cutting than mechanical cutting. More energy can lead to more noise; more energy is used in manual handling then with dollies, than utilizing machinery. Higher energy equipment has more severe consequences in the event of electric shock, or heavy equipment failure. Work at height or in trenches implies a potential energy released through a fall. Responses to 'severity' and 'probability' should avoid consideration of control measures.

81

Table 1: Pre-Job Hazard Analysis Checklist Concept

SEVERITY – Answer reflects hazard potential without controls in place (Yes = 4 points each) Chemical Is there a possibility that asbestos-containing materials will be encountered? Yes ☐ No ☐ Is there a possibility that lead-containing materials will be encountered? Yes ☐ No ☐ Will there be jack hammering, roto-hammering or similar concrete work Yes ☐ No ☐ Will there be breaking or cutting of tiles, masonry or other silica dust work Yes ☐ No ☐ Will the job involve welding, soldering, or torch cutting? Yes ☐ No ☐ Will there be engines running on the worksite? Yes ☐ No ☐ Will work involve chemicals, solvents, painting, brazing or grit blasting? Yes ☐ No ☐ Will work be within vaults, manholes, trenches, or tanks >4 feet deep? Yes ☐ No ☐ Will the workers require personal protective clothing? Yes ☐ No ☐ Will the job involve materials or processes requiring respiratory protection? Yes ☐ No ☐

Physical Is there a potential for manual material-handling of items over 40 pounds? Yes ☐ No ☐ Is there a potential for repetitive tasks for more than 30 minutes a workshift? Yes ☐ No ☐ Is there a potential for repetitive transfer of materials less than 40 pounds? Yes ☐ No ☐ Will workers be exposed to elevated noise levels on this job? Yes ☐ No ☐ Is there high (e.g. jackhammer) or low vibration (e.g. manned cab vehicles) activity? Yes ☐ No ☐

Safety Will work be performed on or near energized equipment, lines, or circuits? Yes ☐ No ☐ Will there be overhead power lines or potential underground or hidden utilities? Yes ☐ No ☐ Will workers be working above 6 feet from ground level? Yes ☐ No ☐ Will scaffolding or ladders be used and worker access be provided? Yes ☐ No ☐ Will there be work cutting, grinding, or breaking of concrete or masonry? Yes ☐ No ☐ Will the job involve steel and/or scaffolding erection? Yes ☐ No ☐ Will floor, wall, and/or roof openings be created during this job? Yes ☐ No ☐ Will crane(s), forklift(s), manlift(s), or other lifting equipment be used? Yes ☐ No ☐ Will there be excavation or trenching in excess of 4 feet? Yes ☐ No ☐ Will the subcontractor be using motor vehicles or heavy equipment on-site? Yes ☐ No ☐

PROBABILITY (Highest = 20 pts, Mid = 10 pts, Lower = 5 pts, Lowest = 0 pts) Number of workers on the jobsite? More than 10 ☐ 6 - 10 ☐ 3 - 5 ☐ 1 - 2 ☐ Length of project in 8-hour days? 10 or more ☐ 6 - 9 ☐ 2 - 5 ☐ 1 ☐ Dustiness on the jobsite (no controls in place). High (clouds) ☐ Moderate (visible) ☐ Low (puff) ☐ None ☐ Fuel-, electrical-, & manual- based energy on jobsite? High ☐ Moderate ☐ Low ☐ None ☐ Will the work be performed indoors? Mostly ☐ Sometimes ☐ Rarely ☐ None ☐

PJHA scores for severity and probability are each individually added. Each score is then found, within a given range, based on a 4 X 4 matrix (Figure 1). Their intersection determines the RL for the construction project and is the first step in this Construction Toolbox design. The RL identifies the level of training and expertise to match the inherent project‘s risk at the earliest 82

stage (Holmes et al.,1999). The RL also indicates the degree of expertise needed for project pre- planning. The next step for the project lead is to identify each of the tasks performed on the jobsite. Securing the appropriate control measures per task is performed using the solutions- based models identified above prior to work commencement. The PJHA RL outcome assists in identifying when control measures for specific tasks require higher expertise then the project lead possesses. As the RL is determined independently from tasks, higher expertise must be obtained for the task or a substitution for a lower risk task approach can be made.

Figure 1. Risk level (RL) Matrix for with the PJHA to determine jobsite training requirements.

Probability Score

Extremely Less Likely Likely Probable

Unlikely (26-50) (51-75) (76-100) (0-25) Very

High RL 3 RL 3 RL 4 RL 4 Severity score (76-100) High (51-75) RL 2 RL 2 RL 3 RL 4 Medium (26-50) RL 1 RL 1 RL 2 RL 3 Low (0-25) RL 1 RL 1 RL 1 RL 2

Jobsite training requirements by risk level: RL 4: Expert required on jobsite RL 3: Competent person required on jobsite RL 2: Hazard awareness expertise required on jobsite RL 1: Basic craft skills sufficient on jobsite

Training and Expertise The 'toolbox approach' takes into account the ‗mentoring relationship‘ that is (or should be) present in construction trades. An inexperienced apprentice is teamed with an experienced craftsman as mentor, to develop the skills - the knowledge base - of the craft. This process can ensure both competence and an understanding of the craft's OSHH dimensions. Skilled workers are differentiated from apprentices by ‗card-carrying‘ status, certifying ‗skill-of-the-craft‘. Many European countries and US states require, or are working to require, card-carrying construction trade workers on jobsites, ensuring they possess the skill-of-the-craft to perform their work to codes and regulations. In GB, the Sector Skills Council promotes OSHH skills for construction employees. Defining ‗levels of training‘ in the Construction Toolbox meets this concept. An apprentice may attain ‗Basic Craft Skills‘ with a minimum of training under a mentor, working to skill-based ‗Hazard Awareness‘ by trade at the card-carrying level. In the US, OSHA regulations require higher level training to become ‗Competent Persons‘, to oversee specific activities with inherently higher risk, fitting regulatory requirements discussed above. RL4 work requires ‗Expert Training‘, to assist in understanding and controlling multiple risks.

83

The key boundary is between RL2 and RL3 - workers will need to be able recognize:  when work is, and is not, within their skill-of-the-craft and  when they meet the ‗Assistance Required‘ band; needing a certified Competent Person to verify that appropriate control measures are evaluated in place.

Control Level (CL) With a project specific RL determined, the final step is to assist the jobsite manager in accessing control solutions utilizing a layered toolkit structure. Each layer reflects a CL in this manner:  CL1 is the universal precautions model. This level deals with the general hazards at the workplace such as tripping and falling as well as head, foot, hand, ear, and body protection. These hazards apply worldwide to everybody at the construction site and everybody at the construction site must have knowledge of these and follow worksite controls identified.  CL 2 is the T2C model. This level deals with general hazards with basic projects in construction and demolition and include simpler tasks involving hazards such as silica, asbestos, noise, awkward postures, and heavy lifting as well as electrical, machinery, and work at height risks.  CL 3 is the CB Model. This level requires the collection of simple input parameters to use existing online toolkits such as COSHH Essentials and Stoffenmanager Construction. At this level the construction company can assess hazards and get guidance to control for specific situations.  CL 4 is specialist advice. This level is for complicated and/or very hazardous situations such as on large asbestos projects, major construction at heights, and extensive concrete demolition, or work involving biohazards, radioactivity, carcinogens, or a vast array of highly hazardous chemicals.

Planning the project prior to initiating work will require the determination of appropriate CLs for each standardized task. Task-related control measures are selected from available options at the specified CL. As indicated above, these are research and solutions-based resources. The advantage of such a layered model is that CL1 and probably CL2 can be filled on a worldwide uniform scale. Knowledge, instruments, factsheets and checklists are all available at these levels. When tasks are at higher levels of risk, such control measures normally need tailoring to project tasks. The CL helps identify the degree of tailoring and type of help necessary for complex situations. Designations CL3 and CL4 mean that workers need help with control measures from a Competent Person or OSHH specialist. CL3 and CL4 are intentionally aligned in this manner to ensure the provision of effective control measures, and their sustained and correct use.

Discussion

The review and analysis of the solutions-based models presents a wealth of potential tools available across the construction industry. The toolbox framework design presented delivers a banding of project risk that also accomplishes pre-project planning. The primary limitation of the Construction Toolbox is it not being field-tested. The PJHA checklist (Table 1) is a proposal based on expert advice and may require refinement. The scores, and expansion or reduction of the checklist need consideration on national, regional, and cultural bases. The Construction Toolbox will, of course, need evaluation and validation, as a whole and of its parts. The critical issue is the effectiveness and sustainability of control implementation. In theory, the Construction Toolbox approach is well placed to assist the development of mentor-apprentice 84

relationships through standardized OSHH information for construction trades. Also in theory, it would be a tool for appropriate, consistent training of qualified workers, competent persons, the development of OSHH experts in construction, in risk control, and risk management.

In practice, the endpoint of this Construction Toolbox model cannot simply be the identification of an RL, CLs, the production of control advice, and the barriers and control measures in place. Rather, Construction Toolbox use should continue through the project, to assist adaptation as the worksite changes. The correct level of expertise can be identified and risks controlled at all stages. The more common and easily controlled hazards are prioritized as such, represented in the lower RL categories (RL1 and RL2). The aim is to resolve the lower risks simply, and focus upon tasks with more severe outcomes. Therefore, a tool for managers' systematic approach to barriers and controls focuses resources on higher RL events and CL requirements. Focusing expert OSHH professional availability, management-related ‗culpability‘ in accident scenarios can also receive appropriate attention. CB is not - yet - a banding of risk management. Good risk management means reducing high frequency, more severe outcomes from construction hazards.

An important point for the use of CB toolkits is the potential to identify the appropriate control measures in the absence of expertise. At a training level, simplification and uniformity reinforce retention, implementation and the sustainability of prevention. The design of the Construction Toolbox also affords the opportunity to consider these OSHH prevention concepts at the planning, design, and engineering stages of construction projects (Howard 2008). Such ‗prevention through design‘ approaches are now available in many countries (CDM 2007, Creaser 2008, Howard 2008, Schulte et al 2008). Some hazards are simply not anticipated. Unnecessary risks may not appear until workers encounter them during the construction process. Therefore, additional risk prevention methods can be found within NIOSH supported research to gather case studies and to provide a conceptual framework for addressing safety and health at the project design phase (Schulte et al. 2008). Banding a project‘s RL in pre-project planning therefore offers a complementary prevention through design within the Construction Toolbox framework.

Requirements. The availability and usefulness of a Construction Toolbox will require:  a web-based format for standardized tasks and related CLs that is continually updated,  free and readily accessible control guidance that is acceptable, comprehensible and usable,  free and readily available web access to existing solutions-based resources, and  a web-based format to share successes and lessons learned by both task and trade.

Sharing control solutions and lessons learned, e.g. post on website, require CGSs that include:  checklists for control measures and barriers, and  a 'Work-site Hazard Analysis' worksheet.

These require testing of formats to optimise, for users, ‗drill-down‘ for access to control advice based on tasks that they recognise. Tasks should include variables such as those in Table 1, e.g. duration, other workers' proximity, the variety of controls that might be available. ‗Drill-downs‘ produce CGSs that include task summary and generic control measures and/or barriers, or specific CGSs as necessary for tasks. Web-based ‗drill-down‘ actively linked to the Construction Toolbox have potential for currency, updating and real-time translations. This would enhance the risk management aspect, communicate hazard-to-control to the worker, and 85

offer field-based advice to others in a participatory format. The PJHA can also be similarly useful. For SMEs, each question can be linked to a brief explanation and additional resources links as they determine the CL, task-by-task. The SME manager can begin to consider all opportunities for hazard or task substitution, and selecting task parameters to reduce the overall expertise required. Users and experts can score online resources such as CGSs and CL-based solutions for utility, simplicity, and effectiveness as feedback, for evaluation. SMEs can also get Construction Toolboxes on CD by trade, having ‗drill-down‘ format, but no web-based benefits.

In seeking to address these issues, the intent is to remain global in scope. However, many of the current control solutions are skewed toward a few countries. There are often national research gaps, so international research and literature requires locating and consideration. In economically developing countries the number of OSHH professionals and technicians need substantial growth. The Construction Toolbox can become a foundation for training OSHH experts and technicians in construction in university programs, and spreading expertise through train-the-trainer campaigns. Developed countries can also use such campaigns, with exchange of successful control information where language barriers exist. The process can work best internationally when key aspects of the CL approaches in the CGS are communicated largely through pictorial formats, to minimise the need for translation and standardizing the control expectations. For example, the HSE is developing a short series of simple icons as an ‗employee checklist‘. But this whole context is reliant on a Construction Toolbox existence, acceptance, dissemination, and use. Therefore, the authors would advocate that ILO or WHO set up an international working group to collect and order existing information and to make this readily available (e.g., internet or booklets). In this manner, as experts are lacking in many developing countries, an initial goal to 'pick the low-hanging fruit' would be a solid step in right direction. If all countries implemented simple, practical strategies to prevent accidents, it would be possible to eliminate 83% of safety-related deaths and 74% of accidents annually (ILO, 2005b).

Conclusion

A constant throughout this discussion is recognition that workers in the construction industry are involved in a dangerous trade. Construction work-related risks are well understood, but it remains a leader for raised injury, illness and fatality rates; and associated costs to business, society and families. It is unconscionable that construction remains hazardous, while resources over decades - statistics, causal factors, and control measures to reduce risk - are known to OSHH professionals. Construction Toolbox development seeks to change the perception that work-related risks in this sector are just safety-related and inevitable; rather it emphasizes that chemical and physical exposures abound and are preventable. The concept is a comprehensive tool for the construction industry, to assess and control occupational risks that currently are segmented between OSHH professions. The Construction Toolbox presents a format to harness multiple solutions-based national programs and publications for controlling construction-related risks across the OSHH professions. CB and Barrier Banding have been united in this RLBMS format using simplified risk assessment and risk management strategies. Multiple OSHH professional expertise unite in this framework to organize, communicate, and implement risk reduction programs at a construction jobsite. Our intent was to propose a simplified risk banding approach for the SMEs that employ over 80% of the world‘s construction workers.

This occupational risk management toolbox strategy:

86

 uses a qualitative risk assessment approach to determine an RL ‗band‘ for a given project,  identifies the appropriate level of training to oversee the work, and  enables identification and implementation of suitable control measures for safe working.

Such a strategy has long been sought, but never presented in this format. Critical elements of the work remain to be done on the research side are:  the validation and verification of this toolbox approach,  implementation and further evaluation of the PJHA checklist,  practicality of integrating enforcement and national regulatory compliance,  a movement back toward solutions-based practical, field research, and  multidisciplinary collaborations at local, national, and international levels.

To appropriately develop the Construction Toolbox components, the needs in practice are:  development of centralized databases at the T2C level expanded across the professions,  creation of training packages for SMEs and train-the-trainer packages for experts,  scaling to the Construction Toolbox model to economically developing countries, and  sharing of successes and limitations as well as field-based feedback and improvement.

Further, the creation of the necessary centralized web-based system to unify an international implementation will require funding, development, and maintenance ensuring the Construction Toolbox availability to the world‘s construction SMEs. Delivering such an essential product to SME managers is the first, but most important, step. However, without further analysis and field implementation it will remain just another publication seeking to reduce the constant stream of work-related injuries and illnesses, an abhorrent mark on the construction industry worldwide.

Construction Solutions-Based Internet Resources HSE Construction Silica Essentials OSHA Construction Industry assistance University of Washington silica solutions University of Washington noise hearing loss, construction NIOSH Construction Health and Safety solutions NIOSH Respirator resource The Center for Construction Research and Training (CPWR) CPWR Simple Ergonomic Solutions for Construction CPWR Construction Solutions eLCOSH Worksafe British Columbia, Canada Worksafe Victoria, Australia

Acknowledgements Tamara Onos for her efforts in support of Stoffenmanager Construction. This work, in part, was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344, LLNL-JRNL-461830.

87

CHAPTER 6 RISK LEVEL BASED MANAGEMENT SYSTEM: A CONTROL BANDING MODEL FOR OCCUPATIONAL HEALTH AND SAFETY RISK MANAGEMENT IN A HIGHLY REGULATED ENVIRONMENT Industrial Health, 48(1): 18 – 28 (2010) Zalk DM1, Kamerzell R1, Paik S1, Kapp J1, Harrington D2, Swuste P3

1) Lawrence Livermore National Laboratory, PO Box 808, L-871, Livermore, CA 94551 USA Corresponding email: [email protected] 2) Consolidated Safety Services, NASA, Ames Research Center, Moffett Field, CA 94035 USA 3) Delft University of Technology, Safety Science Group, P.O.Box 5015, NL-2600 GA Delft, NL

Abstract

The Risk Level Based Management System (RLBMS) is an occupational risk management (ORM) model that focuses occupational safety, hygiene, and health (OSHH) resources on the highest risk procedures at work. This article demonstrates the model‘s simplicity through an implementation within a heavily regulated research institution. The model utilizes control banding strategies with a stratification of four risk levels (RLs) for many commonly performed maintenance and support activities, characterizing risk consistently for comparable tasks. RLBMS creates an auditable tracking of activities, maximizes OSHH professional field time, and standardizes documentation and control commensurate to a given task‘s RL. Validation of RLs and their exposure control effectiveness is collected in a traditional quantitative collection regime for regulatory auditing. However, qualitative risk assessment methods are also used within this validation process. Participatory approaches are used throughout the RLBMS process. Workers are involved in all phases of building, maintaining, and improving this model. This worker participation also improves the implementation of established controls.

Key Words: Control banding, qualitative risk assessment, occupational risk management, occupational health and safety management system, risk level approach, toolbox, participatory.

Introduction

Lawrence Livermore National Laboratory (LLNL) is a Department and Energy (DOE) and National Nuclear Security Administration (NNSA) research and development (R&D) facility that is operated by the Lawrence Livermore National Security (LLNS) LLC. What makes LLNL unique from a majority of organizations with over 6000 employees is its primary R&D focus, rather than a traditional production or manufacturing working environment. Most large-scale enterprises worldwide with ORM models within occupational health and safety management systems (OHSMS) are focused on a finite number of uniform activities.

LLNL‘s R&D focus breaks away from traditional systems by consistently performing unique work. Therefore, the creation of new and potentially hazardous operations and related occupational exposures is standard. Perhaps an expected outcome of this large-scale R&D work is a pervasive regulatory oversight. Though based in California, the California Occupational Safety and Health Administration (OSHA) does not oversee LLNL‘s activities. Instead, Federal OSHA (FedOSHA) has jurisdiction. In addition, as LLNL is run by LLNS for the DOE, additional contractual requirements are in place to ensure workers performing the tasks that 88

benefit national R&D are well protected. For industrial hygiene (IH), the contract requires the lowest established occupational exposure limit (OEL) for a given chemical, physical, or biological exposure. Therefore, the ACGIH® Threshold Limit Values (TLVs) often meet this specification and therefore are treated as regulation. DOE has also developed and enforced standards published in the Code of Federal Regulations for LLNL compliance. To say that LLNL has substantial regulatory oversight is a quintessential understatement.

As an outcome of this regulatory oversight, LLNL has a consistent stream of auditors, both internal and external, offering subject matter expertise on the quality of adherence to requirements. A common by-product of this process is a steady stream of policy implementation and deadlines. To the Environmental Safety and Health (ES&H) multidisciplinary teams assisting clientele in working safely, this regulatory oversight process is often seen as a steady stream of increasing paperwork. Therefore, this escalating paperwork often reduces team member opportunities to get into a field practitioner mode. It is within this working environment that the RLBMS concept was born.

Methods

Building the RLBMS; Managing Risk Although ORM is seen as responsible for reducing injury and illness statistics, the quality of occupational risk perspective should be the primary consideration (Swuste 2008). Numerous OHSMSs are available for managerial use and implementation relating to occupational, safety, hygiene, and health (OSHH) professions. Since their growth in the 1990s, they continue to gain importance in measuring performance and OSHH management success (Redinger et al., 2002). The three main components of an OHSMS common worldwide are establishing input parameters, risk assessment and control, and management and evaluation (Redinger et al., 2002; Levine and Dyjacka 1997; Redinger and Levine 1997; Machida 2001; Kogi 2002). Taken together, they comprise an ORM basis for fundamental and comprehensive OSHH regulatory adherence. However, not all OHSMSs are equal as some measure managerial systems whereas others emphasize the work environment where implemented (Kogi 2002).

ORM programs focusing on risk factors are considered the most essential approach for reducing the economic and sociological burden of work-related illness and injury (Fingerhut et al., 2005). Programs that manage risk appropriately balance all aspects of risk affecting a company‘s operation. Thus, risk is often defined beyond its occupational application, to include economic viability and product quality risk-benefit analyses. These OHSMS approaches are risk-based; however, the risk is not necessarily a graded management decision making so much as competing risk factors. Here, risk-based solutions are in a decision-making and problem-solving context seeking cost effectiveness. Review of ORM principles and national OHSMS approaches reveals worker-based success and measurable reduction of occupational risk factors are often seen when internal, adaptive risk assessment procedures and controls are developed (Kogi 2002). Multidisciplinary research also finds that emphasizing individual task-based controls by risk, rather than on control technology, achieves effective ORM across the OSHH professions (Lingard and Holmes 2001).

Banding Risk In recent years a qualitative occupational risk assessment strategy known as control banding (CB) has gained international attention in its goal of offering a complementary and simplified approach to reduce work-related injury and illness (Zalk and Nelson 2008). CB‘s simplified 89

approach to grouping workplace hazards into stratified risk ―bands‖ based on common hazards and commensurate control approaches offer long-sought unification across OSHH professions (Fingerhut 2008). The term has an IH focus and represents a qualitative instrument to assess risks for chemical substances, generating solutions and implementing control measures (Russell et al., 1998). However, the concept of segregating risk into a simple and discreet order has its roots in occupational safety. A number of qualitative safety risk concepts began in the 1970‘s focusing on stratification, or banding, of central events in a risk matrix describing the likelihood and severity of an explosion, or toxic material release, for use by major chemical companies. In the 1980‘s this risk assessment approach expanded to radiation, lasers, biosafety, and eventually pharmaceuticals in the 1990‘s (Zalk and Nelson 2008; Naumann et al., 1996; Sargent and Kirk 1998). Modern CB approaches utilize simplified strategies directing users to control solutions and control guidance sheets for chemical exposure, or seeking specialist advice for the highest risks (Zalk and Nelson 2008).

Recent CB expansion of range, beyond bulk chemicals and into OSHH professions, uses the basic stratification of practical prevention strategies as earlier risk matrices. This includes barrier banding, a strategy utilizing CB concepts for occupational safety rather than IH (Swuste 2004). Barrier banding explores the practicality of addressing safety accident scenarios, implementing barriers, and managing solutions in a simplified manner to achieve injury reduction (Zalk 2006). CB in ergonomics offers comparable approaches for controlling musculoskeletal disorders within participatory programs (Kogi and Caple 2008; Zalk 2001). Commonality of form and function, toward identifying and reducing OSHH risks and controlling and reducing injuries and illnesses, has built international demand for simplified approaches to banding risk (Fingerhut 2008).

Risk Level Approach Safety sciences have few qualitative tools to assess risks (Zwaard and Passchier 1995). Risk is seen as a numeric variable, as its most simple format equates the combination of unwanted consequences and the probability for their occurrence. Probabilities and consequences are normally divided into groups or classes and provided with a value. By multiplying these values, a risk score is created and used to compare different risks. Because valuing probabilities can be difficult, variations are present which split these probabilities. First a hazard exposure frequency is estimated, then the probability of scenarios occurring is specified for consequences (Zwaard and Goossens 1997). Other tool variations can incorporate the number of people exposed and level of turning away risk. The general term of these tools is ‗relative ranking‘, or ‗rapid ranking‘. Qualitative risk assessment approaches stratifying multidisciplinary risk levels (RLs) must emphasize practicality for OSHH professional acceptance. This simplified transparency was found within publications behind the creation of the Control of Substances Hazardous to Health (COSHH) Essentials CB toolkit. Here, Brooke outlined three criteria for the toxicological basis of this approach: (1) simple and transparent, (2) make best use of available hazard information, and (3) recommend control strategies that vary according to degree of health hazard (Brooke 1998). In writing about the development of the COSHH Essentials model, Maidment stressed the importance of limiting the number of factors in the model to control its complexity and applicability (Maidment 1998). These simplified bands of risk, or RLs, link it with a commensurate control system and risk matrix harkening back to its safety origins. The RL approach at LLNL bands risk across four levels to obtain simplicity (Figure 1). This RL matrix approach has successfully been put into practice at LLNL in a qualitative risk assessment program to prevent nanoparticulate exposure (Paik et al., 2008). Reaching across the OSHH disciplines, this matrix is also the basis for barrier banding23 and a burgeoning multidisciplinary 90

approach for an ORM construction toolbox (Fingerhut 2008; Zalk et al. 2010b). This RL matrix has also been determined to be best suited for the unique needs of LLNL‘s workforce. It provides versatility in addressing both complex issues like nanomaterials and multidisciplinary approaches like with construction. The RL matrix also calibrates LLNL ES&H Team risk perceptions, reducing among-discipline variability or, statistically put, Type II error, when the assumption is that disciplines will respond the same, but in reality do not.

Regulatory Compliance Some aspects of risk assessment, ORM, and regulatory compliance have been around for a long time, but not necessarily as an accepted practice. The US White House Office of Management and Budget recently released risk assessment guidelines in 2006, describing them as ―clear, minimum standards for the scientific quality of federal agency risk assessments.‖ When the National Research Council reviewed these guidelines, it concluded that they were ―fundamentally flawed‖ from a scientific and technical viewpoint and recommended withdrawing them in favor of an approach drawing on existing risk assessment expertise already within federal organizations (Hogue 2007). Another example of this at a government level is with the Precautionary Principle, which has become part of international law. This principle has become the basis for European environmental legislation; however, a high level of uncertainty exists and there is still a standard of proof needed before utilizing this principle (Foster et al., 2000). LLNL ES&H Team disciplines have consistently developed successful expert systems to

Figure 1. Risk level (RL) matrix with Control Documentation as the Output. PROBABILITY

Extremely Less Likely Likely Probable Unlikely Very High (serious injury RL 3 RL 3 RL 4 RL 4 or illness) High (lost work RL 2 RL 2 RL 3 RL 4 time) Medium (recordable) RL 1 RL 1 RL 2 RL 3 Low (up to first RL 1 RL 1 RL 1 RL 2 aid) SEVERITY

RL 1: OK. Employees perform work under bi-annual application or approval. No oversight by OSHH disciplines necessary. RL 2: Log. Established tasks with approved controls, recorded by supervisor. Periodic review of the tasks, procedures, and controls by OSHH disciplines is necessary. RL 3: Permit. ES&H Team and supervisor review of the hazards and controls (1 page). Supervisor and cognizant OSHH disciplines need to formally concur. RL 4: Controlling Document. A thorough review of hazards and controls with the ES&H Team, workers, and supervisors is performed and documented. 91

balance their workload. Recently, with increasing regulatory requirements, related paperwork, and pervasive deadlines, the OSHH disciplines have found limitations in maintaining a field presence. The RL matrix approach was initially developed to provide a consistent method for documentation requirements. The RL control document process as seen in Figure 1 provides a graded approach for ES&H Team involvement based on the RL of tasks and procedures.

All involved parties incorporate the traditional hierarchy of controls in order to provide a consistent hazard elimination process to mitigate employee risk. Tasks with an RL1 designation are primarily work performed by the general public and a common sense procedure approach is presumed. Tasks with an RL2 designation are commonly performed by industry; however, they may require certain standardized controls to ensure regulatory compliance. RL2 tasks do require some level of assurance that these controls are consistently in place. Having a supervisor‘s record of these tasks in a designated logbook works well. ES&H Team technicians and disciplines can audit these activities with the supervisor during workplace visits, ensuring work performed is within scope and established controls are in place. As RL1 and RL2 activities comprise a majority of standardized tasks at LLNL, this cost-effective method ensures health and safety regulatory compliance by focusing on risk (Zalk et al., 2010b).

RL1 and RL2 activities also afford workers an opportunity to take credit for their training and job-specific expertise while minimizing ES&H Team involvement. The RL3 designation is for work relating to either a higher level of potential risk, including regulatory non-compliance, or inappropriately characterized exposure potential. RL3 tasks require a standardized one-page permit as a higher level of control documentation in justifying the implemented controls will reduce risks relative to operative OELs. For primarily safety-related risks, such as roof work, confined space entry, welding, or other hot work activities, the permit ensures that potential risks are addressed, controls are in place, and the worker‘s training is commensurate to the task. For health-related risks, such as potential chemical exposure requiring respirators in addition to other standardized personal protection equipment (or PPE, such as glove selection, clothing, hearing protection, etc.), the permit is also a regulatory compliance document. Health-related permits include tasks with potential exposures to asbestos, lead, silica dust, carcinogens, and other common maintenance and support activities. Once procedures, or combined tasks, are deemed to have an RL4 designation, the controlling documents that are required by LLNL policy are required. RL4 work relates to the highest level of potential health and safety risk where the application of the traditional hierarchy of controls, in isolation, could remain ineffective at reducing or eliminating the multiple risks presented. As these RL4 designations are typically multidisciplinary OSHH activities involving complex work, a number of controlling documents can be utilized, including an Integration Work Sheet (IWS), a Hazard Assessment and Control (HAC) document, or a combination of the two. An IWS is a more thorough documenting process then a HAC as a review and concurrence by responsible parties and authorization for potentially hazardous work is electronically tracked. IWS structure includes a multidisciplinary review of the broader hazards and controls relating to the established scope of work, with the chain of command concurring on this broad scope of work, whereas a HAC may record similar hazards and controls, but is narrower in scope and contains more detail.

This RL process finds its elegance in standardization of the most commonly performed tasks. This in turn standardizes risk, document generation, and approaches beneficial in addressing regulatory compliance. RLBMS processes maximize resources and achieve risk calibration and consistency of controls across OSHH disciplines. This assists planning and initiation of 92

complicated LLNL R&D projects for the clients of ES&H Teams. Historically, ES&H Teams have been coerced by tradition or policy-generated workloads to treat all RLs with the same level of priority. As regulatory compliance issues are consistent and pervasive for ES&H Teams, workload prioritization remained difficult. As a result, the highest priority is either the closest deadline or the loudest client. Prioritizing ES&H Team involvement with work activities by RL and standardized tasks has offered an opportunity to change this dynamic. By involving ES&H Teams in a graded approach as determined by risk of a given activity, the ES&H Team disciplines and technicians can focus their time, expertise, and resources where most needed – the activities that have the highest potential for an adverse health and safety outcome (Figure 2).

Figure 2. Risk Level Approach to Optimizing ES&H Team Capabilities

This RLBMS approach, in theory, has multiple outcomes that support all aspects of a traditional OHSMS with the added benefit of coordinating an approach supporting OSHH disciplines‘ field time, maximizing available resources, and minimizing R&D costs on a consistent basis. Establishing appropriate RLs for their related tasks and ensuring hazards and controls are consistently implemented are now prioritized. To assist this intensive process, we begin by maximizing the input and involvement of the resource most often overlooked in the development of OHSMSs – the worker.

Participatory Approach Successful OHSMS performance is often based on lagging indicators like frequency rates and lost time work-related injury and illness. Often ignored are the attitudes and perceptions of workers in this process, an essential component of a healthy and safe work climate. Harnessing workers‘ input and involvement in OHSMS development can be a powerful management tool (Coyle et al., 1995). Worker attitudes toward accident prevention in relation to management commitment and level of involvement are important for promoting safe workplaces. Appropriate risk perception was found to be significantly correlated to risk behavior and is related to the occurrence of accidents and near misses (Rundmo 1996). Therefore, an employee‘s safe work attitude can be positively influenced when they consider themselves an integral part of a safety culture approach. This attitude as enhanced with a participatory OHSMS program, can itself be a measurable performance indicator of a successful safe and healthy workplace (Guldenmund 2000).

Though worker participation in reducing workplace hazards is often called for in research, underlying ideologies of management control and worker empowerment need to be fully understood to ensure a long lasting participation in practice over time (Zalk 2001; Moir 2005). 93

Participatory approaches focusing on worker input to achieve good practices and acceptable controls have an excellent track record of successful implementation over time and for establishing improvements in technical areas including materials handling, ergonomics, and comprehensive work organization (Zalk 2002; Kogi 2006). Once participatory methods for developing risk assessments are in practice, ongoing facilitation with training tools, checklists, and group involvement assists in maintaining and evaluating these applications positively (Kogi 2006).

The RLBMS is therefore the nexus of the participatory approach and an OHSMS, where OSHH discipline expertise and collective worker input meet. Building an effective corporation requires this comprehensive feedback approach as it tears down existing preconceptions, rules, and institutional customs in order to build a more effective and functional health and safety system (Rosenberg et al., 2001; Ackoff 1999). OSHH field practitioners have an understanding of the workforce that is necessary to identify existing practices and organizational structures to determine the correct organizational direction. OSHH practitioners are also in an appropriate position to promote a positive course for management to enhance their organization. Conversely, with a lack of productivity and profitability identified in an existing health and safety system, OSHH disciplines must also identify these weaknesses to ensure the negative organizational direction does not persist. Participatory methods in the collective redefining and rebuilding of a health and safety organizational structure has been shown to achieve the buy-in of managers and workers alike as this approach assists in achieving a collective vision of the ideals, objectives, and goals of a successful, organizationally-specific, OHSMS.

The roles of workers and expertise of OSHH field practitioners have assisted in the creation of the RLBMS and its bottom-up approach. At LLNL, the ES&H Teams were an essential part in not just raising the consistent issues facing them over the years, but also in their persistence in developing solutions that are positive for everyone involved. The RLBMS is an integration of multiple solutions to regulatory compliance and policy issues that had worked themselves into the crevices of all aspects of R&D operations. Through this, the RLBMS was compiled and sold to upper management as a benefit to operations and a cost-effective approach for highly stressed and shrinking ES&H Teams. With over two thousand IWSs and similar controlling documents in place, unraveling this web of procedures, hazards, and controls appeared insurmountable. At this point, the role of the workers and the development of a teamwork approach to address this project were formulated.

Results

In implementing the participatory approach, ES&H Teams met with their clients and began focusing on the most commonly performed tasks. Teams of affected workers, supporting ES&H Team staff, and management representatives sat down at meetings designed to develop this approach. With procedures identified and tasks isolated, the workers described how these tasks should be performed safely. Delineated during this process were R&D related activities that could not be standardized and required more time and energy in the development of a system to control these hazards. For the more commonly performed activities, ES&H Team staff assisted in aligning hazard-to-control designations for tasks and assigned RLs based on the depth and integrity of supporting data. Management involvement in this process was limited to assuring final product quality and understanding needs and expectations of their workers on a first hand basis. Together, once agreements were reached, next steps were in documenting the approach,

94

training, and procedural requirements to maintain this system over time. This process became invaluable for receiving worker buy-in and implementation.

As might be expected, not all team conclusions were in agreement. Workers would at times prefer less PPE, especially respiratory protection, but ES&H Team regulatory-related data could not support a lower RL without quantified personal monitoring results. In these cases, assigning a higher RL was temporarily necessary. The protocol necessary to make quantitative and statistically-based decisions on requirements to reduce RLs was established when practicable. At times different RLs were derived by ES&H Team disciplines for similar, common tasks. These issues were raised by the cognizant OSHH disciplines in previously established weekly meetings as a topic for discussion and agreement. This process was successful on many levels, but especially in calibrating risk perception within and between the ES&H Teams and their disciplines.

RLBMS In Practice The following topics, with the exception of ‗Safety‘, provide concise, practical examples of our existing excel spreadsheet ‗database‘ of tasks, controls, monitoring, statistically-based results, and our process for assigning RL2s (e.g. taking off the respirators) for various tasks. OSHH field disciplines respond consistently to hazardous tasks by implementing the hierarchy of controls: elimination, substitution, engineering, administrative and lastly, PPE. This process has been applied to each task example given in the topics presented below.

Asbestos Tasks that were most commonly performed were targeted operations for monitoring and development of controls. Workers identified the activities most performed and assisted in standardizing simplified controls in a participatory process including IHs and managers. IH professionals then designed sampling protocol that worked best with the asbestos OEL, fitting the amount of work to be done within a 30-minute period as the Excursion Limit of 1.0 fibers per cubic centimeter (f/cc) is an order of magnitude higher then the 8-h TWA (8-hour Time Weighted Average) of 0.1 f/cc. Therefore, as seen in Table 1, the RL2 designation was given to tasks that were 30 minutes or less with cumulative results below 0.1 f/cc (an order of magnitude below, or 10% of, the effective OEL) as an in-house standard (Zalk 2000).

Table 1. Negative Exposure Assessments for common activities in wallboard with asbestos containing building material (ACBM). Regulatory requirements will not permit an RL1. Task Task Description and Limitations Wallboard Controls Data Obtained Risk (within 30 minutes) Profile Level 1 Hammering up to 12, 2 inch nails ACBM in Shaving Gel over penetration, N = 14 RL2 Joint Mud wet rag to wipe up area 95th UCL = 0.028 fibers/cc 2 Installation of up to 8 lag bolts in ACBM in Shaving Gel over penetration, N = 11 RL2 wallboard Joint Mud wet rag to wipe up area 95th UCL = 0.041 fibers/cc 3 Up to 12, 1 inch hole saw ACBM in Shaving Gel over penetration, N = 12 RL2 penetrations Joint Mud wet rag to wipe up area 95th UCL = 0.027 fibers/cc 4 Twist drilling 11 (3/8 inch) molly ACBM in Shaving Gel over penetration, N = 7 RL2 bolts Joint Mud wet rag to wipe up area 95th UCL = 0.026 fibers/cc 5 Installation of up to 8 lag bolts ACBM in Shaving Gel over penetration, N = 8 RL2 through frames Joint Mud wet rag to wipe up area 95th UCL = 0.025 fibers/cc 6 Twist drilling and installation of 11 ACBM in Shaving Gel over penetration, N = 7 RL2 molly bolts Joint Mud wet rag to wipe up area 95th UCL = 0.026 fibers/cc 7 Patching holes to be exposed in ACBM in Shaving Gel over penetration, N = 9 RL2 work area Joint Mud wet rag to wipe up area 95th UCL = 0.069 fibers/cc 8 Removing of frames allowing 15 ACBM in Shaving Gel over penetration, N = 7 RL2 holes to be exposed in work area Joint Mud HEPA vacuum behind frame 95th UCL = 0.026 fibers/cc

95

It has been determined where the employer has monitored each asbestos job and how data during work tasks, and under workplace conditions closely resembling the processes, was obtained for compliance purposes. Conditions for workplace conditions include the type of material, control methods, work practices, environmental conditions, and prevailing environmental conditions in the employer's current operations. With this information the employer may rely on such monitoring results to satisfy FedOSHA regulations (Oberta and Fischer 2000).

Beryllium Work with beryllium poses an interesting dilemma at DOE with its unique regulatory criteria, but even more so at LLNL where surface levels of beryllium delineate beryllium work, reflecting potential dermal exposure, not airborne levels which are more typically OEL-based criteria. Therefore, the RLBMS for beryllium requires delineating ES&H Team involvement at RL1, rather than at the customary RL2. This conservative approach reflects the potential of beryllium surface levels above the release criteria of 0.2 micrograms per 100 square centimeters (g/100cm2) as is seen in Table 2; therefore RL1 is the category that workers seek to attain.

Table 2. Beryllium (Be) work Risk Level (RL) approach. Beryllium work is defined as the potential for airborne or dermal exposure above the release criteria for all workers. Risk Level Potential Exposure Controls Required Documentation Required Sampling Training RL1 No dermal and no None Designated CL1 Baseline obtained to Be Awareness inhalation task confirm designation RL2 Dermal potential, no Be Work Area (BWA): IH report Air & surface; need Be Worker + inhalation Gloves (Long Sleeves, & surface solid objective data Medical Booties possible) sampling (SOD) to become CL1 Surveillance RL3 Dermal + inhalation, BWA: Gloves (Tyvek & Be Permit, Air & surface; need Be Worker + confidence airborne is Booties necessary), IH report & air, SOD to become CL2 Medical < 0.2 g/m3 8-h TWA Respirator (1/2 mask) surface sampling Surveillance RL4 Dermal + inhalation, no Regulated BWA: HAC document. Air & surface; need Be Worker + confidence airborne is Gloves (Tyvek & IH report & air, SOD to become CL3 Medical < 0.2 g/m3 8-h TWA Booties necessary), surface sampling Surveillance Respirator (Full face)

To achieve RL1, a determination is made that there no potential for dermal exposure; therefore, a quantification of surface measurements is necessary and achieved by collecting wet swipes in and around work areas. Achieving RL2 requires statistically ensuring there is no potential for airborne beryllium levels approaching the Action Level of 0.2 micrograms per cubic meter (g/m3) as an 8-h TWA. If such data is not available, the resulting control designation is RL3, or RL4 if there is a potential to be in excess of the Action Level. These RL delineations reflect DOE regulatory criteria and provide a necessary consistency for IH decision-making. Beneficially, it also offers clarity for workers who may work in proximity to beryllium components or related contamination. Therefore, IHs who routinely quantify surface and personal exposures within established or potential Beryllium Work Areas (BWAs) are more likely to have established controls effectively lowering the RL more consistently. It is important to highlight regulatory oversight specific to beryllium as FedOSHA‘s Permissible Exposure Limit, currently at 2.0 g/m3 as an 8-h TWA and identified within the DOE criteria, is an order of magnitude higher than the Action Level.

Lead RLs were determined for four different types of activities involving potential exposure to lead and are shown in Table 3. The activities were defined as activity Similar Exposure Groups (SEGs). These activities were monitored to satisfy FedOSHA lead monitoring requirements, however the data also assigns appropriate RLBMS classifications. To illustrate this, only the most pertinent information is shown.

96

Table 3. Activity SEG descriptions and corresponding RL determinations for work involving potential exposure to airborne lead, these are examples of classification and not cumulative results. Sampling Results 8-h TWA Eng Admin OEL Risk Level Category Description duration (g/m3) (g/m3) Controls Controls PPE (g/m3) >OEL (RL) (min) Soldering Electric 77, 347, <1.5, <0.24, General None None 50 No RL1 soldering 332, 207, <0.6, <0.43, ventilation using Sn63 235 <0.6, <0.43, (37% Pb) <1.0, <0.72, solder <0.8 <0.58 Brick Transport 111, 117, 1.8, 5, 0.42, 1.2, Outdoors Lead Work C, LG & 50 No RL2 transport of 480 110, 96, 24, 5, 5.5, 1.0, Permit, Lead NG, HF, (Based on bricks from 103, 101 6.4, 6.7 1.4, 1.4 Training, shoe results, old pallets Water spray covers, HF or lead onto new (Land’s and wet safety work pallets “Exact” wipe. prior to glasses, permit not 95% UCL = transport Hudson required 7.0) sprayer with controls) Paint Wet-scrape 94 <2.1 <0.41 HEPA Lead Work C, NG, 50 No RL3 removal paint using vacuum Permit, Lead HF, (More data HEPA Training safety required to vacuum glasses, reduce RL) Dirt Removal of 278, 277, 0.57, 1.5, 0.30, 0.87, Water Lead Work C, NG, 50 No RL2 removal lead- 51, 54, <3.0, <0.32, spray Permit, did HF, (Based on contaminat 172, 170 <2.9, 1.0, <0.33, 0.36, not overfill safety results, ed dirt at <0.92 <0.33 bucket used glasses, HF or pistol range (Land’s for dirt permit not “Exact” transport required 95% UCL = with 0.64) controls) C = Tyvek coveralls, LG = Leather Gloves, NG = Nitrile Gloves, HF = Half-face air purifying respirator

In Table 3, RLs were determined from personal exposure data from the four operations. Brick transport and dirt removal activities were initially characterized as RL 3 operations prior to obtaining exposure data; however, based on a minimum of six samples obtained for each activity and calculation of Land‘s ―Exact‖ 95% upper confidence level (UCL), the RL was downgraded to RL 2. Hence, these activities would no longer require respiratory protection or a lead work permit. For the paint removal activity, as only one data point was obtained, additional exposure data would be required before RL downgrading is possible.

Silica RLs were determined for six different activities (including three pertaining to saw cutting, but on different materials or with different controls) involving potential exposure to respirable crystalline silica (silica) and are shown in Table 4.

97

Table 4. Activity SEG descriptions and corresponding RL determinations for work involving potential exposure to silica dust. These are examples of classification and not cumulative results. Sampling Results 8-h Eng Admin OEL Risk Level Category Description duration (g/m3) TWA Controls Controls PPE (g/m3 >OEL (RL) (min) (g/m3 ) ) Rotoham Drilling holes up to Over 20 < Limit of Not 2 methods: Silica C, LG, HP, 25 No RL 2 -mering 20 holes, up to 4 samples. Detectio Applic- 1) HEPA safety no R inches deep, into a Various n able vacuum at training concrete slab or times drill site; or floor using a from 30 – 2) wet cloth rotohammer. 180 min around hole Jack Breaking up a 205 180 77 Ventilated Silica C, LG, HP, 25 Yes RL 3 ham- concrete floor using enclosure, safety FF-R mering a jackhammer wet training method, local exhaust Saw Cutting concrete 70 60 8 Ventilated Silica C, LG, HP, 25 No RL 3 cutting floor using consaw enclosure, safety FF-R (More data wet training required to method, reduce RL) local exhaust Saw Cutting asphalt 68 90 13 None Silica HF-R 25 No RL 3 cutting sidewalk safety (More data training required to reduce RL) Saw Cutting asphalt floor 85 <47 <8 Wet Silica Double 25 No RL 3 cutting methods safety HP, hard (More data (hoseline training hat, PA-R. required to sprayer) anti- reduce RL) vibration boots and gloves, Chipping Chipping concrete 201, 305 120, 50 68, 31 Ventilated Silica C, HP, LG, 25 Yes RL 3 using chipping gun enclosure, safety FF-R wet training method, local exhaust C = Coveralls, LG = Leather Gloves, HP = Hearing Protection, HF = Half-face, FF = Full-face, PA = Powered air, R = Respirator

For rotohammering, two different control options are available and both are determined as RL2, or no required respiratory protection. For two of the five activities, the OEL for silica was exceeded, even with engineering controls in place. These activities were therefore designated as RL 3 and require silica work permits and further IH monitoring, with an emphasis on upgrading the engineering controls to reduce exposures as practicable.

Safety RLs for life-critical activities are initially determined as RL3. The proactive process begins by evaluating the activity‘s full scope and potential RL3 activities to be performed in line with the project as part of a pre-job briefing (Zalk et al., 2010b). Controls commensurate to these RL3 items are identified in advance and a daily project walk-through with a checklist, as in Table 5, is performed. Since controls related to each item are previously identified as in place, they are considered RL2.

Table 5. Example of an Occupational Safety Assessment Checklist. Potential RL3 items are RL2 when proactively identified, controls determined, and jobsite assessed to ensure compliance.

98

YE YE NO N/A NO N/A S Items Assessed S Items Assessed Fall Prevention and Protection Excavation and Trenching Employees are utilizing 100% fall protection at/above 6 1 22 Before digging, “utility locates” have been performed feet (2 m). 100% tie-off maintained at/above 6 feet (2 m) or when Occupied excavations are adequately protected against 2 23 exposed to a fall hazard. cave-in “Competent Person” daily inspections are completed prior 3 Fall protection in use is in satisfactory condition 24 to excavation entry Employees will not contact a lower level obstruction Adjacent equipment (stationary/mobile) is controlled to 4 25 during an arrest prevent imminent danger to occupants Floor/Wall openings are covered, protected and labeled Employees are hand digging with non-conductive tools 5 26 (i.e., load rating) while locating underground utilities Electrical Safety and Scaffolds Lockout/Tagout (LOTO) A GFCI/CB/Assured Grounding/Earthing program is Scaffolds are installed, maintained and inspected per 6 27 being used where required requirements and possess scaffold tag All exposed conductors are covered by closed electrical Modification, erection and dismantling are performed only 7 28 enclosures by competent scaffold erectors Temporary wiring terminations are protected both Scaffolds are grounded where exposed to 8 29 dielectrically and mechanically induction/electrical conductors Ground prongs are present on extension cords and 9 power tools as required Ladders Proper PPE is being used when working on energized Metal ladders and multi purpose ladders are not being 10 30 circuits used All applicable hazardous energies are isolated with an Straight/Extension ladders are secured against 11 attached LOTO device and tag and all residual/stored 31 displacement energy relieved Zero energy checks are being performed with a volt ohm Ladder positioning is adequate to perform work safely 12 32 meter (proper ladder angle) Each exposed individual has control over the lockout Ladder is suitable for the task (e.g., extension vs. A- 13 33 device frame) LOTO device emergency/absent removal protocols are Employees are not standing on the top two rungs of 14 34 being followed ladders Employees are maintaining 3-point contact while climbing 15 Proper insulated tools are being used for electrical work 35 ladders 16 Is the proper signage in place Lifting/Rigging Operations Only qualified operators are operating hoists (stationary 36 Confined Space and mobile) A full time attendant is present during confined space Rigging operations are performed only by qualified 17 37 entry riggers Confined space is being monitored for potential chemical In-service rigging equipment is in satisfactory condition 18 38 and atmospheric hazards (load limit tags, inspected, defect free, hoists) 19 Adequate rescue equipment is readily available 39 All load hooks are equipped with safety latches The entry permit addresses all imminent dangers for Swing radius has been identified/barricaded with danger 20 40 permit required confined spaces tape or barricaded if needed High voltage lines are shielded when hoisting and rigging 21 Is the proper signage in place 41 operations are within 10 feet

This checklist assessment approach not only identifies compliance and non-compliance with FedOSHA regulations, deficiencies that are rectified simply in the field can be recorded as a ‗near-miss‘. These checklists are also to be available in an electronic format compatible with field inspections. Therefore, these inspections will also serve as a positive reinforcement mechanism trending percentages of field compliance, rather than non-compliance, as a leading indicator. This also serves as a quantified trending of areas requiring improvement (e.g., controls not in place), feeding into a quarterly tailgate training that focuses resources and attention where most warranted. This is intended to replace the more prominent lagging indicator of workplace accidents and injuries, which then can be graded as a percentage of incidents over compliance rather than incidents in isolation.

99

Discussion

The RLBMS risk-based ORM model has begun to successfully focus dwindling OSHH expertise and resources at LLNL onto the highest risk procedures as originally intended. Utilizing the CB strategies in this unique approach has taken a large variety of commonly performed maintenance and support activities and integrated the RL approaches to maximize R&D output. Another additional benefit for R&D activities is the tracking of R&D tasks so, when placed within SEGs, they can be further standardized. Additionally, expanding the RLBMS SEG approach to include co-located workers will establish a quantified exposure connection within and between working environments to bolster medical surveillance selectivity. The majority of RLBMS tasks are routine and well defined within production- oriented companies and can therefore be shared. This is also true for R&D environments with many support and maintenance tasks, but not with programmatic work. Where support and maintenance tasks may differ is in the work location itself, as workers may be exposed to co- located and unique facility hazards. Auditable tracking of these activities through the RL approach has also been successful in standardizing regulatory compliance and is proving to be useful for all levels of management.

Hand-in-hand with the maximization of OSHH prioritization of activities is a more confident workforce. Workers can now focus more on the quality of their product, rather than compliance, as emphasized in their role in developing their task-based controls. As presented, the regulatory documentation of hazard assessments and implemented controls offers a validation of the RLBMS. Consolidation of the measurement and evaluation of exposure control effectiveness for auditing purposes adds to the growing research emphasizing the utility of qualitative risk assessment approaches within a traditional framework (Zalk et al., 2009). The RLBMS‘s participatory approach can be seen building effectiveness from the bottom up, teaming workers with OSHH staff in a collaborative manner that develops partnerships at an accelerated pace. The eventual linking of the RLBMS model into a comprehensive OSHH database, currently delayed due to budgetary constraints and now manually introduced into electronic spreadsheets, will eventually strengthen the calibration of risk-based decisions across OSHH disciplines. Currently, an auditable trail is demonstrating the benefits of the consistency inherent in the RLBMS approach. This consistency is an essential component of ensuring a healthy and safe working environment within a complicated regulatory framework.

Conclusion

RLBMS‘s simplicity belies the complexity of ensuring a consistent implementation strategy while adjusting OSHH resources to regulatory changes. RLBMS‘s auditable tracking of activities, maximization of OSHH professional field time, and standardization of control documentation is proving invaluable. Validation of RLs and evaluation of task-to-control effectiveness continues within a quantitative regime for regulatory auditing to ensure the appropriateness of qualitative risk assessment methods. The creation of an OHSMS protocol that integrates participatory approaches so intrinsically has also created a simplified and consistent risk communication that underscores the overall value of the RLBMS. It must also be made clear that the RLBMS should not be implemented in many organizations prima facie as there are varying levels of confidence in working with qualitative risk assessment techniques in any working environment. Therefore, a thorough review of the RLBMS and its components is highly recommended. National regulatory requirements, enforcement, and cultural acceptance of exposure assessment protocol can vary greatly. Organizations worldwide may find acceptable 100

quantitative risk assessment protocol at 10%, 50%, or 95% of established OELs, and at varying statistical confidence levels as well. Other organizations may feel it appropriate to operate above an OEL with respiratory protection and PPE in use. The RLBMS can be adapted to any of these regimes, however validation and evaluation will always be necessary to ensure worker health and safety remain paramount. Should others see RLBMS‘s utility, validated task-based controls can be captured in international databases and national programs, incorporating them into trade- based ORM toolboxes and further integrating OSHH professions to assist in a multidisciplinary reduction of work-related injury and illness worldwide.

Acknowledgements This work performed, in part, under the auspices of the US Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-JRNL-413441.

101

CHAPTER 7 BARRIER BANDING: A CONCEPT FOR SAFETY SOLUTIONS UTILIZING CONTROL BANDING PRINCIPLES Submitted to Safety Science David M. Zalk1, Paul Swuste2, Andrew R. Hale2 1. Lawrence Livermore National Laboratory, Hazards Control Department, PO Box 808, L-871 Livermore, CA, USA, Email: [email protected] 2. Safety Science Group, Delft University of Technology, PO Box 5015, 2600 GA Delft, The Netherlands

Abstract

The application of banding principles, developed by occupational hygienists, is explored for potential occupational safety applications. Control Banding strategies offer simplified solutions for controlling worker exposures to workplace constituents. The Control Banding process qualitatively generates exposure scenarios according to ‗banding‘ principles. For safety, probably ‗Barrier Banding‘ is a better title. A banding approach for occupational safety could offer a simplified accident prevention approach providing scenario-specific links between risk classification and prevention. Barrier Banding has the potential to offer Small- and Medium- sized Enterprises and developing countries an easy decision-making process for reducing workplace injuries in the absence of safety expertise. Obstacles in achieving this are encountered. Banding of controls in occupational hygiene fits well into the established control hierarchy, however barriers in safety do not ‗band‘ quite as transparently. An evaluation of customary safety models presents an understanding of relationships between hazards, scenarios, and a potential banding approach, however it remains theoretical. Examining falls and electrocution in the construction industry provides an illustration of how a Barrier Banding strategy can be applied. Further research is necessary to test this experimental Barrier Banding strategy with end users and to further develop scenario- and hazard-specific links to achieve accident prevention.

Keywords: Barrier Banding, Control Banding, Safety Solutions, Risk Level, Occupational Risk Management, Occupational Risk Model, Risk Communication, Multidisciplinary

1. Introduction

The history of Control Banding (CB) began with the banding of hazards into three or four levels of risk (Zalk and Nelson, 2008). In the 1970‘s it appears in the safety profession with the stratification of central events related to fire and explosion, the so-called Mond Index (Lewis, 1980). This led to a risk matrix describing the likelihood and probable severity of an explosion or release of toxic material developed for use by chemical companies with major facilities (ICE, 1985). Later evolutions occur in the 1980‘s for delineating risk levels in radiation, lasers, and biosafety (Zalk and Nelson, 2008). The concept of hazard banding emerges in the 1990‘s for the pharmaceutical and chemical industries (Naumann et al., 1996; Sargent and Kirk, 1998). These sector-specific approaches progressed the idea that the banding of hazard classifications could provide the basis for applying generic exposure control standards. The pharmaceutical industry had a unique and appropriate niche for the formal development of CB strategies since research, development, and production requires work with large numbers of new chemical compounds 102

with few toxicity data. The experts reasoned that such compounds could be classified by their toxicity and the need for restriction of exposure into ―bands‖ aligned with a stratified control scheme (Naumann et al., 1996). CB in its modern evolutionary state has its primary focus in the field of occupational hygiene and in the late 1990‘s the United Kingdom Health and Safety Executive developed an early toolkit, known as the Control of Substance Hazardous to Health (COSHH) Essentials. COSHH Essentials represents a qualitative instrument to assess risks for chemical substances into four bands that generate solutions fitting well into the hierarchy of control measures (Russel et al., 1998). The four levels in the hierarchy of controlling exposures to chemicals consist of: (1) good occupational hygiene practices, including personal protective equipment, (2) engineering controls, including local exhaust ventilation, (3) containment, and (4) the need to seek specialist advice. The essential characteristic of this hierarchy is that these four levels are alternative ways of controlling the same exposure scenario. Since the introduction of the COSHH Essentials approach, many other similar CB strategies have emerged internationally (Zalk and Nelson, 2008). CB categorizes the hazards of chemicals based on their so-called R (risk) – phrases and generates exposure scenarios, based on quantities used, the degree to which chemicals can become airborne, and the length of time the task is performed. CB has had a wealth of success, balanced with room for improvement as reflected in recent research (Zalk and Nelson, 2008; Bracher et al., 2009, Tischer et al., 2009; Lee et al., 2009). Although the COSHH Essentials model has also received its share of criticism (see for instance Kromhout, 2002a; Swuste, 2004; Jones and Nicas, 2006 (a and b); ACGIH, 2008), the focus on controls, without a lot of expert input, is a strong point of the tool, and makes it applicable in branches of industries and countries which are deprived of expert support.

In The Netherlands a tool developed for Small- and Medium-sized Enterprises (SMEs); the ‗Stoffenmanager‘ (substance manager) (Feber et al. 2003; Noy, 2004; Tielemans, et al. 2008) is becoming popular in both developing and developed countries. This digital tool (www.stoffenmanager.nl) allows risk assessments of hazardous chemicals. Unlike COSHH Essentials, this model factors in exposure potential through utilization of an interactive chemical risk management approach. The employer completes an exposure assessment, including worker proximity to the exposure source, to determine the chemical‘s exposure class. Using this exposure assessment and R-phrases categorized according to COSSH Essentials, the tool automatically calculates a risk score to complete an initial assessment of health risk. The employer can then calculate the efficacy of various control measures and choose the most effective ones. This combination leads to a relative risk score, and consequently the influence of control measures on risk scores can be calculated. Quantities of chemicals used are not taken into account. Instead work activities and operations (duration, type, frequency) are parts of the exposure scenarios, as well as control measures already installed to control transmission and immission. Sector-specific versions of Stoffenmanager, such as for construction, are being created and are already receiving wide spread acceptance by the intended users, the SMEs that helped in the design and development of the interface (Marquart, et al. 2008, Zalk and Spee, 2009).

To achieve simplicity, CB intentionally forgoes many of the determinant intricacies that comprise exposure scenarios (Maidment, 1998). COSHH Essential‘s preferential selection of simplicity, transparency, and applicability as a main argument justifies its limitation of exposure scenarios to chemical amount, dustiness or volatility, and task length (Brooke, 1998). If we are to apply this to safety, the determinants that will be chosen will need to be scenario specific. In CB, the bands of classification seek to capture relevant determinants of exposure, leading to bands of exposure scenarios. However, CB does not take all determinants of exposure into 103

account. Examples of exposure determinants not addressed include: specificity of tasks and operations, process disturbances, operational control methods, existing control methods, motive power, distance of worker from the exposure source, and work in enclosed spaces.

In developing the concept of Barrier Banding (BB) toward achieving safety solutions, a number of questions emerge. Are banding principles applicable in occupational safety? Are classifications for hazards, determinants of accident scenarios, and controls applicable for a BB application? Can accident scenarios be used in the same way to limit determinants in such a way as to achieve the simplicity that was necessary for CB? What are the pitfalls of utilizing banding principles in occupational safety? Selections need to be made for the determinants, for both exposure as well as accident scenarios. The choices relating to which determinant has relevancy has a theoretical side, which is dominant in both accident and exposure processes. There is a practical side to investigate as well. An evaluation of customary safety models will be presented to provide an understanding of relationships between hazards, scenarios, and a potential banding approach. Scoring of accident scenario determinants requires exploration, as does the classification of these determinants into groups or bands. An examination of falls and electrocution in the construction industry is performed to provide an illustration of how a BB strategy can be applied. Taken together, the strengths and weaknesses of a safety banding approach will be evaluated and opportunities for further research will be presented.

2. Modeling Safety Risk and its Control

Models of occupational risk prevention are essential to understand the causal pathway of accidents. From the late 1800s onwards the engineering approach to safety has been dominant, referring to technical or hardware safety measures, such as the enclosure of moving machine parts. Crystal Eastman developed the first scientific theory in safety in 1910, the environmental hypothesis, explaining accidents by external factors such as hazardous machines and long working hours (Swuste et al., 2010). Psychologists entered the safety field in the early 20th century with the individual hypothesis, leading to the accident proneness theory, also in 1919. The portrayal of the accident process as a sequence of events can be traced back to DeBlois (1926), while Heinrich (1929) expanded on this theory, leading eventually to his Domino Theory (1941); accidents are seen as a formalized sequence of events: social environment, fault of person, unsafe act or condition, accident, and injury. The third domino, unsafe acts and conditions, was considered the central factor of accident prevention and the following four reasons expand this notion: improper attitude, lack of knowledge or skill, physical unsuitability, and lack of a proper mechanical or physical environment. Knowledge on causal pathways of accidents not only provides insight into reasons why accidents occur, but also directs efforts to prevent these accidents through the use of barriers.

One big name in this respect is Haddon, who introduced the energy and barrier concept. This model derived from the medical success to combat the great epidemics in the 19th century, notably cholera. It used the well-known agent – host – environment triangle, which in safety terms became the energy – victim – surroundings triangle. Haddon, a physician, never validated this model for occupational safety, but postulated the safety triangle as a model that could provide better results in accident prevention than psychologists‘ efforts. The concept of barriers has its basis in the energy-barrier model which was presented by Haddon (1968) with his 10 steps towards the prevention of loss. This energy-barrier-victim safety model can be widely applied to all hazardous interactions of energy with vulnerable people and property (Hadden et 104

al., 1964, Haddon 1973). The risk-oriented Management Oversight and Risk Tree (MORT) energy flow process model also addresses work design and includes the management system as part of a causal factor analysis for the choice of prevention measures, safety inspections and accident investigations (Johnson 1973). A fundamental modeling concept representing an accident process is the bow-tie (Hale et al., 2007; Bellamy et al., 2007). A bowtie, which is a combination of a fault tree and an event tree, linked together by a ‗central event‘, presents a further refinement of accident causes. Central events present a state of ‗loss of control‘, whereby the energy content of the hazard is released. Various scenarios can lead to central events. A scenario describes elements and conditions leading to central events, and after them to consequences.

Haddon‘s (1973) 10 prevention strategies can be seen in Table 1. These strategies can be grouped into four energy (E) categories: Eliminate E (prevent creation of the hazard), Reduce E (reduce amount of the hazard), Isolate E (prevent release of the hazard), and Reduce Flow of E (modify and mitigate hazard quantity and quality). These four categories

Table 1. Haddon‘s 10 prevention strategies. Strategy Examples Using Traffic Energy (E) [Bands] Safety 1) Prevent the initiation of Discourage the use of vehicles Eliminate E [Band 1] the energy form 2) Reduce the amount of Setting speed limits, less Reduce E [Band 2] energy aggregated powerful engines 3) Prevent inappropriate Reduce driver‘s ability to Isolate E [Band 3] releases of energy make mistakes 4) Alter rate of energy Appropriate crashworthiness release from source of vehicles 5) Separate energy & Separate lanes, curfew for Reduce Flow of E objects by space/time truck use in cities [Band 4] 6) Place barrier to Placing road dividers on separate energy & highways objects 7) Modify contact surface Softer front bumpers, that to be impacted motorcycle helmet use 8) Strengthen humans Osteoporosis treatment for from energy transfer older drivers 9) Quickly detect damage Emergency care, management Not Applicable and ameliorate of crash sites 10) Measures from Long-term victim care and emergency to rehabilitation stabilizing are also useful for classifying safety control principles. In considering these as ‗principles‘ there is the ability to draw a relationship with CB, as local exhaust ventilation and containment are control principles that require adapting to local circumstances. Although Haddon‘s main focus in publications is traffic safety, his published principles for prevention are also applicable to occupational safety. The first three of his 10 strategies for classifying the transfer of energy include: preventing energy from forming, reducing energy, and preventing release of energy. 105

The last three strategies of the ten reveal both his medical background (focus on host‘s condition) and its automobile accident derivation: strengthening the structure against damage by energy transfer, reduce the losses not prevented by this strengthening, and all the stages between emergency response to the damage which has occurred and final stabilization. As these last three strategies are all principles after the energy transfer, they are not applicable here. Haddon‘s strategies 4-7 can be combined to the category ‗Reduce Flow of E‘, resulting in four bands of control for BB, as seen in Table 1, that appear commensurate to the CB approach.

2.1 Risk Level Matrix Risk is seen as a numeric variable, because its most simple formulation equates with the combination of unwanted consequences and the probability for their occurrence. Fine and Kinney (1971) present an early model for quantifying risk due to a safety hazard by multiplying factors relating to probability that risk is present, exposure length (frequency), and consequences (severity) of exposure. Probabilities and consequences are normally divided into groups or classes and provided with a value. By multiplying these values, a risk score is created and used to compare different risks. Because valuing probabilities can be difficult, variations are present which split these probabilities. First a hazard exposure frequency is estimated, then probability of scenarios occurring is specified for the range of consequences (Zwaard and Goossens 1997). Risk matrices are also quantitative risk assessment techniques used in safety management for assigning a risk class to a potential accident or hazard with consequences of exposure and probability predicted for a given accident or hazard (Cook 2007). Risk matrices that have only two dimensions of probability and consequences are used by process industry companies for prioritization (e.g. Visser 1998). However, these matrices are primarily used to make rather simple decisions, such as risk tolerability, amount of effort or money spent getting rid of it, or the speed with which it should be tackled. However, linking these categories to types of prevention in safety is lacking, because there is no obvious dimension of prevention measures to match with the categories of risk.

In occupational hygiene the same risk matrix approach paints a different picture as it is considered a qualitative risk assessment and numerous successful approaches of this sort link with prevention measures. COSHH Essentials presents a risk matrix that has become the basis for over 20 CB toolkits worldwide (Zalk and Nelson, 2008, Fingerhut, 2008, NIOSH, 2009). What they have in common, stressing multidisciplinary potential, is that they typically use a common language in determining a risk level (RL). Simplified bands of risk presented within a matrix show a common footing between occupational hygiene and safety (Visser 1998). This approach bands risk as an outcome of severity and probability across four levels to obtain simplicity and resulting RLs are tied to commensurate controls (Figure 1).

106

Figure 1. Risk level (RL) matrix as a function of severity and probability that can be used for the banding of construction projects. Probability Score

Extremely Less Likely Likely Probable Unlikely

Very RL 3 RL 3 RL 4 RL 4 High Severity Score High RL 2 RL 2 RL 3 RL 4

Medium RL 1 RL 1 RL 2 RL 3

Low RL 1 RL 1 RL 1 RL 2

As this RL matrix approach is considered a qualitative risk assessment, occupational hygiene experts have been hesitant in their acceptance of this model as quantitative methods are considered the gold standard of the profession. A milestone has been achieved in shifting expert opinion internationally with the successful implementation and evaluation of a qualitative risk assessment program utilizing the risk matrix seen in Figure 1 for the prevention of nanoparticulate exposure (Paik et al., 2008; Zalk et al., 2009). This CB ‗Nanotool‘ offers a simplified process for controlling worker exposures to nanomaterial. Exposure to nanomaterials has toxicological and quantitative measurement uncertainties, however, this CB model may in fact be superior to traditional methods. This model addresses severity factors using multiple engineered nanomaterial composition parameters (shape, size, surface area and surface activity) and probability factors based on exposure parameters (dustiness, amount in use, and number of employees). These indices are linked by the resulting RL to bands with the four corresponding control approaches of the traditional hierarchy. Extensive evaluation of the Nanotool has proven it a useful tool for risk communication with non-experts (Zalk et al., 2009). This was surprising to many experts as the multiple parameters of severity and probability that were determinants for exposure were thought to add complication for risk matrix use. Non-experts quickly learned that reducing the risk of tasks, by adjusting input parameters, could reduce scores within the RL matrix and lessen the corresponding control outcome required. This is an important consideration for BB. The determinants of accident scenarios may also have multiple input parameters, however this should not necessarily be seen as a deterrent for non-expert use.

3. Banding of Barriers

Occupational safety risk matrices are used to make rather simple decisions, such as tolerability of risk, amount of effort or money to be spent getting rid of it, or speed or expertise with which it should be attacked. Often missing is a decision that links categories of risk to types of barriers, 107

as there is not immediately an obvious dimension of prevention measures to match these. Some exceptions can be given where there is a dimension (e.g., machine guarding). Barriers can be on the hazard (the machine), between the hazard and the person (a perimeter fence or pressure pad on the floor) or on the person (personal protective equipment – PPE). Barriers are hardware control measures and human interventions, which will block or influence the development of a scenario. When barriers are failing, scenarios can become active. Barriers are divided into primary barriers, which are able to prevent the development of a central event, and management factors, which influence the quality of primary barriers. Management factors have a direct relation with the safety management system. Hardware barriers can be either passive barriers (e.g. machinery guarding, thickness of reactor vessels), or active barriers like relief valves, automatic shut downs etc. (Goossens and Hourtolou, 2003; Hollnagel, 2006), or they can require operators to use the hardware before it is effective (e.g., push sticks for circular saw use). When hazards are seen as energy content, classifications used in a so–called ‗energy analysis‘ will be useful, and distinctions can be made here (CCPS, 1989; Hale and Swuste, 1997; Harms- Ringdahl, 1993). Hazards in safety are generally seen as energy: potential, kinetic, electrical and the like. In occupational hygiene, when limited to chemical exposure, the chemical itself with its physical and toxicological properties is the hazard. In Table 2, consideration is given for what sort of dose-response relationship can be developed based on the consequences dimension from Fine & Kinney.

Is it valuable to think of specifying risk control measures on these consequence dimensions? Clearly in some cases this is done. Precautions on scaffolds are linked to the height from the ground that a person is working and this can be seen in many national safety regulations. In traffic, the barriers used alongside roads are adapted to the speed (and to a lesser extent the mass) of the traffic that might contact them – see for example the Swedish Zero Accident risk control measures (Rosencrantz et al. 2007). But this all applies only to the sorts of barriers linked to reducing consequences and not really to other dimensions of reducing exposure and probability that the risk will manifest itself (the probability and exposure dimensions in Fine & Kinney). There are safety fields that link the size of probability and exposure to different levels of rigor for risk control barriers. As an example, railways long decided on the type of level crossing based on the number of vehicles passing across per hour and the speed of trains on the track. Barrier choices, by risk level, were between: flashing light – no gates, single carriageway barriers, double carriageway barriers and replacement with a tunnel or bridge.

108

Table 2. Linking hazard types to dose relationships

Type of hazard Dose relationship Fall from ladder (steps, stairs) Height, Surface factors Fall from height roof/floor/platform Height, Surface factors Fall from scaffold Height, Surface factors Fall though hole/ opening Height, Surface factors Fall from heights other Height, Surface factors Fall on same height Surface factors Struck by vehicle Speed & mass Dropped object Height & mass Flying Object Speed & mass Contact with swinging/ hoisted object Speed & mass Contact rolling/sliding/moving object Speed & mass Contact with object person is carrying/ using/ holding Speed & mass, surface factors Contact with hand held tools Power, sharpness/ penetration Contact with moving part of fixed machinery Power, sharpness entanglement Speed of person, sharpness of Movement against or into surface/ object object Time, engulfment, entrapment, Confined spaced (21) or oxygen deprival (22) (incl. suffocation (10) occupational hygiene Vehicle speed or exit velocity, In or on moving vehicle surface factors Electricity Voltage, current Contact with extreme hot/ cold surfaces or open flame Time & Temperature Already covered in Spill/splash of hazardous substances in open container Occupational hygiene Person contacts hazardous substance incl. biological agent Occupational hygiene Loss of Control hazardous substance (active) Occupational hygiene Loss of Control hazardous substances (passive) Occupational hygiene Temperature, speed of spread, Fire & Explosion pressure wave Exposure to damaging noise or radiation (19) Occupational hygiene Speed and mass/ sharpness of Aggression blow or object used to attack Depth for diving, speed of Drowning & Diving related hazards (24) surfacing Posture, force, repetition, Extreme muscular exertion loading

3.1 Design Analysis and Banding Proposed classifications of process or activity are derived from design analysis, a descriptive analysis of the structure of a production process. It basically answers the questions; what has to be produced and by which variety of methods (Kroonenberg and Swiers, 1983; Eekels, 1987; Swuste et al., 1993). According to design analysis a production process consists of three decision levels: production function, production principle, and production form. Design analysis

109

can be interrelated and organized hierarchically (Table 3).

Table 3. Linking Design Analysis Principles to Haddon‘s Energy Flow and Solutions.

Haddon Energy Solution Decision Design Analysis Activities (E) & Flow Approaches

Elimination & Production Core activity, unit What Eliminating E Substitution to Function operation Prevent Hazard

General process, Dose/Response, Production Motive power, How Reduce E Reduce Amount Principle operational control of Hazard measure

Detailed design, Prevent Hazard Isolate E & By use of Production Form machines, tools, Release, Reduce Flow E preventive measures Modify/Mitigate

The production function is the highest level in order and divides the production process into its core activities. Second in order is the production principle, specifying the general process (e.g. bulk feed through pipes vs. sacks opened and poured into the process), motive power and operational control methods by which the function is, or can be achieved. Motive power relates to the energy source in use. Operational control methods determine the distance of the worker from the source of exposure, and can be classified (a) manual, (b) mechanically driven, (c) remote controlled or (d) automated method. The first two, manual and mechanically driven operations, find the worker close to the machine or equipment in use. Remote controlled and automated operations increase distance from the machine, at least for the operator, though not for maintenance work. The lowest level is represented by the production form which determines the detailed design used in carrying out the production principle; it describes the materials, tools and machines in use and the additional measures applied to prevent accidents or exposure.

The concept of production functions is closely related to the classification introduced in 1936 in the chemical industry, dividing a production process into so-called unit operations (Badger and McCabe, 1936). The idea behind unit operations is the assumption that, although the number of individual processes is great, each one can be broken down into a series of steps which appear in process after process. These operations have common techniques and are based on the same scientific principles. The concept of unit operations is not restricted to the chemical industry, but has been used in the process industry in general. A classification according to unit operations is helpful for decision making about risk and prevention and is incorporated in the proposed classification of production processes (Shreve, 1956; Blackadder and Nedderman, 1971; Geankoplis, 1978; Hovinga and Deurloo, 1984; Ghosh and Mallik, 1986).

When analysing solutions we always start at the production form. Here you have the exposure, the safety risks. This means that the production form is the first target to address prevention at the lowest level. It describes the materials, tools and machines in use and the additional 110

measures applied to prevent accidents or exposure (Swuste 1996). Technical safety measures also provide examples of changes in production form. To achieve more fundamental changes in risk and exposure it is necessary to move up from form to principle and eventually to function. Eliminating the production function is the ultimate solution, because then all levels below will change. Returning to the central position of the principle, this fits very well with the energy source to dose relationships seen in Table 1. Potential energy links to height and mass, kinetic energy to speed and mass, electrical energy to voltage and current and so on. The combination of the energy source and the mechanical principle defines the hazard level, and the operational method the worker‘s distance to the source. Taken together, it can be seen how design analysis and the ten countermeasure strategies of Haddon (1973) present, in theory, a way for banding principles to be applicable in occupational safety. The classifications for hazards, an overview of accident scenario determinants, and solution approaches to control hazards are shown as applicable for banding barriers. However, this theoretical approach requires a practical application to consider how BB can be applied.

4. Barrier Banding in Construction

Design analysis in construction can fit well with production principles containing the main determinants of scenarios, providing information on energy, mechanical principles, and operational control methods. A prominent accident scenario in the construction industry with potentially fatal outcomes is when a worker falls from a height (BLS, 2007; Bentley et al., 2006; Chi et al., 2005; Ale et al., 2008). The hazard is the potential energy in the height of the fall translated into kinetic energy when the person does fall and it scales well by ‗banding the hazard‘ (minimizing velocity dose) across four risk levels: < 2 m; 2 – 5 m; 5 - 10 m;  10m (Aneziris et al., 2008). These four classifications of the hazard of height represent a dose, which are based on the determinants of accident scenarios. For the purpose of analyzing a BB approach, the accident scenario to be evaluated will be a worker‘s fall from a ladder. Aneziris et al. (2008) present examples of what can be used for input factors within a BB approach for work on ladders. Focusing on prevention, falls from ladders require sequential considerations of the ladder‘s strength, stability, type, placement and protection from collision with e.g. vehicles or other objects, and of the users‘ abilities to work on the ladder and their stability on the ladder. Inputs to consider for the strength of a ladder include the weight of the worker and the equipment and material (load) that is carried as well as the strength of the ladder elements. Stability applies to the ladder providing both worker and load a stable equilibrium to support the user and the user‘s stability both in climbing and performing work. The type of ladder requires a proper selection that includes addressing the strength and stability factors described above as well as the right length corresponding to work height. Placement and protection includes the same input factors affecting stability, ensuring additional friction between ladder and support surfaces, and the ladder‘s location and positioning to protect it from external forces. The user‘s ability to stay on the ladder includes all the rules and requirements that determine the probability that the worker will complete a task without losing balance, including fitness to perform the work (e.g. no blackouts or fainting fits). A risk model that combines the height hazard with these accident scenario determinants can be made for four ladder types: placement, step, fixed, and rope. The last two do not require the input of stability determinants, thus affording limitations in the determinants to simplify the model.

111

At workplaces generally more than one scenario will be active. Research is necessary to develop rules to distinguish and rate active scenarios and to investigate the interdependency of determinants. The design analysis approach is helpful here, for example:

 Function (what): maintenance of a construction  Principle (how): painting at height, manual operation, human force as motive power  Form (by use of): ladder, user abilities, balance of worker and load, craftsmanship

To assess the operation sketched above there are three important questions to be answered; what function is carried out, how is it carried out (which principles are applied), and by use of which equipment (form), including preventive measures taken and a potential ladder selection outcome. The only mitigating factor is the energy absorbing potential of the impact surface. Here we see how falls from ladders do fit the Haddon model as a dose-response as potential energy links to heights and mass. Regarding for instance the central event ‗falling from heights‘, BB should also offer alternative solutions to Eliminate E (provide alternative to performing task at height, such as spray painting), to Reduce E (‗choose ladder‘ or ‗choose scaffold‘ (or choose a mobile elevating work platform – MEWP) commensurate to task parameters), and Isolate and Reduce Flow of E (modify or mitigate hazard) when applicable. Although the energy level does not guide you toward this end, it can be relevant in choosing between a ladder and a scaffold, as scaffolds afford additional mitigation options that task parameters may require. With the example of 'falling from heights', determinants of accident scenarios or determinants of tasks performed is the only way to link a risk classification to prevention. Determinants of the scenario falling from height are related to the quality of the ladder, the operational control method, and the motive power.

Electrical hazards are common in the construction industry and electrocution is another accident scenario with potentially severe or fatal consequences. Narrowing the focus for the sake of exploring banding scenarios, the abundant use of electrical hand tools (e.g. lighting, saws, drills, presses) in construction makes this an appropriate selection. The electrical energy is one of the hazards for hand tools and their increasing current requirements band well across four risk levels (minimizing voltage dose): < 12 V; 12 – 50 V; > 50 – 120 V; >120 V. To prevent electrical shocks with hand tools, both direct and indirect contact can be evaluated for their determinants. Direct contact with an electrical source requires a sequential consideration of insulation, obstacles, and enclosures. Indirect contact requires evaluation of grounding, electrical protection (i.e., safety switches), and double insulated apparatus. Indirect contact also includes environmental conditions including work in wet conditions and proximity to metallic parts. Selection, or absence, of these measures can be applied in isolation or in combination. When appropriately applied they can afford effective electrical engineering strategies. However, user abilities also require consideration to ensure these strategies are implemented appropriately. User abilities affect the quality of these determinants including ensuring they are properly maintained, inspected, and applied in a safe manner.

As with ‗falling from heights‘, determinants of electrical accident scenarios or of hand tool related tasks performed links a risk classification to prevention. Determinants of the electrocution scenario are related to the operational control method of the complete electrical system powering the hand tool, including extension cords and power sources, as well as motive power of both the tool and human force in its use. For both height and electricity, it is easy to understand how to use a ranking, or banding principle, resulting in a qualitative assessment. For 112

BB, scoring these two risk assessment elements should lead you to a risk level (RL), which will lead you to a risk reduction control strategy based on the ‗ladder quality‘ or the ‗hand tool quality‘. Perhaps, as in CB, hazard bands and scenario determinants can lead directly to control strategies, leaving risk assessment an implicit part of the hazard-exposure route. An example of this can be seen as part of the Construction Toolbox presented above. If work will be performed at heights with a ladder as indicated on the Pre-Job Hazard Analysis, the parameter of this task should be evaluated to ensure the appropriate control strategy would be used. By inputting the parameters outlined above, such as the height of the work, the ground on which the ladder will be place, the level of training the worker, the weight of the worker and the load carried, and the energy involved in the task at elevation, the appropriate type and strength of the ladder that will be required would be an appropriate control strategy outcome. If the task involved the use of electrical hand tools as well, then these accident scenario parameters can also be included in the same input process and common scenario considerations can be addressed once, such as the use of water in or around the task location

4.1 Addressing Barrier Banding Simplicity

In achieving accident prevention, both SMEs and developing countries are hindered by an absence of safety expertise. According to ILO, this lack of expertise is a direct cause of a large percentage of preventable accidents due to the presence of unnecessary safety risks. ILO notes that if all countries implemented simple and practical accident prevention strategies, then it would be possible to eliminate 83% of safety-related deaths and 74% of accidents each year (ILO 2005b). BB is not a one-to-one translation of CB to safety. But, as with occupational hygiene, where SMEs do not understand, or are unwilling to reduce exposure, a similar situation is present for safety. Here measurements do not complicate the introduction of controls, but there is a combination of competition with other company goals, a management system which overlooks safety, a lack of easily applicable tools which makes safety problems and their controls understandable for a non-expert, and probably a resistance against an overregulated field. Internationally, construction SMEs are especially vulnerable to accidents as enforcement of regulatory requirements are weak at best, they often perform higher risk tasks for less money, and the least expensive controls are often chosen, if chosen at all (Haslem et al., 2005; Watterson, 2007).

CB strategies like the COSHH Essentials have success with SMEs and in developing countries due in part to the simplicity of their design for reducing work-related illness in the absence of occupational hygiene expertise. The introduction of BB can be of assistance to companies who are willing to make efforts to work safely, but do not exactly know how. The number of determinants relating to accident scenarios can at times be quite extensive and utilizing all of them as part of a BB application may be perceived as a complicating factor. However, the CB Nanotool and its multiple input requirements to determine severity and probability scores, extensive application and evaluation did not find this to be a detriment to the non-expert. When they were motivated to reduce task-related risks or the costs of controls, the process of using the Nanotool taught non-experts how to mitigate or eliminate process-related risks or to modify the way tasks were performed to reduce their overall RL. This effect was also seen in the pre- planning of future projects. For construction, promotion of pre-project planning is often the focal point of an effective occupational risk management program (see also, for example, the British Construction (Design and Management) Regulations 2007 (HSC 2007).

113

There may be benefits in considering two potential models for BB that differ in their level of complexity to achieve the common end of achieving pre-project planning. The first model would be for SME managers and project planners, using hazard-based risk matrices that are more complex contain determinants as an excellent training tool. As most SMEs specialize in specific construction industry sectors, continual use of a few different hazard-specific risk matrices can assist in familiarity with project-specific factors that will increase the level of risk due to task demands or environmental factors. This can benefit managers as a contracting specification to ensure additional funds for a project with higher risk tasks and aid project planners in ensuring control strategy requirements and the appropriate level of expertise are available at the worksite. The second BB model would be for workers and be simpler to use. This simpler strategy would employ a user interface of hazard bands and scenario determinants that would lead directly to the control strategy, leaving the risk assessment an implicit part of the hazard-exposure route. An example of how this might work can be found within the structure of a Construction Toolbox framework that is currently in development (Zalk et al., 2010). The Construction Toolbox utilizes a multidisciplinary probability and severity risk matrix to determine a project‘s RL in the pre-planning phase. The RL of the project determines the level of expertise required on the worksite. This tool is called the Pre-Job Hazard Analysis Checklist and it determines the severity of a project‘s chemical, physical, and safety parameters and probability for worker exposure based on common exposure and accident scenarios prevalent in the construction industry. Chemical examples include work involving asbestos- or lead- containing materials, tasks requiring the cutting or breaking of concrete, welding and torch cutting, and the expected use of PPE. Physical indicators include repetitive tasks, heavy lifting, noisy work, tasks with vibration. Safety considerations include work with or around energized equipment, tasks performed at elevation, use of scaffolding or ladders, steel erection, excavation, and use of heavy equipment. Probability factors include the number of workers, length of project, dustiness at the jobsite, and comprehensive energy needs. Both the severity and probability scores are tabulated and the outcomes scores use a risk matrix to determine the RL. The next steps require a task-based evaluation for each task in the project. These tasks are also RL-based commensurate to the level of expertise required to properly implement the task‘s multidisciplinary controls. The simpler BB model could be used at this phase, with input factors bounded by the RL of the project and the tasks to be performed. Therefore, scenario determinants that are not factors in the project design are not necessary in the model‘s decision making. In this phase there is also an opportunity for the simpler BB model to account for tasks with RLs higher than the project‘s RL, offering options for elimination or substitution of task parameters.

Discussion

The BB approach presented takes classification of hazards and seeks to offer controls to achieve risk reduction based on accident scenario determinants. Benefits are seen at the SME level where banding strategies in safety present an opportunity to reduce accidents. A tool that can lead decision makers to make a priori determinations for the proper hazard-based safety controls to protect workers is valuable for accident prevention. The construction industry examples presented offer two theoretical BB models for SME use, a place where CB models have already had success with communicating exposure reduction approaches to the non-expert. With BB and CB as parallel risk communication models, the worker can have a comprehensive understanding of task-specific controls. BB also offers teaching opportunities for other professions. Occupational health professionals can learn a lot in improving risk communication from the 114

safety bow-tie model for presenting accidents. (Figure 2). For safety, risk prevention, or exposure elimination, it is necessary to stay left of central events.

Figure 2. Bow–tie presented in a multidisciplinary format

This left side preference can also be seen in ergonomics, where the multiple strategies for controlling musculoskeletal risk factors can be seen as left-side barriers (Zalk 2001; Zalk 2003). Occupational hygiene does not see risk prevention in the same manner. Chemical exposures are seen as the central event and in most instances their emission, the right side, is expected and elimination and substitution are the only left side components. Occupational medicine is also applied most often on the right side of the bow-tie, except for its influence on job selection and allocation which are examples of preventative actions. Therefore, when the bow-tie is viewed as a multidisciplinary model the role of barriers can be seen as a progressive process in which disciplines unite to seek prevention and mitigation of unwanted work-related injury, illness, or disease consequences.

BB can also be seen to have a number of weaknesses. First and foremost, as presented here, it still remains theoretical. Although two examples are given, they are both narrowly selected within their hazard classification in order to minimize the number of accident scenario determinants and keep things simple. For ‗falling from heights‘ and ‗electrical shock‘, if scaffolding and wiring had been chosen rather than ladders and hand tools, there would be considerably more scenario determinants to be listed. Dealing with the long list of determinants can be perceived as a more complicated process, whereas BB should strive for a more simplistic strategy comparable to CB. However, if accident scenario determinants are limited to achieve simplicity, this could also be seen as a weakness of the model and a failure to cope adequately with the complexity of reality. Further hazard classifications also need to be presented in a banding format to further develop the BB strategy as a practical field tool for SMEs. This could best be developed first for higher risk construction industry hazards associated with commonly performed tasks. Subsequently, the concerns about the level of simplicity for SME usability and about BB being only a theoretical concept could both be addressed by a practical application of the two BB models proposed.

115

Conclusions

Principles for banding in occupational safety have been presented and their potential for utility in practice has been given. Although this concept has not yet been presented in publications, the topic has been one of high interest globally as a companion for the occupational hygiene CB strategies. The principles presented remain theoretical, although a number of opportunities for further research are presented to formalize a standardized approach for occupational safety. Two examples, prevention of falls and electrocution, are given for a BB application in the construction industry and expanding the number of hazard classifications can be accomplished through research. For the examples presented, the hazard classifications have been banded for levels of risk, the determinants of accident scenarios presented, and controls for these accidents as risk reduction outcomes are given. Research is necessary to develop rules to distinguish and rate active scenarios. Investigation into the interdependency of determinants is another area of further research. Additional research is also necessary to determine the utility of banding for additional hazard classifications for accidents. Two models were presented in theory to address the need for simplicity of a BB approach. The manager BB model should be the first to receive further research as a full compliment of accident scenario determinants is not a deterrent for use based on experience with the CB Nanotool. The worker model will also require further research as the number of determinants will require limitation. There are two modes of research required here, one as a research-derived reduction of determinants that are deemed acceptable and the second for a worker interface that simplifies the entries necessary while performing the risk assessment intrinsically within the model.

The pitfalls for utilizing banding principles in occupational safety can first be compared to those by occupational hygienists in CB‘s early incarnation. The regulatory framework is seen by many as a sufficient guide to required barriers and safety risk reduction controls and BB may be seen as only confusing these requirements and even complicating adherence to them. Occupational safety experts are the best option for protecting workers and BB will never be as good as using professional expertise. This is especially true when workplace variations are present as the occupational safety profession takes these into account, whereas BB cannot account for all these. This would be especially true should safety banding principles afford the limitation of accident scenario determinants. Research to address the occupational safety professional‘s input to BB models, their strengths and limitations, the effect of limiting determinants, and accounting for workplace variations will all be necessary. In addition, research into appropriate BB model designs will assist in determining maximum utility for SMEs and developing countries as part of a multidisciplinary framework. These research parameters that need to be addressed are essential in ensuring the simplicity necessary to achieve a global reduction in accidents where expertise is in short supply.

Acknowledgements

This work, in part, was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344, LLNL- JRNL-461829.

116

CHAPTER 8 DISCUSSION AND CONCLUDING REMARKS

Thesis Answers The main thesis issue was presented in Chapter 1, however it can be summarized as two questions. Can Control Banding‘s simplified, qualitative risk assessment models be expanded beyond bulk chemicals and identify solutions comparable with quantitative models? Can multidisciplinary ORM models based on Control Banding strategies be developed to identify solutions commensurate to risk for the reduction of work-related injury and illness? The summaries of the sub-questions are presented below:

1) Beyond Bulk Chemicals: Can the application of Control Banding and Hazard Banding strategies expand beyond bulk and liquid chemicals to establish a foundation for a unified ORM model across the OSHH professions? 2) Qualitative vs. Quantitative: Can qualitative risk assessment models identify the appropriate level of risk and select the appropriate controls at a level comparable with traditional quantitative models? 3) Multidisciplinary Control Banding: Can multidisciplinary ORM models that integrate a qualitative risk assessment approach be developed within an occupational trade and within a regulatory framework?

Beyond Bulk Chemicals There appears to be an affirmative answer to this question of the applicability of applying Control Banding and Hazard Banding strategies beyond bulk and liquid chemicals, however further research remains necessary. As part of this research Control Banding through history was thoroughly reviewed and evaluated for its strengths and weaknesses. In doing so, the potential for a broad reach beyond bulk and liquid chemicals was presented and the common elements distilled to establish a foundation for how it can be applied in other OSHH professions. This research also presented original research in the development of a Hazard Banding approaches for ergonomic risk factors which led to the establishment of a framework to implement appropriate controls, the creation of an internationally accepted Control Banding method for nanomaterials, and presented the basis for a Control Banding approach in occupational safety. When the concept of this PhD was initially put forth, Control Banding was relatively new within the occupational hygiene profession. The primary established toolkit was the UK HSE‘s COSHH Essentials and its focus was on controlling chemical exposures relating to professions that utilize bulk liquids and solids as part of a production process. This quite limited parameter hampered initial Control Banding acceptance. Many occupational hygienists believed COSHH Essential and Control Banding were synonymous and therefore Control Banding was simply too narrow to be useful and too simplified to be effective. In order to best assess this belief, a thorough review of the existing Control Banding-related literature was a necessary foundation for any proposed future efforts. This review established the strengths and weaknesses of these Control Banding strategies, approaches, and techniques that were originally geared toward the reduction of occupational bulk chemical exposures. This evaluation of the research relating to the COSHH Essentials and related toolkits also assisted in the development of qualitative risk assessments to reduce work-related exposures and determine a future path for other OSHH-related models. As these toolkits focus on bulk chemical use, a further investigation was necessary to identify the opportunities and limitations for addressing point source exposures. By extracting information in this regard, the literature review assisted in the distillation and standardization of the components that are within the construct of the most successful models. These components are the collective 117

tools of practical primary prevention and include Control Banding strategies, task-to-control methods, the solutions initiatives, and participatory processes. Taken together, these are the tools developed by OSHH experts for the non-experts, as opposed to the early Control Banding methods that were developed for use by experts.

Tools for the non-experts have a difficulty in clarifying precisely the level at which they are intended to be applied as both SME managers as well as workers can be seen as a non-expert. When participatory approaches are intrinsic to the tool, it is clear that workers are intended as primary users. Utilizing solutions initiatives and participatory processes forms a logical bridge, since these are tools developed by experts directly for workers as the non-experts, as can be seen in uniting Hazard Banding to control processes to achieve prevention of occupational exposure and reduction of work-related risk for ergonomics (Chapter 3). A review of the existing Ergonomics literature revealed a wealth of research that supports the use of participatory methods as an integral part of solutions initiatives. This highlights that the development of practical approaches to identify and reduce occupational risks can be created directly for worker use and that similar simplified strategies can also be developed to reduce chemical exposures. The Ergonomics literature review therefore provides a learning opportunity for the occupational hygiene profession. Successful applications of qualitative and semi-quantitative tools for assessing and reducing risks to MSDs are maximized when Hazard Banding leads to utilizing participatory methods and including workers in all aspects of multidisciplinary risk assessment, ORM, and exposure control. Offering participatory occupational hygiene in conjunction with existing practical primary revention models also serves to present a foundation in how to achieve sustainable effectiveness in a manner comparable to traditional quantitative models.

Qualitative vs. Quantitative There are many instances identified that can positively answer the question whether qualitative risk assessment models identify the appropriate level of risk and select the appropriate controls at a level comparable with traditional quantitative models, however there remains additional needs for the evaluation and validation of Control Banding strategies. This research revealed a primary strength for Control Banding for affording controls in the absence of OELs, which was agreed to by NIOSH, ACGIH, and many other international experts. The research basis for Control Banding is showing the potential to be effective in line with quantitative models and to offer a simplified control of worker exposures when there is an absence of firm toxicological and exposure information. The range of Control Banding‘s strategies has expanded amazingly well in both breadth and depth. It has grown beyond its initial role as a toolkit for SMEs and has begun to gain a level of profession acceptance as an essential complement for field occupational hygiene practitioners utilizing both qualitative and quantitative approaches while working with limited budgets and staff (AIHA 2007, NIOSH 2009). However, the development and implementation of practical and simplified prevention strategies for the reduction of occupational risk factors still requires a quantitative validation process; yet this process has also grown well in the past few years. This professional review of existing qualitative and semi-quantitative occupational risk assessment strategies includes statistical comparisons between the outcomes from these models, with COSHH Essentials and Stoffenmanager as the primary examples, and those from the traditional scientific quantitative methods. In addition, what had once been limited efforts to critique the existing toolkits and to validate the Control Banding strategies for effectiveness has now become a steady stream of national and international publications. These publications address all the essential aspects originally outlined in this thesis and include the evaluation of emerging toolkits and Control Banding strategies, the expansion of the range of practical approaches to identify and reduce occupational risk in line with quantitative risk 118

assessment approaches as part the implementation and evaluation of comprehensive Control Banding programs (Bracker et al. 2009. Fingerhut 2008, NIOSH 2009, ACGIH 2008, AIHA 2007).

The nanotechnology industry best addresses the thesis sub-question as it has posed a significant challenge to the quantitative paradigm in occupational hygiene, as there are overwhelming uncertainties of work-related health risks posed by nanomaterials. By creating, evaluating, and validating the Control Banding Nanotool as part of this thesis, a solution for these issues has been an achievement for qualitative risk assessment strategies as it helped to overcome the significant limitations seen with the traditional, quantitative approach to risk assessment and personal exposure to nanoparticulate. The current international use of the Control Banding Nanotool as part of national programs in Canada and South Korea and its consideration for a national Control Banding regulation in Australia reflects on the acceptability of Control Banding approaches within these countries‘ scientific communities (Safe Work Australia 2010; Zalk and Paik 2010d). This success can be seen in not only providing a commensurate level of control as determined from the level of risk for an evaluated task, it is also a facilitation to ensure the allocation of resources to the activities that most need them by focusing the need for quantitative measurements within specific parameters. Although the Nanotool does not appear prima facie to be developed as a participatory process, it has in great part been implemented as one. Key to this understanding is that the Nanotool is developed to protect researchers and, in most instances, they are the workers utilizing this approach. However, it is important to note that nanomaterial researchers have a higher educational background then, for example, workers in the manufacturing sector. While researchers tend to be experts in their field of interest, they are likely to be non-experts with respect to knowledge and tools use by OSHH professionals. In addition, this initiation within a participatory process has created a risk communication with these non-experts, which is now firmly established. In contrast, with only quantitative measurements these risk discussions with the non-expert workers might not exist. It is this effective and invaluable risk communication with the non-expert workers that was an unintended consequence of utilizing Control Banding strategies.

Multidisciplinary Control Banding The multidisciplinary ORM models presented can integrate qualitative risk assessment approaches and be developed within an occupational trade and within a regulatory framework provide an optimistic outlook for future research, however even with an optimistic outlook for future research much work remains necessary. The firm foundation for, and acceptance of, Control Banding that has been developed presents an excellent opportunity for putting these strategies together across the OSHH professions. Integration of these multidisciplinary OSHH approaches into an ORM process, derived from the distillation of toolkit models, has aided in the development of comprehensive practical primary prevention approaches for the identification and reduction of worker exposures to known occupational health and safety risk factors. An example of this comprehensive OSHH strategy was found in the established trades that perform their work within the construction sector. By focusing on uniform work categories within the construction industry, the individual chemical, physical, and safety risk assessment strategies and models were unified and applied within the framework for a Construction Toolbox. Construction is a uniquely appropriate trade to focus on for this thesis because it is performed throughout the world with approximately 85% employers being SMEs. Therefore, the development of a unified Control Banding strategy that brings together the tools across the practical primary prevention spectrum and across the OSHH professions has given an opportunity to unite the wealth of solutions-based initiatives and participatory approaches with a depth of control-oriented research that until now 119

remained formatted within individual OSHH professional disciplines. Utilizing an international panel of experts to develop the Construction Toolbox framework, an opportunity has been given to SME managers to better protect their workers from both injuries and illnesses common to this industry sector. In addition, the need for implementation of this simplified ORM process derived by experts will aid in making it possible for other common international trades to have similar Control Banding toolboxes created. This outcome does well in establishing a practical primary prevention strategy that can positively affect health and safety programs that also need to address standardized tasks and trades within other well-established industrial sectors.

Achieving an overall effectiveness within comprehensive OSHH programs also requires joining the scientific aspects of practical prevention with more traditional ORM processes. When taken together, a path forward is then established so that a more formalized occupational health and safety management systems (OHSMS) can be created. This multidisciplinary ORM model, when combined with Control Banding‘s simplicity, has assisted in achieving a primary prevention of work-related diseases, illnesses, and injuries over time by utilizing proactive indicators rather then passive. The cumulative goal of this macro focus on the most common industry work- related exposures was to develop a Control Banding model that can be effectively integrated within a highly regulated working environment. Although the overall effectiveness of the Risk Level Based Management System (RLBMS) has shown that comprehensive ORM approaches have proven to be effective for both OSHH professionals and workers when they work together in parallel, it should be noted that it was developed by and for a US national research laboratory and has not yet been tested in a manufacturing enterprise. The RLBMS was effective as a platform for risk communication that was essential in the highly regulated oversight at the Lawrence Livermore National Laboratory and assisted in their development of this unique OHSMS process. However, there may therefore be limitations in directly applying the RLBMS to other large enterprises. The ability of utilizing Control Banding strategies within a framework that includes extensive regulatory oversight reflects on the positive impacts on the non-expert workforce that qualitative risk assessment approaches can bring to larger industrial sectors and enterprises. This Control Banding OHSMS model has also shown the potential role that qualitative risk assessment can play in reducing occupational exposures utilizing practical ORM processes. The RLBMS, in a global context, is an example of how practical primary prevention assists in ensuring that appropriate controls are implemented across all sizes of industry. When successful Control Banding methods can be applied in such a complex working environment, the RLBMS standardized framework affords an excellent opportunity for these strategies to be further developed by multiple national entities to be shared, implemented, evaluated, and re- evaluated for their effectiveness to benefit an expanded range of workers globally.

One key challenge for Control Banding in achieving a true multidisciplinary approach was in its application into the occupational safety profession. Barrier Banding explores the possibilities in banding of accident scenarios to make the decision-making processes simpler for SMEs to make the appropriate control selections. Although both the construction toolbox and the RLBMS address safety in their multidisciplinary framework, neither addresses specifically the principles for banding in occupational safety into parallel approaches. However, using the basic stratification of prevention strategies used in Control Banding‘s simplified risk assessment approaches has shown much promise for the development of comparable injury prevention tools and methods. Barrier Banding was developed in a manner that implements barriers and manages solutions in a simplified manner, similar to controlling chemical exposures, in such as way as ultimately to reduce occupational risk for work-related injuries. It is intended that this enhanced approach to reduce occupational risk for work-related injuries is another practical prevention 120

approach that must also be evaluated for its ability to achieve effectiveness in a manner comparable to traditional quantitative models. Engaging injury prevention into this context is essential for the integration of true multidisciplinary simplified risk assessment models for multiple occupational health and safety risk factors.

Control Banding’s Growth The central question asked whether Control Banding‘s simplified, qualitative models and solutions initiatives can be integrated within the multidisciplinary requirements of ORM to achieve a practical primary prevention to address the appropriate controls are implemented commensurate to risk, to assist in reducing work-related injury and illness. Taken together, the sub-questions combine optimistically to the overall question, but much research remains and only time can give solid evidence of Control Banding‘s ability to reduce work-related injury and illness. However, evidence is beginning to compile that trends are tending toward this direction. As these collective qualitative risk assessment and risk management strategies are primarily designed to target small businesses and can now address a variety of work-related risks, a status check of Control Banding effectiveness is essential. Toolkit approaches for chemicals now include specialized strategies for various chemicals, silica, asbestos, allergens causing bakers‘ asthma, and dermal exposure. Sector specific toolkits and toolboxes in development include Construction, Health Care Workers, and Health Care Wastes. Ergonomic and participatory methods have been employed and are being considered in the development of an international Hazard Banding approach and basic models covering a wider range of professions that soon will include agricultural ergonomics. Control Banding developmental efforts and implementation can now be seen in countries throughout the world. This goes beyond the US and the EU, and now includes: Singapore, India, China, South Africa, Korea, Colombia, Brazil, Peru, Chile, Mozambique, Australia, and Canada to name a few and collaborative initiatives are in progress throughout Latin America, Africa, and Asia (Fingerhut 2008). A growth curve has also been seen in developing country university programs that are emphasizing occupational hygiene in both safety and occupational health programs. This process is beginning to see a growing number of trained and qualified OSHH professionals and appropriately equipped technicians worldwide. In part, Control Banding strategies have served in this growth as both a fuel to spread the need for occupational hygiene‘s prevention of work-related disease and illness and a function of offering a simplified risk assessment approach for students and graduates to use immediately in applying their training directly into the workplace. Most importantly, by having graduate programs within non-EME countries, the growing professional base is remaining within these countries as practitioners. This serves to develop the homegrown sustainability that was elusive with students going to the US and EU to learn, but has always been deemed essential for the growth of the OSHH professions where they are most needed. This cumulative effort has assisted in maximizing available resources to synergistically address the global burden of work-related disease, illness, and injury (Fingerhut 2008).

Recommendations and Further Research As numerous components of Control Banding approaches have been presented, it is important to discuss each of its components as part of a practical primary prevention program. The core component is simplicity and both Control Banding models and solutions initiatives seek to achieve this. Control Banding is optimized when work-related risks are practically addressed within an easy to use four-by-four qualitative risk matrix that leads directly to an appropriate control. This simplicity is best emphasized when these tools, designed for non-experts, can be applied at the worker level. In one aspect, achieving an application at the worker level is accomplished when participatory processes with the worker are either included, or better yet 121

integral, within the Control Banding model. In another aspect, the creation of a Control Banding model that can be both used and implemented by workers is the most optimal overall practical primary prevention approach. However, this second aspect is more illusory since Control Banding approaches as designed cannot always take into account the educational level of workers in certain sectors. In Chapter 5 we see how the multidisciplinary needs of the construction sector require all the Control Banding approach components to be included: qualitative risk matrices, task-to-control approaches, solutions initiatives, and participatory approaches. The latter is qualified by the limitations inherent within the range of worker qualifications required for each of the construction sector professions. Although the construction toolbox approach is designed for SME managers, there are certainly many workers who are indeed capable of using this process based on the simplicity inherent in the model‘s design. However, integral to this design is the need for higher levels of expertise as the risk level of the task increases. The requirement of professional expertise at the highest level of risk can therefore be seen as a fundamental baseline for all Control Banding approaches.

Although the number of Control Banding models continues to increase, prospective research on these models is still inadequate. It is difficult to quantify Control Banding‘s ability to reduce occupational exposures and work-related injury and illness if operational analyses are not performed for these models as implemented. In addition, the R-phrases that are integral to an expanding number of Control Banding chemical models remain unevaluated. This evaluation process is essential if SMEs are expected to rely on their control outcomes beyond their use as qualitative screening tools and for work in the absence of OELs. The expansion of Control Banding beyond chemical exposures has shown tremendous progress, however formal qualitative risk assessment toolkits for ergonomics and safety have yet to be developed. Further, as mentioned in Chapter 1, a true banding of ergonomic exposure scenarios in a manner comparable to occupational hygiene has yet to be fully developed in literature. Barrier Banding has presented both the potential for banding in safety in both theory and practice, however further development and implementation will require research projects to further refine the techniques and the toolkits to make the approach readily accessible to SMEs.

The validation of qualitative risk assessment models continues to expand in literature. This need for the growth of validation processes will continue as an ongoing research requirement. However, an international standardization of validation parameters and national support for these initiatives will be essential as the growth of Control Banding accelerates. Positive steps have been seen for Control Banding models toward this direction with their implementation and evaluation in support of validating their use as REACH Tier 1, and soon to be Tier 2, models (Zalk and Heussen 2010). Arguably, the Control Banding Nanotool has been an essential step in providing for professional acceptance of the potential of qualitative risk assessment strategies in selecting controls to be comparable to, or better than traditional quantitative models. Further research is underway for Control Banding validation to control nanomaterial exposures at the international level, including as national regulation, however fruition from these efforts remains in progress. The standardization of toxicological parameters remains an issue for nanoparticle research at many levels, however the need to provide safety data sheets with nanomaterial severity rankings for use in the CB Nanotool will go further in protecting workers in this industry. In many ways this can be seen as Control Banding influencing the direction of future nanotoxicological research.

The work in developing the construction toolbox has come far in showing the potential for multidisciplinary collaboration to address the needs of a very dangerous industry; however much 122

work remains. Validation and verification of this toolbox approach remains a research priority and this will require an evaluation of its practicality for integrating into regulatory compliance in a manner comparable to the RLBMS. Although the process for validating the RLBMS is more robust, because its implementation in an industry with increased regulatory oversight requires this, the further validation of RL outcomes and effectiveness in application remains necessary. In addition, the variability of national regulations leads to additional research needs to evaluate the ability for the RLBMS framework to apply in addressing enforcement requirements and the acceptance of qualitative to quantitative OHSMS models in relation to enforcement.

Lacking for many of these research recommendations is a consistent method for funding these initiatives both nationally and internationally. To date, much of the advancement on the topic of Control Banding has been due to the tireless efforts of a handful of volunteers. Although international organizations like the WHO, ILO, and ISO are emphasizing the need for a centralized database of Control Banding models and their input parameters, they are also requesting proof of validation and effectiveness. Making these requests in the absence of funding for the necessary research efforts and the staff to oversee these processes renders these requirements closer to a panacea than a potential reality. The outcomes of this thesis have been accomplished in the absence of this necessary funding and oversight. Unfortunately, the Control Banding frameworks developed by this research have the potential for being stifled without some continuous stream of funding to achieve the needs of the international community of OSHH professionals in conducting research to validate them. The CB Nanotool, Construction Toolbox, and RLBMS all have presented the opportunity for extensive reach to workers internationally. This reach to workers can be seen with the integration of participatory methods as practicable as well as in their simplicity in use, however as previously noted this simplicity is relative as there is a broad range of expertise found within different work sectors. The specific needs for a centralized database to both collect and distribute information have been clearly delineated in the previous chapters. It is perplexing that this funding remains elusory even though the need for these Control Banding strategies is unanimously called for by international and national entities as well as reflected in the statistics of the global burden of work-related injury and illness. The US has a unique opportunity to play a key role in addressing these needs through the development of their own national regulatory framework for Hazard Banding and, eventually, Control Banding. The EU REACH regulations have presented an impetus for this US role through the new, burgeoning requirements for international trade and transport of chemicals. This situation also affords US the opportunity to address a longstanding need for an update to occupational and environmental regulations and, in this process, address the challenge of utilizing a singular Hazard Banding framework to create a unified environmental and occupational risk management model. It is the hope of this author that this thesis serves to present the cumulative foundation necessary to progress the international dialogue to obtain the necessary funding and oversight to a achieve this overarching future vision for the unification of the OSHH and environmental professions. Otherwise, the billions of workers that will never meet an OSHH professional will have the deck stacked against them, facing work-related exposures and risks that are by and large preventable. In particular, the highest risk activities, with the most devastating consequences for individuals, families, communities, national economies, and common sensibilities, can be drastically reduced if not avoided altogether.

123

APPENDIX A Application of a pilot control banding tool for risk level assessment and control of nanoparticle exposures Annals of Occupational Hygiene 52(6):419 – 428 (2008) Samuel Y. Paik1, David M. Zalk1*, and Paul Swuste2 1Lawrence Livermore National Laboratory, P.O.Box 808, L-871 Livermore CA 94551, US 2Delft University of Technology - Safety Science Group, PO Box 5015, Delft 2600 GA, NL *Author to whom correspondence should be addressed. Email [email protected]

Abstract

Control Banding (CB) strategies offer simplified solutions for controlling worker exposures to constituents that are found in the workplace in the absence of firm toxicological and exposure data. These strategies may be particularly useful in nanotechnology applications, considering the overwhelming level of uncertainty over what nanomaterials and nanotechnologies present as potential work-related health risks, what about these materials might lead to adverse toxicological activity, how risk related to these might be assessed, and how to manage these issues in the absence of this information. This study introduces a pilot CB tool or ‗CB Nanotool‘ that was developed specifically for characterizing the health aspects of working with engineered nanoparticles and determining the level of risk and associated controls for five ongoing nanotechnology-related operations being conducted at two Department of Energy (DOE) research laboratories. Based on the application of the CB Nanotool, four of the five operations evaluated in this study were found to have implemented controls consistent with what was recommended by the CB Nanotool, with one operation even exceeding the required controls for that activity. The one remaining operation was determined to require an upgrade in controls. By developing this dynamic CB Nanotool within the realm of the scientific information available, this application of CB appears to be a useful approach for assessing the risk of nanomaterial operations, providing recommendations for appropriate engineering controls, and facilitating the allocation of resources to the activities that most need them.

Key words: nanotechnology, nanoparticle, nanomaterial, control banding, risk assessment, risk level, exposure control, toolkit, CB Nanotool.

Introduction

The traditional industrial hygiene (IH) approach to controlling exposures to harmful particles in the workplace is to measure the air concentrations of the particles of interest from the worker‘s breathing zone, compare those concentrations to exposure limits determined for those particles, and implement control measures to reduce concentrations below the exposure limits. This assumes the following: 1) the sampled concentrations are representative of what the worker is actually breathing; 2) the appropriate index of exposure is known; 3) analytical methods are available to quantify that index; and 4) the exposure levels at which those particles produce adverse health effects are known. If any of these is not well characterized, the measurements taken may have limited value as it would be difficult to perform a valid risk assessment. In addressing worker exposures to engineered nanoparticles, the first requirement can be satisfied by obtaining an air sample from the worker‘s breathing zone using a sampling pump, where forces such as particle inertia and gravity have minimal impact on the ability of the nanoparticles (defined as having 2 or 3 dimensions less than 100 nanometers (ASTM, 2007)) to follow the sampled air into the sampler since nanoparticles approach molecular size. The second 124

requirement - an appropriate index of exposure - has not yet been satisfied for nanoparticles with no international scientific community consensus on what the relevant index of exposure is (NIOSH, 2006; ISO, 2007). For example, a number of studies are suggesting that total surface area concentration may be a better exposure index than mass concentration (Oberdorster et al., 1994; Tran et al., 2000). Particle number concentration has also been suggested as an alternative to mass concentration (NIOSH, 2006). This lack of consensus directly affects the third requirement, since sampling and analytical methods rely on knowledge of what needs to be measured. Commercially available instruments can measure surface area concentration, number concentration, or mass concentration, but these generally measure larger particles in addition to nanoparticles, introducing potentially large biases (summarized in ISO, 2007 and NIOSH, 2006). For example, both the CPC Model 3007 (TSI, Shoreview, MN), which measures particle number concentration, and the Model 3550 Nanoparticle Surface Area Monitor (TSI, Shoreview, MN), do not have cut-offs at the upper limit of what is defined as a nanoparticle. The fourth requirement may be the largest barrier to assessing the risk of working with nanomaterials. Very little toxicological data for determining exposure limits for nanoparticles, and virtually no human studies, are available (Maynard and Kuempel, 2005). This is due to the lack of consensus on the appropriate index of exposure and the relative novelty of nanotechnology and the new materials used in this technology. Therefore, there are numerous barriers to overcome before traditional IH can produce useful data.

A plausible alternative to the traditional IH approach is the utilization of control banding (CB). Control Banding (CB) strategies offer simplified solutions for controlling worker exposures to constituents that are found in the workplace. Historical progression has shown that CB is a framework for managing occupational risks in the face of uncertainty (summarized in Zalk and Nelson, 2008 and Money, 2003). The CB concept developed by the U.K. Health and Safety Executive (HSE) in 1999 as the COSHH Essentials model (HSE, 1999; Oldershaw, 2001), has seen widespread use in the U.K. and elsewhere. CB makes business sense because chemical companies are constantly synthesizing new chemicals, and developing occupational exposure limits for all experimental chemicals is not feasible as most will never become commercialized. This very aspect of decision-making based on incomplete information makes CB an attractive option for controlling nanoparticle exposures.

Like its counterparts in the pharmaceutical and microbiological industries, the nanotechnology industry also has to achieve a risk management program with an insufficient basis for traditional IH quantitative risk assessment approaches. While nanotechnologies show incredible promise in such areas as materials science, cancer treatment, and environmental remediation, they have created a heightened level of concern for research and development (R&D) and manufacturing workers due to the overwhelming level of uncertainty over what nanomaterials and nanotechnologies present as potential work-related health risks, what about these materials might lead to adverse toxicological activity, how risk related to these might be assessed, and how to manage these issues in the absence of this information (Maynard, 2007). In theory, CB has been proposed as a practical approach to address exposure to nanoparticles and achieving exposure control in the absence of this data (Zalk and Nelson, 2008; Schulte et al., 2008; Maynard, 2007; and Nelson et al, 2007). A conceptual CB model was presented by Maynard (2007) which offers the same four control approaches of the COSHH Essentials model as stratified by corresponding ‗impact‘ and ‗exposure‘ indices. This model combines engineered nanomaterial composition parameters (shape, size, surface area, and surface activity) with their exposure availability (dustiness and amount in use) and links these indices to bands with corresponding control 125

approaches. This model is presented in a historical progression of pragmatic approaches to exposure control considered a complement to traditional IH risk assessment.

Objective

While CB appears to be an appropriate methodology for controlling exposures to nanomaterials in concept, very few, if any, comprehensive tools are currently available for ongoing nanotechnology operations. The goal of this study, therefore, was to further explore the feasibility of using CB for controlling exposures to nanomaterials by developing and introducing a pilot CB tool or ‗CB Nanotool‘ based on existing knowledge of nanomaterial toxicology and utilizing the CB framework proposed in earlier publications. As part of this effort, the CB Nanotool was used to determine the risk and controls associated with five ongoing operations at two Department of Energy (DOE) research laboratories.

Methods

This study can be divided into two phases: 1) development of the CB Nanotool for nanotechnology operations; and 2) application of the tool to determine risk levels and controls for five different operations.

Development of the CB Nanotool for nanotechnology operations Maidment (1998) stressed the importance of limiting the number of factors in the CB model to reduce its complexity and increase its applicability for non-experts. To achieve this balance of simplicity and effectiveness, Maidment suggested four categories, or ―bands‖, to assist in preventing exposure to chemicals. These four control strategies are a grouping of three levels of engineering controls based on sound IH principles, with professional IH expertise as a fourth category. The control band for a particular operation is based on the overall risk level (RL) determined for that operation. The RL is determined by a ‗severity‘ score and a ‗probability‘ score, which are analogous to the ‗impact‘ index and ‗exposure‘ index described in Maynard (2007). The biggest challenge in developing any pilot CB tool is deciding how these scores are to be determined. Fig 1 provides the matrix for overall RL determination.

Figure 1. Risk level (RL) matrix as a function of severity and probability. Control bands are based on overall RL.

126

This matrix is similar to that used in the implementation of CB through the HSE‘s COSHH Essentials program (HSE, 1999; Garrod and Rajan-Sithamparanadarajah, 2003); however, for simplicity, it contains one less column and row in line with comparable model development parameters (Maidment 1998). It should be noted that for several of the factors described below, 0 points were assigned to the lowest rating for a given factor. This does not in any way imply that no adverse health effects are anticipated at these levels; the 0 points were assigned as an indication of low ‗relative‘ severity or probability.

Severity Determination It was anticipated early in the development of this tool that for many of the factors that are considered important for determining the severity score, the information for that factor would not be known due to the reasons stated above. While the most conservative approach would be to treat an unknown hazard as equivalent to a high hazard, the authors felt this was over- conservative and would likely place an unnecessary burden on those managing the work. For this reason, it was decided that when the information for a given factor was ―Unknown‖, 75% of the point value of ―High‖ would be given for that factor. What this translates to is that for a hypothetical nanotechnology operation for which nothing was known (other than it involves nanoparticles), the resulting RL would be ―RL3‖ and the required control would be ―Containment‖. In this scenario, if just one rating for any of the factors was later determined to be ―High‖, with all other ratings remaining as ―Unknown‖, the tool would assign this activity as ―RL4‖ and require the maximum control.

Based on what is known about the toxicological effects of nanoparticles in the current literature, the authors believe the following are factors that should be considered in determining the overall severity of the nanoscale materials. While it is recognized that different groups may disagree on what the most important factors are, the intent of the CB Nanotool was to account for all the major factors that the current literature suggests is important in determining nanomaterial toxicity. These factors influence the ability of particles to reach the respiratory tract, their ability to deposit in various regions of the respiratory tract, their ability to penetrate or be absorbed through skin, and their ability to elicit biological responses. It was recognized that particles entering the respiratory tract can cause adverse effects by remaining in the respiratory tract (primarily the lungs) or by entering the blood circulation.

1. Surface chemistry. Surface chemistry is known to be a key factor influencing the toxicity of inhaled particles (Maynard and Kuempel, 2005). Crystalline silica, for example, elicits a much stronger response than titanium dioxide, even when normalized for surface area or mass. Particle surface free radical activity is the primary factor that influences the material‘s overall surface reactivity. Research studies should be consulted, when available, to make a judgment of whether the surface reactivity of the nanomaterial is high, medium, or low. For example, free radical activity is associated with the generation of reactive oxygen species and oxidative stress responses in the lungs. Reactive oxygen species and oxidative stress responses can be quantified by analyzing the bronchoalveolar lavage fluid (BALF) from rats used in toxicological studies. The BALF may be analyzed for markers of inflammation, levels of pulmonary oxidants, antioxidant status, and markers of lung tissue damage (Albrecht et al., 2005). These types of information need to be consulted in determining the surface reactivity of the nanomaterial. A rating of ―High‖ results in 10 points; a rating of ―Medium‖ results in 5 points; a rating of ―Low‖ results in 0 points; and a rating of ―Unknown‖ results in 7.5 points. 127

2. Particle shape. Studies have shown that exposure to fibrous particles like asbestos have long been associated with increased risk of fibrosis and cancer (Doll, 1955). Tubular structures, like carbon nanotubes, have also been shown to cause inflammation and lesions in rat lungs (Lam et al., 2004). Based on this information, the highest severity score is given to fibrous or tubular-shaped particles. Particles with irregular shapes (other than tubular or fibrous) are given a medium severity score because they typically have higher surface areas relative to isotropic (e.g. compact or spherical particles) particles. A rating of ―Tubular or fibrous‖ results in 10 points; a rating of ―Anisotropic‖ results in 5 points; a rating of ―Compact or spherical‖ results in 0 pts; and a rating of ―Unknown‖ results in 7.5 points.

3. Particle diameter. Based on the particle deposition model developed by the International Commission of Radialogical Protection (ICRP, 1994), particles in the 1-10 nm range have a greater than approximately 80% chance of depositing in the respiratory tract. Particles in the 10-40 nm range have a greater than approximately 50% possibility of depositing in the respiratory tract and particles in the 41-100 nm range have a greater than approximately 20% possibility of depositing in the respiratory tract. Since deposition is the first step in producing potential adverse health effects, regardless of which region of the respiratory tract the particles deposit in, the severity score was based on the particles‘ ability to deposit anywhere in the respiratory tract. Based on this modeling, a rating of ―1-10 nm‖ results in 10 points; a rating of ―11-40 nm‖ results in 5 points; a rating of ―<41-100 nm‖ results in 0 points; and a rating of ―Unknown‖ results in 7.5 points.

4. Solubility. A number of studies have shown that poorly soluble inhaled nanoparticles can cause oxidative stress, leading to inflammation, fibrosis, or cancer (Castranova, 1998; Donaldson et al, 1998). Since soluble nanoparticles can also cause adverse effects through dissolution in the blood, severity points are assigned to soluble nanoparticles as well, but to a lesser degree than for insoluble particles. A rating of ―Insoluble‖ results in 10 points; a rating of ―Soluble‖ results in 5 points; and a rating of ―Unknown‖ results in 7.5 points.

5. Carcinogenicity. Points are assigned based on whether the nanomaterial is carcinogenic or not, regardless of whether the material is a human or animal carcinogen. Very few nanomaterials (e.g., titanium dioxide) have been identified as potential carcinogens (IARC, 2006). A rating of ―Yes‖ results in 7.5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 5.625 points.

6. Reproductive toxicity. Points are assigned based on whether the nanomaterial is a reproductive hazard or not. This information is not readily available for most nanomaterials. A rating of ―Yes‖ results in 7.5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 5.625 points.

7. Mutagenicity. Points are assigned based on whether the nanomaterial is a mutagen or not. This information is not readily available for most nanomaterials. A rating of ―Yes‖ results in 7.5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 5.625 points.

128

8. Dermal toxicity. Points are assigned based on whether the nanomaterial is a dermal hazard or not. This is understood to encompass both dermal absorption and cutaneous toxicity. This information is not readily available for most nanomaterials. A rating of ―Yes‖ results in 7.5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 5.625 points.

9. Toxicity of parent material. The bulk materials of some nanoparticles have established occupational exposure limits. While it is known that the toxicity of particles at the nanoscale can differ significantly from their larger counterparts, this provides a good starting point for understanding the toxicity of the material. Points are assigned according to the OEL band of the bulk material. A rating of ―0-10 mg/m3‖ results in 10 points; a rating of ―11-100 mg/m3‖ results in 5 points; a rating of ―>100 mg/m3‖ results in 2.5 points; and a rating of ―Unknown‖ results in 7.5 points.

10. Carcinogenicity of parent material. Points are assigned based on whether the parent material is carcinogenic or not. A rating of ―Yes‖ results in 5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 3.75 points. The National Toxicology Program, International Agency for Research on Cancer, and the American Conference of Governmental Industrial Hygienists provide lists of suspected and confirmed human carcinogens.

11. Reproductive toxicity of parent material. Points are assigned based on whether the parent material is a reproductive hazard or not. A rating of ―Yes‖ results in 5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 3.75 points.

12. Mutagenicity of parent material. Points are assigned based on whether the parent material is a mutagen or not. A rating of ―Yes‖ results in 5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 3.75 points.

13. Dermal hazard potential of parent material. Points are assigned based on whether the parent material is a dermal hazard or not. As stated before, this is understood to encompass both dermal absorption and cutaneous toxicity. A rating of ―Yes‖ results in 5 points; a rating of ―No‖ results in 0 points; and a rating of ―Unknown‖ results in 3.75 points.

A number of studies show that the particle surface area is closely associated with lung responses, including tissue damage and inflammation in rat lungs (Oberdorster et al., 1994; Tran et al., 2000). This factor is accounted for by assigning higher severity scores to smaller particles (which would have a higher surface area compared to larger particles at the same mass concentration) and anisotropic particles (which generally would have higher surface-to-volume ratios). This factor is also accounted for by assigning higher probability scores to operations that have higher ―dustiness‖ levels (see next section), which would invariably have higher overall surface area concentrations relative to operations with lower dustiness levels.

The overall severity score is determined based on the sum of all the points from the severity factors. The maximum score is 100. Since nanoparticles usually behave much differently than their parent material due to their small scale, which is what makes engineered nanoparticles so useful and potentially much more toxic, greater consideration 129

was given to the nanomaterial characteristics (70 possible points out of 100) than to the parent material characteristics (30 possible points out of 100). Since the parent material and nanomaterial are both considered in determining the severity score, it should be understood that the parent material ratings should not influence the ratings that are given for the same factor at the nanoscale (e.g., carcinogenicity), i.e., each factor should be rated independently of another. An overall severity score of 0-25 was considered low severity; an overall severity score of 26-50 was considered medium severity; an overall severity score of 51-75 was considered high severity; and an overall severity score of 76- 100 was considered very high severity.

Probability Determination In order to determine a probability score that can be combined with the severity score to determine the overall RL of the operation, the authors believe the following factors should be considered when determining the overall probability score. These factors determine the extent to which employees may be potentially exposed to nanoscale materials. The probability score is based on the potential for nanoparticles to become airborne. This primarily affects exposure by inhalation; however, it also influences the potential for dermal exposure because the likelihood of skin contact with the nanomaterials increases with more nanoparticles becoming airborne and depositing on work surfaces.

1. Estimated amount of nanomaterial used during task. When all else is constant, the amount of the nanomaterial used during an operation increases the likelihood of the material being available to interact with the user. For nanomaterials embedded on substrates or suspended in liquids, the amount should be based only on the nanomaterial component itself, not to include the substrate or liquid portion. Therefore, points are assigned based on the total amount of nanomaterial used during a single operation. A rating of ―>100 mg‖ results in 25 points; a rating of ―11-100 mg‖ results in 12.5 points; a rating of ―0-10 mg‖ results in a rating of 6.25 points; and a rating of ―Unknown‖ results in 18.75 points.

2. Dustiness/mistiness. Since employees are potentially exposed to nanoparticles in either dry or wet form, this factor encompasses both dustiness and/or mistiness of the nanomaterial. For the same mass concentration, however, non-agglomerated dry nanoparticles should be given a higher dustiness/mistiness rating than agglomerated or liquid-suspended nanoparticles. While not required, quantitative measurement devices would be particularly useful in determining the dustiness/mistiness level. A condensation nuclei counter that provides number concentration, for example, would provide insight into the overall dustiness level. Knowledge of the operation (e.g., handling dry powders versus liquid suspensions of nanoparticles) and observation of work surfaces (e.g., cleanliness of surfaces pre- and post- handling of nanomaterials) would be another means to qualitatively estimate dustiness/mistiness. Due to the size of nanomaterials, visibility may not a reliable means to estimate overall dustiness/mistiness. Until further guidance is provided on the appropriate means to quantify exposure to nanoparticles, points will be assigned based on an estimate of ‗relative‘ dustiness/mistiness level. One design feature of the CB Nanotool is that a rating of ―None‖ for dustiness/mistiness level (and only for this factor) automatically causes the overall probability score to be ―Extremely Unlikely‖, regardless of what the other probability factors are, since the other factors will not be relevant if no dust or mist is being generated. Examples of operations that would result in a ―None‖ rating are handling of carbon nanotubes embedded on fixed substrates and 130

working with non-agitated liquid suspensions. This feature was specifically incorporated into the tool for this reason and represents the only departure from the ‗rules‘ that govern the tool. The dustiness/mistiness factor is the most important one in determining the overall probability score, and as such, relatively high numbers of points are assigned to the ratings in this category. A rating of ―High‖ results in 30 points; a rating of ―Medium‖ results in 15 points; a rating of ―Low‖ results in 7.5 points; a rating of ―None‖ results in 0 points; and a rating of ―Unknown‖ results in 22.5 points.

3. Number of employees with similar exposure. For this factor, points are assigned according to the number of employees assigned to this activity. With higher numbers of employees engaged in the activity, there is a higher probability of employees being exposed. A rating of ―>15‖ employees results in 15 points; a rating of 11-15 points results in 10 points; a rating of ―6-10‖ results in 5 points; a rating of ―1-5‖ results in 0 points; and a rating of ―Unknown‖ results in 11.25 points.

4. Frequency of operation. Points are assigned based on the frequency of the operation, as more frequent operations are more likely to result in employee exposures. A rating of ―Daily‖ results in 15 points; a rating of ―Weekly‖ results in 10 points; a rating of ―Monthly‖ results in 5 points; a rating of ―Less than monthly‖ results in 0 points; and a rating of ―Unknown‖ results in 11.25 points.

5. Duration of operation. Points are assigned based on the duration of the operation, as longer operations are more likely to result in employee exposures. A rating of ―>4 hours‖ results in 15 points; a rating of ―1-4 hours‖ results in 10 points; a rating of ―30-60 min‖ results in 5 points; a rating of ―Less than 30 min‖ results in 0 points; and a rating of ―Unknown‖ results in 11.25 points.

131

Table 1. Severity and Probablity Factors and Maximum Points Per Factor (NM = Nanomaterial; PM = Parent Material)

The overall probability score is based on the sum of all the points from the probability factors. The Severity Factor Maximum Pts Maximum Severity Score Surface Chemistry (NM) 10 Particle Shape (NM) 10 Particle Diameter (NM) 10 Solubility (NM) 10 Carcinogenicity (NM) 7.5 Reproductive Toxicity (NM) 7.5 Mutagenicity (NM) 7.5 100 Dermal Toxicity (NM) 7.5 Toxicity (PM) 10 Carcinogenicity (PM) 5 Reproductive Toxicity (PM) 5 Mutagenicity (PM) 5 Dermal Hazard Potential (PM) 5 Probability Factor Maximum Pts Maximum Probability Score Estimated Amount of Nanomaterial 25 Dustiness/Mistiness 30 Number of Employees With Similar 15 100 Exposure Frequency of Operation 15 Duration of Operation 15 maximum score is 100. An overall probability score of 0-25 was considered extremely unlikely; an overall probability score of 26-50 was considered less likely; an overall probability score of 51-75 was considered likely; and an overall probability score of 76-100 was considered probable. Based on the severity score and probability score for an operation, the overall RL and corresponding control band is determined by the matrix shown previously in Figure 1.

Application of the CB Nanotool for five different operations. In order to pilot test the CB Nanotool, information was gathered from five different operations in two DOE research laboratories. CB Nanotool inputs for these operations can be seen in Tables 2 – 5 below. Four operations are being performed at the Lawrence Livermore National Laboratory (LLNL) and one operation was performed at the Stanford Linear Accelerator Center (SLAC). A nanotechnology information field-based form was developed to appropriately collect data. Field visits were initiated at LLNL through the cognizant IHs for those operations with principal researchers participating in reviews. The field visit at SLAC was initiated by their ES&H Division Office and principal researchers for their operation participated in the review along with ES&H Division staff.

Results

132

Operation descriptions are summarized below, mostly in general terms, and the results of the CB Nanotool are shown in the appendix.

Synthesis of nanoporous metal foams (Activity 1) Nanoporous metal foams are synthesized by mixing metal nanoparticles with polystyrene spheres and water. These components are weighed and combined into a vial inside a glove box and the mixture is transported to a sonicator. After sonication is complete, the sample is pipetted into a tube where water is removed from the sample using a water-absorbing medium. Once the sample is removed from the tube, it is placed inside a furnace and the polystyrene spheres are vaporized, producing a nanoporous metal foam. Based on knowledge of the nanomaterial characteristics and a thorough review of the operation in the field, the CB tool indicated that the overall RL was 3. The required engineering control, therefore, would be containment. The portion of the activity that had the highest likelihood of exposure was during the initial weighing and mixing phase, and this was performed inside a glove box with a HEPA-filtered exhaust system. The current controls, therefore, were consistent with what was recommended from the CB Nanotool.

Flame synthesis of ceramic nanoparticles (Activity 2) Ceramic nanoparticles (e.g., lutetium oxide, lutetium aluminum garnet) are synthesized by injecting carrier liquids into a flame inside a fume hood which are consumed through combustion. The resulting nanoparticles are produced and collected onto a filter plate. Based on knowledge of the ceramic nanoparticle characteristics and a thorough review of the operation, the CB Nanotool indicated that the overall RL was 2. The required engineering control, therefore, would be a fume hood or local exhaust ventilation, which was in fact what was utilized during this operation.

Synthesis of carbon nanotubes (Activity 3) Carbon nanotubes are synthesized by passing a mixture of an inert carrier gas (Ar), hydrogen, and hydrocarbon precursor gas (e.g., ethylene, acetylene) over catalyst particles deposited on silicon substrates within a horizontal tube furnace. Trace amounts of water are added to the gas mixture to enhance the growth process. The carbon nanotubes are fully attached to the substrates when they are removed from the tube furnace using forceps. The samples are then transferred into plastic containers for further characterization. Based on knowledge of the carbon nanotube characteristics and a thorough review of the operation in the field, the CB Nanotool indicated that the overall RL was 2. The required engineering control, therefore, would be a fume hood or local exhaust ventilation. In this particular operation, the carbon nanotubes were synthesized within an enclosed tube furnace and therefore the level of control achieved was containment. This level exceeded the required control as determined from the CB Nanotool.

Consolidation of ceramic nanoparticles (Activity 4) Ceramic nanoparticles are weighed inside a chemical fume hood. An organic solvent (e.g., ethanol) is added to the powder mixture inside a ball mill jar and milled for several hours. The mixture is pressed into a die inside the fume hood and the compacted material is heated in a burn oven inside the fume hood to remove the organics and other residues. The material is then sintered inside a vertical tube furnace and quenched as it is dropped into a bucket located below the furnace. The cooled material is transferred into a plastic container. Based on knowledge of the ceramic nanoparticle characteristics and a thorough review of the operation in the field, the CB Nanotool indicated that the overall RL was 3. The required engineering control, therefore,

133

would be containment. A fume hood, in fact, was used throughout this operation; therefore, the level of control was not adequate and would need to be upgraded.

Preparation of a single dry bacteriogenic uranium dioxide sample (Activity 5) A sample of uranium dioxide in a container is opened inside an anaerobic chamber. The sample is allowed to dry out inside the chamber and then transferred into a vanadium metal canister for shipment to another research facility. Based on knowledge of the uranium dioxide nanoparticle characteristics and a thorough review of the operation, the CB Nanotool indicated that the overall RL was 3. The required engineering control would be containment. The current controls, therefore, were consistent with what was recommended from the CB Nanotool, as all the operations were performed inside an enclosed chamber with HEPA filtered exhaust.

Table 2. Nanotechnology activity descriptions (Row 1 in Table 2 corresponds to Row 1 in Tables 3 – 5 and similarly with the other rows.)

Table 3. Severity factors of the parent material.

134

Table 4. Severity factors of the nanomaterial

Table 5. Probability factors, RL and recommended control

Discussion The understanding of structure- and chemistry-related health effects from exposures within all aspects of the nanoparticle technology industries comes together into a burgeoning toxicological research field. Traditional IH sampling for nanoparticles at this point in time may very well miss an appropriate exposure index unless a complete collection of associated number, surface area, and mass concentrations is simultaneously measured. The stratification of health risk within professional IH teachings begins to lose footing when the appropriate toxicological endpoint, biologically available concentrations, and its effective dose potential are not fully understood. From the practical aspect of protecting the worker as a primary objective, the toxicological ―wait and see‖ approach begins to lose ground to the ―band and control‖ method of primary prevention.

The CB approach for controlling nanoparticle exposure is given leeway from its most popular requestor. In order to work safely with nanomaterials, Maynard has said that existing IH ―will get us 60 to 70 percent of the way‖, leaving ―a gap that has to be filled with this strategic, targeted research‖ (Cable, 2006). CB offers a method to bridge this gap while remaining dynamic in adjusting to new, available research. While the determination of severity and probability were dependent on factors that are known or suspected to be important in characterizing risk from nanoparticle exposure, the relative importance of one factor compared to another may change as more knowledge on the adverse effects of nanoparticles becomes

135

available. Ranges of values corresponding to discrete scores given for each factor may also be modified according to the level of risk one is willing to accept and ranges of values relevant to the organization utilizing the tool. Thus, some level of expert judgment should be used to ensure recommended controls produced from the CB Nanotool are in fact the most appropriate for the activity in question. In this study, the ranges of values used in the CB Nanotool correspond to those ranges that one would expect in small-scale research-type operations. For large-scale manufacturing of nanoparticles, ranges of values may be quite different than those utilized for small-scale R&D work, particularly with respect to the probability factors‘ ranges. Large-scale manufacturing processes also typically involve several steps, each of which would likely need to be assessed as a separate line item using the tool.

The CB Nanotool was developed in a Microsoft Excel® spreadsheet allowing automatic RL calculations and corresponding control band based on the operational review. While this tool can be used without obtaining specific field measurements, the tool can be used in conjunction with quantitative measurements as they become available. For example, dustiness may eventually be defined in terms of overall particle surface area or particle number and be measurable. The CB Nanotool therefore is dynamic and can potentially be utilized as effective measurement techniques become available. It should be recognized, however, that any CB tool, must be used with some degree of caution. The different factors considered, weighted, and influencing the overall RL and control band are determined as educated ‗guesses‘ as to factor importance and range delineation. Any CB tool utility requires frequent use, validation, and evaluation of recommended control effectiveness. The authors, therefore, strongly encourage the further utilization of this or other similar tools for a wide range of applications as these efforts will undoubtedly improve and refine the tool.

Conclusion With investment increasing the global value of nanotechnology products to 2.5 trillion dollars by 2014 (Lux Research, 2004), health and safety professionals must strive to protect employees involved in technological development and product manufacture, as well as eventual consumers. Engineering controls remain the most important and effective means for preventing or limiting employee exposures. Based on the application of the CB Nanotool, four of the five operations evaluated in this study were found to have implemented controls consistent with what was recommended by the CB Nanotool, with one operation even exceeding the required controls for that activity. The one remaining operation was determined to require an upgrade in controls. The fact that the CB Nanotool produced recommendations that were largely consistent with the IH expert opinions that dictated the existing controls can be viewed as a further validation of the CB Nanotool. By developing this dynamic CB Nanotool within the realm of scientific information available, this application of CB appears to be a useful approach for assessing the risk of nanomaterial operations, providing recommendations for appropriate engineering controls, and facilitating appropriate resource allocations.

Acknowledgement This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344 as LLNL-JRNL-401987. The authors would like to thank the Stanford Linear Accelerator Center for providing a venue for one of the nanotechnology operation reviews. The authors would like to acknowledge Remko Houba for his expertise and review, and Zak Islam, Janeen Robertson, Jennifer Kapp, Octavio Cervantes, Joshua Kuntz, Aleksandr Noy, and Howard Hoyt for their assistance in the activity reviews. 136

APPENDIX B GRASSROOTS ERGONOMICS: INITIATING AN ERGONOMICS PROGRAM UTILIZING PARTICIPATORY TECHNIQUES Annals of Occupational Hygiene, 45(4):283 – 289 (2001) D. M. Zalk University of California, Hazards Control Department, LLNL, Livermore, CA U.S.A.

Abstract

Introduction of ergonomics programs throughout the world requires an easy to understand and inexpensive process. Participatory ergonomic intervention techniques have proven to be beneficial in the prevention of musculoskeletal disorders. The participatory approach to ergonomics has also been found to be a useful application within industrialized (developed) countries and industrially developing countries (IDCs). Grassroots Ergonomics principles utilize expertise within a workforce that focus on participatory ergonomics interpretations of quantitative and qualitative risk and exposure assessment information that in turn result in a peer-developed ergonomics training. Regardless of the intricacy of the exposure assessment tools, workers should fully assist in gathering and analyzing data, then in identifying and implementing solutions. A coordinated and multidisciplinary application of this approach within IDCs would succeed in the creation and sharing of job-specific ergonomics training information for high physical exposure professions, such as agriculture, fishing, forestry, mining, and small- scale enterprises, to initiate ergonomics programs regionally.

Introduction

Reduction in the occurrence of musculoskeletal disorders (MSDs) is essential to the improvement of occupational health in both industrialized (developed) countries (DCs) and industrially developing countries (IDCs) (Jafry and O‘Neill, 2000; Buckle and Devereux, 1999; Partanen et al, 1999; Bernard, 1997; GAO, 1997). Currently 40% of the world‘s occupational and work-related health costs are attributed to musculoskeletal diseases, so the concern can be considered to be distributed throughout both DCs and IDCs (Takala, 1999). Efforts to introduce ergonomics programs within IDCs have focused primarily on large-scale industries and since the majority of the working population of IDCs are involved in the agricultural trades, ergonomic interventions need to have the adaptability to go beyond the factory environment and into rural villages (Jafry and O‘Neill, 2000). This requires a programmatic process that is low cost, easy to understand, and sensitive to the social, cultural, and political considerations of a given population (Shahnavaz, 2000; Kawakami et al, 1999; Rubio, 1995).

An ergonomics program should utilize intervention techniques that focus on a method of achieving prevention. Ergonomic interventions have been successful in reducing the number of MSDs by over 50%, especially in professions that expose employees to a high level of work risk factors (Rosskam, 1997; Hagberg and Wegman, 1987; Fine et al, 1987). Acknowledged hazardous work in IDCs, with a high level of physical demand, include agriculture, mining, construction, fishing, and logging (Takala, 1999). Participatory ergonomics, utilizing worker involvement as part of an intervention, has been a successful technique for the prevention of MSDs in many of these professions (Jafry and O‘Neill, 2000; Kawakami et al, 1999; Koda et al, 1997, Moir and Buchholz, 1996; Rainbird and O‘Neill, 1995).

137

Effective interventions must actively involve the worker, must reduce exposure to the stressor, and must affect the organizational culture (Westgaard and Winkel, 1997). When the worker is involved in ergonomic interventions it offers a greater likelihood of reducing musculoskeletal problems (Buckle and Devereux, 1999). Recent recommendations from the National Institute of Occupational Safety and Health (NIOSH), European Agency for Safety and Health at Work (EASHW), and the General Accounting Office (GAO) include participatory ergonomics as an important method for controlling MSDs and initiating an ergonomics program.

The GAO report highlights the components of effective ergonomics programs. Five companies were reviewed and the synopsis emphasizes a core set of six elements necessary to ensure the identification and control of ergonomic hazards to protect workers: (1) management commitment, (2) employee involvement, (3) identification of problem jobs, (4) development of solutions (controls) for problem jobs, (5) training and education for employees, and (6) appropriate medical management. The application of these elements has resulted in the reduction of injuries and illnesses as well as reduced worker compensation costs. There were also reports of improved worker morale, productivity, and product quality. The GAO goes further to recommend that federal and state-operated Occupational Safety and Health Administration (OSHA) programs utilize a similar approach in developing a framework for worksite ergonomic programs that emphasizes the need to develop and implement flexible site- specific efforts that effectively address hazards. (GAO, 1997).

The NIOSH publication recommends a focus on controlling work-related MSDs. This publication provides information to identify, correct, and prevent MSDs by following seven steps: (1) determining if musculoskeletal problems exist in a workplace, (2) developing roles for managers and workers in an ergonomic program, (3) recognizing and filling the training needs, (4) gathering and analyzing the data to define the scope and characteristics of ergonomics concerns, (5) developing control solutions, (6) establishing health care management, and (7) creating a proactive ergonomics program. NIOSH adds two very important criteria to the GAO recommendations, the gathering and analysis of data and the emphasis on the creation of a proactive program (Bernard, 1997).

The EASHW report defines a scope for the prevention of work-related upper limb disorders that includes risk assessment, health surveillance, employee information, training, ergonomic work systems, and the prevention of fatigue. Exposure and risk assessment methods are put forward while giving acknowledgement that these methods are often in competition with the realities of field applications. The report also emphasizes the use of an ergonomic intervention that focuses on the whole workplace and work system as an integrated approach. Participatory ergonomics has the potential to assist in the development of this integrated approach (Buckle and Devereux, 1999).

Principles and Processes

The Grassroots Ergonomics (GE) approach to initiating an ergonomics program utilizes participatory ergonomics (PE) principles as a part of its integrated process. As a necessary part of the GE process, PE principles are combined with the gathering and analysis of exposure assessment (EA) data as well as the creation of training (CT) in the following manner:

GE = PE + EA + CT 138

This combination of the elements essential to creating a successful ergonomics program also assist in the maximization of benefits obtained from a given intervention.

Participatory Ergonomics (PE) Participation at work is a general technique of giving employees an opportunity to control the design of their workplace and plan their work activities. The premise is that workers know their workplace better than anyone else does, and that this knowledge allows them to develop a more comprehensive approach to their work. This is the fundamental benefit of PE. The participatory process, when applied to the study of the musculoskeletal system and its disorders, creates a more thorough understanding of ergonomic problems and a more diagnostic approach to their solutions (Nora and Imada, 1991).

The amount of control which workers are given over their workplace is an important element, as well as a potential limitation, for the effectiveness of the participatory process. By definition, managers have a level of control over their workplace that is not available to the workers. Without an appropriate incentive, management is not usually willing to truly empower workers to determine their own solutions. Additionally, a situation in which workers do not have a climate of trust between themselves and upper management can, in and of itself, add to their ergonomic and psychosocial risk factors (Israel et al, 1989). This problem exists as much in developed as in developing countries. Unfortunately, lack of worker empowerment is not consistent with the fact that workers are usually the ones who know their job and their peers well enough to identify and create solutions that will persist (Nora and Imada, 1991). However, management can be in a position to give a certain level of control to the workers when their production costs increase due to MSDs. These increased costs can often be related to decreased working efficiency, absenteeism, medical care, and worker‘s compensation (Buckle and Devereux, 1999; Bernard, 1997; GAO, 1997). In these instances, the need to reduce MSDs can be seen as directly linked to the need to increase production. Therefore, both management and workers have a vested interest in achieving these objectives. It is important to all the parties that the involvement and support of top management is often related to the reduction in workers‘ exposure to ergonomic and psychosocial risks (Israel et al, 1996).

PE is one intervention strategy that can simultaneously address both ergonomics and the psychosocial risk factors in the work environment (Haims and Carayon, 1998; May and Schwoerer, 1994). Psychosocial factors, and their relationship to MSDs, are characterized as being associated with work organization factors, the external work environment, and the characteristics of the individual worker (Buchanan et al, 1998; Bernard, 1997). Cumulative research has indicated that MSDs may act through a variety of physical or biological mechanisms, with psychosocial factors also playing an important role in the onset and development of MSDs (National Research Council, 1999). The very complex nature of psychosocial implications, and its relationship with ergonomics, underlies the benefits of the participatory approach in both DCs and IDCs. There are often cultural differences in how work is performed between companies as well as between countries. PE can assist in developing an approach that takes cultural work method differences into account (Buchanan et al, 1998).

Addressing the social, psychological, and cultural needs of a given working population has been an important aspect of PE. Within both DCs and IDCs, the PE approach has been successfully applied by maximizing the role of the employee within the company. PE principles have been applied to reduce MSDs in an extensive variety of occupational fields within DCs. Examples 139

include: carpentry, components-parts manufacturing, construction, custodians (janitors), health care industry, meat packing, newspaper industry, and waste container handling (Rosecrance and Cook, 2000; Zalk et al, 2000; Evanoff et al, 1999; Albers et al, 1997; Bohr et al, 1997; Moore and Garg, 1997; Zalk et al, 1997; Moir and Buchholz, 1996; Moore and Garg, 1996; Israel et al, 1989). This approach has also been applied within IDCs in Asia for assessing the needs of small enterprises and agriculture by building on local practices within the limitations of locally available resources (Kawakami et al, 1999). Small-scale enterprises within Thailand have also emphasized participation to improve working conditions, materials handling methods, and productivity with a special focus on the locally invented improvements (Tandhanskul et al, 1995). Collaborative participatory ergonomic efforts have also emphasized the skills of the local population to create practical, inexpensive solutions that have resulted in development and implementation of training packages (Kogi, 1998).

Exposure Assessment (EA) During a participatory ergonomic intervention, employees become aware of their ergonomic problems and can identify work procedures that may have ergonomic risk factors associated with them. There is an important need to quantify the level of exposure in these procedures (Buckle and Devereux, 1999; Bernard, 1997). This quantification, or ergonomics EA, is crucial to GE and is essential to determine the scope and characteristics of ergonomics concerns for everyone involved (Zalk et al, 1997). This is one point that is emphasized in the NIOSH and EASHW publications, but not addressed by the GAO report. Ergonomic analysis tools are used to gather and analyze data on identified procedures to form an EA. Tools for an EA can be as inexpensive as checklists and video for posture analysis and targeting or as intricate as three- dimensional motion evaluation systems and surface electromyography (Zalk et al, 2000; Li and Buckle, 1999).

Identifying the ergonomic risk factors, including posture, force, and repetition, within a work procedure is useful for EA of MSDs (Burdorf, 1992). If the EA data will be utilized for in-depth research on a given profession‘s exposure, then more intricate measures should be considered. However, limitations of funds and the lack of availability of measuring equipment within DCs and IDCs require an emphasis on the inexpensive approaches. This reality presents a difficult dilemma facing ergonomics research in the field. The feasibility of a measurement device, level of detail required, and variation of exposures needs to be weighed against practicality and ease of use of obtaining this information from the field (Burdorf and van der Beek, 1999; Buckle and Devereux, 1999). This may necessitate compromise, especially if monetary funding is as scarce as EA tool availability everywhere, but especially in IDCs.

If the goal is to collect less detailed data for practical work situations, an EA tool should be chosen with consideration given to its use and practicality in the field. Li and Buckle (Li and Buckle, 1999) have performed a review of the current techniques available for assessment. The strengths and limitations of pen and paper observational methods, videotaping, computer-aided analysis, and self-report assessment are considered. Most of the existing tools are found to be research-oriented rather than field-oriented. They are designed to maximize the information obtained for a job and its procedures. The field practitioner‘s need during an intervention may not require a researcher‘s level of information. Most important, according to Li and Buckle, is the information that helps decide whether an intervention is necessary and a measurement to determine whether the intervention was effective. The GE process takes this into account with an additional consideration that the workers fully assist in gathering and analyzing data, utilizing

140

the data to address training needs, then identifying and implementing solutions (Zalk et al, 2000).

An ergonomics EA should be aimed at collecting information that accurately reflects the exposures and behaviors of the employees in the field. This approach is recommended in the NIOSH publication within the step of gathering and analyzing data to define the scope and characteristics of ergonomics concerns (Bernard, 1997). GE takes this recommendation two steps further in that it emphasizes that the gathering of data should be performed in the field under actual working conditions and that the data be quantifiable to afford replication after the ergonomic intervention (Zalk et al, 1997). This extra effort allows the opportunity to quantify the ergonomic exposures before and after the intervention has been completed and controls have been implemented. This concept is similar to traditional occupational hygiene sampling theories and is a more appropriate measure of the intervention‘s success rather than the use of often transient injury statistics.

Checklists have frequently been the ergonomic tool of choice within participatory ergonomic interventions (Kawakami et al, 1999; Moir and Buchholz, 1996; Tandhanskul et al, 1995). Regardless of the intricacy of the EA tools, workers should fully assist in gathering and analyzing data, then in identifying and implementing solutions. A team of custodial (janitorial) employees designed and implemented a GE intervention that included video-based posture analysis and posture targeting applied in the field (Zalk et al, 1997). Using the same PE approach, waste container handling employees used the above video-based analyses and added surface electromyography and lumbar motion monitor direct measuring instruments to ensure that the field analysis optimized the EA information pertaining to their profession (Zalk et al; 2000). Each employee‘s interpretations of musculoskeletal quantitative and qualitative data are an essential part of the GE process. These interpretations can be developed into a training program that focuses on the prevention of MSDs within the professions analyzed (Zalk et al; 2000; Kawakami et al, 1999; Tandhanskul et al, 1995).

Creation of Training (CT) Emphasis on ergonomics training has been an historical focus for assisting in the creation of ergonomics programs. Over thirty years ago the World Health Organization made this an important element in a plan to develop an inter-regional course on ergonomics for developing countries (Singleton and Whitfield, 1968). The creation of a training program, resulting from the application of PE principles, is a relatively new concept. When participatory techniques are utilized in the CT, the resultant training proves most useful for the promotion of intervention successes that have come from the local workplace population (Kogi, 1998).

CT is the culmination of the GE process, which often begins with employees and occupational health and hygiene staff working together to discuss ergonomics-related problems within a given company, or profession, and its population. Schemes to address these problems, by developing a framework for the research and intervention, have the potential to enhance the relevance and utilization of results when employees are involved as researchers to address the causes and the symptoms of the problems identified (Baker et al, 1994, Israel et al, 1989). Occupational health and hygiene staff are useful in the collection of background and ergonomics exposure data to be used at the discussion sessions, which can serve for the foundation for the CT (Zalk et al, 2000). Interventions for particular musculoskeletal issues raised during these discussions are most effectively initiated with consensus from the entire group (Moore and Garg, 1997). 141

Once these musculoskeletal issues are identified, a participatory ergonomic intervention team should be created that consists of a small, representative group of affected workers, a health and safety professional (ergonomist), and a management representative (GAO, 1997; Bernard, 1997; Zalk et al, 1997; Israel et al, 1996). In the GE process, the workers are in control of the intervention team, so one of these members serves as the team‘s leader. The workers‘ initial role is to teach the ergonomist about their work and its procedures. The ergonomist‘s role is teaching basic principles of ergonomics, how these principles may apply to the work‘s procedures or cycles, and consulting on behalf of the workers.

The consulting role of the ergonomist is extremely important to the CT because this person not only serves as a conduit of introductory, as well as technical, ergonomics information, but they may also have to serve as a mediator between workers and their management. Management should have a representative on the team to understand the inner workings of the PE approach and provide guidance on management issues (Bernard, 1997; GAO, 1997; Koda et al, 1997; Israel et al; 1996; Tandhanskul et al, 1995). Management representatives need to know the right people to go to when obtaining funding, making changes in the workplace, and implementing the training. They also need to become an advocate for the needs of the workers, as determined by the intervention team, to assist in the implementation of the overall ergonomics program. This begins a process for sustaining an ergonomics program after the initial participatory ergonomic intervention (Zalk et al, 2000; Kawakami et al, 1999; Tandhanskul et al, 1995).

For an effective ergonomic program to be initiated, training needs to be developed in hand with ergonomic research that is both flexible and dynamic (Kawakami et al, 1999; Baker et al, 1994). An appropriate goal of an ergonomic intervention team is to create and implement training for the affected workforce that includes discussion of the ergonomic EA study and its results (Zalk et al, 2000; Kawakami et al, 1999; Kogi, 1998). Since workers are in control of the GE process, they assist in selecting the EA techniques that best address their job task and most likely reduce their exposure to musculoskeletal hazards during their work. It is necessary for an ergonomist to take the process into the field, working with the employees to obtain and analyze the data. This includes full worker participation in the measurement, processing, and interpretation of ergonomics EA information. This process is known as in-the-job training (IJT). The IJT obtained by the team‘s workers while going through the GE process is far more intricate and involved than typical classroom and hands-on training.

IJT intrinsically educates the workers in ergonomics because they are fully involved in applying solutions that they helped develop. Thus, they are trained within the parameters of their own job. The information obtained from the EA study within the GE process will assist the intervention team in developing their own training, their own training manual, and applying appropriate controls for their own co-workers. This training also becomes an IJT for workers who are not part of the intervention team because their peers‘ interpretations of the GE process are presented in work-related language they can understand. Additionally, when training comes from within the trainees‘ own ranks there is a much greater participation, acceptance, retention, and application of the goals from the training. Workers from the team can also become qualified to assist in presenting the IJT, creating their own periodic training, and reinforcing ergonomic lessons learned (Zalk et al, 2000, Kawakami et al, 1999).

142

Results Office and production workers in a newspaper company teamed with ergonomists to reduce MSD risk factors by developing and implementing solutions for very low intervention costs (Rosecrance and Cook, 2000). PE principles in the meat packing industry were implemented with ergonomic intervention teams, utilizing workers as researchers, with a problem solving method that began with obtaining background, exposure, and effects data. The teams then approached solutions to the problems with brainstorming sessions and selected interventions by consensus that effectively addressed and reduced MSDs (Moore and Garg, 1997; Moore and Garg, 1996). A similar approach with health care workers proved, for some, to be highly effective in identifying problems and implementing solutions (Bohr et al, 1997). By including management on the team and focusing efforts on designing and implementing changes in training, hospital orderlies showed a decline in MSD symptoms and improvements in job satisfaction and psychosocial stressors (Evanoff et al, 1999).

Participatory training, using hands-on exercises and ―learner-centered‖ instruction, was found to be extremely useful in acquiring and retaining ergonomic knowledge among apprentice carpenters, when compared to a control group (Albers et al, 1997). Applications in the construction industry have utilized participatory ergonomic interventions for not just the practical needs associated with training, but for political reasons as well (Moir and Buchholz, 1996). These construction advisory groups served to evaluate intervention ideas and compare safety systems within the company. A ten-year follow-up study in a waste management bureau that found significant reductions in compensation claims for low back pain demonstrates the endurance of the participatory process (Koda et al, 1997). This participatory ergonomic intervention began with a revision of the safety procedure manual and an emphasis on continual participatory training classes.

A custodial (janitorial) workforce of 150 employees, 25 of whom are developmentally disabled, used the GE process and its principles, including EA, to address their consistent rise in MSDs (Zalk et al, 1997). The custodians on the intervention team revised their training manual and created an ergonomics training based on their EA results. This resulted in a training video and manual package that has been shared in over 20 countries. Substantial reduction of MSDs and a measured decrease in EA results were found three years after the initial training (Tolley and Zalk, 1998). Use of more intricate EA techniques, such as surface electromyography and lumbar motion monitor, did not deter waste container handling employees from also interpreting the results into their own training that they assisted in presenting to their peers (Zalk et al, 2000). The training methods developed by both of these GE interventions have not only resulted in a decrease in ergonomics-related injuries, more importantly they have reduced employee exposure to the ergonomic stressors identified through the PE process.

Participatory ergonomic interventions have also been a successful approach for the reduction of MSDs in IDCs (Jafry and O‘Neill, 2000; Buckle and Devereux, 1999). Successful applications of this intervention method within IDCs have come from utilizing local talent, skills, and available resources (Jafry and O‘Neill, 2000; Kawakami et al, 1999; Koda et al, 1997, Rainbird and O‘Neill, 1995; Tandhanskul et al, 1995). Participatory ergonomic interventions within IDCs have done extremely well in keeping costs to a minimum, adapting to the customs, traditions and politics of a given region, and achieving the management buy-in which is essential for the in-house development of a PE program (Rubio, 1995).

143

Ergonomic programs in Asia have grown out of successful participatory interventions. Assessment of the local needs in small-scale enterprises and agriculture have resulted in the development and implementation of widely-applicable ergonomic improvements and the sharing of training methods (Kawakami et al, 1999). These improvements include materials handling, design of workstations, and work organization - all achieved by utilizing local resources. It is the participatory process that assisted in the creation of the improvements and training methods that meet the diversity of a given local population. Locally invented ergonomic improvements were utilized to initiate ergonomic programs within small-scale enterprises in Thailand (Tandhanskul et al, 1995). The improvements from this intervention included workstation redesign and material handling changes that were derived from the results of a checklist for ergonomic risk assessment. Field study interventions, use of practical assessment methods, and voluntary efforts within a participatory framework are all part of the collaborative research and training approach that has been successful within Asian IDCs (Kogi, 1998).

Discussion When a training program is derived from the workers, both the process of learning and the process of teaching will assist in the reinforcement of ergonomics as an essential element in work procedures. Program initiation occurs when the training is presented to the employees‘ peers, and the feedback incorporated into future ergonomic training sessions. By utilizing the GE principles, the nature of the program incorporates the psychosocial and cultural issues that are virtually inseparable within working populations. Further, understanding of ergonomic concepts is best described from the perspective of local interpretation and application (Jafry and O‘Neill; 2000; Kogi; 1998, Wisner, 1989). Therefore, resulting training packages and workplace redesign applications are well situated to be introduced to other workers within similar professions both regionally and worldwide (Kawakami et al, 1999; Kogi, 1998; Tandhanskul et al, 1995). This is especially important for the rural, or informal, working populations throughout the world where the application of GE is most difficult or impossible due to the limited size of the workforce, minimal finances, and the lack of professional ergonomics expertise. Workers whose employment is not connected with companies or factories, such as small-scale enterprises, independent farmers, or artisans, have the most to gain from the sharing of GE training packages.

Most of the ergonomics research performed in IDCs, especially when supported or implemented at the management level, has been in the industrial sector and has focused primarily on maximizing work efficiency and increased productivity (Jafry and O‘Neill, 2000, LaDou, 1996). Technological advancement has brought new sources of musculoskeletal stresses and requirements for rapid production schedules as well as an increase in chemical exposures and related occupational diseases (Shahnavaz, 2000; Cory, 1999; Takala, 1999). Perhaps utilizing the GE process and its principles can assist in achieving maximum benefits for the working population that can balance out the continually increasing demands for production. Participatory principles, like those within the GE process, are an essential part of the collaborative, multidisciplinary approach necessary to reduce occupational health diseases (Stubbs, 2000). To lower the occupational health cost of musculoskeletal diseases internationally over time, this approach must assist in standardizing criteria so international organizations can add to the global database on accurate statistical musculoskeletal information and share job-specific ergonomic training information to initiate ergonomic programs within and between DCs and IDCs.

144

APPENDIX C PARTICIPATORY OCCUPATIONAL HYGIENE; A PATH TO PRACTICAL SOLUTIONS David M. Zalk, President, International Occupational Hygiene Association Asia-Pacific Newsletter 9:51 (2002)

The formality of addressing a participatory application in our profession requires an understanding that occupational hygiene is defined as both a science and an art. Our devotion is to the prevention, recognition, evaluation, and control of hazards or stressors in the workplace, that may cause illness, injury, or significant discomfort. In many of the Industrialized countries we are fortunate to have both the opportunity and resources to delve into our daily activities with the luxury of the scientific and technological tools and aspects of our profession. To construct a path to practical solutions within the Industrially Developing Countries (IDCs) we need to recognize the artistic aspects of application that lie within the cultural understanding of the workers we have a responsibility to protect.

The concept of participatory occupational hygiene is presented as somewhat analogous to the successful and accepted practice, performed under our common occupational health umbrella, known as participatory ergonomics. Participation at work is a technique of giving employees an opportunity to control the aspects of their workplace that are associated with exposures resulting in work-related illness and disease. The assumption is that workers know their workplace better than anyone else does and that this knowledge allows them to develop a more comprehensive approach to their work. With the majority of the working population of IDCs involved in the agricultural trades and small-scale enterprises, practical solutions need to have the adaptability to go beyond the factory environment and into rural villages. This requires a programmatic approach that is low cost and easy to understand. The introduction of the International Labour Organisation (ILO) Toolkit, with the assistance of the International Occupational Hygiene Association, may be an important step to the implementation of such an approach. Utilizing the ILO Toolkit with the assistance of trained professionals, workers can participate in the gathering and analysing of necessary occupational hygiene information, then in identifying and implementing practical solutions.

An occupational hygiene programme should utilize intervention techniques that focus on a method of achieving prevention. For an intervention to be effective and for practical solutions to be achieved, the intervention must actively involve the worker, must reduce exposures to the stressors, and must affect the organizational culture. There are often cultural differences in how work is performed between countries, and participatory occupational hygiene can assist in developing an approach that takes this into account. Addressing the social, psychological, and cultural needs of IDCs has been, and continues to be, an important aspect of achieving prevention. Within IDCs, the participatory approach has been successfully applied by maximizing the role of the employee within the diversity of the local population. This approach has been applied within developing countries in Asia for assessing the needs of small enterprises and agriculture by building on local practices within the limitations of locally available resources. Field study interventions, use of practical assessment methods, and voluntary efforts within a participatory framework are all part of the collaborative research and training approach

145

that has been successful within Asian-Pacific IDCs.

Technological advancements have brought an increase in chemical exposures and related occupational diseases. Participatory occupational hygiene offers a technique that can be an essential aspect for a multidisciplinary approach to address these changes in the workplace and reduce occupational diseases. Many of the technological advancements have created industrial problems where an occupational hygienist, ergonomist, and occupational health physician can work together with workers to develop understanding and create solutions. The implementation of this method can be extended by a greater level of participation within a given community or village. Information learned and practical solutions created need to be shared with other IDCs within an established participatory occupational hygiene protocol, which can be associated with the ILO Toolkit. This protocol would also assist in the development of accuracy in reporting occupational health statistics. A collaborative, multidisciplinary approach is necessary in order to strengthen the occupational health and safety foundation and assist IDCs in initiating preventive programmes. This approach must aim at reducing work-related exposures for employees in large industry, small businesses, and agricultural professions. Utilizing local talent in training, developing, and implementing participatory occupational hygiene programmes and practical solutions within IDCs will aid in ensuring costs are minimized and cultural, ethical, and psychosocial aspects are addressed.

146

APPENDIX D The ICOH and IOHA Declaration David M. Zalk Asian-Pacific Newsletter on Occup Health and Safety, 10:42–45 (2003)

On 27 February, 2003, at the 27th ICOH Congress in Iguassu Falls, Brazil, a 26-point document entitled ―ICOH and IOHA Declaration to Strengthen the Position of Occupational Hygiene‖ was ratified. This document was signed by myself, on behalf of the International Occupational Hygiene Association (IOHA), and Bengt Knave for the International Commission on Occupational Health (ICOH). In light of the international nature of this document, it was also signature approved at the event by both Jukka Takala, head of occupational safety and health at the International Labour Organisation (ILO), and Maged Younes, head of occupational health at the World Health Organization (WHO). The importance of this document is immediately being felt worldwide, so it is essential that it be shared with all those in the greater occupational health professions in a timely fashion. This opportunity to do so is very much appreciated.

The genesis for the Declaration began at the Workshop on Occupational Hygiene in São Paulo, Brazil from 4–6 December 2002. This workshop was put together thanks to the ever-ongoing work of Berenice Goelzer and the efforts and support of SENAC São Paulo and Fundacentro. The purpose of the workshop was to emphasize the importance of occupational hygiene as a profession in Brazil and focus on the skills necessary to truly achieve prevention under the guise of occupational health. In addition to myself, in attendance at this workshop on behalf of IOHA were Past Presidents Kurt Leichnitz and Paul Oldershaw. Discussions during this workshop included the development of graduate school curriculum, professional certification, and the state of the occupational hygiene profession in the greater occupational health arena within Brazil. A key focus underlying these discussions was the ongoing efforts in Brazil and the region to officially recognize the profession of occupational hygiene.

In São Paulo, Brazil, there is a massive outdoor sculpture located in Ibirapuera Park which is in the heart of their cultural centre. It is known as the Bandeirantes Monument, a tribute to the banner carrying pioneers of Brazil who fought through the wilderness of South America to claim the western boundaries that now comprise their nation. There is currently a group of professionals in Brazil who are considered to be the Bandeirantes of Occupational Hygiene in that region. This workshop assisted in crystallizing the issues facing this group of professionals. It served to adequately assess their needs for them to begin carrying forth their profession‘s banner toward an official recognition of their preventive role in ensuring occupational health for the labour force. This is not only important for the worker‘s health we are all dedicated to protect, but also for the continuation of Brazil‘s burgeoning economy.

IOHA pursued this lack of recognition in concert with ICOH to ensure this issue would be addressed in time for the ICOH Congress to be held in Brazil a few months later. It is essential to recognize the historical aspects of the co-operative efforts between IOHA and ICOH that afforded moving forward so efficiently. At the 26th ICOH Congress in Singapore, then ICOH President Jean-François Caillard worked with former IOHA Presidents Michel Guillemin and Linnea Lillienberg to address avenues of co-operation for these international Non-Governmental Organizations (NGOs). The solution was a formal Letter of Agreement signed by then IOHA President Paul Oldershaw and Jean-François Caillard of ICOH in September of 2000. This Declaration was created upon the key points of that agreement and put into practice their words of co-operation. It was therefore both timely and appropriate that this Declaration was agreed 147

upon and ratified at the next ICOH Congress with many of the original players in this cooperation present.

There is fervent optimism that this document will indeed lead to the recognition of occupational hygiene in Brazil soon. It is IOHA‘s goal that this precedence set in Brazil will continue throughout Latin America and simultaneously the rest of the world. The Declaration was drafted with the intent to be utilized by countries throughout the world for this purpose. It is most appropriate that this document be considered for the utilization by countries in a similar position within the Asian-Pacific region. IOHA is taking a lead role in representing the occupational hygiene profession and infusing prevention into the global concepts of occupational health. International organizations that represent the occupational health professions, such as WHO, ILO, ICOH, and IOHA, are in a position to cooperate towards common occupational health needs worldwide. This interplay has improved understanding and communication between these organizations for the benefit of the workers we are committed to protect. The Declaration embodies this co-operative effort.

The 26 points within this document emphasize the need for occupational physicians and occupational hygienists to team together in truly achieving prevention of work-related diseases. That is the goal of this Declaration. We at IOHA look forward to working in concert with ICOH to build on this Declaration and put it into practice. The current President of ICOH, Jorma Rantanen, put forward this notion as well in his President‘s Address. As he stated, in both the ILO Convention No. 161 and the WHO Global Strategy on Occupational Health for All, there is a call for every working individual to be provided occupational health services. Professor Rantanen is working to put forward a new concept known as Basic Occupational Health Service. We at IOHA applaud this effort and would very much like to assist as it is developed internationally. Perhaps this Declaration can serve as a launching point for putting some of these wonderful cooperative ideas and initiatives into practice. This Declaration to Strengthen the Position of Occupational Hygiene serves as an excellent skeletal structure on which to build. Both ICOH and IOHA are looking forward to working together in fielding initiatives that put some muscle onto it.

ICOH/IOHA Declaration to Strengthen the Position of Occupational Hygiene

(1) We welcome WHO and ILO activities in occupational health and occupational hygiene; these are the right concepts for improvements at workplaces.

(2) We call on WHO and ILO to continue to support occupational hygiene efforts.

(3) We request WHO and ILO, to take steps – with the involvement of NGOs – to promote the development of a comprehensive, easily accessible network of data bases on occupational hygiene issues.

(4) WHO‘s Health for All Policy provides a positive framework for making further progress.

(5) We request WHO and ILO to further develop close cooperation with the European Commission.

(6) Countries are invited to carry out hazard surveillance at workplaces and the impact of proposed policies and programmes. 148

(7) Special assistance is needed for countries which face more severe occupational health problems.

(8) International financial institutions are invited to multidisciplinary support occupational health programmes, including occupational hygiene.

(9) We welcome efforts to involve relevant and well recognized NGOs at the earliest possible stage in the implementation and further development of occupational hygiene and occupational health.

(10) We will strive to implement measures aimed at attaining the occupational health targets.

(11) We welcome the integration of occupational hygiene into national policies which includes legislation and finance.

(12) We welcome the development of national communication and public information strategies in matters affecting occupational health.

(13) We welcome the implementation and further development of occupational hygiene through measures on national and international level – under consideration of local requirements - and with the support of occupational hygiene professionals with adequate training and resources and capacity building in OHS management systems.

(14) Hazard prevention and control in the work environment is a multidisciplinary task which may involve occupational health professionals, such as occupational physicians, occupational hygienists, safety engineers, ergonomists, and nurses. The WHO document on ―Hazard prevention and control in the work environment (WHO/OCH95.3)‖ should be considered.

(15) Occupational health professionals should strive for public awareness and ensuring access to information about work related hazards and their prevention.

(16) Effective measures for monitoring and assessing situations have to be implemented.

(17) We will continue to provide policy advice and guidance on occupational health initiatives on international, national and local level.

(18) Hazard prevention should replace hazard control.

(19) Prevention of occupational health damage is many times more effective than to attempt to cure it after it has occurred.

(20) We recognise that economic analysis helps with setting priorities with regard to risk- reduction by assessing the cost-effectiveness of such measures.

(21) It is important to systematically include the interaction between the human factors and the design of workplaces.

149

(22) We invite all concerned national and international organisations in occupational health to promote a holistic concept of health-, environment-, and safety management systems in enterprises.

(23) We will promote good practice on occupational health in enterprises.

(24) The comprehensive scope of occupational health management should be multidisciplinary.

(25) Indicators for assessing the performance of OSH management systems should be introduced.

(26) An OHS management system should be implemented, along the lines of the ―ILO Guidelines on Occupational Safety and Health Management Systems (ILO/OSH-MS2001)‖.

Figure 1. Signatories of the Declaration

150

APPENDIX E CARPENTER SHOP WOOD DUST CONTROL - PRACTICAL EXPERIENCE TO REDUCE HARDWOOD DUST EXPOSURES BELOW THE ACGIH® TLV® James Martin and David M. Zalk Applied Occupational and Environmental Hygiene, 12:595 - 605 (1997)

Abstract

This paper describes the process that was necessary to reduce hardwood dust exposures in a carpenter shop to levels below the current American Conference of Governmental Industrial Hygienists (ACGIH®) Threshold Limit Value (TLV®). Air monitoring included traditional personal sampling for "total dust" in accordance with the National Institute for Occupational Safety and Health (NIOSH) Method 0500. Achieving the ACGIH TLV of 1 mg/m3 for hardwood dust in a typical carpenter shop has been demonstrated to be possible but very difficult. The implemented shop improvements focused on engineering controls and work practice improvements. Reducing most personal wood dust exposures below 2 mg/m3 was accomplished relatively easily, however reducing the exposures below 1 mg/m3 was considerably more difficult and expensive. A preliminary quantitative comparison of available dust control equipment with hand-held power sanders was also performed. Sanding tool use guidelines were developed from these data to reduce airborne wood dust until engineering controls could be fully implemented. The key to further reducing airborne hardwood dust concentrations was to implement effective local exhaust control for portable tools, especially sanders. The results of this study support the use of the current ACGIH TLV for hardwood dust of 1 mg/m3 as a conservative guideline, and an occupational exposure limit of 2 mg/m3 including consideration of feasibility.

Key words: Wood dust, Hardwood, Total dust, Dust control

Introduction

The focus of this study was to reduce hardwood dust exposures in a carpenter shop to levels below the current American Conference of Governmental Industrial Hygienists (ACGIH®) Threshold Limit Value (TLV®). This carpenter shop services facilities and responds to specific job needs on a work order basis. Most of the work involves the construction of cabinets, book cases and tables. The most commonly used woods are oak, birch, pine, fir and plywoods (generally a mixture of hardwoods and softwoods). The carpenter shop has two rooms referred to as the High Bay and the Low Bay. The High Bay has a very high ceiling and is separated from the Low Bay by a hallway. The available major woodworking equipment are listed in Table 1. The table saws have been the most frequently used major equipment and some of the equipment, such as the panel router, have been rarely used. The carpenter shop also contains other smaller and/or portable tools such as grinders and sanders. The selected equipment for use varies substantially depending on the job request and worker preference.

There is one local exhaust ventilation system which services most of the major shop equipment in both the High and Low Bays. The portable tools, such as sanders, were not initially connected to a local exhaust ventilation system.

151

There are usually 3-6 carpenters working in the shop at any one time. Most of these carpenters work continually in the shop. Carpenters involved in this study typically performed project- based operations. This includes all aspects of furniture assembly from surface preparation to fine hand sanding. Therefore sources of wood dust exposure were not uniform and these activities are not typical of a production line assembly type operation. The majority of employees who performed work in the carpenter shop wore hearing protection, safety glasses, and safety shoes.

Background

Wood dust includes any wood particulate that is generated during the processing or handling of wood. Wood is composed of cellulose, hemicellulose, lignin and extractives. Extractives are high and low molecular weight organic compounds found primarily in the heartwood area of the wood. Extractives makeup approximately 5 - 30% of wood on a mass basis (DEC 1988). The two basic classes of woods are hardwoods and softwoods. Common hardwoods include oak, beech, birch, maple and walnut. Worker exposures to hardwoods usually occur in furniture and cabinet-making industries. Common softwoods include fir and pine. Worker exposures to softwoods usually occur in the building (construction), lumber and sawmill industries. Worker exposures to mixtures of softwood and hardwood are common.

Potential health effects from exposure to wood dust include skin and eye irritation, rhinitis and nasal dryness, pulmonary function changes, allergic respiratory responses (asthma) and cancer of the nasal cavity and paranasal sinuses. Most of the literature has addressed respiratory hypersensitivity and sinonasal cancer. Both acute and chronic occupationally-related asthma and acute respiratory effects have been linked to Western red cedar, California redwood, Teak, and some African woods. Allergic alveolitis, organic dust toxic syndrome, non-asthmatic chronic airflow obstruction, simple chronic bronchitis, rhinitis, and dyspnea are all accepted as wood dust exposure related Enarson and Chan-Yeung 1990; Goldsmith and Shy 1988; ACGIH 1991; Pisaniello et al., 1991).

Cancers of the nasal cavity and paranasal sinuses in the general population are rare with an approximate annual incidence rate of one per million (Andersen et al., 1977; Acheson 1976). A study by Anderson found that more than two-thirds of the population affected were woodworkers in the furniture industry (Andersen et al., 1977). The risk of adenocarcinoma of the nose among Danish and English furniture makers was found to be in the range of 0.5 - 0.7/1000/year. The average latency period for adenocarcinoma is 41 years (range of 4 - 69 years) from the first occupational exposure to wood dust (Andersen et al., 1977; Acheson 1967; Acheson 1976). There is a substantial increased risk of nasal adenocarcinoma in workers exposed to hardwood dust (Andersen et al., 1977; Acheson 1976) and several studies have also implicated softwood as a possible human carcinogen (Briton et al., 1984; Hernbert et al., 1983; Ironside and Mathews 1975). The strongest link between nasal cancer and exposure to hardwood dust has been in occupations where chemical additives (such as preservatives, lacquers and resins) were not used (Nylander et al., 1993).

Mucostasis has been postulated to be a contributing factor for the development of nasal adenocarcinoma in furniture workers due to the prolonged retention of wood dust in the nasal cavity (Goldsmith and Shy 1988; ACGIH 1991; Andersen et al., 1977; Guney et al., 1987). Anderson found that the number of woodworkers with nasal mucostasis was directly proportional to the dust concentration, and that 1 of 9 woodworkers exposed to a mean airborne 152

concentration of 2.2 mg/m3 wood dust developed mucostasis (Andersen et al., 1977). Pisaniello concluded that wood dust exposures below 2-3 mg/m3 may protect against mucostasis (Pisaniello et al., 1991).

The airborne concentration and size distribution of wood dust vary depending on many factors including the type of wood, the type of machinery and tools used, the accessories used (e.g. sand paper coarseness), the type of work being performed (e.g. sanding, sawing, milling), and the water content of the wood. Published wood dust monitoring data have been previously summarized by others (ACGIH 1991; Guney et al., 1987; NIOSH 1987; Scheeper et al., 1995). One set of total wood dust measurements in the breathing zone of furniture industry workers found airborne concentrations for machine and hand sanders at 14.3 mg/m3 and drilling, planing, or sawing at 5.2 mg/m3. No percentage of time performing these procedures was given. Overall, 37% of the sample results were < 5 mg/m3 and 63% were > 5 mg/m3 (Andersen et al., 1977).

The major portion of the mass of wood dust has been found in particles > 10 µm, which agrees with studies indicating that most of the airborne dust will deposit in the nasal region (Pisaniello et al., 1991; Andersen et al., 1977). Sanding is expected to generate more airborne particulates than other types of woodworking because the mechanistic action of sanding shatters the lignified wood cells (Hinds 1988; Holliday et al., 1985; Thorpe and Brown 1995). Sanding of hardwoods is expected to generate lower dust concentrations than softwoods and the particulate size from hardwood sanding is generally smaller than with other types of woods because hardwoods have a more tightly bound cell structure (Hamill et al., 1991). The grade of sandpaper has not been found to substantially alter the rate of airborne dust production during sanding even though coarse sandpaper removes up to four times as much wood as fine sandpaper (Thorpe et al., 1995).

Table 1. Carpenter Shop Major Wood Working Equipment High Bay Low Bay 1 Table Saw 3 Table Saws 1 Radial Arm Saw 1 Radial Arm Saw 1 Panel Saw 2 Band Saws 1 Drill Press 2 Planers 1 Band Saw 1 Horizontal Sander 1 Boring Machine 1 Speed Belt Sander 1 Disc Sander and Grinder 2 Lathes 2 Miter Saws 1 Vertical Sander 2 Drill Presses 1 Disc Sander and Grinder 1 Horizontal Boring Machine 2 Table Routers 1 Panel Router

Occupational Standards and Guidelines

The ACGIH TLV for exposure to hardwood dust is one milligram of dust per cubic meter of air (1 mg/m3) based on an 8-hour time-weighted-average (TWA). The ACGIH TLVs for exposure 153

to softwood dust are 5 mg/m3 based on an 8-hour TWA, and 10 mg/m3 as a short-term exposure limit (STEL) based on a 15-minute TWA. The ACGIH TLVs for wood dust were established on the expectation that chronic exposures at or below the TLVs would not lead to adverse health effects for most workers (ACGIH 1991). The current TLV for hardwood dust is based primarily on a study which indicated that mean exposures of 2.2 mg/m3 (range of 1.0 - 2.9 mg/m3) had an increased likelihood of causing wood-induced nasal mucostasis, a condition that may be a contributing factor in the development of nasal adenocarcinoma in furniture workers (ACGIH 1991; Andersen et al., 1977).

The Occupational Safety and Health Administration (OSHA) does not currently have a Permissible Exposure Limit (PEL) for hardwood or softwood. Before 1980, OSHA regulated wood dust under the nuisance dust standard of 15 mg/m3. However, in a 1985 enforcement proceeding before the Occupational Safety and Health Review Commission, the Administrative Law Judge ruled that wood dust was not an inert mineral dust and therefore not covered under the nuisance dust standard (ACGIH 1991; NIOSH 1987). In 1989, OSHA proposed a 5 mg/m3 PEL for hard and softwood based on an 8-hour TWA, a 10 mg/m3 PEL for hard and softwood as a 15 minute short term exposure limit (STEL), and a 2.5 mg/m3 PEL for western red cedar based on an 8-hour TWA (OSHA 1989). These proposed PELs were subsequently rejected by the U.S. Court of Appeals (US 1992).

The National Institute for Occupational Safety and Health (NIOSH) established a Recommended Exposure Limit (REL) of 1 mg/m3 based on an 8-hour TWA for hardwood and softwood dust as an interim level to be followed by future rule making (NIOSH 1992). OSHA considered the NIOSH support for a PEL and the ACGIH TLV of 1 mg/m3 for hardwood dust. OSHA determined that a 5 mg/m3 PEL for hardwood dust was feasible (OSHA 1989). OSHA is required to consider the feasibility of each PEL, whereas the ACGIH TLV Committee develops guidelines which are not intended to be interpreted as legal standards and are primarily based on the potential for adverse human health effects without specific regard for feasibility (ACGIH 1991). Many industry representatives have argued that complying with a 1 mg/m3 criteria for hardwood dust would be technologically and/or economically infeasible and at least extremely difficult.

Most carpenter shops in the United States are privately owned and, in the absence of a new OSHA standard for wood dust, are not currently mandated to maintain personnel exposures to hardwood or softwood dust below the ACGIH TLVs. This carpenter shop performs work under contract with the Department of Energy (DOE) and is mandated by DOE Orders to comply with OSHA standards and ACGIH guidelines (whichever are most stringent). Therefore the target criterion for personal hardwood dust exposures in this carpenter shop is less than 1 mg/m3 based on an 8-hour TWA.

Sampling and Analytical Methodologies

Personal "total dust" sampling was performed using 37 millimeter diameter sampling cassettes containing 5.0 micrometer pore size polyvinyl chloride (PVC) filters for the collection of wood dust and gravimetric analysis in accordance with NIOSH Method 0500 (NIOSH 1984). The filters were placed in line with SKC® and Gilian® brand air sampling pumps which were pre- calibrated to a flow rate of 2.0 liters per minute using a precision rotameter which had been calibrated against a primary standard. Post-calibration was also performed with the same

154

rotameter to check if wood dust build-up may have changed the initial flow rates. Each sample filter cassette was located in the worker's breathing zone approximately three inches from the worker's mouth for the duration of the work shift. Personal sampling was conducted without the use of cyclones or other particulate sizing devices due to the fact that the target occupational exposure criteria for wood dust is based on total inhalable wood dust. Inhalable dust samples were also collected in the later stages of this study using an inhalable particulate mass sampler for comparison with the total dust sample results, however these data have been submitted for separate publication at a future date. All samples were analyzed by an American Industrial Hygiene Association accredited laboratory. Area samples were also collected during the latter rounds of sampling in order to characterize work area airborne wood dust concentrations. Short-term (15 minute) samples were not collected for comparison with the ACGIH TLV-STEL for softwood because this study focused primarily on hardwoods and most of the operations which generated substantial airborne wood dust, such as sanding, were performed on hardwoods.

A preliminary quantitative comparison of hand-held sanders and available dust control equipment was also performed to determine which hand-held sanders and wood dust capturing devices were most effective in reducing wood dust exposure levels. Sanders were compared by collecting air samples in a carpenter's breathing zone while different sanders and dust control equipment were used. This sampling (hereafter referred to as sanding tests) was conducted using the total dust sampling methodology during 20 minute test periods while a carpenter operated test sanders on the same type of hardwood in a manner consistent with normal work practices.

Progression Towards "Compliance"

Initial Personal Sampling Initial baseline personal sampling for wood dust was conducted on two separate days. Five different workers were monitored at least once during the two day period. Sample results are listed in Table 2 including TWA results. The personal 8-hour TWA sampling results for the initial sampling ranged from 0.096 mg/m3 to 4.3 mg/m3. The majority of work performed during this sampling period was with hardwood (walnut, birch and oak) but some of the work used relatively small amounts of plywood which contained softwoods (such as fir and pine). Four of the sample results exceeded the ACGIH TLV of 1 mg/m3 for hardwood dust. Two of the four high sample results were from workers who only performed work with hardwood. Workers were observed using compressed air for personal and area cleaning. Housekeeping was infrequently performed and floor accumulations of wood dust were cleaned up using dry sweeping methods.

Corrective Actions The use of compressed air for cleaning was discontinued and housekeeping was improved. The capacity of the local exhaust ventilation, which services 14 stationary woodworking tools in both the high and low bay, was increased by 14%. This increase from 4900 CFM to 5600 CFM was accomplished by increasing the speed of the blower, changing out the existing 20 HP motor with a 25 HP motor, and installing an adjustable motor pulley.

Second Round of Personal Sampling The second round of personal sampling was conducted on one day to measure the effectiveness of the corrective actions to reduce airborne wood dust concentrations. Most of the work performed during this sampling period was with hardwood (primarily oak and other hardwoods) but some of the work used oak-faced plywood which 155

contained softwoods (such as fir and pine) covered by a thin oak veneer. The amount of actual woodwork performed on this day was observed to be relatively light to average compared with other work days. Personal sampling during this period, and during subsequent sampling periods, included separate morning and afternoon samples to better characterize the relationships of work activities with the airborne dust concentrations. The morning and afternoon sampling results are representative of airborne concentrations which could potentially occur if the same work activities were to be performed for an entire work shift.

The personal 8-hour TWA air monitoring results for the second round of sampling, summarized in Table 2, ranged from 0.095 mg/m3 to 1.2 mg/m3. The non-time-adjusted sampling results ranged from 0.16 mg/m3 to 2.0 mg/m3. The personal TWA sampling results for the second round of sampling did indicate airborne dust concentrations substantially less than the Round 1 monitoring results. Two workers' TWA sample results exceeded the TLV of 1 mg/m3 for hardwood. However, some of the work during their sampling period involved softwood so these sampling results could not be considered as necessarily over the TLVs since the results were all well below the TLV for softwood (5 mg/m3). The morning and afternoon sample results indicated a continued potential to exceed the TLV for hardwood during sanding and table saw cutting for long periods of time. Overall, this round of sampling results indicated the need for additional corrective actions to ensure personal exposures below the TLV for hardwood dust.

Table 2. Carpenter Shop - Wood Dust Monitoring Results Total Dust Method

SAMPLE 8 hour Sampling Sample TIME ACTIVITIES2 RESULT TWA3 3 Round Type1 (min.) (mg/m3) (mg/m ) Round 1 P 441 - Sanding and cutting oak 0.91 0.84 Day 1 and oak-faced plywood P 485 - Cutting and sanding walnut 2.1 2.1 P 428 - Sanding pine, Cutting fir 0.48 0.43 plywood P 360 - Cutting and sanding birch 1.6 1.2 Round 1 P 467 - Sanding walnut and 4.4 4.3 Day 2 walnut-faced plywood - Sanding, cutting and planing pine and fir plywood P 477 - Cutting fir and pine 0.86 0.85 plywood - Cutting Formica P 404 - Cutting and sanding oak 2.9 2.4 and birch 460 - Assembling cabinets, 0.10 0.096 Cutting Formica, Sanding birch

1 P = Personal; A = Area 2 Activities are listed in decending order based on the amount of time spent during the sampling period. 3 TWA = Time-weighted-average 156

Round 2 P 200 - 1/2 sheet hand-held power 1.1 (morning) sander with no bag, oak; 16 inch table saw, oak. 10 inch table saw , oak and oak- faced plywood, Hand-held belt sander with bag , oak P 209 - Approximately same as 1.7 1.2 (afternoon) morning P 184 - Table Saw, birch 1.2 (morning) P 203 - Table Saw, birch and oak- 1.6 1.1 (afternoon) faced plywood P 191 - 10 Inch table saw, oak 0.16 (morning) P 174 - 10 Inch table saw, oak and 2.0 0.79 (afternoon) oak-faced plywood; Table Planer, oak; Table Joiner, oak P 147 - Vibration sander with bag, 0.31 0.095 (afternoon oak; Belt sander with bag, only) oak; Routing, oak; Hand- held Jig Saw, oak faced plywood Round 3 P 280 - Light hand sanding table 0.70 Day 1 (morning) edge, birch; Limited Dynabrade Orbital Sander, birch; Cut 1 small piece of plexiglass; Delta table saw, oak- faced plywood; Ritter sander, oak- faced plywood P 160 - Face frame construction 0.63 0.62 (afternoon) (nailing/gluing); Table saw, oak-faced plywood; Limited hand routing, birch P 276 - Measuring; Table saw, fir 0.49 (morning) plywood (2 sheets); Limited sanding wheel, fir; Makita 14" miter saw P 159 - Gluing; Routing, fir 0.60 0.48 (afternoon) plywood; Vacuum cleanup P 230 - Porter Cable edge routing, 0.91 (morning) fir plywood; Table saw, fir plywood; Measuring, gluing and nailing; Steibig Standard cutting, fir plywood; Makita 14" miter saw, fir plywood P 202 - Nailing and gluing; Makita 0.40 0.60 (afternoon) 14" miter saw, fir plywood; Limited table saw, oak faced plywood Round 3 P 295 - Cabinet assembly, oak and 1.5* Day 2 (morning) oak-faced plywood; Machine rotary sanding, oak and oak- faced plywood; Delta table saw, oak-faced plywood

* Sampling cassette was found to have slipped down from the breathing zone location for a portion of the sampling period. The cassette location during this time over-estimates breathing zone concentrations. 157

P 221 - Cabinet assembly; 1.3 1.3 (afternoon) Routing, oak; Delta table saw, oak-faced plywood P 236 - Dynabrade palm sander, 1.1 (morning) birch

P 221 - Dynabrade palm sander, 1.2 1.1 (afternoon) birch; Routing, birch; "Hand scraper" on edges, birch P 295 - Birch cabinet assembly; 0.51 (morning) Speed belt sanding, birch; Routing, birch P 157 - Birch cabinet assembly; 0.25 0.40 (afternoon) Speed belt sanding, birch; Routing, birch Round 4 A 458 0.28 --- Day 1 A 445 0.32 --- P 420 - Porter Cable palm sander, 2.4 2.1 birch and birch plywood; Routing, birch and birch plywood; Nail gun assembly P 395 - Table saw, pine; Routing, 3.6* 3.0 pine; Lathe, pine Round 4 A 400 0.36 --- Day 2 A 398 0.34 --- A 395 0.56 --- P 382 - Area cleanup with vacuum 1.2 0.96 and broom; Table saw, plywood; Hand sanding, plywood; Miscellaneous shop help Round 5 P 270 - Job preparations and 0.33 Day 1 (morning) material ordering; Streibig Standard cutting - Dynabrade palm sander4; Saber saw; All work with birch-faced plywood P 144 - Dynabrade palm sander4; 0.17 0.24 (afternoon) Routing; One hour meeting and quit early; All work with birch-faced plywood P 240 - Miscellaneous non-wood 0.79 (morning) work; Striebig Standard and table saw cutting; Makita belt sander, edge sanding; Used compressed air to clean table once; All work with oak-faced plywood P 236 - Dynabrade palm sander4, 0.15 0.47 (afternoon) flat and edges; Gluing; Limited chop saw and hand sanding; All work with oak- faced plywood

4 Used the new ventilation system for portable sanders 158

P 281 - Miscellaneous non-wood 0.60 (morning) work; Striebig Standard and table saw cutting; Makita belt sander, edge sanding; Used compressed air to clean table once; All work with oak-faced plywood P 188 - Dynabrade palm sander4, 0.080 0.38 (afternoon) flat and edges; Gluing; Limited chop saw and hand sanding; All work with oak- faced plywood P 118 - Miscellaneous shop help 0.22 0.054 (afternoon including handling of scrap only) with dust accumulations; Limited table saw and routing, plywood; Quit early A 226 0.11 (afternoon only) Round 5 P 259 - Porter Cable Palm 0.25 Day 2 (morning) sanding4, oak; Hand sanding, oak; Table saw, oak and oak-faced plywood; Machine planing, oak P 78 - Porter Cable Palm 0.45 0.21 (afternoon sanding4, oak; Machine planing, oak; Quit early P 244 - Measuring and planning, 0.10 (morning No wood work P 82 - Measuring and planning, < 0.061 0.06 (afternoon) No wood work - Quit early P 235 - Dynabrade palm sanding4; 0.70 (morning) Streibig Standard cutting; Table saw cutting; Drill press; Bandsaw cutting; All work with oak-faced plywood P 219 - Table saw cutting; 1.5 1.0 (afternoon) Dynabrade palm sanding4; All work with birch-faced plywood P 236 - Job planing and field work, 0.51 (morning) all non-wood work; Limited hand sanding, birch; Limited table saw cutting, walnut P 103 - Job planning, all non-wood < 0.048 0.26 (afternoon) work; Limited Table saw cutting, birch-faced plywood;Glue laminating/equipment repair; Quit early A 214 0.047 (morning) A 211 0.071 --- (afternoon)

159

Corrective Actions Industrial hygiene personnel conducted a one hour meeting with all carpenter shop personnel, including supervision. An important aspect of the meeting was to solicit employee ideas for methods to further reduce airborne dust concentrations. Workers had numerous suggestions including several very constructive ideas which had not already been identified by industrial hygiene or facility management personnel.

New table saw guards were purchased and installed which allowed for the local exhaust system to be connected to the top of each table saw and collect dust above the saw blade. All other local exhaust connections on the table saws were from the bottom of the saw. It was expected that the majority of the largest and heaviest wood dust particulate would fall below the table saw. The relatively smaller particulate, which would be more likely to be inhalable, would not necessarily be well controlled without dust collection above the table level.

Inexpensive pneumatic cleaning devices were purchased and installed for easier surface dust collection. These devices consisted of a switch-activated portable wand which operated off the compressed air system utilizing the venturi effect. Each cleaner was equipped with a bag collector for the wood dust. The size and configuration of each cleaner was similar to a hand- held vacuum cleaner. These were intended as a better alternative for small area clean-up than dry sweeping. However, these cleaning devices were rarely used because shop personnel preferred vacuums with larger collection surface areas.

One horizontal sander was found to have been modified in such a way as to project wood dust away from the local exhaust system collection area. This sander was reconfigured back to its original operating position to maximize dust collection.

Workers were advised to regularly inspect bag collection systems for possible connection leaks, holes or tears in the bag, or poorly designed or otherwise inefficient bags. They were also instructed in the proper method of emptying and replacing bags and vacuum canisters to avoid unnecessary wood dust exposure.

Several carpenters suggested that they would be more likely to use vacuum attachment devices if they were more conveniently available. Repositioning the vacuum unit and hose is often necessary, especially for work on large work pieces. Many carpenters also preferred to use the sanders without any bag or local exhaust collection because the connection obstructed their view of the work piece. A pull-down vacuum hose system with one vacuum hose located directly above each major sanding work station was needed to improve this situation. Each carpenter could attach and detach the vacuum hose to compatible sanding equipment and continue each sanding task with minimal setup labor, and the ventilation controls would be out of the way. This suggestion was seriously considered along with other engineering control options. The initial cost estimate for such a system was approximately $80,000 and was not considered to be financially feasible. This suggestion was eventually implemented, at a much lower cost, but not until after the fourth round of sampling as described below.

Third Round of Personal Sampling The third round of personal sampling was conducted over a period of two non-consecutive work shifts representative of different work levels and activities. The woodworking activity performed on the first day was relatively low compared with other workdays and most of the woodwork involved softwoods (such as fir and pine). The 160

woodworking activity performed on the second day was relatively high and most of the woodwork involved hardwoods (such as oak and birch). Some of the woodwork on both days involved oak-faced plywood which contained softwoods covered by a thin oak veneer. Based on all the corrective actions implemented up to this round of sampling, the expected result was to document personal exposures well below the TLVs for wood dust.

Air monitoring results (Table 2) from air samples collected on the first day were all less than 1 mg/m3 and well below the TLV for softwood of 5 mg/m3. Personal 8-hour TWA sample results ranged from 0.48 mg/m3 to 0.62 mg/m3. Considering that the work level on this day was low and much of the workers' time was spent measuring and assembling cabinets or talking to co-workers, the air sampling results were higher than expected.

The personal 8-hour TWA sample results for the second day (Table 2) ranged from 0.40 to 1.3 mg/m3. Sampling on this day was representative of potential worker exposures to wood dust during relatively high woodworking activity. One worker's TWA sample result of 1.1 mg/m3 was slightly higher than the applicable hardwood TLV. This work was performed with hardwood only (birch). This carpenter spent approximately 6 hours sanding with the Dynabrade™ palm sander, so this monitoring was a good test case for the use of similar palm sanding equipment for long periods of time.

A second worker's TWA sample result of 1.3 mg/m3 was higher than the hardwood TLV and well less than the softwood TLV. This worker performed work on this day with hardwood and softwood. However, the morning sample cassette was found to have slipped down from the breathing zone location for an unknown period of time. The cassette location during this time was closer to the source of dust generation and therefore may have over-estimated the actual breathing zone concentration.

Sample results which were not TWA adjusted (listed in the "sample result" column of Table 2) 3 ranged from 0.25 to 1.3 mg/m . These results are representative of airborne concentrations which could potentially occur if the same work activities were to be performed for an entire work shift. These sample results indicated a continued potential to exceed the TLV for hardwood during long-term sanding and table saw cutting.

Overall, this round of sampling results still indicated the need for additional corrective actions to ensure personal exposures below the TLV for hardwood dust. This round of sampling results also indicated a possible "background" airborne dust concentration of approximately 0.4 mg/m3 based on sample results from carpenters performing the least amount of actual woodworking. This level included the modification of existing engineering controls, work practices and cleanup practices, and even a relatively low level of woodworking activity. At this point in the study, there were no area samples to confirm this estimate of background dust concentrations.

Housekeeping throughout the carpenter shop was observed to be very good during this round of sampling. Carpenter shop workers and maintenance personnel regularly cleaned their work areas. Area cleanup was performed by dry sweeping but this method did not generate much visible airborne dust, possibly because the dust was usually collected on the same surface and was not swept from table tops to the floor. There was also less dust to be cleaned-up

161

throughout the shop due to the implemented dust control improvements. Compressed air and the new pneumatic vacuums were not used for area cleanup.

Corrective Actions Custodial cleanup using dry sweeping was minimized by providing a portable Nobles Typhoon wet/dry vacuum system that is currently equipped with a paper bag rated at 35% filter efficiency at 1.0 micrometer.

The personal sampling results through round three sampling showed enough agreement with the sanding test results (described below) to support using the sanding test data as a basis for equipment use decisions. "Use guidelines" (shown in Table 3) were developed based on the sanding test results and were to be used as an administrative tool to minimize potential worker exposures to wood dust, particularly hardwood dust, during periods of relatively high woodworking activities and long-term sanding tasks. The use guideline is the maximum length of time during any one 8-hour work shift that the specific tool should be used to remain in compliance with the TLV without additional controls. A background concentration of 0.4 mg/m3 was assumed whenever the specified tool was not in use in order to calculate the use guidelines. The use guidelines were intended as a temporary measure until additional engineering controls for the sanding equipment could be researched, selected and implemented. These guidelines were considered to be too restrictive to meet production goals but were acceptable on an interim basis.

Table 3. Sanding Tool Use Guidelines

USE GUIDELINE1 Sanding Tool Control Hardwood Makita Belt None 15 minutes Bag Attachment 2.2 hours Porter Cable 1/2 Sheet Vacuum attachment on but not in use 13 minutes None 21 minutes Vacuum No limit Porter Cable Palm None 1 hour Bosch Palm Bag Attachment 42 minutes Bosch 1/2 Sheet Bag Attachment 51 minutes Makita Palm None 2 hours Dynabrade Orbital Palm Vacuum 2.2 hours Bag Attachment 2.8 hours Fein 1/2 Sheet Vacuum 4 hours 1 Use guideline is the maximum length of time during any one 8-hour work shift that the specific tool should be used to remain in compliance with the ACGIH® 8-hour TWA guidelines without additional controls.

The efficiency of the local exhaust system was improved by the addition of automatic duct dampers. Each damper was wired to allow air flow to individual shop tools only when the tool was turned on. At all other times, the damper would close and allow for a larger exhaust volume at other active tool locations. An example of this improvement is best portrayed by comparing face velocity measurements of the hood located at the shop's drum sander. When combined with the previous improvement in the capacity of the local exhaust system, an overall 26% increase in the average face velocity, from 2844 fpm to 3578 fpm, is now

162

available when the drum sander is the sole stationary tool in use. These modifications also measurably reduced the noise level in the shop area. Carpenter shop personnel built a prototype sanding table in an attempt to better control the wood dust from the portable sanders at a lower cost than installing another local exhaust ventilation system, or attempting to further modify the existing ventilation system. The sanding table had a one horsepower fan and a peg board top configured to draw air immediately from the table surface. Relatively small and flat cabinet components, such as door panels, can be placed on the sanding table top for sanding. Personal sampling was conducted using the total dust sampling methodology to evaluate the effectiveness of the sanding table to reduce the worker's personal exposure to airborne wood dust during sanding tasks. The sanding test methodology was repeated, without any changes, using the same carpenters except that the sanding table was used with the associated ventilation operating. Three different sanders were used and the results are illustrated as Figure 1. The airborne wood dust concentrations were reduced by more than 85% with the Porter Cable half sheet sander and the Porter Cable palm sander, and 56% with the Makita palm sander. These results indicated that the prototype sanding table was an effective dust control option for small and flat work pieces. Sanding on larger work pieces further than 6 - 12 inches from the sanding table surface was not observed to provide effective dust control.

Figure 1. Sanding Table Tests

Fourth Round of Sampling The fourth round of sampling included personal and area sampling on two non-consecutive days. Work on these days involved both hardwoods (primarily birch) and softwoods (primarily pine), and the woodworking activity on both days was relatively high. Personal 8-hour TWA air monitoring results (Table 2) for the fourth round of sampling ranged from 0.96 mg/m3 to 3.0 mg/m3. The non-time-adjusted sampling results ranged from 1.2 mg/m33 to 3.6 mg/m3. The higher personal sampling results were affected by two factors; 1) one worker 163

used a palm sander with no dust control device (2.1 mg/m3 TWA), and 2) another worker performed a lathe operation which projected large softwood particulates directly towards him at the same time that his sampling cassette was found to be positioned below his actual breathing zone (3.0 mg/m3 TWA). This improper positioning would tend to over-estimate the actual wood dust exposure. Area air sampling results ranged from 0.28 mg/m3 to 0.56 mg/m3. It is important to note that the ventilated sanding table, described above, was not used during this sampling period. Carpenters generally preferred their own work benches or the work pieces were too large for the sanding table.

Overall, this round of sampling results indicated personal exposures to wood dust below the ACGIH TLVs for wood dust since most of the woodwork was performed with softwood. However, the results still indicated the potential to exceed the hardwood TLV during hardwood activities, particularly sanding.

Corrective Actions Two supplemental local exhaust systems (4.5 and 6.7 HP) were installed (one in each Bay) for dedicated local exhaust of portable hand tools, particularly sanders. These systems allow each carpenter to attach existing portable tools to a drop-down line at their work stations. The overhead positioning of the exhaust connection avoids damage to the work piece (e.g. cabinet or panel) and exhaust line by reducing contact between them. The overhead positioning also provides the carpenter with a clearer view of the sanding surface. These systems were fully operational during the fifth round of sampling except that the overhead positioning feature had not yet been installed. The ventilation performance of this system tested after installation provided an air flow velocity of 1100 - 1300 feet per minute (fpm) through each of ten (10), 1.5 inch diameter tool hook-up locations (approximately 15 cubic feet per minute for each location). The effective performance of the system was dependent on the configuration and characteristics of the selected woodworking tool.

Work benches were modified to allow for the powered raising and lowering of the bench table. This modification was made for ergonomic reasons, primarily to reduce back injuries, and was not related to efforts to reduce airborne wood dust. It is not known if the use of such work benches will tend to increase or decrease breathing zone concentrations of wood dust. Consideration of this factor is recommended for future study.

Fifth Round of Sampling The fifth round of sampling included personal and area sampling on two non-consecutive days. Work on the first day involved both softwoods and hardwoods and the woodworking activity was relatively low. Work on the second day involved primarily hardwoods (oak and birch) with some softwoods (primarily pine), and the work level was relatively high. Personal 8-hour TWA air monitoring results (Table 2) for the fifth round of sampling ranged from 0.054 mg/m3 to 1.0 mg/m3. The non-time-adjusted sampling results ranged from < 0.048 mg/m3 to 1.5 mg/m3. The highest sample result of 1.5 mg/m3 was collected from a carpenter who primarily performed table cutting of birch-faced plywood. Area air sampling results ranged from 0.047 mg/m3 to 0.11 mg/m3.

Sanding Tests

Sanding tests were conducted between the second and third round of personal sampling to initially characterize the relative dust control for different portable sanders. The need to characterize sanders was determined by their frequency of use at the carpenter shop, by studies indicating that sanding produces relatively high levels of airborne dust (Hinds 1988; Holliday et 164

al., 1985; Thorpe and Brown 1995) and by field observations of visible airborne dust generated during sanding work. A total of eight (8) different hand-held power sanders were used by two (2) carpenters for twenty (20) minutes each on the same type of hardwood surface and in the same manner of operation. Three (3) sanders were tested with and without wood dust collecting devices. The results for all of the sanding tests are illustrated as Figure 2. The results are considered preliminary in nature due to the limited number of tests and repetitions conducted. Additional sampling would be necessary in order to more accurately predict the actual airborne dust concentrations when each sander is used.

Figure 2. Sanding tests: comparison of sanders and wood dust control devices.

The Makita belt sander and the Porter Cable half sheet sander with no dust collection devices generated the highest airborne dust concentrations. Significant amounts of dust were thrown onto the carpenter's pants during the Makita Belt sander test. A bag collection device used with the Makita belt sander reduced the airborne dust concentration from 19 mg/m3 to an average of 2.6 mg/m3 (two repetitions). Based on these results, limited use of the Makita belt sander was recommended unless it could be fitted with an effective dust control device. The Porter Cable half sheet sander was intended to be used with a vacuum canister attachment but the attachment was rarely used. The vacuum attachment used with the Porter Cable half sheet sander reduced the dust sample result from an average of 18 mg/m3 (two repetitions) to less than 0.25 mg/m3. The sander was actually effective in collecting existing wood dust off the work station surface when the vacuum attachment was used. The vacuum attachment, however, extends out from the perimeter of the sander which limits the carpenter's visibility of the work surface. Therefore, this attachment is best suited for surface preparation rather than fine finish sanding.

The Dynabrade™ orbital palm sander (with vacuum or bag attachment) and the Fein vibrating half sheet sander (with vacuum attachment) generated the least amount of airborne wood dust in the carpenter's breathing zone. The Dynabrade™ orbital palm sander was designed to be used with a collection bag rather than with a vacuum attachment, even though the hose makes it tempting to use a vacuum. The air sample results for the Dynabrade™ orbital palm sander, with

165

collection bag (2.0 - 2.2 mg/m3) verses vacuum attachment (1.8 - 3.5 mg/m3), indicated a small improvement in dust collection efficiency when used with the collection bag attachment.

Good work practices can also help to minimize wood dust exposure. Some sanders were observed to throw dust primarily to one side of the sander. Therefore, certain sanders were found to be more appropriate for right or left handed tool position depending on the direction of wood dust exhaust. Positioning the sander's exhaust as far away from the breathing zone as possible will help to lower the concentration of wood dust inhaled. In the same manner, straightening one's arms to keep the head and breathing zone away from the work surface will reduce wood dust exposures.

The sanding test results indicated that using appropriate engineering controls and good work practices with hand sanders can dramatically reduce the user's wood dust exposure. This information was then used as a basis for equipment use decisions. The lessons learned from the sanding tests were discussed with carpenter shop personnel, and workers were encouraged to consider tool selection and work practice improvements each time they performed a woodworking task. It was emphasized that their choice of tools, tool position and body position during the work has a significant impact on their personal exposures to wood dust. Industrial hygiene personnel did not want to unnecessarily restrict the tools that could be used in the shop. However, the sanding test results clearly indicated that some sanders could not be used for very long without exceeding the TLV for hardwood and excursion values, especially if they were not equipped with a bag attachment or vacuum device. Carpenter shop management personnel decided to remove selected sanding equipment, such as a Makita Belt sander with a damaged bag attachment, from the shop area so that carpenters would use other available sanding tools.

Discussion and Conclusions

Personal air monitoring results for Rounds 4 and 5 samples were all less than the applicable ACGIH TLVs for wood dust. The samples collected in Round 5 after installation of the new exhaust system for portable sanders indicated reasonably effective wood dust control. None of the Round 5 sampling results would have exceeded the ACGIH TLV for hardwood dust even if all the work had only involved hardwoods. This is an important finding because the relative woodworking activity on the second day of Round 5 sampling was high and a substantial portion of this work involved hardwood and sanding activities. Future rounds of air monitoring will be necessary to gain statistical confidence of effective dust control.

Achieving the current ACGIH TLV of 1 mg/m3 for hardwood dust in a typical carpenter shop has been demonstrated to be possible but very difficult. Improvements focused on engineering controls and work practice modifications in order to reduce airborne concentrations of wood dust to personal exposure levels consistently below the TLVs. Reducing most personal wood dust exposures below 2 mg/m3 was accomplished relatively easily, however reducing the exposures below 1 mg/m3 was considerably more difficult and expensive. The key to further reducing the hardwood dust concentrations was to implement effective local exhaust control for portable tools, especially sanders.

Achieving the current TLV-TWA of 5.0 mg/m3 for softwood dust was not difficult due to the presence of local exhaust ventilation at many major tool locations such as table saws and planers. However, with increasing evidence that softwood dust may be a suspect human carcinogen, the TLV for softwood may be reduced and both hardwood and softwood may have 166

to be controlled to similar levels. The use of respiratory protective equipment is not considered a desirable method for worker protection from wood dust in a carpenter shop due to; 1) the expectation that safety-related incidents may become more frequent as a result of respirator usage, 2) worker comfort would be compromised, especially during hot summer months, and 3) the desire to provide a safe and healthful work environment without the use of personal protective equipment.

A recent English study involving the evaluation of airborne dust generation from hand sanders (Thorpe and Brown 1995) found that orbital sanders provided relatively good dust control on flat wood surfaces but not on edges. The local exhaust systems installed in this carpenter shop for portable sanding tools did have sufficient capture velocity to provide good dust control even on edges for most sanders.

The current ACGIH TLV of 1 mg/m3 for hardwood dust, while not considering feasibility, is regarded as a conservative guideline to protect workers from the development of mucostasis, adenocarcinoma and respiratory irritation. An occupational exposure limit of 2 mg/m3 for hardwood dust (measured as total dust) is considered reasonable, including consideration of feasibility. This level still adequately reduces airborne hardwood dust to concentrations below those expected to cause mucostasis, adenocarcinoma and respiratory irritation. This is consistent with the conclusions of previously published studies (Pisaniello et al., 1991; Andersen et al., 1977; Whitehead et al., 1981; Whitehead 1982). These study results and practical experience also illustrate the potential error in a policy adopting ACGIH TLVs as mandatory occupational exposure criteria, rather than as the guidelines they are intended to be.

Acknowledgement

The authors would like to thank Richard E. Fairfax, CIH for technical review assistance in preparation of this manuscript and the related manuscript to follow.

167

ACRONYM GLOSSARY

ABPI Association of the British Pharmaceutical Industry BAuA Federal Institution for Industrial Safety and Medicine (Germany) BOHS British Occupational Hygiene Society CB Control Banding CEFIC European Chemical Industry Council CEMAS CEFIC Exposure Management System CGS Control Guidance Sheets CIA Chemical Industries Association CHIP Chemical Hazardous Information and Packaging CL Control Level COSHH Control of Substances Hazardous to Health CPWR Center for Construction Research and Training EASE Estimation and Assessment of Substances Exposure ECETOC European Centre for Ecotoxicology and Toxicology of Chemicals ECHA European Chemical Agency ECP Exposure Control Practices EF Exposure Factor EINECS European Inventory of Existing Substances eLCOSH Electronic Library of Construction Occupational Safety and Health EU European Union GB Great Britain GHS Globally Harmonized System of Classification and Labeling of Chemicals GTZ German Technical Cooperation (Germany) HHE Health Hazard Evaluation HPD Hearing Protection Devices HSE Health and Safety Executive (UK) HSDB Hazardous Substances Data Base ICBW International Control Banding Workshop ICCT International Chemical Control Toolkit (new name of the ILO Toolkit) IH Industrial Hygiene ILO International Labor Organization IOHA International Occupational Hygiene Association ISSA International Social Security Association LE Large Enterprises LEV Local Exhaust Ventilation LLNL Lawrence Livermore National Laboratory LOAEL Lowest Observed Adverse Effect Level MAK Maximum Allowable Concentrations (Germany) MoE Margins of Exposure MSD Musculoskeletal Disorder NIOSH National Institute for Occupational Safety and Health (US) NL The Netherlands NOAEL No Observed Adverse Effect Level NORA National Occupational Research Agenda OEB Occupational Exposure Band OEL Occupational Exposure Limit ORM Occupational Risk Management 168

OSHA Occupational Safety and Health Administration (US) OSHH Occupational Safety, Health and Hygiene PACE Prevention and Control Exchange PB-ECL Performance-Based Exposure Control Limits PJHA Pre-Job Hazard Analysis PPE Personal Protection Equipment PPM Parts per million REACH Registration, Evaluation, and Authorization of Chemicals RF Reduction Factor RL Risk Level RLBMS Risk Level Based Management System RTECS Registry of Toxic Effects of Chemical Substances SME Small- and Medium-Sized Enterprises SOC Skill-of-the-Craft SQRA Semi-Quantitative Risk Assessment T2C Task-to-Control TLV® Threshold Limit Value TWA Time Weighted Average UK United Kingdom US United States UW University of Washington WHO World Health Organization

169

REFERENCES

ABPI (1995). Association of the British Pharmaceutical Industry (ABPI): Guidance on setting in-house occupational exposure limits for airborne therapeutic substances and their intermediates. London: ACGIH (1991). American Conference of Governmental Industrial Hygienists (ACGIH).: Documentation Of The Threshold Limit Values And Biological Exposure Indices, Sixth Edition, Cincinnati, OH. ACGIH (2008) Control Banding: Issues and Opportunities. A report of the ACGIH Exposure/Control Banding Task Force, ACGIH Publishing, Cincinnati, OH Acheson, E.D.; Hadfield, E.H.; Macbeth, R.G. (1967). Carcinoma of the nasal cavity and accessory sinuses in woodworkers. Lancet, 1:311-312. Acheson, E.D. (1976). Nasal cancer in the furniture and boot and shoe manufacturing industries. Preventive Medicine, 5:295-315. Ackoff RL (1999). Re-creating the corporation, a design of organizations for the 21st Century. Oxford University Press. NY, NY, pgs 63-80. Adelmann, K. Chemical Safety and Development. Society for Technical Cooperation (GTZ) [German], 1-10 (2001). AIHA (2007) Guidance for conducting control banding analyses. American Industrial Hygiene Association (AIHA) Control Banding Working Group, AIHA Publications, Fairfax, VA. Albers, J. T., Li, Y., Lemasters, G., Sprague, S., Stinson, R., Bhattacharya, A. (1997) An ergonomic education and evaluation program for apprentice carpenters. American Journal of Industrial Medicine 32(6), 641-646. Ale B, Bellamy L, Baksteen H, Damen M, Goossens L, Hale A, et al. (2008). Accidents in the construction industry in the Netherlands: an analysis of accident reports using storybuilder. Reliab engin and syst safety. 93;1523- 1533. Andersen, H.C.; Andersen, I.; Solgaard, J. (1977). Nasal cancers, symptoms and upper airway function in wood workers. British Journal of Industrial Medicine. 34:201-207. Aneziris, O.N., Papzpglou, I.A., Baksteen, H., Mud, M., et al. (2008). Quantified risk assessment for fall from height. Safety Science, 46:198-220. Annals (1998) Special issue: Exposure limits and control banding. Ann Occup Hyg 42(6):355-407. Annals (2003) Chemical control in small and medium sized enterprises, control banding and other approaches. Papers from a meeting in London, November 2002. Ann Occup Hyg 47(7):531-577. ANSI (2000) Risk assessment and reduction: a guide to estimate, evaluate and reduce risks associated with machine tools. Association for Manufacturing Technology (B11, TR3-2000). McLean, VA Armstrong, T.J., et al. (1993). A conceptual model for work-related neck and upper-limb musculoskeletal disorders, Scandinavian Journal of Work, Environment, and Health 19:73-74. ASTM International (2007) Terminology for Nanotechnology. West Conshohocken, PA: ASTM International, E2456-06. Bałazy A, Toivola M, Reponen T, Podgorski A et al (2006) Based Performance Evaluation of N95 Filtering- Facepiece Respirators Challenged with Nanoparticles. Ann Occup Hyg 50(3):259-269. Baker, E. A., Israel, B. A., Schurman, S. J. (1994) A participatory approach to worksite health promotion. Journal of Ambulatory Care Management 17(2), 68-81. Balsat, A., P. Mairiaux, and J. De Graeve (2002). A global approach for assessing and managing chemical risks at the workplace: Advantages, limits and perspectives. Presented at 1ICBW, London, 4-5 November. Balsat, A., J. de Graeve, and P. Mairiaux (2003). A structured strategy for assessing chemical risks, suitable for small and medium-sized enterprises. Ann. Occup. Hyg. 47: 549-56 . Balsat, A., P. Mairiaux, and J. de Graeve (2004). Control banding for assessing chemical risks Belgian companies experience. Presented at 2ICBW, 1-2 March 2004, Cincinnati, OH. Baluyutt, R., et al. (1995) Use of visual perception in estimating static postural stresses: magnitudes and sources of errors. Ergonomics 38(9):1841-1850. Bellamy, LJ, Ale, BJM, Geyer, TAW, Goossens, LHJ, Hale, AR, Oh, J, Mud, M, Bloemhof, A, Papazoglou, IA & Whiston, J.Y. (2007). Storybuilder - A tool for the analysis of accident reports. Reliability engineering & system safety, 92, 735-744. Bentley T, Hide D, Tappin D, et al. (2006). Investigating risk factors for slips, trips and falls in New Zealand residential construction using incident-centred and incident-independent techniques. Ergon. 49: 62-77. Bernard, B. P. (1997) Elements of ergonomics programs: a primer based on workplace evaluations of musculoskeletal disorders. Cincinnati, National Institute of Occupational Safety and Health. US Department of Health and Human Services Publication Number 97 – 117. Blackadder, D. and Nedderman, Rl, (1971). A handbook of unit operations. Academic Press, London. BLS (1994). Bureau of Labor Statistics, Bureau of Labor Statistics (BLS): Annual Survey of Occupational Injuries and Illnesses (US Department of Labor).

170

BLS (2007). Bureau of Labor Statistics occupational injury, illness, and fatality data http://www.bls.gov/news.release/cfoi.toc.htm Bohr, P. C., Evanoff, B. A., Wolf, L. D. (1997) Implementing participatory ergonomics teams among health care workers. American Journal of Industrial Medicine 32(3), 190-196. Borm P, Houba R, Linker F (2008) Omgaan met nanodeeltjes op de werkvloer. Survey naar goede praktijken in omgaan met nanomaterialen in de Nederlandse industrie en kennisinstellingen. (Dealing with nanoparticles at the shop floor). Ministry of Social Affairs and Employment and Ministry of Housing, Spacial planning and Environment Bracker AL Morse TF Simcox NJ (2009). Training health and safety committees to use control banding: lessons learned and opportunities for the United States. J Occup Env Hyg, 6:307-314. Briton, L.A.; Blot, W.J.; Becker, J.A. (1984). A Case Control Study of Cancer Of The Nasal Cavity And Paranasal Sinuses. Am. J. of Epid. 119(6):896. Brooke, I.M. (1998). A UK scheme to help small firms control health risks from chemicals: toxicological considerations. Ann. Occup. Hyg. 42:377-390. Boussenna, M, Corlett EN, Pheasant ST. (1982). The relation between discomfort and postural loading at the joints. Ergonomics ;25:315-322. Buchanan, D., Cressey, P., Hiba, J. C., Scmid, F., Wilson, J. (1998). Work organization and ergonomics. Geneva, International Labour Office. Buckle, P., Devereux, J. (1999) Work-related neck and upper limb musculoskeletal disorders. Bilbao, European Agency for Safety and Health at Work. 114p. Burdorf, A. (1992) Exposure assessment of risk factors for disorders of the back in occupational epidemiology. Scandinavian Journal of Work, Environment, and Health 18(4), 1-9. Burdorf A., van der Beek, A. (1999) Exposure assessment strategies for work-related risk factors for musculoskeletal disorders. Scandinavian Journal of Work, Environment, and Health 25(4), 25-30. Burstyn, I. and K. Teschke (1999). Studying the determinants of exposure: a review of methods. Am. Ind. Hyg. Assoc. J. 60:57-72. Cable, J. (2006). Nanotech expert calls for more investment in EHS research. Occupational Hazards, November 2006. Castranova, V. (1998). Particles and the airways: Basic biological mechanisms of pulmonary pathogenicity. Applied Occupational and Environmental Hygiene; 13(8): 613-616. CCPS (1989), Centre for Chemical Process safety (CCPS) Guidelines for chemical process quantitative risk analysis. American Institute of Chemical Engineers, New York CDC (1993). Centers for Disease Control and Prevention (CDC): Biosafety in Microbiological and Biomedical Laboratories. 3rd Rev. J.Y. Richmond and R.W. McKinney eds. [HHS pub. No. (CDC) 93-8395] Washington, DC: US Government Printing Office 1993. pp. 6-43, 138-145. CDM (2007). Great Britain regulatory framework in Construction (Design and Management) Regulations, 6 April 2007. http://www.hse.gov.uk/construction/cdm.htm Cherrie, J. and T. Schneider (1999) Validation of a new method for structured subjective assessment of past concentrations. Ann. Occup. Hyg. 43:235-245. Chi CF, Chang TC, Ting HI. (2005). Accident patterns and prevention measures for fatal occupational falls in the construction industry. Appl Ergon. 36(4):391-400. CIA (1992) Chemical Industries Association (CIA): Safe handling of potentially carcinogenic aromatic amines and nitro-compounds. London: CIA. CIA (1993). Safe handling of colourants 2. London: CIA. CIA (1997). The control of substances hazardous to health regulations: guidance on allocating occupational exposure bands. (Regulation 7). London: CIA. CPWR (2004). Silica Hazard Alert http://www.cpwr.com/pdfs/pubs/hazard_alerts/KFsilica.pdf Collis E. Greenwood M. (1921). The health of the industrial worker. Churchill, London. Cook, R. (2007). Simplifying the creation and use of the risk matrix. In: Redmill F. and Anderson T. (Eds.) Improvements in systems safety. Springer. London. Corlett, E.N., Madeley, S.J., Manenica, I. (1979). Posture targeting: a technique for recording working postures. Ergonomics 22(3)357-366. Cory, H. (1999) The effects of production changes on the musculoskeletal disorders in Brazil and South America. International Journal of Industrial Ergonomics 25, 103-104. Coyle IR, Sleeman SD, Adams N (1995). Safety climate. J of Safety Res 26:247-54. Crabtree, H.C., D. Hart, M.C. Thomas, B.H. Witham, I.G. McKenzie, and C.P. Smith (1991) Carcinogenic ranking of aromatic amine and nitro compounds. Mutation Res. 264:155-162. Creaser W (2008). Prevention through design (PtD) safe design from an Australian perspective. J Saf Res. 39(2):131-134. 171

Croteau G., S. Guffey, M. Flanagan, and N. Seixas (2002). The Effect of Local Exhaust Ventilation Controls on Dust Exposures During Masonry Activities. AIHA Journal 63:458–467. De Blois (1926). Safety in the chemical industry. Annals Amer Acad Polit & Soc Sci, 123:127-131. De Jong AM, P Vink (2002). Participatory ergonomics applied in installation work. Appl Ergon 33(4):39-46. De Looze MP, IJM Urlings, P Vink, et al (2001). Towards successful physical stress reducing products: an evaluation of seven cases. Appl Ergon 32(5):525-534. DEC (1988). Dutch Expert Committee (DEC) for Occupational Standards. Wood Dust [Draft Report], WGD, Voorburg, Netherlands. Doll, R. (1955) Mortality from lung cancer in asbestos workers. British Journal of Industrial Medicine; 12:81-86. Donaldson, K, Li, X, and MacNee, W. (1998). Ultrafine (nanometer) particle mediated lung injury. Journal of Aerosol Science; 29(5-6): 553-560. Donaldson K, Aitken R, Tran L, Stone V, et al (2006) Carbon Nanotubes: A review of their properties in relation to pulmonary toxicology and workplace safety. Toxicol Sci 92(1):5-22. EC (1998). European Commission (EC): Council Directive 98/24/EC on the protection of the health and safety of workers from risks related to chemical agents at work. Off. J. Eur. Commun. L131. ECA (2007). European Chemicals Agency. http://ec.europa.eu/environment/chemicals/reach/reach_intro.htm. REACH. Accessed 15 June 2007. ECETOC (2002). European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC): Outline concept of a pragmatic approach to risk assessment. Brussels: http://www.ecetoc.org. Eekels, J. (1987). Design methods: possibilities and limits (in Dutch). Delft University of Technology. Eijkemans, G. and I. Fedetov (2005) Global implementation strategy of the occupational risk management toolbox (control banding). Presented at 3ICBW, 21 September 2005, Pilanesberg, South Africa.. EINECS (1987). European inventory of existing commercial chemical substances (EINECS Vols. I - VIII. Commission of the European Communities, EEC, Brussels, Belgium. Enarson, D.A.; Chan-Yeung, M. (1990). Characterization of health effects of wood dust exposures. American Journal of Industrial Medicine. 17:33-38 . Evanoff, B. A., Bohr, P. C., Wolf, L. D. (1999) Effects of a participatory ergonomics team among hospital orderlies. American Journal of Industrial Medicine 35(4), 358-365. Evans, P. (2005a). COSHH Essentials. Presented at the National Control Banding Workshop, 9-10 March 2005, Washington, DC. Evans, P. (2005b). Silica Essentials and the Chemical Control Toolkit. Presented at the Third International Control Banding Workshop (3ICBW), 21 September 2005, Pilanesberg, South Africa. Evans, P. (2005c). COSHH Essentials – Reducing Silicosis. Presented at 3ICBW, 21 September 2005, Pilanesberg, South Africa. Evans, P. and A. Garrod (2006). Evaluation of COSHH Essentials for vapour degreasing and bag-filling operations. Letter to the Editor. Ann. Occup. Hyg. 50:641. Farris, J. P., A.W. Ader, and R.H. Ku (2006). History, implementation, and evolution of the pharmaceutical hazard categorization and control system. Chemistry Today 24:5-10 . Fernholm A (2008) ―Nanoparticles scrutinized for health effects‖, San Francisco Chronicle, May 12, 2008. Feber, le F., Marquart, J., Brouwer, D., Tielemans, E., Tijsen, S., (2003). Model to assess inhalatoire exposure in SME‘s (in Dutch). TNO RAPPORT v5520 Fine, L.J., Punnett, L., Keyserling, W.M. (1987). An epidemiological study of postural risk factors for back disorders in industry, in P. Buckle (ed.) Muscular Disorders at Work (Taylor & Francis, London) 108 -109. Fine, W.T., Kinney W.D. (1971). Mathematical evaluation for controlling hazards. J Safety Res, 3(4):157-166. Fingerhut M, Driscoll T, Imel Nelson D, Concha-Barrientos M, Punnett L, Pruss-Ustin A, Steenland K, Leigh J, Corvalan C. (2005) Contribution of occupational risk factors to the global burden of disease – a summary of findings. Scandinavian Journal of Work, Environment & Health Suppl (1): 58-61 http://www.sjweh.fi/show_abstract.php?abstract_id=939 Fingerhut M (2008). Global qualitative risk management (control banding) activities. Industrial Health 46:305-307. Flanagan M.E., N. Seixas, M. Majar, J. Camp, M. Morgan (2003). Silica Dust Exposures During Selected Construction Activities. AIHA Journal 64:319–328. Flanagan M.E., N. Seixas, P. Becker, B. Takacs, J. Camp (2006). Silica Exposure on Construction Sites: Results of an Exposure Monitoring Data Compilation Project. J Occup Environ Hyg, 3(3):144-52. Flynn, M.R. and P. Susi (2003). Engineering controls for selected silica and dust exposures in the construction industry--a review. Appl. Occup. Environ. Hyg. 18:268-277. Foster KR, Vecchia P, Repacholi MH (2000). Risk management: science and the precautionary principle. Science 288:979-81. GAO (1997). General Accounting Office (GAO) Private Sector Ergonomics Programs. Washington, D. C., Health, Education, and Human Services Division. Publication number 97 – 163. 172

Gardner, R.J. and P.J. Oldershaw (1991). Development of pragmatic exposure-control concentrations based on packaging regulation risk phrases. Ann. Occup. Hyg. 35:51-59. Garrod, A. and R. Rajan-Sithamparananadarajah (2003). Developing COSHH essentials; dermal exposure, personal protective equipment and first aid. Ann. Occup. Hyg. 47:577-588. Geankoplis, G., (1978). Transport processes and unit operations. Allyn and Bacon, Inc., Boston. Genaidy, A.M., Al-Shedi, A.A., Karwowski, W. (1994). Postural stress in industry. Applied Ergonomics 25(2):77- 87. Genaidy, A., Barkawi, H., Christensen, D. (1995). Ranking of static non-neutral postures around the joints of the upper extremity and the spine. Ergonomics 38(9):1851-1858. Geraedts R, Wamelink H. (2008). Dutch status report: voluntary arrangements for collaborative working in the construction sector. EU Tendor No ENTR/07/008. Ghosh, A. and Mallik, Al (1986). Manufacturing science. Ellis Horwood Limited, Chichester. Ghosh, M. (1993) Ergonomic job analysis made easy. Safety & Health ;78-81. Goldsmith, DF.; Shy, C.M. (1988). Respiratory health effects from occupational exposure to wood dusts. Scandinavian Journal of Work and Environmental Health. 14:1-15. Goldsmith, D.F. (1997). The link between silica dust levels, risk assessments, and regulations. J. Expo. Anal. Environ. Epidemiol. 7:385-395. Goossens, L., Hourtolou, D. (2003). What is a barrier? Report ARAMIS project Sectie Veiligheidskunde, Technische Universiteit Delft Grandjean, E., Hunting, W. (1977). Ergonomics of posture: review of various problems of standing and sitting posture. Applied Ergonomics 8:135-140. Guest, I. (1998). The Chemical Industries Association Guidance on Allocating Occupational Exposure Bands. Ann. Occup. Hyg. 42:407-411. Guldenmund FW (2000). The nature of safety culture: a review of theory and research. Safety Sci 34:215-57. Guney, E.; Tanyeri, Y.; Kandemir, B. et. al. (1987). The Effect of Wood Dust On The Nasal Cavity And Paranasal Sinuses. Rhinology. 256:273. Haddon W. Suchman E. Klein, D. (1964). Accident research: methods and approaches. Harper and Row, New York Haddon, W. (1968). The changing approach to epidemiology, prevention and amelioration of trauma. Am J. Public Health, 58(8):1431-1438. Haddon, W. (1973). Energy damage and the ten countermeasure strategies. Human Factors, 15(4):355-366. Hagberg, M. (1981) Work load and fatigue in repetitive arm elevations. Ergonomics, 24(7):543-555. Hagberg M., Wegman, D. H. (1987) Prevalence rates and odds ratios of shoulder-neck diseases in different occupational groups. British Journal of Industrial Medicine 44, 602-610. Haims, M. C., Carayon, P. (1998) Theory and practice for the implementation of ‗in-house‘, continuous improvement participatory ergonomic programs. Applied Ergonomics 29(6), 461-472. Hale, A., Swuste, P., (1997). Avoiding square wheels: international experience in sharing solutions. Safety Science 25(1-3):3-14 Hale A., Ale, B., Goossens, L., Hiejer, T., et al., (2007). Modeling accidents for prioritizing prevention. Reliability Engineering and Systems Safety. 92: 1701-1715. Hallock MF, Greenley P, DiBerardinis L, Kallin D (2008) Potential risks of nanomaterials and how to safely handle materials of uncertain toxicity. J Chem H&S, 16(1):16-23. Hamill, A.; Ingle, J.; Searle, S.; Williams, K. (1991).. Levels Of Exposure To Wood Dust. Ann. Occup. Hyg. 35(4) 397-403. Harms-Ringdahl, L., (1993). Safety Analysis. Principles and practices in occupational safety. Elsevier Applied Sciences, London Haslem FA Hide SA et al (2005). Contributing factors in construction accidents. Appl Ergon. 36:27-31. Health and Safety Commission (2007). ‗Managing Health and Safety in Construction: Construction (Design and Management) Regulations 2007‘, Approved Code of Practice, L144, HSE Books. Heinrich H. (1929). The foundation of a major injury. National safety News January 19(1);9-11, 59 Heinrich H. W. (1941). Industrial Accident Prevention, New York, McGraw-Hill. Helland A, Scheringer M, Siegrist M et al (2008) Risk assessment of engineered nanomaterials: a survey of industrial approaches. Environ Sci Technol 42(2):640-6 Henry, B.J. and K.L. Schaper (1990). PPG's safety and health index system: a 10-year update of an in-plant hazardous materials identification system and its relationship to finished product labelling, industrial hygiene, and medical programs. Am. Ind. Hyg. Assoc. J. 51:475-484. Hernbert, S. et. al. (1983) Nasal and Sinonasal Cancer: Connection Between Occupational Exposure In Denmark, Finland, and Sweden. Scan. J. Work Environ. Health. 9:315. Hignett S, JR Wilson, W Morris (2005). Finding ergonomic solutions – participatory approaches. Occup Med. 55:200-207. 173

Hinds, W.C. (1988). Basis for Particle Size-Selective Sampling for Wood Dust. Appl. Ind. Hyg. 3:67. Hogue, C (2007). National Research Council blasts risk guidelines. Chem & Eng News January 12, 2007. [http://pubs.acs.org/cen/news/85/i03/8503nrc.html] Accessed 20 July 2009. Holliday, M. G. (1985). Dranitsoris, P.; Strhlendorf, P.W.; et. al.: Wood Dust in Ontario Industry: The Occupational Health Aspects. Report to the Occupational Safety and Health branch of the Ontario Ministry of Labour. Hollnagel, E., (2006). Risk + Barrier = Safety? 3rd International Conference Working on Safety, The Netherlands (www.wos2006.nl) Holmes N, Lingard H, Yesilyurt Z, De Munk F (1999). An exploratory study of meanings of risk control for long term and acute effect occupational health and safety risks in small business construction firms. J Safety Res. 30(4):251-261. Holsapple MP, Farland WH, Landry TD, Monteiro-Riviere NA, et al (2005) Research strategies for safety evaluation of nanomaterials, part II: toxicological and safety evaluation of nanomaterials, current challenges and data needs. Toxicol Sci 88(1):12-17 Hovinga, H. and Deurloo, A. (1984). Manufacturing techniqes (in Dutch). Delft University of Technology. Howard J (2008). Prevention through design – introduction. J Saf Res. 39(2):113. HSE (1983). 100 Practical applications of noise reduction methods. United Kingdom Health and Safety Executive (HSE). HMSO. London. HSE (1993). 100 Practical applications of noise reduction methods. HSE. HMSO. London (1993). HSE (1999) Health and Safety Executive, COSHH Essentials. London, U.K.: HSE. HSE (2005). Health and Safety Commission paper HSC/05/55 United Kingdom Health and Safety Executive (HSE). HMSO. London, http://www.hse.gov.uk/aboutus/meetings/hscarchive/2005/050405/c55.pdf. HSE (2007). Control of Substances Hazardous to Health (COSHH) Essentials guidance publications, United Kingdom HSE, http://www.hse.gov.uk/pubns/guidance/index.htm. HSE (2008a). Health and Safety Statistics 2007/8, United Kingdom HSE, http://www.hse.gov.uk/statistics/overall/hssh0708.pdf HSE (2008b). Control of Substances Hazardous to Health (COSHH) Essentials guidance publications. United Kingdom United Kingdom HSE. Hudspith, B. and A.W. Ha (1998). : Information needs of workers. Ann. Occup. Hyg. 42:401-406. IARC (2006). International Agency for Research on Cancer (IARC) Titanium dioxide (IARC Group 2B), Summary of data reported, Feb. 2006. ICE (1985). The assessment and control of hazards. Institute of Chemical Engineers. Rugby, 1985. ICRP (1994) Human respiratory tract model for radiological protection. International Commission of Radiological Protection Publication 66. ILO (1996). Ergonomic Checkpoints: Practical and easy-to-implement solutions for improving safety, health and working condition. ISBN 92-2-109442-1, International Labour Organisation (ILO): Geneva, 1996 ILO (2005a). Facts on Safety at Work. International Labour Office, Geneva [http://www.ilo.org/global/about-the- ilo/media_and_public_information/Factsheets/lang—en/docnamer—WCMS_067574/index.htm]. ILO (2005b). Introductory Report; Decent Work - Safe Work. XVIIth World Congress on Safety and Health at Work. Orlando, Geneva, Geneva www.ilo.org/public/english/protection/safework/wdcongrs17/intrep.pdf Ironside, P.; Mathews, J. (1975). Adenocarcinoma of the Nose and Paranasal Sinuses in Woodworkers in the State of Victoria, Australia. Cancer. 36:1115-1121. IRSST (2009) Best practices guide to synthetic nanoparticle risk management. Report R-599, Institut de recherche Robert-Sauvé en santé et en sécurité du travail (IRSST), Montréal, Québec, Canada. http://www.irsst.qc.ca/en/_publicationirsst_100432.html Israel, B. A., Schurman, S. J., House, J. S. (1989) Action research on occupational stress: involving workers as researchers. International Journal of Health Services 19(1), 135-155. Israel, B. A., Baker, E. A., Goldenhar, L. M., Heaney, C. A., Schurman, S. J. (1996) Occupational stress, safety, and health: conceptual framework and principles for effective prevention interventions. Journal of Occupational Health Psychology 1(3), 261-286. ISO (2007). International Organization for Standardization (ISO) Workplace atmospheres—ultrafine, nanoparticle and nanostructured aerosols—inhalation exposure characterization and assessment, ISO/TR 27628: 2007(E). Geneva: ISO. ISSA (2009). The Declaration of Brussels, International Social Security Association (ISSA), section non prevention in construction, Belgium 25 November 2009. http://www.issa.int/prevention-construction Jackson, H. (2002) Control Banding – Practical tools for controlling exposure to chemicals. Asian-Pacific Newsletter 9:62-63. Jafry, T., O‘Neill, D. H. (2000) The application of ergonomics in rural development: a review. Applied Ergonomics 31(3), 263-268.

174

Jayjock, M.A., J.R. Lynch, and D.I. Nelson (2000) Fundamentals of Risk Assessment for Practicing Industrial Hygienists. Fairfax, VA, American Industrial Hygiene Association. Jensen, C., et al. (1993). Trapezius muscle load as a risk indicator for occupational shoulder-neck complaints. International Archives of Occupational and Environmental Health 64:415-423. Johnson, W. (1973). The managment oversight and risk tree-MORT (SAN 821-2). Germantown, MD: U.S. Atomic Energy Comission. Jones, R.M. and M. Nicas (2004). Evaluation of the ILO Toolkit with regards to hazard classification and control effectiveness." Presented at 2ICBW, 1-2 March 2004, Cincinnati, OH. Jones, R.M. and Nicas M. (2006a). Margins of Safety Provided by COSHH Essentials and the ILO Chemical Control Toolkit. Ann. Occup. Hyg. 50:149-156. Jones, R.M. and Nicas M. (2006b). Evaluation of COSHH Essentials for Vapor Degreasing and Bag Filling Operations. Ann. Occup. Hyg. 50:137-147. Jones, R.M. and Nicas M. (2006c). Author‘s reply. Ann. Occup. Hyg. 50: 643-644. Karlqvist, L., Hagberg, M., Selin, K. (1994). Variations in upper limb posture and movement during word processing with and without mouse use. Ergonomics 37(7):1261- 1267. Kawakami, T., Batino, J. M., Khai, T. T. (1999) Ergonomic strategies for improving working conditions in some developing countries in Asia. Industrial Health 37(2), 187-198. Kawakami T, Kogi K (2001). Action-oriented support for occupational safety and health programs in some developing countries in Asia. International Journal of Occupational Safety & Ergonomics. 7(4):421-34. Kellogg P. (1909). The Pittsburgh Survey. Charities Publications Committee, New York. Koda, S., Nakagiri, S., Yasuda, N., Toyota, M., Ohara, H. (1997) A follow-up study of preventive effects on low back pain at worksites by providing a participatory occupational safety and health program. Industrial Health 35(2), 243-248. Kogi, K. (1998) Collaborative field research and training in occupational health and ergonomics. International Journal of Occupational Health and the Environment 4(3), 189-195. Kogi K (2002). Work Improvement and Occupational Safety and Health Management Systems: Common Features and Research Needs. Indust Health. 40:121-133. Kogi, K. (2006). Advances in participatory occupational health aimed at good practices in small enterprises and the informal sector. Ind. Health, 44:31-34. Kogi K, Caple D (2008). Developing ergonomic checkpoints for facilitating practical improvements in small-scale workplaces. SY20-03. XVIII World Congress on Safety and Health at Work, Seoul, Korea. Kogi K (2010). Integrating occupational hygiene and health: the effectiveness in improving small-scale workplaces. Italian J Occupational and Environmental Hygiene, 1(2):69-75 Kreuter MW, Bernhardt JM (2009). Reframing the dissemination challenge: a marketing and distribution perspective. Am J Public Health. 99(12):2123-2127. Kristenson TS (2005). Intervention studies in occupational epidemiology. Occup Environ Med. 62:205-10. Kromhout H, Vermeulen R (2000). Long term trends in occupational exposure: Are they real? What causes them? What shall we do with them? Ann Occup Hyg 44:325-327 Kromhout, H. (2002a). Author's reply. Occup. Environ. Med. 59:788-789. Kromhout, H. (2002b). Design of measurement strategies for workplace exposures. Occup. Environ. Med. 59:349- 354. Kroonenberg, H. van der, and Siers, F., (1983). Design Methods (in Dutch). Vakgroep Ontwerp- en Constructieleer, Technishce Universiteit Twente, Enschede. LaDou, J. (1996) The role of multinational corporations in providing occupational health and safety in developing countries. International Archives of Occupational and Environmental Health 68(6), 363-366. Lam, C, James, J, McCluskey, R and Hunter, R. (2004) Pulmonary toxicity of single-walled carbon nanotubes in mice 7 and 90 days after intratracheal instillation. Toxicological Sciences; 77: 126-134. Lee, E.G., Harper, M., Bowen, R.B., Slaven, J. (2009). Evaluation of the COSHH Essentials: methylene chloride, isopropanol, and acetone exposures in a small printing plant. The Annals of Occupational Hygiene. 53 (5):463-474. Levine SP, Dyjacka DT (1997). Critical features of an auditable management system for an ISO 9000-compatible occupational health and safety standard. Am Ind Hyg Assoc J 58:291-8. Lewis, D.J., (1980). The Mond fire, explosion & toxicity index. In Proceedings of the 15th annual loss prevention symposium. New York: AIChE, 1980. LI (2008). Projectrapportage kwarts in de bouw (project report quartz in Construction, in Dutch). Labor Inspectorate (LI), the Hague, the Netherlands. http://docs.minszw.nl/pdf//38/2008/38_2008_6_18355.pdf Liao C, Chiang Y, Chio C (2008) Assessing the airborne titanium dioxide nanoparticle-related exposure hazard at workplace. J Haz Mat, 162:57-65.

175

Li, G., Buckle, P. (1999) Current techniques for assessing physical exposure to work-related musculoskeletal risks, with emphasis on posture-based methods. Ergonomics 42(5), 674-695. Lingard H, Holmes N (2001). Understandings of occupational health and safety risk control in small business construction firms: barriers to implementing technological controls. Const Manag & Econ, 19:217-226. Lumens M, Spee T (2001). Determinants of exposure to respirable quartz dust in the construction industry. Ann Occup Hyg 45(7):585-595. Lux Research (2004) Sizing nanotechnology‘s value chain. New York, NY: Lux Research Inc. Machida S (2001). Guidelines on occupational safety and health management systems (ILO/OSH 2001). Asian- Pacific Newslett Occup Health Safety. 8:72-3. Maidment, S.C. (1998). Occupational hygiene considerations in the development of a structured approach to select chemical control strategies. Ann. Occup. Hyg. 42:391-400. Malchaire JB (2004). The SOBANE risk management strategy and Deparis method for the participatory screening of risks. Int Arch Occup Environ Health 77: 443-450. Marquart H, Heussen H, Le Feber M, Noy D et al (2008) ‗Stoffenmanager‘, a web-based control banding tool using an exposure process model. Ann Occ Hyg 52(6):429-442 May, D. R., Schwoerer, C. E. (1994) Employee health by design: using employee involvement teams in ergonomic job design. Personnel Psychology 47, 861-876. Maynard, A and Kuempel, E. (2005) Airborne nanostructured particles and occupational health. Journal of Nanoparticle Research; 7: 587-614. Maynard, A. (2007). Nanotechnology; the next big thing or much ado about nothing?. Ann. Occup. Hyg. 51: 1-12. Mitchel T, Else D (1983). Noise control in mining: seventy-five noise control solutions. Victorian Institute of Occupational Safety and Health, Ballarat University College, Victoria, Australia. Mitchel, T., Else, D. (1993). Noise control in mining: seventy-five noise control solutions. Victorian Institute of Occupational Safety and Health, Ballarat University College, Victoria, Australia. Moir, S., Buchholz, B. (1996). Emerging participatory approaches to ergonomic interventions in the construction industry. American Journal of Industrial Medicine 29(4), 425-430. Moir S (2005). Ideological influences on participatory research in occupational health and safety: a review of the literature. New Sol: J of Env and Occup Health Policy, 15:15-28. Money, C. D. (1992). A structured approach to occupational hygiene in the design and operation of fine chemical plant. Ann. Occup. Hyg. 36:601-607. Money, C.D. (2001). Data solutions for the 21st century: CEFIC's vision and intentions. The European Chemical Industry Council. Appl. Occup. Environ. Hyg. 16:329-330. Money, C.D. (2003). European experiences in the development of approaches for the successful control of workplace health risks. Ann. Occup. Hyg. 47:533-540. Money, C., S. Bailey, M. Smith, A. Hay, B. Hudspith, D. Tolley, et al. (2006). Evaluation of the utility and reliability of COSHH Essentials. Letter to the Editor. Ann. Occup. Hyg. 50:642-643. Moore, J.S., Garg, A. (1995). The strain index: a proposed method to analyze jobs for risk of distal upper extremity disorders. American Industrial Hygiene Association Journal, 56:443-458. Moore, J. S., Garg, A. (1996) Use of participatory ergonomics teams to address musculoskeletal hazards in the red meat packing industry. American Journal of Industrial Medicine 29(4), 402-408. Moore, J. S., Garg, A. (1997) Participatory ergonomics in a red meat packing plant. Part II: Case studies. American Industrial Hygiene Association Journal 58(7), 498-508. Murie F (2007), Building safety – an international perspective. Int J Occup Environ Health 13 :5-11. Nasterlack M, Zober A, Oberlinner C (2008) Considerations on occupational medical surveillance in employees handling nanoparticles. Int Arch Occup Environ Health. 81(6):721-6. Naumann, B.D., E.V. Sargent, B.S. Starkman, W.J. Fraser, G.T. Becker, and G.D. Kirk (1996).Performance-based exposure control limits for pharmaceutical active ingredients. Am. Ind. Hyg. Assoc. J. 57: 33-42 . Nelson, D., Chiusano, S., Bracker, A., Erickson, L., Geraci, C., Zalk DM. et al. (2007). Guidance for conducting control banding analyses. AIHA Guideline 9 – 2007, AEAG07-726. Neitzel R, NS Seixas, J Camp, M Yost (1999). An assessment of occupational noise exposures in four construction trades. Am Ind Hyg Assoc J. 60:807-817 Neitzel R, D William, L Sheppard, et al (2009). Comparison of perceived and quantitative measure of occupational noise exposure. Ann Occup Hyg 53(1):41-54. Nelson DI, Nelson RY, Concha-Barrientos M, Fingerhut M (2005). The global burden of occupational noise- induced hearing loss. Am J Ind Med 48: 446-458. Nelson DI, Zalk DM (2010). Control banding: background, critique, and evolution. In: Patty‘s Industrial Hygiene, Sixth Edition as Chapter 28, Wiley publishers. NIOSH (1984). National Institute for Occupational Safety and Health (NIOSH) Manual Of Analytical Methods, Third Edition, Cincinnati, OH . 176

NIOSH (1987). Health Effects of Exposure to Wood Dust - A Summary of the Literature. U.S. Department of Health and Human Services; Public Health Service, Centers for Disease Control; NIOSH Division of Standards Development and Technology Transfer (1987). NIOSH (1992). NIOSH Recommendations For Occupational Safety and Health, Compendium of Policy Documents and Statements, Pub. No. 92-100,U.S. Dept. of Health and Human Services, Center for Disease Control, Cincinnati OH, (January 1992). NIOSH (2006). Approaches to Safe Nanotechnology: An Information Exchange with NIOSH Version 1.1. NIOSH (2007). Simple Solutions: Ergonomics for Construction Workers NIOSH Publication No. 2007-122. http://www.cdc.gov/niosh/docs/2007-122/floor.html. NIOSH (2008). National Occupational Research Agenda: National Construction Agenda for Occupational Safety and Health Research and Practice in the U.S. Construction Sector. Developed by the NORA Construction Sector Council, revised October 27, http://www.cdc.gov/niosh/nora/comment/agendas/construction/pdfs/ConstOct2008.pdf . NIOSH (2009). Qualitative risk characterization and management of occupational hazards: Control Banding (CB); a literature review and critical analysis. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, DHHS (NIOSH) Publication No. 2009–152. Noro, K., Imada, A. (Eds.). (1991) Participatory Ergonomics. London, Taylor & Francis. Noy, D., (2004). Development of a software tool for a hazardous substance policy in SME‘s (in Dutch). TNO Chemie, Arbo Unie NRC (1999), National Research Council (NRC) Work-related musculoskeletal disorders: report, workshop summary, and workshop papers. Washington DC. NRC (2008) ‗Toxicoloog Paul Borm is ongerust over huidige nanoproducten. (Toxicologist Paul Borm is worried about the present nanoproducts), NRC Handelsblad, PCM Uitgevers Pub, NL 16 Aug 2008. NRC (2008) ‗Toxicoloog Paul Borm is ongerust over huidige nanoproducten. (Toxicologist Paul Borm is worried about the present nanoproducts), NRC Handelsblad, PCM Uitgevers Pub, NL 16 Aug 2008. Nylander, L.A.; Dement, J.M. (1993). Carcinogenic effects of wood dust: review and discussion. American Journal of Industrial Medicine, 24:619-647. Oberdorster, G, Ferin, J, and Lehnert, B. (1994) Correlation between particle-size, in-vivo particle persistence, and lung injury. Environmental Health Perspectives; 102 (S5): 173-179. Oberta AF Fischer KE (2000). Negative Exposure Assessments for Asbestos Floor Tile Work Practices. Within Advances in Environmental Measurement Methods for Asbestos. Beard ME and Rook HL ed. Published by ASTM International, 1999 ISBN 0803126166, 9780803126169 pgs 193-208. Oldershaw, P. (2001). Control banding – a practical approach to judging control methods for chemicals. J. Prev. Med. 9:52-58. Oldershaw, P.J. (2003) Control banding workshop, 4-5 November 2002, London. Ann Occup Hyg 47:531-532. Oliver T. (1902). Dangerous Trades: the historical, social, and legal aspects of industrial occupations as affecting health, by a number of exports. T. Oliver ed., John Murray Publisher, London Onos T, Hoonakker PLT, Spee T (2003) Nulmeting in het kader van het arboconvenant kwarts in de afbouwsector. (baseline survey for the working conditions agreement on quartz in the finishing sector, in Dutch) Arbouw, Amsterdam. Orthen B (2008) Approaches for the definition of threshold limit values for nanomaterials. OECD Workshop on Exposure Assessment and Exposure Mitigation, 20 Oct 2008, Frankfurt Germany. OSHA (1989). Occupational Safety and Health Administration. (OSHA) Code of Federal Regulations, Title 29, Part 1910, "Air Contaminants; Proposed Rule", Federal Register, 54(12):2920, January 19, 1989. Paik S, Zalk DM, Swuste P (2008) Application of a Pilot Control Banding Tool for Risk Level Assessment and Control of Nanoparticle Exposures. Ann Occup Hyg 52(6):419-428 Partanen, T. J., Hogstedt, C., Ahasan, R., Aragon, A., Arroyave, M. E., Jeyaratnam, J., Kurppa, K., Loewenson, R., Lundberg, I., Ngowi, A. V., Mbakaya, C. F., Stayner, L., Steenland, K., Weiderpass, E., Wesseling, C. (1999) Collaboration between developing and developed countries and between developing countries in occupational health research and surveillance. Scandinavian Journal of Work, Environment, and Health 25(3), 296-300. Pheasant, S. (1986). Bodyspace: Anthropometry, Ergonomics and Design. (Taylor & Francis, London). Pisaniello, D.L.; Connell, K.E.; Muriale, L. (1991) Wood Dust Exposure During Furniture Manufacture - Results From An Australian Survey And Considerations For Threshold Limit Value Development. Am. Ind. Hyg. Assoc. J. 52(11):485-492. Poland C, Duffin R, Kinloch I, Maynard A et al (2008) Carbon nanotubes introduced into the abdominal cavity of mice show asbestos like pathogenicity in a pilot study. Natur Nanotech 20 May p 1-6.

177

Powers KW, Brown SC, Krishna VJ, Wasdo SC et al (2006) Research strategies for safety evaluation of nanomaterials part VI: characterization of nanoscale particles for toxicological evaluation. Toxicol Sci 90(2):296-303. Punnett, L, et al. (1991). Back disorders and nonneutral trunk postures of automobile assembly workers. Scandinavian J ournal of Work, Environment, and Health 17:337-346. Rainbird, G., O‘Neill, D. H. (1995) Occupational disorders affecting agricultural workers in tropical developing countries. Applied Ergonomics 26, 185-187. Rappaport, S.M., M. Goldberg, P. Susi, and R.F. Herrick (2003). Excessive Exposure to Silica in the US Construction Industry. Ann. Occup. Hyg. 47:111-122. Redinger CF, Levine SP (1998). Development and evaluation of the Michigan Occupational Health and Safety Management System assessment instrument: a universal OHSMS performance measurement tool. Am Ind Hyg Assoc J. 59:572-81. Redinger CF, Levine SP, Blotzer MJ, Majewski MP (2002). Evaluation of an Occupational Health and Safety Management System performance measurement tool—II: scoring methods and field study sites. Am Ind Hyg Assoc J, 63:34-40. Robichaud CO, Tanzil D, Weilenmann U, Wiesner MR (2005) Relative risk analysis of several manufactured nanomaterials: an insurance industry context. Environ Sci Technol. 15;39(22):8985-94. Roelofs CR Barbeau EM, et al (2003). Prevention strategies in Industrial Hygiene: a critical literature review. American Industrial Hygiene Association Journal 64:62-67 Rosenberg BJ, Barbeau EM, Moure-Eraso R, Levenstein C (2001). The work environment impact assessment: a methodologic framework for evaluating health-based interventions. Am J Ind Med 39:218-26 Rosecrance, J. C., Cook, T. M. (2000) The use of participatory action research and ergonomics in the prevention of work-related musculoskeletal disorders in the newspaper industry. Applied Occupational and Environmental Hygiene 15(3), 255-262. Rosencrantz, H., Edvardsson, K., Hansson, S.O. (2007). Vision zero – is it irrational? Transport Research, 41(6):559-567. Rosskam, E. (1997) Preventing workplace injuries and illnesses through ergonomics. Geneva, International Labour Organization. World at Work 21, 5-8 RSC (1996)..Royal Society of Chemistry (RSC): COSHH in laboratories, 2nd edition. London: Rubio, C. A. (1995) Ergonomics for industrially developing countries: an alternative approach. Journal of Human Ergology 24(1), 119-123. Ruden, C. and S.O. Hansson (2003). How accurate are the European Union's classifications of chemical substances. Toxicol. Lett. 144:159-172. Rundmo T (1996). Associations between risk perception and safety. Safety Sci 24:197-209. Russell, R.M., S.C. Maidment, I. Brooke, and M.D. Topping (1998). An introduction to a UK scheme to help small firms control health risks from chemicals. Ann. Occup. Hyg. 42:367-376. Ryman-Rasmussen JP, Riviere JE, Monteiro-Rieviere NA (2006) Penetration of intact skin by quantum dots with diverse physicochemical properties. Toxicol Sci 91(1):159-165 Safe Work Australia (2010). Engineered nanomaterials: feasibility of establishing exposure standards and using control banding in Australia. Commonwealth of Australia, http://www.safeworkaustralia.gov.au Sargent, E.V. and G.D. Kirk (1998). Establishing airborne exposure control limits in the pharmaceutical industry. Am. Ind. Hyg. Assoc. J. 49:309-313. Scheeper, B.; Kromhout, H.; Boleij, J. (1995). Wood-Dust Exposure During Wood-Working Processes. Ann. Occup. Hyg. 39(2):141-154. Schneider T (2008) Relevance of dustiness and aerosol dynamics for personal exposure. OECD Workshop on Exposure Assessment and Exposure Mitigation, 20 Oct 2008, Frankfurt Germany. Schulte P, Geraci C, Zumwalde R, Hoover M, Kuempel E (2008) Occupational risk management of engineered nanoparticles. J Occup Environ Hyg 5(4):239-249. Schulte, P. Rinehart, R. Okun A. et al (2008). National Prevention through Design (PtD) Initiative. Journal of Safety Research. 39(2):115-121. Seixas NS, L Sheppard, R Neitzel (2003). Comparison of task-based estimates with full-shift measurements of noise exposures. Am Ind Hyg Assoc J. 64:823-829. Seixas NS, B Goldman, L Sheppard, R Neitzel, et al (2005a). Prospective noise induced changes to hearing among construction industry apprentices. Occup Environ Med 62:309-317. Seixas NS, R Neitzel, L Sheppard, B Goldman (2005b). Alternative metrics for noise exposure among construction workers. Ann Occup Hyg 49(6):493-502. Shahnavaz, H. (2000) Role of ergonomics in the transfer of technology to industrially developing countries. Ergonomics 43(7), 903-907. Shreve, R., (1956). The chemical process industry. McGraw-Hill Books Inc., New York. 178

Silk J. (2004). The Globally Harmonized System (GHS) of Classification and Labeling of Chemicals: Harmonized Phrases for International Control Banding. Presented at 2ICBW, 1-2 March 2004, Cincinnati, OH. Simmonds, RH. (1992). Estimating safety costs. Accident Prevention Manual For Business And Industry , 10th edition (National Safety Council, Chicago, Illinois) 291. Singleton, W. T., Whitfield, D. (1968) The organisation and conduct of a World Health Organisation inter-regional course on ergonomics for developing countries. Human Factors, 10 (6), 633-640. Smedbold, H.T. (2004). KjemiRisk - The Norwegian offshore approach to control banding. Presented 2ICBW, 1-2 March 2004, Cincinnati, OH.. Sommerich, C.M., McGlothlin, J.D., Marras, W.S. (1993). Occupational risk factors associated with soft tissue disorders of the shoulder: a review of recent investigations in the literature. Ergonomics 36(6):697-717. Spee, T. (2005). Risk assessment from toxic substances and control measures in the Dutch construction industry. Presented at 3ICBW, 21 September 2005, Pilanesberg, South Africa.. Staatscourant (2001) Staatscourant (Government Gazette in dutch) 19 October 2001, p 14. http://www.staatscourant.nl/ Stern ST, McNeil SE (2008) Nanotechnology safety concerns revisited. Toxicol Sci 101(1):4-21. Stetson, D.S., et al. (1991). Observational analysis of the hand and wrist: a pilot study. Applied Occupational and Environmental Hygiene 6(11):927-937. Stewart, K., M. Kiefer, P. Johnson, and D.M. Zalk (2005). Establishing Ergonomics in Industrially Developing Countries. Proceedings SS (K1), International Occupational Hygiene Scientific Conference, 21 September 2005, Pilanesberg, South Africa.. Stubbs, D. A. (2000) Ergonomics and occupational medicine: future challenges. Occupational Medicine 50(4), 277- 282. Suter AH (2002). Construction noise: exposure, effects, and the potential for remediation; a review and analysis. Am Ind Hyg Assoc J. 63:768-789 Swuste P., Kormhout, H., and Drown, Dl, (1993). Prevention and control of chemical exposures in the rubber manufacturing industry in The Netherlands. Annals of Occupational Hygiene 37:117-134. Swuste, P and A Hale (1994). "Databases on measures to prevent occupational exposure to toxic substances." Applied Occupational and Environmental Hygiene 9(1): 57-61. Swuste, P., M. Corn, and B. Goelzer (1995). Hazard prevention and control in the work environment; Report of a WHO meeting. African Newsletter 1:20-21. Swuste P (1996). Occupational hazards, risks and solutions. Delft University Press. Swuste, P, A Hale, S Pantry (2003). "Solbase: a databank of solutions for occupational hazards and risks." Annals of Occupational Hygiene 47(7): 541-7. Swuste P. (2004). Occupational hazards, risks, and solutions. Presentation at the 2nd International CB Workshop: Validation & Effectiveness of Control Banding. Cincinnati March 1-2. Swuste, P. (2005). Control banding, expansion of range; safety. Presented at 3ICBW, 21 September 2005, Pilanesberg, South Africa. Swuste P (2007). Qualitative methods for occupational risk prevention strategies in safety, control banding – safety. Safety Science Monitor 11(3). Online journal http://www.monash.edu.au/muarc/ipso/vol11/Issue3/8 Swuste.pdf Accessed 23 Jan 2009 Swuste P. (2008). You will only see it, if you understand it‘, or occupational risk prevention from a management perspective. Human Factors and Ergonomics in Manufacturing, 18:438-53. Swuste P, Gulijk C, van Zawaard W, (2010). Safety metaphors and theories, a review of the occupational safety literature of the US, UK, and The Netherlands till the first part of the 20th century. Safety Science, 48(8): 1000-1018. Takala, J. (1999) Introductory report of the International Labour Office, International Occupational Safety and Health Information Centre, Geneva, International Labour Office, 10p. Takala, J. (2010). Present, new and emerging risks. Italian J Occupational and Environmental Hygiene, 1(2):60-62. Tait, K. (2004). Control banding: an improved means of assessing and managing health and safety risks at Pfizer. Presented at the Second International Control Banding Workshop (2ICBW), 1-2 March 2004, Cincinnati, OH. Tandhanskul, N., Duangsa-Ad, S., Pongpanich, C., Pungok, A., Punpeng, T., Juengprasert, W., Kawakami, T. (1995) Experiences of successful action programmes for occupational health, safety, and ergonomics promotion in small scale enterprises in Thailand. Journal of Human Ergology 24(1), 105-115. Thackrah, C.T. (1832). The effects of arts, trades, and professions. Baines and Newsome Publishing, Leeds. Thienen G van, Spee T (2008). Health effect of construction materials and product. J Applied Occup Sci. 1:2-23. Thinksafe (2007). Noise management in the construction industry: a practical approach. Worksafe. http://www.docep.wa.gov.au/Worksafe/Content/Safety_Topics/Noise/Further_information/Noise_managemen t_in_the_constr.html

179

Thomas K, Aguar P, Kawasaki H, Morris J et al (2006) Research strategies for safety evaluation of nanomaterials, part VIII: international efforts to develop risk-based safety evaluations for nanomaterials. Toxicol Sci 92(1):23- 32 Thorpe, A.; Brown, R.C. (1995). Factors Influencing The Production Of Dust During The Hand Sanding Of Wood. Am. Ind. Hyg. Assoc. J. 56:236-242. Tielemans, E. (2007). Evaluation of the Stoffenmanager. Presented at the Occupational Hygiene Conference, 17-19 April, Glasgow. Tielemans E, Noy D, Schinkel J, Heussen H et al. (2008) Stoffenmanager exposure model: development of a quantitative algorithm. Ann Occup Hyg 52(6):443-454 Tijssen, S., M. le Feber, H. Heussen, J. West, and D. Noy (2004). A new tool for small and medium enterprises to work safely with hazardous substances. Presented at 2ICBW, 1-2 March 2004, Cincinnati, OH. Tischer, M. (2001). Helping SMEs to manage risks from chemicals in the workplace. Proceedings from an international workshop, 26-27 November 2001, Amsterdam, The Netherlands. Zeist: TNO Nutrition and Food Research. Tischer, M. (2002). Current BAuA/GTZ research on occupational exposure and control strategies: recent results from various industrial areas and from Indonesian SMEs. Presented at the First International Control Banding Workshop (1ICBW), 4-5 November 2002, London, UK. Tischer, M., S. Bredendiek-Kamper, and U. Poppek (2003). Evaluation of the HSE COSHH Essentials exposure predictive model on the basis of BAuA field studies and existing substances exposure data. Ann. Occup. Hyg. 47:557-569. Tischer, M. and S. Scholaen (2003). Chemical management and control strategies: experiences from the GTZ pilot project on chemical safety in Indonesian small and medium-sized enterprises. Ann. Occup. Hyg. 47:571-575 (2003). Tischer, M., Bredendiek-Kamper, S., Poppek, U., Packroff, R. (2009).. How safe is control banding? Integrated evaluation by comparing OELs with measurement data and using Monte Carlo simulation. The Annals of Occupational Hygiene. 53 (5):449-462. Tjoe Nij E, Heederik D (2005). Risk assessment of silicosis and lung cancer among construction workers exposed to repsirable quartz. Scand J Work Environ Health. 31(2):49-56. Tolley, J. C., Zalk, D. M., Custodian Quality Improvement Team (1998) A custodian journal on grassroots ergonomics. Proceedings of the American Society of Safety Engineers (ASSE) International Professional Development Conference 1998, Seattle, 14 – 17 June, 247-254. Topping, M. (2002a). Design of measurement strategies for workplace exposures: Letter to the editor." Occup. Environ. Med. 59:788. Topping, M. (2002b). COSHH Essentials from concept to one stop system." Presented at the First International Control Banding Workshop, 4-5 November 2002, London, UK. Tran, C, Buchanan, D, Culen, R, Searl, A, Jones, A and Donaldson, K. (2000) Inhalation of poorly soluble paritlces. II. Influence of particle surface area on inflammation and clearance. Inhalation Toxicology; 12(12): 1113-1126. Tsuji JS, Maynard AD, Howard PC, James JT, Lam C (2006) Research Strategies for Safety Evaluation of Nanomaterials, Part IV: Risk Assessment of Nanoparticles. Toxicol Sci 89(1):42-50 UIC (1999). Chemical Industry Union (UIC): [Tool for evaluation of the risks to the chemicals], [Document in French] technical document DT 63. Paris. US (1992). U. S. Court of Appeals for the Eleventh Circuit: AFL--CIO v. OSHA, CA 11, No. 89-7185 (July 7, 1992) USCB (2004). Country Business Patterns (NAICS), U.S. Census Bureau (USCB). http://censtat.census.gov/cgi- bin/cbpnaic/cbpdetl.pl Vedder J, Carey E (2005). A multi-level systems approach for the development of tools, equipment and work processes for the construction industry. Appl Ergon, 36(4):471-80. Vincent, R. and F. Bonthoux (2000). [Evaluation of chemical risk : hierarchisation of ―risk potential‖.] [Article in French] I Cahiers de notes documentaires - Hygiene et securite du travail 179:29-34. Visser J.P. (1998). Developments in HSE management in oil and gas exploration and production. in Hale A.R. & Baram M. (Eds.) Safety management: the challenge of organisational change. Pergamon. Oxford Waehrer GM. XS Dong, T Miller, et al (2007). Costs of Occupational Injuries in Construction in the United States. Accid Anal Prev. 39(6): 1258–1266. Wagner GR (1995). The inexcusable persistence of silicosis. Am Public Health Assoc. 85(10):1346-1347. Warheit DB, Borm PJ, Hennes C, Lademann J. (2007a) Testing strategies to establish the safety of nanomaterials: conclusions of an ECETOC workshop. Inhal Toxicol 19(8):631-43. Warheit DB, Webb TR, Reed KL, Frerichs S, Sayes CM (2007b) Pulmonary toxicity study in rats with three forms of ultrafine-TiO2 particles: Differential responses related to surface properties. Toxicol 230(1):90-104.

180

Warheit DB, Sayes CM, Reed KL, Swain KA (2008) Health effects related to nanoparticle exposures: Environmental, health and safety considerations for assessing hazards and risks. Pharm & Therap, 120(1): 35- 42. Waterson A. (2007). Global construction health and safety – what works, what does not, and why? Int J Occup Environ Health. 13(1):1-4. Weinstein MG, Hecker SF, et al (2007). A roadmap to diffuse ergonomic innovations in the construction industry. Int J Occup Environ Health 13:46-55. Westgaard, R. H., Winkel, J. (1997) Ergonomic intervention research for improved musculoskeletal health: a critical review. International Journal of Industrial Ergonomics 20, 463-500. Whitehead, L.W.; Fruend, T.; Hahn, L. (1981). Suspended Dust Concentrations and Size Distributions and Qualitative Analysis of Inorganic Particles from Woodworking Operations. Am. Ind. Hyg. Assoc. J. 42:461. Whitehead, L. W. (1982). Health Effects Of Wood Dust - Relevance For An Occupational Standard. A. Ind. Hyg. Assoc. J. 43(9):674-678. WHO (2003), Review of the WHO Activities in Occupational Safety and Health, 13th Session of the Joint ILO/WHO Committee on Occupational Health, Geneva. Wisner, A. (1989) Variety of physical characteristics in industrially developing countries – ergonomic consequences. International Journal of Industrial Ergonomics 4, 117-138. Woodward, K.N., A. McDonald, and S. Joshi (1991). Ranking of chemicals for carcinogenic potency--a comparative study of 13 carcinogenic chemicals and an examination of some of the issues involved. Carcinogenesis 12:1061- 1066. Yang W, Peters JI, Williams RO (2008). Inhaled nanoparticles—A current review. Int J of Pharma, 56(1-2):239-247 Yap, S.M. (2004). Assessing the utility of the ILO Toolkit in Singapore. Presented at 2ICBW, 1-2 March 2004, Cincinnati, OH. Zalk, D. M., Tolley, J. C., Kim, Y. (1997) Grassroots ergonomics to modify custodial training procedures. Professional Safety 42, 21-25. Zalk, D. M., Biggs, T. W., Perry, C. M., Tageson, R., et al. (2000) Participatory ergonomics approach to waste container handling utilizing a multidisciplinary team. Proceedings of the International Ergonomics Association 14th Triennial Congress/HFES 2000, San Diego, 31 July – 4 August, 324-327. Zalk DM (2000) Exposure assessment techniques to reduce occupational uptake of asbestos and lead for maintenance activities. IOHA Cairns 2000 Abstract 60, Cairns, Australia [http://nla.gov.au/nla.cat-vn2019109]. Zalk DM (2001). Grassroots ergonomics: initiating an ergonomics program utilising participatory techniques. Annals of Occupupational Hygiene 45(4):283-289. Zalk DM (2002). Participatory occupational hygiene; a path to practical solutions. Asian-Pacific Newsletter 9:51 Zalk, DM (2003). Control banding principles to reduce musculoskeletal disorders; the ergonomics toolkit." Proceedings V5(327), International Ergonomics Association Congress, Seoul, South Korea. Zalk, DM (2006a). Control banding‘s human health-based risk management strategies. Presentation at American Industrial Hygiene Conference Roundtable 221, Chicago, IL ( Zalk, DM (2006b). Practical prevention in safety; from control banding to barrier banding. Proceedings 17(5), International Working on Safety Conference, 15 September 2006 Eemhof, The Netherlands. Zalk DM, T Spee (2008) Barrier banding and the construction toolbox. Proceedings SY20-04, XVII World Congress on Safety and Health at Work, 1 July 2008 Seoul, South Korea. Zalk DM, Nelson DI (2008). History and evaluation of control banding: a review. J Occup Environ Health 5(5):330- 346 Zalk DM, Paik S, Swuste P (2009). Evaluating the Control Banding Nanotool, a qualitative risk assessment approach for controlling nanomaterial exposure. J Nanopart Research. 11(7):1685-1704. Zalk DM and Heussen H (2010). Outcomes of the 6th International Control Banding Workshop. IOHA 2010 Conference Proceedings, Rome, Italy, 28 September. Zalk DM, Kamerzell R, Paik S, Kapp J, Harrington D, Swuste P (2010a). Risk level based management system: a control banding model for occupational health and safety risk management in a highly regulated environment. Industrial Health, 48(1):18 - 28. Zalk DM, Swuste P, Hale A (2010b). Barrier banding: a concept for safety solutions utilizing control banding principles. Submitted to Safety Science, May 2010. Zalk DM, Spee T, Gillan M, Lentz TJ, Garrod A, Evans P, Swuste P (2010c). Review of qualitative approaches for the construction industry; designing a risk management toolbox. Submitted to Ann Occup Hyg, May 2010. Zalk DM, Paik S (2010d). Control Banding and nanotechnology. The Synergist, 3:26-29. Zwaard W. Passchier W. (1995). Risicobepaling en risicobeheersing. Tijdschrift voor toegepaste Arbowetenschap 8;8-12 [in Dutch] Zwaard W. Goossens L. (1997). Relatieve Ranking als hulpmiddel voor risico-evaluatie. Tijdschrift voor toegepaste Arbowetenschap 10;10-15 [in Dutch] 181

Further ICBW information and research can be obtained from the following websites: ICBW1. First International Control Banding Workshop London, Nov. 4-5, 2002. http://www.bohs.org/mod.php?mod=fileman&op=view_cat&id=14 ICBW2. Second International Control Banding Workshop Cincinnati Ohio, 1-2 March 2004. http://www.acgih.org/events/course/controlbandwkshp.htm. ICBW3. Third International Control Banding Workshop Pilanesburg South Africa 21 September 2005. http://www.saioh.org/ioha2005/Proceedings/SSI.htm ICBW4. Fourth International Control Banding Workshop Seoul South Korea, 1 July 2008. http://www.ioha.net ICBW5. Fifth International Control Banding Workshop Cape Town South Africa, 25 March 2009. http://www.ioha.net ICBW6. Sixth International Control Banding Workshop Rome Italy, 27 September 2010. http://www.ioha.net

182

SUMMARY

The main objective of this thesis is determining whether the simplified, qualitative risk assessment models and solutions initiatives of Control Banding (CB) can be integrated within the multidisciplinary requirements of occupational risk management. This was achieving practical primary prevention that assists non-experts in obtaining the appropriate controls. Investigative methods include a literature review of CB and solutions initiatives approaches. Research methods include the development, implementation, and evaluation of qualitative risk management models for the health hygiene and safety professions. The results of this thesis include the development of qualitative risk assessment models for nanomaterials and safety, participatory approaches for creating ergonomics programs, and multidisciplinary CB strategies for risk management in the construction sector as well as an occupational health and safety management system. This research unveils an excellent potential for application of CB beyond bulk chemicals, though validated toolkits for ergonomics and safety have not yet been created. CB has been shown to be effective in line with quantitative models when there is an absence of firm toxicological and exposure information, although expertise is required for both validation of control effectiveness and for higher risk tasks. The multidisciplinary CB models presented show tremendous opportunity to protect workers, though more research and validation is necessary for implementation. Outcomes of this research, such as the CB Nantool, have been applied internationally through the WHO and ILO, implemented by national institutes and university programs, and directly transferred into national regulations. Collectively, this thesis is in a position to address the needs of the 2.5 billion workers around the world without access to health and safety experts and increase abilities to achieve prevention of work-related illnesses, diseases, and safety-related accidents and injuries.

SAMENVATTING

De belangrijkste doelstelling van het proefschrift is vast te stellen of een vereenvoudigd model van risicobepaling en de daaraan verbonden oplossingen middels Control Banding geïntegreerd kan worden in een multidisciplinaire management benadering van beroepsmatige risico‘s. Hiermee wordt praktische en primaire preventie mogelijk voor niet-deskundigen door geschikte beheersmaatregelen te selecteren. Om dit te bereiken zijn een aantal stappen gezet. Relevante literatuur over Control Banding is samengevat, evenals literatuur over benaderingen om oplossingen te genereren en de ontwikkeling, de implementatie en de evaluatie van kwalitatieve modellen voor risico bepaling en risico management binnen het domein van arbeidsveiligheid, -hygiëne en gezondheid (sentence too long). Het resultaat van het proefschrift is een uitvoerig overzicht van de ontwikkeling van kwalitatieve risicobepalingen voor nano-deeltjes en voor veiligheid, een participatieve benadering voor ergonomische programma‘s, multidisciplinaire Control Banding methoden voor risico management in de bouw en een arbeidsveiligheid en -gezondheid management systeem. Dit onderzoek laat zien dat Control Banding zeer goed toepasbaar is buiten het terrein van de bulk chemicaliën, hoewel een gevalideerde versie voor ergonomie en veiligheid nog niet is ontwikkeld. Control Banding heeft, vergelijkbaar met kwantitatieve modellen, bewezen effectief te zijn in situaties met beperkte informatie over toxiciteit en blootstellingn. Deskundigheid blijft echter vereist voor zowel effectiviteit van maatregelen als voor hoog risico taken. De gepresenteerde multidisciplinaire Control Banding modellen bieden enorme mogelijkheden om werknemers te beschermen. Meer onderzoek en validatie is echter nodig voordat deze modellen gebruikt kunnen worden. De resultaten van het (which?) onderzoek zijn internationaal gebruikt door de International Labor Office en de World Health Organisation, en ook door nationale instituten en universiteiten evenals opgenomen in nationale wetgeving. Dit proefschrift kan over de wereld 2,5 miljard werknemers ondersteunen, die geen toegang hebben tot arbeidsveiligheid, -hygiëne en gezondheidsdeskundigen (more explanation needed here). Het vergroot de mogelijkheden van preventie voor arbeidsgerelateerde aandoeningen en ziektes en veiligheidsgerelateerde ongevallen en letsel. 183

CURRICULUM VITAE

David M. Zalk was born in Boston, Massachusetts in 1964 and moved to northern California in 1973. Living in San Jose, for most of his life, David has watched this part of the world transition from fruit orchards into what is now known as Silicon Valley. He attended the University of California, Santa Barbara for five years, completing two Bachelor‘s degrees in Chemistry and Environmental Studies in 1987. He was introduced to Industrial Hygiene as a Manager of an asbestos abatement company, but left the business to marry his wife Janice and they were amazingly fortunate to spend a year-long honeymoon traveling around the globe. These travels opened his eyes to the cultures and peoples of the globally and left him wanting to spend part of his career working to make a difference in the world. He then entered the University of California, Berkeley Industrial Hygiene program and graduated with a Masters of Public Health in Environmental Health Sciences in 1994. During his time in Berkeley, David began his career at the Lawrence Livermore National Laboratory (LLNL) as an Industrial Hygienist performing both as a field practitioner and a researcher. In 1996 he was awarded the ACGIH John J. Bloomfield Award for excellence in Industrial Hygiene for a professional under 40 years old. This award opportunity opened the doors for his future in international work when in 1998 he became Chair of the newly established ACGIH International Committee. In 1999 he became the ACGIH representative on the International Occupational Hygiene Association (IOHA) Board of Directors, a position he would hold through 2005. In 2002 David was honored to become IOHA President at the ripe young age of 35. In his IOHA leadership roles he was not hesitant in taking on both the challenges and responsibilities that come with having the potential to elevate the reputation and position of Industrial Hygiene on a global level.

David became aware of Control Banding in 1998 as an alternative to address relative exposure risks and controls when OELs were not available. His work with Control Banding created both leadership and research opportunities that included being a host of the 1st International Control Banding Workshop (ICBW) as IOHA President, co-chair of the next 5 ICBWs held around the world as well as the U.S. National Control Banding workshop in 2005. David has continued working with IOHA as their Envoy to the WHO. He has also been a co-leader for the WHO Collaborating Centers in Occupational Health since 2001 for initiatives leading to the development and dissemination of Control Banding strategies. He is currently the Vice President of the Foundation for Occupational Health and Safety and the creator of their Worldwide Outreach Program, established to give grants to university programs in developing countries that work to develop the Industrial Hygiene programs where they are needed most. To date, over 14 grants have been awarded to all corners of the world. Back at work during this time period, David had become the EHS Manager at LLNL‘s Site 300 Experimental Test Facility and he continues in this position to date. David has published over 35 articles covering a wide variety of Industrial Hygiene and Ergonomics research projects and is co-author on 3 Control Banding related book chapters, including the latest edition of Patty‘s Industrial Hygiene. David‘s latest of many awards was in 2010, when he became a Fellow of the American Industrial Hygiene Association. David began his PhD efforts at the Delft University of Technology in 2006, after his IOHA Board of Director obligations subsided, moving with the kind guidance of Paul Swuste toward this amazing accomplishment. His most important accomplishment, however, is being a father to his three outstanding and wonderful boys: Joshua, Jacob, and Jesse. Together with his wife of over 20 years, Janice, they find laughter and love in all they do with their boys, truly realizing the blessings in all they have been given and in the wondrous, amazing, and adventurous path that continues to unfold before them.

184