Musculoskeletal Disorders in the Workplace Principles and Practice 2nd Edition - nordin

1600 John F. Kennedy Blvd. Ste 1800 Philadelphia, PA 19103-2899 MUSCULOSKELETAL DISORDERS IN THE WORKPLACE: ISBN-13: 978-0-323-02622-2 PRINCIPLES AND PRACTICE ISBN-10: 0-323-02622-2 Copyright © 2007, 1997 by Mosby Inc., an affiliate of Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permissions may be sought directly from Elsevier’s Health Sciences Rights Department in Philadelphia, PA, USA: phone: (+1) 215 239 3804, fax: (+1) 215 239 3805, e-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://www.elsevier.com), by selecting ‘Customer Support’ and then ‘Obtaining Permissions’. Notice Knowledge and best practice in this field are constantly changing. As new research and experience broaden our knowledge, changes in practice, treatment, and drug therapy may become necessary or appropriate. Readers are advised to check the most current information provided (i) on procedures featured or (ii) by the manufacturer of each product to be admin- istered, to verify the recommended dose of formula, the method and duration of administra- tion, and contraindications. It is the responsibility of the practitioner, relying on their own experience and knowledge of the patient, and to make diagnoses, to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions. To the fullest extent of the law, neither the Publisher nor the Editors assume any liability for any injury and/or damage to persons or property arising out of or related to any use of the material contained in this book. The Publisher Library of Congress Cataloging-in-Publication Data Musculoskeletal disorders in the workplace: Principles and practice / [edited by] Margareta Nordin, Gunnar B.J. Andersson, Malcolm H. Pope. —2nd ed. p. ; cm. Includes bibliographical references and index. ISBN 0-323-02622-2 1. Musculoskeletal system—Diseases. 2. Occupational diseases. I. Nordin, Margareta. II. Pope, M. H. (Malcolm Henry), 1941- III. Andersson, Gunnar, 1942- [DNLM: 1. Musculoskeletal Diseases—therapy. 2. Biomechanics. 3. Human Engineering. 4. Musculoskeletal Diseases—prevention & control. 5. Occupational Diseases—etiology, WE 140 M9854 2006] RC925.5M8783 2006 616.7—dc22 2006043830 Acquisitions Editor: Rolla Couchman Project Manager: Bryan Hayward Printed in United States of America Last digit is the print number: 9 8 7 6 5 4 3 2 1

Foreword Workers’ health priorities are now driven by demographics. to his or her previous level of performance, accomplishing All of the First World Nations are facing a future with a rapidly a particular task within a reasonable time frame. Although diminishing work force, an aging population, and a growing the average orthopaedic surgeon may be well-versed with the number of pensioners. The prevention and efficient man- conditions of the gridiron, he or she may not be familiar with agement of work place injuries and disabilities has become the requirements and limitations of the industrial playing a priority. field. In industrial medicine, it is not only necessary to \"fix\" the worker; one must have an idea about how to fix the work- Medical professionals work in an increasingly specialized place to prevent further injury. Like sports medicine, the world brought on by an explosion of knowledge, the demand management and prevention of industrial injury demands from society for \"the very best\" in services, and the need for a dedicated and knowledgeable cadre of physicians, surgeons, expertise to keep pace with technological change and innova- and therapists who are able to apply modern knowledge and tion. Modern medical history, in particular, is replete with expertise to a successful medical program. examples of sudden bursts of information that challenged the growth of new domains and abilities. The period of World This volume brings together chapters authored by the War II saw an explosion of medical knowledge, rapidly divid- most knowledgeable group of surgeons, physicians, scientists, ing General Surgery into numerous subspecialties. Similarly, ergonomists, and therapists currently addressing the preven- the 1970s was a critical period in orthopaedic surgical practice tion and management of workplace injury. The editors have as many subspecialties developed that allowed greater expert assembled a most versatile and practical tool for the many use of modern technology. allied-health professionals involved with work-related injuries. This updated text will have a strong impact within industry Occupational orthopaedics is a relatively young specialty and on the management of patients well into the 21st century. that is still evolving rapidly. As in sports medicine, we have learned that it is not sufficient to examine and treat injuries Victor H. Frankel KNO, MD, PhD alone. After recovery, an injured football player is expected to Professor of Orthopaedic Surgery, NYU return to the game and perform at his previous level of athletic President Emeritus, Hospital for Joint Diseases ability. The injured industrial worker is also expected to return v

Contributors K. N. An, Ph.D. Amit Bhattacharya, Ph.D., C.P.E. Craig J. Della Valle, M.D. John and Posy Krehbiel Professor of Professor Assistant Professor of Orthopaedic Surgery Biomechanics-Ergonomics Research Laboratories Rush-Presbyterian-St. Luke's Medical Center Orthopaedics Department of Environmental Health Chicago, IL 60612 Maylo Clinic College of Medicine University of Cincinnati Medical College Orthopaedics Biomechanics Lab Cincinnati, OH 45267-0056 James A. Dewees, M.S., C.P.E., C.E.E.S. Rochester, MN 55905 ERGO Accommodations Inc Anthony M. Buoncristiani, M.D., L.T. P O Box 499 Gunnar B. J. Andersson, M.D., Ph.D. Orthopaedics Department Union, KY 41091-0499 Professor and Chairman Naval Medical Center Department of Orthopaedic Surgery San Diego, CA 92134 Jiri Dvorak, M.D., Ph.D. Rush-Presbyterian- Department of Neurology Linda Carroll, M.D. Schulthess Hospital St. Luke’s Medical Center Associate Professor 8008 Zurich Chicago, IL 60612 Department of Public Health Sciences Switzerland University of Alberta Thomas J. Armstrong, Ph.D. Edmonton, Alberta T6G 2E1 Ulf Eklund M.D. Professor Canada Orthopedic Surgeon Industrial and Operations Engineering Department of Orthopedics Director J. David Cassidy, M.D. Molndal Hospital Center for Ergonomics Department of Public Health Sciences Molndal, Sweden University of Michigan University of Alberta Ann Arbor, MI 48109-2117 Edmonton, Alberta T6G 2E1 Freddie H. Fu, M.D. Canada Department of Orthopaedic Surgery Federico Balagué University of Pittsburgh Médecin Chef Adjoint Christine Cedraschi Pittsburgh, PA 15203 Division of Rheumatology, Physical Medicine Division of General Medical Rehabilitation & Douglass Gross, Ph.D. and Rehabilitation Multidisciplinary Pain Center Assistant Professor Hôpital Cantonal Division of Clinical Pharmacology and Department of Physical Therapy 1708 Fribourg University of Alberta Switzerland Toxicology Edmonton, Alberta T6G 2G4 and Geneva University Hospitals Canada Adjunct Associate Professor 1211 Geneva 14 Department of Orthopaedic Surgery Switzerland Robert Gunzburg, M.D., Ph.D. New York University School of Medicine Senior Consultant New York, NY 10014 Mark S. Cohen, M.D. Department of Orthopaedics Department of Orthopaedic Surgery Centenary Clinic Michele Crites Battié, Ph.D. Rush-Presbyterian- 2018 Antwerp Professor Belgium Department of Physical Therapy St. Luke’s Medical Center University of Alberta Chicago, IL 60612 Daniel J. Habes, M.S.E., C.P.E. Edmonton, Alberta T6G 2G4 Industrial Engineer Canada Pierre Côté, D.C., Ph.D. Industrial Hygiene Section Scientist Hazard Evaluations and Technical Assistance Jane Bear-Lehman, Ph.D., OTR, FAOTA Institute for Work and Health Associate Professor Toronto, Ontario M5G 2E9 Branch New York University Canada Division of Surveillance, Hazard Evaluations, Steinhardt School of Education Occupational Therapy Department Benjamin Crane, M.D. and Field Studies New York, NY 10012 Resident National Institute for Occupational Safety and Department of Orthopaedic Surgery David P. Beason, M.S. Rush University Medical Center Health Research Engineer Chicago, IL 60612 Cincinnati, OH Laboratory Manager McKay Orthopaedic Research Laboratory James N. DeBritz, M.D. Robert H. Haralson, III, M.D., M.B.A. vii University of Pennsylvania Assistant Instructor Executive Director of Medical Affairs Philadelphia, PA 19104 Department of Orthopaedics American Association of Orthopaedic Georgetown University Hospital Washington, DC 20007 Surgeons Rosemont, IL 60018

viii Contributors Rudi Hiebert, B.S. Margareta Nordin, Dr.Sci. David Rempel, M.D., M.P.H. Interim Director Professor Professor Musculoskeletal Epidemiology Unit Departments of Orthopaedics and School of Medicine –Ergonomics Program Occupational & Industrial Orthopaedic Center Division of Occupational and Environmental NYU Hospital for Joint Diseases Environmental Medicine New York, NY 10014 School of Medicine Medicine New York University Program Director University of California, San Francisco Beat Hintermann, M.D. Program of Ergonomics and Richmond, CA 94804 Chief Orthopaedic Clinic University of Basel Biomechanics Michiel Reneman, Ph.D, P.T. Kantonsspital Graduate School of Arts and Science Center for Rehabilitation CH-4410 Liestal New York University University Medical Center Groningen Switzerland Director University of Groningen Occupational and Industrial P.O. Box 30002, 9750 RA Haren David M. Kalainov, M.D. The Netherlands Clinical Assistant Professor Orthopaedic Center (OIOC) Department of Orthopaedic Surgery NYU Hospital for Joint Diseases Per A.F.H. Renström, M.D., Ph.D. Northwestern University New York University Medical Center Professor Chicago, IL 60611 New York, NY 10014 Department of Molecular Medicine and Dennis D.J. Kim, M.D. Mooyeon Oh-Park, M.D. Surgery Associate Professor Clinical Associate Professor Section of Orthopaedics and Sports Department of Physical Medicine and Rehabilitation Department of Rehabilitation Medicine Montefiore Medical Center Montefiore Medical Center Medicine Bronx, NY 10467 Bronx, NY 10467 Karolinska Hospital SE-171 76 Stockholm Stephan Konz, Ph.D., P.E. Rita M. Patterson, Ph.D. Sweden Professor Associate Professor and Deputy Director Department of Industrial Engineering Orthopaedics Biomechanics Laboratory Mana Rezai, H.B.Sc., D.C., M.H.Sc. Candidate Kansas State University Division of Research Research Associate Manhattan, KS 66506 Department of Orthopaedic Surgery and Institute for Work & Health University of Toronto Vicki Kristman, B.Sc., M.Sc. Rehabilitation Toronto, Ontario M5G 2E9 Ph.D. Candidate, Epidemiology University of Texas Medical Branch Canada Department of Public Health Sciences Galveston, TX 77555 University of Toronto Tonu Saartok, M.D., Ph.D. Research Associate David I. Pedowitz, M.S., M.D. Department of Surgical Sciences Institute for Work & Health Chief Resident Section of Sports Medicine Toronto, Ontario M5F 2E9 Department of Orthopaedic Surgery Karolinska Institute Canada University of Pennsylvania SE-171 76 Philadelphia, PA 19004 Stockholm, Sweden Shrawan Kumar, Ph.D., D.Sc., F.Erg.S., F.R.S.C. Professor Anthony Petrizzo, M.D G. James Sammarco, M.D. Department of Physical Therapy c/o Ronald Moskovich, M.D. The Center for Orthopaedic Care, Inc. Faculty of Rehabilitation Medicine 301 East 17th Street Cincinnati, OH 45219-2906 University of Alberta New York, New York 10003 Edmonton, Alberta T6G 2G4 Peter Sheehan, M.D. Canada Derek Plausinis, M.D. Director Shoulder & Elbow Surgery Fellow Diabetes Center of Greater New York Marianne Magnusson, R.P.T., Dr.Med.Sci. Department of Orthopaedic Surgery Cabrini Medical Center Senior Lecturer NYU Hospital for Joint Diseases New York, NY 10003 Liberty Safe Work Research Centre New York, NY 10003 Department of Economy and Technology Ali Sheikhzadeh, Ph.D., C.I.E. Halmstad University Malcolm H. Pope, Dr.Med.Sci., Ph.D. Research Assistant Professor SE-301 18 Halmstad Professor Departments of Orthopaedic Surgery and Sweden Liberty Safework Research Centre Department of Environmental & Occupational Environmental Medicine Paul H. Marks, M.D. New York University School of Medicine Associate Professor Health Associate Director of Research Department of Surgery Foresterhill Occupational and Industrial Orthopaedic University of Toronto Aberdeen, Scotland AB 25 2ZD Toronto, ON M4Y 1H1 United Kingdom Center Canada NYU Hospital for Joint Diseases Laura Punnett, Sc.D. New York, NY 10014 Ronald Moskovich, M.D. Professor Assistant Professor Department of Work Environment Mary-Louise Skovron, Dr. PH. Department of Orthopaedic Surgery University of Massachusetts Lowell Group Director, Pharmaco-epidemiology NYU Hospital for Joint Diseases Lowell, MA 01854 Global Epidemiology New York, NY 10003 Bristol–Myers Squibb Robert G. Radwin, Ph.D. 311 Pennington-Rocky Hill Road Professor and Chair Pennington, NJ 09534 Department of Biomedical Engineering University of Wisconsin Madison, WI 53706

Contributors ix Louis J. Soslowsky, Ph.D. Ross Taylor, M.D. Sam W. Wiesel, M.D. Professor of Orthopaedic Surgery and Coastal Orthopaedic Associates Professor and Chair Conway, SC 29526 Department of Orthopaedic Surgery Bioengineering Georgetown University Medical Center Vice Chair for Research Victor Valderrabano, M.D., Ph.D. Washington, D.C. 20007 Director, McKay Orthopaedic Research Human Performance Laboratory and Laboratory Harriët Wittink, Ph.D., M.S., P.T. University of Pennsylvania Orthopaedic Department Head Philadelphia, PA 19104 University of Calgary Physical Therapy Professional Master Program Calgary, Alberta T2N1N4 Hogeschool Utrecht Dan M. Spengler, M.D. Canada 3508 AD Utrecht Professor and Chair also The Netherlands Department of Orthopaedics and Orthopaedic Department University Hospital of Basel Joseph D. Zuckerman, M.D. Rehabilitation 4031 Basel Professor and Chair Vanderbilt Orthopaedics Institute Switzerland Department of Orthopaedic Surgery Nashville, TN 37232 NYU Hospital for Joint Diseases Tapio Videman, M.D., D.Med.Sci. New York, NY 10003 Marek Szpalski, M.D. Professor Associate Professor and Chair Faculty of Rehabilitation Medicine Department of Orthopaedics University of Alberta IRIS South Teaching Hospitals Edmonton, Alberta T6G 2G4 Free University of Brussels Canada 1190 Brussels Belgium Sherri Weiser, Ph.D. Research Assistant Professor of Environmental James B. Talmage M.D. Occupational Health Center Medicine Cookeville, TN 38501 Occupational & Industrial Orthopaedic Center NYU Hospital for Joint Diseases New York, NY 10014

1C H A P T E R Introduction to Epidemiologic Concepts in Musculoskeletal Disorders Mary Louise Skovron and Rudi Hiebert The literature on the epidemiology of occupational musculoskele- Health Administration (OSHA) by employers, workers’ compen- tal disorders is often confusing because of conflicting evidence on sation records, records of visits to the workplace health facility, the importance of various potential risk or causal factors. This and surveys of the work force.11 In clinical practice, the simple chapter describes basic epidemiologic methods so the reader can case count is usually derived by chart review (retrospectively) or evaluate critically the published literature on occupational mus- by enrollment of patients seen during a given period (prospec- culoskeletal disorders. Most examples are drawn from the litera- tively). The frequency of the disorder can also be expressed as a ture on occupational low back pain, but the reader should be proportionate ratio, a ratio of cases of a particular disorder to aware that similar methodologic standards must be applied to cases of all disorders in the population of interest. In 1985 for the literature on upper extremity disorders. example, occupational back injuries accounted for 26% of all closed compensation cases in a sample of nine states. Epidemiology is the study of the distribution and determi- nants of diseases and injuries in human populations. It consists By itself, numerator data cannot provide useful information of a developed methodology for testing scientific hypotheses in regarding the risk or probability of acquiring the disorder. The groups of individuals rather than in a laboratory setting. With case frequency has to be related to the underlying population knowledge of the intrinsic strengths and limitations of the design that could have potentially developed the disorder. For example, and execution of studies reported in the literature, it is possible to the U.S. Bureau of Labor Statistics estimated 303,750 OSHA- evaluate the strength of the evidence derived from these studies reportable occupational injuries involving the back in 2003.4 and even to make sense of conflicting results from different studies Without reference to the number of people at risk, it is not pos- on the same topic. In this chapter we present an overview of the sible to estimate the risk of back injury in the population or to test basic terminology used in epidemiology and the characteristics and hypotheses regarding risk factors for occupational back injury. For generic strengths and limitations of analytic (hypothesis testing) this reason, rates are used when the objective is to assess the risk study designs, with an emphasis on observational study designs. of the disorder or determinants of disorders or their outcomes. There are several types of epidemiologic studies. Descriptive Rates and ratios epidemiology is a means of monitoring the health of a population, identifying health problems, and compiling information that can Rates describe the frequency of a disorder or disorder per unit be used for the development of causal hypotheses. Analytic epi- size of the population per unit time of observation. The rates demiology is a set of epidemiologic study methods used to test commonly used in epidemiology are morbidity and mortality specific hypotheses. rates. The general form of a morbidity or mortality rate is MEASURES OF DISORDER FREQUENCY Number of cases risk × 100 (1000, etc.) per unit time Number of persons at The fundamental strategy of epidemiology is the analysis of rel- ative and absolute measures of frequency and a comparison of The most frequently used morbidity rates in epidemiologic the characteristics of individuals with and without disorder. The research are the incidence rate and the prevalence rate. The inci- most obvious measures of frequency are case counts and their dence rate is based on new cases of a disorder or disorders (or new variations, which are often referred to as numerator data. They disorder events), whereas the prevalence rate is based on existing describe the frequency of the disorder without reference to the cases. Because they are based on new versus existing cases, underlying population at risk. Examples of sources of case count data include back injury reports to the Occupational Safety and

4 Chapter 1 ● Introduction to epidemiologic concepts in musculoskeletal disorders incidence and prevalence rates have different uses and different stable and the duration of a disorder is also stable, it is possible limitations. to estimate prevalence from incidence and vice versa according to the following approximation: In a sense, the incidence rate is a rate of change, the frequency with which people change from healthy to injured, sick, or dis- Prevalence ~~ incidence × duration abled. Therefore the appropriate denominator is the population at risk of acquiring the disorder (i.e., those who are free of the Thus a change in prevalence may reflect changes in the inci- disorder at the start of the time interval). The incidence rate may dence rate, duration, or both. For example, the prevalence of low be quantified in a number of ways, for example, as the number back pain in a population may change because of alterations in of new events per 1000 persons per year, when the population is individual, work-related, or other environmental risk factors stable and the number of new events is counted each year. affecting incidence rate or because treatment changes alter the Alternatively, it may be quantified as the number of new events duration of back pain episodes and risk of chronicity. It is occa- per 1000 person-years, as is done in prospective studies where a sionally the case that improved treatment extends the duration fixed population is followed until the disorder, the end of the of a disorder, with the result that the prevalence increases in the study, or loss to follow-up occurs. In practice, although the best face of a decreasing incidence, as occurred some decades ago denominator for incidence rates is the number of people free of with Down syndrome. The survival of infants with Down syn- the disorder at the start of the time interval, surveillance incidence drome improved because of improved medical and surgical man- rates (and prevalence rates) that are based on case reports often use agement of their associated disorders. The prevalence of Down the total population derived from census data or from work-force syndrome increased, although the incidence declined as a result estimates. The U.S. Bureau of Labor Statistics’ estimate of 303,750 of prenatal screening programs. OSHA-reportable occupational injuries involving the back repre- sents an incidence of 3.46 new cases per 1000 workers.4 DESCRIPTIVE EPIDEMIOLOGY The prevalence rate is the number of existing cases of a disor- The first step often undertaken in epidemiology is development der in a given population in a given time period. For example, of the descriptive epidemiology of a disorder or disorders. the 1-year prevalence of disabling back pain is as high as 25%.14 Descriptive epidemiology supports the development of causal Point prevalence is the number of cases per unit population hypotheses but does not in itself support conclusions about disor- at one moment of counting, for example, all persons receiving der causality or about any hypotheses. In descriptive epidemiol- disability because of back pain in the work force of a metropoli- ogy the frequency of a disorder in the population is characterized tan electrical utility company on January 1, 2005, expressed per in terms of person (e.g., age, sex, ethnicity-specific incidence rates, 1000 population. For point prevalence, the unit of time is often economic, behavioral, occupational, and other factors), place not expressed because the period of time is effectively instanta- (rural versus urban, type of housing, national variations, type of neous. Period prevalence is the number of cases existing at one industry, job requirements), and time (long-term trend, seasonal- time or another during a definable time interval such as 1-year, ity, occasionally day of the week or time of day). 5-year, or lifetime prevalence. Some epidemiologists do not express prevalence as a rate because in practice it is often derived The need to explain variation in descriptive studies drives the from surveys that are difficult to assign to a specific time interval. formulation of causal hypotheses. Drawing on current available information from various fields (such as anatomy, physiology, A number of factors other than the risk factor under study psychology, behavioral science, etc.), specific hypotheses are may affect the incidence and prevalence rates. These include developed by inductive reasoning to explain observed patterns of demographic characteristics of the underlying population, most variation and then evaluated using specific study designs to test obviously age distribution6 because age is known to be associated these hypotheses. Studies that test specific hypotheses are called with the onset of almost all disorders. Gender and ethnicity dis- analytic. As the results of hypothesis-testing (analytic) studies are tributions must also be taken into account when incidence rates accrued, they are added to the basis for causal inference, depend- are interpreted. Other influences can distort the apparent inci- ing on their strengths and generalizability, and hypotheses are dence rate, including certain company policies, workers’ com- supported, modified, or negated. pensation claims, and health care system influences that affect the likelihood of seeking medical attention, of being diagnosed In interpreting the evidence from all scientific sources, the with a given disorder or disorders, or of having the disorder rules of causal inference are applied.7 Briefly, the hypothesized reported. These factors should be considered when measures of cause must be demonstrated to have preceded the disorder by a disorder frequency are evaluated, particularly when changes are length of time sufficient to allow disorder development and assessed over time or different populations are compared. expression (time sequence of events). The disorder should be more common in those with the hypothesized cause than in To eliminate the effects of differences in these factors, the rates those without it (increased risk in those exposed to the hypothe- may be adjusted or standardized algebraically. The adjusted rates sized cause), and as the intensity or duration of exposure to the express the risk of acquiring the disorder in the populations being hypothesized cause increases, the frequency of the disorder compared as if they had the same age, sex, and ethnicity distribu- should increase (dose-response relationship). The association tions. Alternatively, if it is not necessary to have a single summary between the hypothesized causal factor and the disorder should index of disorder risk, the morbidity rates within population be consistently demonstrated in methodologically sound studies strata defined by age, sex, and ethnicity may be compared. and should be biologically plausible. In addition, the specificity of an association (i.e., the extent to which the hypothesized The number of existing cases of a disorder or disorders at any time is a function of both the rate of new cases (incidence) and the duration of that disorder. Therefore, when a population is

Chapter 1 ● Analytic epidemiology 5 causal factor is associated with only one disease or disorder) adds bias because the health behavior and health status of people who weight to a causal hypothesis, but it is not necessary for causal volunteer for research are well documented to be better than inference; for example, cigarette smoking is accepted as a cause those of refusers. No characteristics of the individuals should of lung cancer, although the association is not specific. Cigarette affect the likelihood of selection for the study, including their smoking is also associated with a number of other cancers, knowledge of the question at issue; their beliefs about the risk obstructive pulmonary disorder, heart disorder, and a variety of factors or about the cause of the disorder being studied; or any disorders, including osteoporosis, low back pain, and, in particu- characteristic such as age, sex, or education that could be inde- lar, herniated intervertebral disks. pendently associated with both the disorder and the hypothe- sized causal factor. ANALYTIC EPIDEMIOLOGY It is important for the internal validity of the study results Analytic, or hypothesis-testing, epidemiology relies on two types that the information collected is accurate and complete. If there of study designs: observational and experimental. In observa- is inaccuracy (measurement error) in the information collected, tional studies, exposure to the hypothesized causal factor and the ability to detect the association of interest is reduced. If the development of the disorder in the population under study accuracy of the information is worse for one exposure group occur in the natural course of events; the investigator does not than for another, the effect on the study results may not be cause them to occur. The study is designed and executed to max- predictable. For this reason, an evaluation of the accuracy (or imize the extent to which it can be seen as a natural experiment, validity) of measurements is necessary for any study. Research that is, the extent to which all extraneous sources of variation are reports should describe the validity of the sources of information. eliminated and only the exposure to the putative cause and the Questionnaires or reporting methods that have been validated in frequency of disorder vary between populations being com- the study population or in similar populations or circumstances pared. It is often the case that once substantial observational should be used. The problem of validity of information is partic- evidence has accrued, causality is widely accepted. However, it is ularly important in research on occupational musculoskeletal desirable in etiologic epidemiology and almost universally disorders because the methods of both case diagnosis13 and required in evaluations of treatment that the final test of the measurement of work exposure17 have substantial limitations. hypothesis is in interventional or experimental studies. Before specific study designs can be discussed, the term con- In experimental studies, the investigator causes individuals or founding must be defined. Confounding occurs when the study groups of individuals in the population to receive the treatment results can be explained by a factor extraneous to the hypothesis in question. To demonstrate ethically the causal role of a risk fac- being tested. A potential confounding factor must be associated tor for which there is only observational evidence, the investiga- with both the disorder in question and the hypothesized causal tor would prevent exposure to the risk factor for a group of factor. That is, the proportion of persons with the disorder hav- people. In both types of interventional design strategies, a com- ing the confounding exposure must be different from the pro- parison group that does not receive the intervention is necessary. portion of persons without the disorder with the confounding All other factors that might influence the outcome of the study exposure. It is also necessary that the proportion of those with (potential confounding factors) can be eliminated or controlled the hypothesized causal factor who have the confounding expo- by the investigator. Because the conditions of the study are much sure are different from the proportion of those not exposed to more under control of the investigator, interventional studies can the hypothesized causal factor who have the confounding factor. more closely approximate true experiments than can observa- For example, a study that found an association between job sat- tional studies. When such studies are well designed and executed, isfaction and the risk of occupational back injury could be con- they provide very strong support (or negation) for a hypothesis. founded by the physical requirements of work if heavy work was a risk factor for back injury and was also associated with lack of All analytic study designs have potential problems of internal job satisfaction in the studied population. Potential confounding and external validity that must be solved by the investigator factors can be eliminated in the design of the study by restricted either in the study design or in the data analysis. Internal validity or matched sampling or, in the data analysis phase, by stratified is the extent to which a study is a true test of the specific hypoth- or multivariate analysis, for example. If in the study just esis, that is, the extent to which all possible biases of measure- described the statistical analyses controlled for physical require- ment or information and all possible confounding variables are ments of work or if the researchers conducted an exploratory eliminated as explaining the observed study result. External analysis and found no association between job satisfaction and validity is the extent to which the study results can be general- the physical requirements of work, the potential for confound- ized to the population of interest, namely, whether the study ing would be eliminated. In experimental studies, potential con- subjects are representative of the population at risk. If the poten- founding should be successfully eliminated by truly random tial validity problems have been solved in either the design or blind assignment of subjects to the different treatments under analysis of the study, the study evidence is strengthened. study. Comparability of the treatment groups should be con- firmed by presentation of the baseline characteristics of each Because it is not possible to study the entire universe of group on entry to the study. potentially eligible subjects, epidemiologic studies are conducted on samples of the population of interest. Even a study of an entire Confounding invalidates a study as a test of the hypothesis. city or the work force of a company constitutes a sample. The The study’s results cannot be taken as evidence of causality or method of sampling should not introduce selection biases. For efficacy of treatment. Lack of generalizability, as opposed to con- example, a volunteer study is potentially susceptible to selection founding, does not invalidate a study’s results but merely restricts inference to populations similar to those under study.

6 Chapter 1 ● Introduction to epidemiologic concepts in musculoskeletal disorders Observational study designs are applicable in both clinical Risk factor Disease and etiologic epidemiology. In etiologic epidemiology the present or researcher tests whether a hypothesized factor is a determinant or cause of disorder in previously healthy people, whereas in clini- Target Sample outcome occurs cal epidemiology one tests whether particular characteristics, risk population factors, or clinical interventions are determinants of the progno- Disease or sis or outcome. The classic observational analytic study designs Risk factor outcome are the cohort study, the case-control study, and the cross-sec- absent does not occur tional study. Disease Cohort study (Prognostic study) or The cohort study is the observational design that, when well outcome occurs designed and executed, produces the soundest results in terms of incidence rates and disorder etiology or prognostic determinants Disease or of all the observational study designs. The hallmark of a cohort outcome study is that a population initially free of the outcome of interest does not occur is identified and characterized with respect to the hypothesized risk factor, important covariates, and potential confounders. The Figure 1.1 Cohort study. population is observed for a period of time adequate for devel- opment of the disorder, and the new cases (incident cases) are In this case, the observed relative risk would need to be very large recorded. Rates of disorder development are compared between to support the causal hypothesis. For example, consider a cohort those exposed and those not exposed to the hypothesized risk study examining the causal role of occupational repetitive factor. motion in carpal tunnel syndrome. New workers hired in 1985 through 1990 are enrolled and followed forward for 10 years, A study of prognostic factors related to return to work after with information on new cases of carpal tunnel syndrome com- episodes of absence due to work-related low-back-pain sickness is ing from the company medical department records. If 30% of an example of a cohort study. The cohort consisted of all those the workers retire, take disability pensions, die, get another job, first presenting to an occupational health clinic at a large munic- or leave the company for other reasons, there is a substantial loss ipal transportation agency for medical clearance for sick leave to follow-up. A bias in loss to follow-up occurs if the workers from work because of a complaint of work-related low back pain. who leave the company are those with the highest exposure to These individuals were asked to complete a questionnaire on repetitive work movements and those who leave because upper function, pain, satisfaction with work, and beliefs about pain. The extremity problems consistent with preclinical carpal tunnel syn- occupational physicians conducting the sickness absence clear- drome are making it more difficult for them to do the job. The ance examinations included assessment of gain, posture, and dis- observed relative risk is an underestimate of the true relative risk tribution of painful symptoms specific for back pain. Participants because the detected incidence of carpal tunnel syndrome among in the study were followed for 3 months, at which time the those with repetitive-motion jobs is lower than the true incidence participant’s return to work status was determined. To identify and the detected incidence among those not exposed is not which factors best predicted return to work, rates of return to affected. Biased loss to follow-up leading to underestimates of work were compared between those with high and low scores on incidence in the unexposed would produce an inflated observed clinical signs and symptoms, function, pain, work satisfaction, relative risk. High proportions lost to follow-up or higher propor- and pain beliefs. Predictors that showed large differences in rates tions lost in one exposure category than another (selective loss to of return to work were interpreted as being strongly predictive.12 follow-up) leave open the possibility of biased loss to follow-up with consequent distortion of the study findings. Cohort studies can be prospective in nature, meaning that a disorder-free population or group is initially identified and then Another form of selection bias can occur. This bias, called subsequently tracked over time (Fig. 1.1). This same model can selective survival or selective attrition, occurs when people who also be used with historical records. Employment records, for have both the exposure and the disorder have a different proba- example, can be used to identify a group of new employees at a bility of dropping out of the population available to be included company. Job status and medical records can then be linked to in the study than do people who are not exposed and get the these employment records to identify work exposures and the disorder. This type of bias can easily occur in cross-sectional and development of the disorder of interest. Studies that use histori- case-control studies. It can also occur in a particular variant of cal records are called retrospective. the cohort study called the prevalent cohort study. For example, a prevalent cohort study examining occupational repetitive Loss to follow-up is a potential problem in cohort studies. If motion as a risk factor for carpal tunnel syndrome that enrolled a substantial proportion of subjects are lost to the study for any workers who were first employed between 1985 and 1990 and reason, for example, having moved out of the region, it would be were still actively employed in 2005 could be affected by selec- expected that fewer cases of the disorder in question would arise tive attrition if carpal tunnel syndrome by and large developed in the study than originally planned. The number of study cases within 15 years of employment and workers tended to leave the may ultimately be too small to yield stable estimates of the company when carpal tunnel syndrome developed. incidence rates and, consequently, estimates of the relative risk.

Chapter 1 ● Analytic epidemiology 7 Some diseases or disorders take many years to develop after and ergonomic factors related to work. Data on these factors the initial exposure to the presumed causal factor or take many were categorized. To analyze these data, the ratio of those exposed years of exposure. The duration of time between the time of to the risk factor to those not exposed was calculated once for exposure to the presumed causal risk factor and the development those with back pain and again a second time for those without of disease or disorder is called latency. Another problem concerns back pain. A risk factor was interpreted to be associated with the prevalence of the disorder in the population. If a disorder is the back pain when the ratio of having the risk factor was higher rare, many thousands of subjects may be required to identify and among those with back pain as compared with those without collect enough cases where the disorder occurs to be suitable for back pain. statistical analysis. Consequently, the cohort study design is not optimal in situations where the disorder of interest is both very Case-control studies frequently suffer from information rare and also has a long latency period. It can be more efficient biases. For example, if information on exposure to the risk factor in terms of time and the number of subjects studied to address of interest comes from a different source for case and control the hypothesis by means of a case-control study, as described in subjects, biased exposure information is possible. Recall bias, the next section. in which a case subject is more or less likely to recall an event in the past than is a control subject, is also possible. There is also Case-control study the problem of unbiased recall failure, in which subjects are asked to recall events or conditions that took place so long The essential feature of the case-control study that differentiates ago they cannot be remembered. Establishing that exposure to it from the other observational study types is that individuals are the factor of interest took place long enough before the outcome selected for the study on the basis of the presence of the disor- to be a biologically plausible determinant is difficult for certain der in question (cases) and compared with individuals selected types of hypotheses; for example, a case-control study examining for the study on the basis of the absence of the disorder under preexisting degenerative disk disorder as a determinant of study (control subjects). The presence or absence of the hypoth- chronicity (symptom duration greater than 6 months) in workers esized causal factor is then ascertained in both case and control with chronic back pain could not establish that the disk problem subjects. Although this appears on its face to be a simple under- predated chronicity based on clinical or imaging examinations at taking, case-control studies present a number of methodologic the time of study. These problems are avoided if the case-control challenges that must be solved for the study results to be valid study uses exposure or prognostic information that was recorded, (Fig. 1.2). for example, in medical or prescription records, long enough before the disorder condition being studied to be a biologically A study of ergonomic risk factors for work absence due to onset plausible cause and to obviate recall problems. of low back pain–related sickness conducted among Baltimore City workers is an example of a case-control study.10 Two hundred Well-designed and well-conducted case-control studies may cases of absence due to back pain–related sickness were identified provide evidence as robust as that of cohort studies at consider- from the city’s occupational health department. Four hundred ably less cost and in considerably less time. However, because individuals without back pain but matched on gender, job classi- of the difficulty in avoiding the problems just described, case- fication, and department served as control subjects. In-person control studies often produce weaker causal evidence than do interviews were conducted to collect data on demographics, cohort studies. work history, psychosocial and work organization characteristics, Cross-sectional study Disease Cases Risk Factor Cross-sectional studies simultaneously ascertain exposure to present? present? risk factors (or the presence of prognostic factors) and the pres- Yes ence of the disorder or outcome in question in a population Yes Sample sampled without regard to the presence of either. This type of No sampling is sometimes called naturalistic sampling. In contrast Population to a cohort study, which follows subjects over time and ascertains Yes incidence, a cross-sectional study ascertains conditions present No Sample Controls at the moment of study, that is, the prevalence of the disorder No or outcome in question at the time of the study. The estimates of relative risk derived from cross-sectional studies are therefore Figure 1.2 Case-control study. estimates of prevalence relative risk. Population-based cross- sectional studies of low back pain often address, among other factors, the association of the type of work (occupation, physical requirements, and so forth) with low back pain (Fig. 1.3).2 Cross-sectional or survey studies are often undertaken because, unlike case-control studies, they require few a priori decisions with regard to the selection of subjects and, unlike cohort studies, it is not necessary to wait for the study outcome. These advantages are offset by their susceptibility to some of the prob- lems of both cohort and case-control studies. When uncommon

8 Chapter 1 ● Introduction to epidemiologic concepts in musculoskeletal disorders Risk factor present Table 1.1 Strengths (+) and limitations (-) of the disease present observational study designs Study Risk factor present Feature Cohort Case- Cross- population disease absent Study Control Sectional Study Study Risk factor absent Sample disease present Selective Survival ++ − Risk factor absent Recall bias +− − disease absent Loss to follow-up −+ + Time sequence of events + − − Time to complete −+ + Expense − + +/− Figure 1.3 Cross-sectional study. disorders or exposures are being studied, a large number of people cause the exposure to the causal factor or treatment for the pur- must be included, as in cohort studies. If information on expo- poses of the study, whereas in interventional designs the investi- sures or on determinants of interest is collected at the time of the gator does cause subjects to be exposed to different factors or study rather than from previously existing records, there can be treatments. Observational study designs are susceptible to treat- recall biases, recall failure, and problems in establishing the time ment assignment biases in which the treatment the patients sequence of events, just as in case-control studies. Nevertheless, receive is influenced by certain patient characteristics (e.g., life- for relatively common disorders (outcomes) and risk factors style or clinical severity) that can confound the results. Clinical (determinants), cross-sectional studies may be a useful first step trials, in which the treating physicians or the investigators con- in exploring a hypothesis. Because of their many limitations, trol which treatment patients receive, are also susceptible to such however, cross-sectional studies rarely produce robust results for biases. For this reason, randomized controlled trials, where only evaluating the importance of causal or prognostic factors. chance influences which treatment eligible patients receive, are the preferred method of evaluating therapeutic interventions. When the literature on a problem consists predominantly of cross-sectional studies, it is often the case that the analytic epi- The validity of randomized controlled trials depends on all demiology of that problem is in its infancy. Until recently, much the methodologic features described for the observational study of the epidemiologic information on occupational low back pain designs and more. The study must be confined to those patients was derived from descriptive and cross-sectional studies.16 In the who have agreed to participate. Comparisons of treatment out- past 5 or 6 years there has been a substantial advance in the quan- comes in patients who agree to participate with those in patients tity and quality of observational analytic studies of work-related who refuse to participate are not valid. Assignment of patients to back pain. The epidemiologic investigation of upper extremity treatments must be done by using accepted methods of random- disorders began later than that of low back pain. Consequently, ization, which are described in the report, and the resulting com- knowledge of the risk factors for work-related upper extremity parability of the treatment groups on important covariates disorders is less developed. The intrinsic strengths and limita- should be described, usually in a table summarizing the baseline tions of the basic observational study designs are summarized in characteristics of the treatment groups. On the occasions when, Table 1.1. by chance, randomization does not result in comparable groups, potential confounding must be controlled in the statistical analysis. Experimental study designs: clinical trials Figure 1.4 is a schematic representation of appropriate design in a randomized controlled trial. The distinction between observational and interventional study designs is that in observational designs the investigator does not Ordinarily, in randomized trials the treating physician and the patient are blind to which treatment group the patient has been assigned. If this is not possible, assessment of the study Randomize Experimental Successes/ treatment Failures Patients Eligible Agree Figure 1.4 Randomized control trial. Standard Successes/ treatment Failures Ineligible Refuse

Chapter 1 ● Statistical Issues 9 outcome should be done by an independent evaluator to avoid case-control study). Instead, an alternative to the relative risk needs observer and participant biases in assessing the outcome. This is to be used. The odds ratio is a measure of association that, in cer- particularly important when the outcome being assessed is tain circumstances, can be used to estimate relative risk. The odds subjective. Information should be collected in the same way ratio is the ratio of the odds of the disorder in those exposed to and with the same frequency in all treatment groups. Eligibility the odds of the disorder in those unexposed. It also has valuable and exclusion criteria should be described and be appropriate to statistical properties because it can be estimated by using logistic the question being addressed. Treatments should be clearly regression. The effects of confounding variables can be controlled described, and patient compliance, dropouts from the study, or the simultaneous effects of several causal variables or covariates and complications should be described and equivalent in both can be estimated by using multiple logistic regression. groups. Another measure of association can be found in studies that Finally, the outcomes studied should be appropriate to the examine the rate of the development of a disorder in a population treatment or condition in question. A number of general health over time. In these studies, the risk of developing a disorder within status assessment measures are used, for example, the SF-36, a cohort changes for each point in time. As members of the cohort a standardized multidimensional assessment instrument that develop a disorder, the total number of individuals in the cohort includes functional capacity, pain, locomotion, mental status, still free of the disorder becomes smaller and the calculation of and affect. There are numerous assessment instruments for back risk changes. A survival curve shows the cumulative proportion of pain disability, including the Oswestry,5 the Roland-Morris ques- cohort members remaining free of the disorder on the vertical axis tionnaire,15 the Quebec Back Pain Disability Scale,8 the Maine- and time on the horizontal axis (Fig. 1.5). Typically, survival curves Seattle back pain disability questionnaire,1 and others.3 show an exponential relationship between cumulative proportion remaining free of the disorder and time. We can examine whether STATISTICAL ISSUES the survival experience is different between members of a cohort with different exposure profiles by using statistical tests specific for Methods of analysis this type of analysis, such as the log-rank test.9 The statistical analysis of any study result should be appropriate to Hazard is a term that expresses the rate of change of the the hypothesis and to the structure of the data collected. When, cumulative proportion surviving with time. In prognostic studies for instance, the study examines the difference in Oswestry it is possible to compare whether the rate of change of survival scores associated with a conditioning program as compared with (hazard) is different between exposure groups. A useful property usual care for subacute low back pain, comparisons of the mean of hazards is that this term can be modeled using logistic regres- scores in the treatment groups may be appropriate. If it is neces- sion techniques. The Cox proportionate hazards model is used sary to control for pretreatment differences between the groups, to evaluate differences in hazard between exposure groups. The the analysis uses multivariate methods such as analysis of covari- hazard ratio is interpreted much like the relative risk or odds ance or multiple regression. Occasionally, because of the statistical ratio. When the hazard ratio is equal to 1, then the survival expe- characteristics of the outcome being assessed, it may be necessary rience is interpreted to be the same among exposure groups. to transform it (e.g., log transformation, square root transforma- When the hazard ratio is not equal to 1, the interpretation is that tion) and analyze the transformed variable. It is often the case there is an association between exposure to the risk factor and that the outcome variable distribution or the conditions of the survival experience related to the disorder of interest.9 study do not conform to the requirements of the usual statistical hypothesis tests such as t-tests, analysis of covariance, and regres- An advantage of survival analysis is that all study subjects con- sion analysis. In these cases, a nonparametric method of statisti- tribute information for as long as they remain in the study. The cal analysis such as the Wilcoxon method is appropriate. reader should be aware, however, that if the number of dropouts during the course of the study is substantial, estimates of the When the hypothesis addresses the relative frequency of an hazard ratio toward the end of the follow-up period are based on event such as a back injury rate, a ratio can be formed consisting relatively small numbers and are consequently unstable. of the risk of development of the disorder among those exposed to the risk factor compared with the risk of the disorder among Estimates and confidence limits those not exposed to the risk factor. This ratio is called a relative risk. When the relative risk is 1, then the risk of the disorder is Research is conducted on a sample of persons or other units of the same among those exposed to the risk factor as those not observation drawn from a target population. The results of any exposed. However, when the relative risk diverges from 1, then given study are estimates of the true means, proportions, relative the risk is not the same between the exposed and unexposed risks, and so forth in the population from which the samples were groups. This is interpreted as evidence for an association between drawn. The precision of a study estimate of the population value, the risk factor and the disorder. The relative risk may be adjusted or parameter, of a measurement is described by the standard for important covariates or to eliminate potential confounding. error of estimate. The standard error (SE) is the square root of the Relative risks can be calculated only in those studies where the ratio of the variance (s2), or variability of the measurement in the entire study population is tracked or a representative sample is sample, to the number of subjects (N) in the study. For example, identified, as would be the case in a cohort study. SEmean = s2 A relative risk cannot be calculated in situations where the N study sample is not representative (e.g., as would be the case in a

10 Chapter 1 ● Introduction to epidemiologic concepts in musculoskeletal disorders 100% Percent returning back to any work 75% 50% 90 180 270 Figure 1.5 Return to any work from Time lost from work because of LBP (in days) sick absence because of nonspecific 25% low back pain (n = 225). (From Hiebert No restrictions R, Skovron ML, Nordin M: Work 0% Some restrictions restrictions and outcome of nonspecific 0 low back pain. Spine 28(7): 722-728, 2003. Reprinted with permission.) 364 Variance is affected by a number of factors, including or differences between treatments reflect the true situation in the interindividual variability, intraindividual variability (such as diur- population from which the samples were drawn. By convention, nal variations), and instrument variability. Designing or executing the cutoff for rejecting the null hypothesis is usually set at 0.05. a study to reduce any of these components reduces the variance of Then if the probability (p value) that the observed results are due the measurement, thus reducing the standard error and increasing to sampling error is less than 0.05, that is, less than α, the null the stability of the estimate of the population parameter. The hypothesis is rejected and the results are declared statistically signif- larger the number of subjects on whom the estimate is based, the icant. Thus, statistically significant results are simply results that smaller the standard error and the more confident we can be in its we have decided, within an acceptable margin of error, probably representation of the population parameter. did not occur by chance. Further, the larger the observed associ- ation relative to the underlying variability of the outcome being Because sample results are estimates of population parameters, measured, the more likely that it will be declared statistically it is increasingly becoming the standard of reporting to describe significant. the precision of the estimates as a range within which the popu- lation parameter probably lies. This is the confidence limit Statistical power and sample size around the estimate and is by convention expressed as the 95% confidence limit. For example, the 95% confidence limit for a Statistical hypothesis tests actually involve two probabilities. The mean is approximated by probability of making a type I error by incorrectly rejecting the null hypothesis, that is, by declaring an observed association to 95% confidence interval = mean ± 2(SEmean) be statistically significant when in fact it is the result of sampling error, is referred to as α, as described in the preceding paragraph. Statistical hypothesis testing There is also the probability of incorrectly accepting the null hypothesis; that is, declaring that the study results are due to Because there is always sampling error, estimates may be sampling error (not statistically significant) when in fact they expected to vary from sample to sample. Consequently, study reflect a true association in the population from which the study results must be subjected to statistical hypothesis testing; that is, subjects were drawn. This is the type II error and its probability is study results must be tested to determine the probability that the beta (β). The complementary probability that a study will be able observed results from a specific study could have occurred by to correctly reject the null hypothesis when it is false, that is, chance alone. correctly detect an association when there is one in the popula- tion at large, is referred to as statistical power (1 – β). Table 1.2 The statistical hypothesis test evaluates the null hypothesis illustrates the different conditions and possible results of a statis- that the observed study results occurred because of sampling tical hypothesis test. error when there was no true association in the population from which the study subjects were sampled. The probability of mak- In the planning phase of research the investigators should ing this type of error is designated as alpha (α). If the observed make a determination of how strong an association would be association is large enough that this kind of error is improbable, clinically significant, that is, how large an estimated relative risk the null hypothesis is rejected. The investigators then accept the alternative hypothesis, that the observed estimates of relative risk

Chapter 1 ● References 11 Table 1.2 Population conditions, statistical result should be specified, the plan of statistical analysis deter- hypothesis test results, error types, and designations mined, and the necessary number of study subjects defined. Study management should avoid the introduction of differential Hypothesis Population Condition loss to follow-up, unblinding, and other potential problems. The Test Result statistical analysis should be appropriate to the structure of the data and to the hypothesis. Finally, although the discussion Accept null hypothesis NULL HYPOTHESIS TRUE NULL HYPOTHESIS FALSE should place the study in the context of other work and what is No association NO ASSOCIATION ASSOCIATION EXISTS already known about the question, the specific conclusions Correct Type II error (beta) should not go beyond what was actually tested. Reject null hypothesis Type I error (alpha) Correct Association exists REFERENCES or how big a difference between treatments. Because the validity 1. Atlas SJ, Deyo RA, van den Ancker M, Singer DE, Keller RB, Patrick DL: The of the study requires that it be a true test of the research hypoth- Maine-Seattle back questionnaire: a 12-item disability questionnaire for evaluating esis, it is important to design the study so that a clinically signif- patients with lumbar sciatica or stenosis: results of a derivation and validation icant association will have a good chance of being declared cohort analysis. Spine 28(16):1869-1876, 2003. statistically significant, that is, so that the study has sufficient power to detect a clinically significant association. The larger the 2. Bernard BP, ed: Musculoskeletal disorders and workplace factors: a critical review sample size, the more power the statistical test has to detect asso- of epidemiologic evidence for work-related musculoskeletal disorders of the neck, ciations; in other words, as expected differences or relative risks upper extremity, and low back. U.S. Dept. of Health and Human Services, Public get smaller, the number of subjects studied must increase to have Health Service, Centers for Disease Control and Prevention. Cincinnati, OH, 1997, adequate power to test the hypothesis. Conversely, with very National Institute for Occupational Safety and Health. large numbers of study subjects it is possible to declare trivial associations statistically significant. When studies with small 3. Bombardier C: Outcome assessments in the evaluation of treatment of spinal sample sizes report results that are not statistically significant, disorders: summary and general recommendations. Spine 25(24):3100-3103, they should also report how large an association would have 2000. been required for there to have good power to detect it. The reader should also evaluate whether the observed difference and 4. Bureau of Labor Statistics, US Labor Department, Annual Survey 2003, its upper confidence limit, although not statistically significant, Tables R19, 22. are clinically significant. When studies with huge numbers of subjects report statistically significant results, the reader should 5. Fairbank JC, Pynsent PB: The Oswestry Disability Index. Spine 25(22):2940-2952; decide whether the differences are trivial in clinical terms, even discussion 2952, 2000. though they are statistically significant. 6. Gyntelberg F: One-year incidence of low back pain among male residents of SUMMARY Copenhagen aged 40-59. Danish Med Bull 21:30-36, 1974. The validity of clinical research depends on a number of factors. 7. Kelsey JL, Thompson D, Evans AS: Methods in observational epidemiology. New York, The hypothesis must be formulated specifically enough to be 1986, Oxford University Press. testable. The appropriate study subjects should be eligible, and there should not be differential participation. The information 8. Kopec JA, Esdaile JM, Abrahamowicz M, et al: The Quebec Back Pain Disability Scale: collected should be appropriate to the hypothesis and accurate. measurement properties. Spine 20(3):341-352, 1995. The study design and information sources should avoid potential information biases. Potential confounders should be eliminated 9. Lee E: Statistical methods for survival data analysis. Belmont, CA, Lifetime Learning in the study design or controlled in the statistical analysis. At the Publications, 1980, Wadsworth Inc. time the study is designed, a clinically significant hypothesized 10. Myers AH, Baker SP, Li G, et al: Back injury in municipal workers: a case-control study. Am J Public Health 89(7):1036-1041, 1999. 11. National Research Council and the Institute of Medicine. Musculoskeletal disorders and the workplace: low back and upper extremities. Panel on Musculoskeletal Disorders and the Workplace. Commission on Behavioral and Social Sciences and Education. Washington, DC, 2001, National Academy Press. 12. Nordin MN, Skovron ML, Hiebert R, et al: Early predictors of delayed return to work in patients with low back pain. J Musculoskel Dis 5(2):5-27, 1997. 13. van den Hoogen HM, Koes BW, van Eijk JT. On the accuracy of history, physical exam- ination, and erythrocyte sedimentation rate in diagnosing low back pain in general practice. Spine 20:318-327, 1995. 14. Praemer A, Furner S, Rice DP: Musculoskeletal conditions in the United States. Rosemont, IL, 1999, American Academy of Orthopaedic Surgeons. 15. Roland M, Morris R: A study of the natural history of back pain. Part I. Development of a reliable and sensitive measure of disability in low-back pain. Spine 8(2): 141-144, 1983. 16. Skovron ML: Epidemiology of low back pain. Baillieres Clin Rheumat 76:559-573, 1992. 17. Winkel J, Mathiassen SE: Assessment of physical work load in epidemiologic studies: concepts, issues and operational considerations, Ergonomics 37:979-988, 1994.

2C H A P T E R Psychosocial Aspects of Work-Related Musculoskeletal Disorders: Clinical Implications Sherri Weiser Until recently, attempts to unravel the mystery of work-related to vulnerable organs and body systems and may even cause musculoskeletal disorders (MSDs) concentrated on the physical death. Today, the relationship between the mind and the body demands of the job and the physical vulnerabilities of the is studied extensively. Psychologic factors have been shown to worker. It became clear that this problem did not easily lend affect a spectrum of diseases involving virtually all bodily organs itself to reductionistic approaches. Although certain physical and systems.8 First described by Engel in 1977,5 the biopsychoso- loads and tasks have been associated with the development cial model extends beyond mind and body to include the impact of musculoskeletal problems, the strength of these associations of social factors on illness. Influenced by models of stress and has been disappointingly weak.29 Furthermore, a determination illness, Engel concluded that any illness must be viewed from a of physical vulnerability before injury is nearly impossible. Even multidimensional perspective that takes biologic, psychologic, when a physical defect such as disk herniation has been estab- and social factors into consideration.5 Furthermore, these factors lished in conjunction with a report of pain, the correspondence are not independent, as a medical model would suggest, but with actual disability is often low. interrelated. It was by understanding this perspective, Engel believed, that physicians would be prepared to take on the Like all human conditions, MSDs can be infinitely complex. complexities of health and illness. A thorough understanding of these disorders requires a consid- eration of influences beyond the physical. The worker’s psycho- The biopsychosocial model soon became the prevailing view logic attributes and social reality may have significant bearing among those involved in pain research. In 1965 the gate control on the onset, progression, and outcome of occupational MSDs. theory outlined the channels of pain transmission through A biopsychosocial perspective offers much in the way of under- neurologic pathways.19 This theory further asserted that pain standing these disorders. transmission may be modulated by cognitive and affective states. This model was fundamental in explaining the disparities Adoption of the biopsychosocial model in pain and disability between physical findings and the phenomenologic experience research has resulted in the identification of specific psychologic of pain. The gate control theory laid the groundwork for modern and social aspects of MSDs. Recently, this research has been approaches to treating chronic pain. summarized in a number of excellent review articles and treat- ment guidelines, conclusions from which are discussed. The In 1992, Waddell28 presented a detailed analysis of low back biopsychosocial model implies not just the inclusion of psycho- pain with the use of a biopsychosocial model (Fig. 2.1). The many logic and social considerations but a new perspective on the part levels on which pain is experienced are defined in the model. The of all stakeholders involved in the prevention and treatment of physical injury is interpreted cognitively, resulting in a corre- MSDs. This chapter describes how to achieve the goal of pre- sponding emotion. A man who believes that he has a herniated venting and treating the disorder in practical terms. disk, for example, is more apt to be anxious and depressed than a man who believes that he has a sprain. The man with the “her- THE BIOPSYCHOSOCIAL MODEL niated disk” may also display more avoidance behavior initially than the man who believes he sprained his back. This process The notion of a link between the mind and body has existed takes place within a social context that is constantly providing throughout history. It was Walter Canon,2 however, who sub- feedback and modifying the individual’s response. stantiated this idea with his scientific explanation of the “fight or flight” response. His research identified the organism’s physical Waddell also pointed out that feedback loops exist among all reactions to psychologic stress. Selye24 later showed how these levels in the model and that a change in one component affects responses, if left unchecked, can over time cause severe damage the others. If the man with the supposed herniation sees a physi- cian and is assured that the problem is only a sprain, his belief has changed. As a result, his affect and pain behavior will change,

14 Chapter 2 ● Psychosocial aspects of work-related musculoskeletal disorders Social environment PSYCHOLOGIC AND PSYCHOSOCIAL FACTORS ASSOCIATED WITH MSDs Illness behavior Psychologic factors refer to cognitions or beliefs about pain and disability and affective or emotional responses, whereas psy- Affective chosocial factors reflect an individual’s perceptions of others and the environment. Recent systematic evidence-based reviews have Cognitive concluded that even when measured early in the injury, both these factors are stronger predictors than physical factors of out- Sensory comes such as work status. Figure 2.1 A cross-sectional analysis of the clinical findings and The Clinical Guidelines for the Management of Acute Low assessment of low back pain and disability at one point in time. (From Back Pain, published in 2001, state that in addition to social and Waddell G: Biopsychosocial analysis of low back pain. In M Nordin, economic factors, psychologic factors play an important role in TL Vischer, eds: Common low back pain: prevention of chronicity. the development of chronic low back pain and disability and London, 1992, Bailliere Tindall.) influence a patient’s response to treatment and rehabilitation.30 Specifically, beliefs that activity and work will make pain worse and nociception may be experienced as less severe. These feed- (fear-avoidance beliefs) and that the patient is not responsible for back loops are particularly important in chronic low back pain, the pain or treatment along with behaviors intended to commu- wherein the original injury is often resolved. In these cases, psy- nicate that the patient is in pain (illness behaviors) are associated chologic and social variables are even greater determinants of with poor outcome. functional status than in cases of acute back pain when nocicep- tion is responsible for much of the illness behavior. Using strict inclusion criteria, in 2001 Koes et al13 published a comparison of clinical guidelines for the management of low Occupational musculoskeletal injuries are clearly amenable to back pain from 11 different countries that appeared from 1994 biopsychosocial analysis. They either occur at work or are until 2000. They found that in most regards the content of the believed to be its result. The cognitive, affective, behavioral, and guidelines appeared to be quite similar and that all recognized social elements of such an injury are therefore inextricably related the importance of psychologic and psychosocial factors, includ- to the workplace. Injuries or disorders that did not occur at work ing work perceptions, in the development and maintenance of may also be considered work-related if attitudes or beliefs about low back pain. work affect recovery. If an injured worker fears that returning to work will exacerbate an injury, for example, recovery may be Indeed, the New Zealand Guidelines for assessing acute low delayed. backpain, updated in 2003, likewise finds good agreement that beliefs, mood states, and behaviors, such as those identified by Recently, a number of studies have demonstrated the impact Waddell et al,29 consistently predict poor outcome.21 One of of psychologic and social factors on occupational musculoskele- these is the belief that work will make the pain worse, evidence tal injuries. What has emerged is evidence that these factors have for which is so compelling that investigators have labeled this as much, and in some cases more, predictive value as physical risk factor one of the “yellow flags.” Yellow flags are defined as and environmental factors. factors that may increase the risk of developing or perpetuating long-term disability and work loss associated with low back pain. The main findings from recent critical review articles and cur- The New Zealand group recommends their assessment as early as rent studies are discussed below. Most of the articles refer to non- 2 weeks after injury. specific disorders of the spine; research has centered on these conditions because they account for most of the associated costs Occupational health guidelines for low back pain were pub- and suffering caused by the disorder. Some reviews include the lished in 2001. Although the effect size was small, strong evi- upper extremities as well. To date, no high quality study of work- dence was found that psychosocial factors are associated with the related lower extremity injuries that explores psychosocial factors risk of onset of low back pain in symptom-free workers. In the has been found. workplace, psychosocial factors are stronger predictors of care seeking and disability than clinical symptoms or physical work demands. Specifically, negative work perceptions were risk fac- tors for symptom reports, health care utilization, and work loss. The most important factors associated with delayed recovery and response to treatment were beliefs that work caused the pain and expectations about recovery.29 A comparison of international occupational low back pain guidelines shows them in complete agreement that the early assessment of these risk factors was important if the patient exceeds the expected recovery time.25 Other reviews of factors affecting low back pain outcomes show similar results but call for more stringent scientific investi- gation of the interrelationships. For example, Hoogendoorn et al10 reviewed the literature and found that low workplace social support and low job satisfaction were related to reports of

Chapter 2 ● What is the relationship between psychosocial factors and work-related MSDs 15 back pain, but most studies failed to adjust for psychosocial work they found an independent and robust relationship between characteristics and physical load at work. Crook et al4 deter- depressive symptoms and onset of a pain episode. They con- mined from their review that among other factors, psychologic cluded that depression is a strong and independent predictor for distress and work environment characteristics were important the onset of an episode of intense and/or disabling neck and low prognostic factors. They noted, however, that most studies failed back pain. to investigate interactions among psychologic, social, and physi- cal variables. WHAT IS THE RELATIONSHIP BETWEEN PSYCHOSOCIAL FACTORS AND Likewise, Pincus et al23 noted a lack of rigorous prospective WORK-RELATED MSDs? studies that evaluated the role of psychologic factors in the development of chronicity. Of 25 publications, only 6 met their There are a number of ways to explain the effect of psychosocial acceptability criteria for review. Depression, distress, and to a factors on MSDs. Perhaps the most straightforward explanation lesser extent somatization were the psychologic factors most is that workers who are depressed or dissatisfied at work simply strongly linked to low back pain disability. Catastrophizing by report more symptoms and disability than those who are content disabled workers as a coping strategy received weak support. The at their jobs. Distress in life and work may cause extra attention role of coping in conjunction with psychologic factors remains to be paid to symptoms, and therefore the signs are experienced unclear. as more troubling. This does not imply that workers are inten- tionally making more of symptoms than is warranted but that Although there are fewer studies on upper extremities than on they are more likely to notice symptoms that can distract them back pain, thus far similar factors seem to be related to both disor- from their daily routines. ders. In another review article, higher levels of perceived job demands and stress were the psychosocial factors most consistently Another popular explanation is that stress results in increased linked to upper extremity disorders. Epidemiologic reviews muscle tension, causing spasms and ischemia, a painful condi- revealed also that psychosocial factors unrelated to work, such as tion resulting from oxygen reduction and the release of pain- general worry/psychologic tension, depression/anxiety, general producing chemicals. Over time this pain may cause more distress, coping style, and response to pain, are likewise associated with leading to a chronic cycle of stress and pain. Some laboratory both back and upper extremity disorders.15 Systematic reviews studies have demonstrated an increase in muscle activation on this topic have not, however, been conducted. under stress,18 but findings have been inconsistent among patients with chronic pain. This may be because stress-induced muscle An interesting study by Ijzelenberg et al11 investigated tension has a role in the development of chronic pain, but once whether individual, work-related physical, and psychosocial risk patients are chronic their reactivity to stress becomes altered in factors involved in the occurrence of musculoskeletal complaints ways that are unclear at present. also determined musculoskeletal sickness absence. Using a cross- sectional study design and self-administered questionnaire, inves- The National Research Council and the Institute of Medicine tigators collected data on individual and work-related risk factors reviewed evidence for the relationship between psychologic and and the occurrence of musculoskeletal complaints and muscu- physical factors and concluded that data exist to support a direct loskeletal sickness absence among 373 employees of laundry-works role of the central nervous system.20 This may help to explain and dry-cleaning businesses. Results show that both work-related how psychologic processes such as attention and emotion influ- physical factors and psychosocial factors showed strong associa- ence pain and its tolerance, an explanation supported by studies tions with low back pain and upper extremity complaints. Work- showing that pain is linked to a tendency to somatize or focus related physical factors did not influence sickness absence, on symptoms.17 however, whereas psychosocial factors showed some associations with it. This finding supports the notion that illness behaviors Other theories mentioned in the review include physiologic such as work absence are modified by cognitive and affective changes that occur under stress and result eventually in muscu- responses to physical symptoms in neck and back disorders. loskeletal pain. Stress hormones may facilitate the transmission of painful stimuli; physically stressful work tasks are associated Linton15 systematically reviewed numerous studies of variable with an increased release of stress hormones and slower recovery methodologic quality that looked at the effects of various psy- times. One reason for this may be that stress hormones affect the chologic factors on neck and back pain and concluded that there health of muscles, tendons, and ligaments and impede inflam- is a clear link. Prospective studies indicate, furthermore, that psy- matory or immune responses. chologic variables are related to the onset of pain and to all of its phases: Certain factors such as distress, anxiety, cognitive func- Other factors such as behavior or life-style may moderate the tioning, and pain behavior were found to be significant at all effects of stress on muscle tension and biochemical changes. The stages from acute to chronic pain. Still, psychologic factors concept of “work style” has been proposed to define how indi- account for only a portion of the variance, underscoring the viduals interact with work demands.6 Originally proposed to importance of a multidimensional view. explain work-related upper extremity disorders, work style is defined as cognitive, behavioral, and physiologic components of Luo et al16 found a variety of factors, including general stress, the stress response expressed behaviorally in movement, posture, depression, and anxiety, to be associated with neck pain disability. and activity. Workers who respond to physical or psychologic Carroll et al3 confirmed the role of depression as a risk factor for workplace demands with a “high-risk” work style display excessive pain onset. They followed a population-based random sample of attempts to cope with stress and are susceptible to its negative adults and assessed them at 6 and 12 months. Controlling for demographic and socioeconomic factors, health status, comor- bid medical conditions, and injuries to the neck or low back,

16 Chapter 2 ● Psychosocial aspects of work-related musculoskeletal disorders consequences such as continuous arousal or reactivity. Similarly, Physicians should share with distressed patients information Marras et al18 found that characteristic ways of responding to the about treatment philosophies and their possible outcomes. environment were predictive of spinal loading. During a lifting At the very least, physicians should be aware of the treatment task, introverts tend to exhibit muscle coactivation and alter- philosophy of any facility to which they are referring patients, ation in movement patterns higher than those of extroverts. ideally one that adheres to current guidelines for the treatment Stress-related behaviors such as these have been linked to symp- of acute low back pain. tom severity, functional limitations, and work disability.9 The occupational health guidelines for the management of It is clear that the relationship between psychologic distress and low back pain summarize recommendations for physicians to physical responses is not simple. More research is needed to minimize the likelihood of chronic problems29: develop detailed models of the stress–pain relationship. Although ● Reassure the worker and provide adequate information about various pathways have been proposed, there is a need to under- stand causal pathways and interactions among stresses and the self-limiting nature and good prognosis of lower back pain. responses. Most likely there is no simple explanation of this rela- ● Advise the worker to continue normal activity and work or tionship, and these proposed pathways are not mutually exclusive but work together in an ongoing dynamic way to result in MSDs. to return to them as soon as possible, even if there is still It is possible also that the stress pathways discussed here have vari- some pain. able influences in different stages of pain. What began as pain ● Because most workers with lower back pain manage to return related to muscle tension, for instance, may over time develop to more or less normal duties quite rapidly, consider tempo- into neuropathic pain. In addition, more clarity is needed in defin- rary adaptations of work duties or hours only when necessary. ing outcome measures, because different factors may moderate ● If a worker fails to return to work within 2-12 weeks (different stress and physical complaints, sick absences, and disability. guidelines vary considerably about the time scale), refer him or her to a gradually increasing exercise program or multidis- PSYCHOSOCIAL INTERVENTIONS FOR MSDs ciplinary rehabilitation (exercises, education, reassurance, and pain management following behavioral principles). These Most individuals at risk for MSDs can be effectively managed by rehabilitation programs should be embedded in an occupa- their usual treatment provider, without the need for referral to a tional setting. psychologist. Appropriate information and advice from the physician can reduce anxiety and improve patient satisfaction Based on the conclusion that active care is superior to passive ther- with care. Every encounter with health care professionals and apy, treatment guidelines recommending active and goal-oriented medical management systems provides an opportunity for inter- physical therapy should be the first course of treatment. Research vention. Waddell28 astutely observed the powerful effects a has shown that the specific type of exercise prescribed is less physician can have on the patient. Information that is vague, important that the overall therapy philosophy. The objective incomplete, or incomprehensible to a lay person can render the should be to increase gradually the individual’s capacity for work patient confused, helpless, and afraid. In the mind of a patient, with a balanced strength, flexibility, and cardiovascular program. the diagnosis of a herniated disk, for example, can be a sentence to disability. Diagnostic film results are often held up as proof by Specific goal setting ensures that improvement is monitored, patients fearful of permanent disability. It is the health care whereas positive feedback rewards productive behavior. The easiest provider’s responsibility to explain the implications of any diag- program to implement early in treatment, a behavioral approach nosis in terms the patient can understand. In the case of disk her- pioneered by Fordyce,7 consists of time-contingent care in which niation, the patient needs to know that pain and disability are medication levels, exercise goals, and duration of treatment are often self-limiting and that many people with this condition preset and not determined by the patient. The idea is to provide maintain their usual life-styles. environmental contingencies that reinforce “well” behavior and ignore “pain” behavior. In a benchmark study, Fordyce7 com- Making sure that the patient has a realistic picture of the diag- pared this approach with traditional care in which patients were nosis and the prognosis is the first step. Patients who seem overly told to “let pain be your guide.” Subjects with back pain for less distressed can be questioned about other life circumstances. than 10 days were randomly assigned to one of the two proto- When the injury is work related, questions about work are partic- cols and were followed at 6 weeks and 9 to 12 months after treat- ularly important. Often, patients feel residual anger toward the ment. Although no differences were noted at 6 weeks, long-term employer, sometimes believing that it is the employer’s respon- follow-up showed significantly greater improvement in function- sibility to make sure they get well. Even if not responsible for ing for the treatment group. their injuries, patients who understand the normal course of low back pain and are encouraged to take responsibility for their Like MSDs, physical therapy may be viewed from a biopsy- recovery from the onset may be spared the ordeal of becoming a chosocial perspective. In contrast to passive modalities, an active compensation failure. approach shifts the responsibility to patients and provides tech- niques that they can use themselves to maintain fitness. As such, Primary health care professionals have a responsibility also to it increases the patient’s sense of control and may be instrumen- make appropriate referrals. Information given by the physician tal in changing negative beliefs about pain. With improvement, can be reinforced or contradicted by secondary caregivers. One the patient’s fears of returning to work should subside. of the major causes of distress, contradictory information can lead to endless doctor shopping and prolongation of recovery. After 2 weeks, patients who are not achieving weekly goals should be referred for a psychosocial evaluation. If warranted, they can begin a multidisciplinary program that addresses psy- chosocial issues. By their very nature these types of programs address physical, psychologic, and social concerns simultaneously.

Chapter 2 ● Evidence 17 A multidisciplinary approach in treating patients with chronic occupational MSDs in general are few. Nonetheless, in line pain is fairly standard,31 but the application of a biopsychosocial with international opinion, expert panels have recommended model in the acute and subacute phases of treatment as well is the use of multidisciplinary team management for episodes of gaining popularity. acute low back pain that is unresolved between 2 and 6 weeks, when patients risk chronic disability. EVIDENCE Evidence clearly shows that multidisciplinary teams or net- No randomized controlled studies have assessed the use of works are effective in managing chronic back pain,14,22,27 but their psychologic interventions alone for MSDs, and high-quality effectiveness for returning chronic patients to work remains studies that assess the effectiveness of multidisciplinary care for unclear. Some studies showed good outcomes for chronic low back pain patients who are highly motivated to return to work,1 but randomized trials have yet to be done. Onset of pain Two weeks after • Explain the biopsychosocial model • Address concerns again, old and new • Address concerns about pain, including • Address social factors including attitudes about work, the course, diagnosis, and prognosis as well as family and friends' responses to pain • Emphasize the patient's role in recovery • Reinforce patient's role in recovery again • Make appropriate referral if indicated • Make appropriate referral if indicated Four weeks after Seven weeks after • Address concerns again, old and new • Address concerns again • Address social factors again • Address social factors again • Reinforce patient's role in recovery again • Reinforce patient's role in recovery • Refer for multidisciplinary evaluation • Refer for multidisciplinary evaluation and possible treatment and treatment Figure 2.2 Application of the biopsychosocial model: general guidelines for primary health care providers.

18 Chapter 2 ● Psychosocial aspects of work-related musculoskeletal disorders Karjalainen et al12 conducted a Cochrane review of multidisci- 4. Crook J, Milner R, Schultz IZ, Stringer B: Determinants of occupational disability plinary programs for subacute low back pain patients. They con- following a low back injury: a critical review of the literature. J Occup Rehabil cluded that “there is moderate evidence of positive effectiveness 12(4):277-295, 2002. of multidisciplinary rehabilitation for subacute low back pain and that a workplace visit increases the effectiveness.” This endorsement 5. Engel GL: The need for a new medical model. Science 196:129-136, 1977. is guarded because the only two articles that fit their inclusion 6. Feuerstein M: Workstyle: definition, empirical support, and implications for criteria were only of moderate quality. However, more support for this treatment in the subacute phase was provided by van den prevention, evaluation, and rehabilitation of occupational upper-extremity disorders. Hout et al.26 In a randomized study, they showed that adding In SD Moon, SL Sauter, eds: Beyond biomechanics: psychosocial aspects of training in problem solving to a graded activity program musculoskeletal disorders in office work. Bristol, PA, 1996, Taylor & Francis, improved outcome. pp. 177-206. 7. Fordyce WE: Acute back pain: a control group comparison of behavioral vs. The Work Loss Data Institute considered but did not recom- traditional management methods, J Behav Med 9:127-140, 1986. mend multidisciplinary or cognitive-behavioral treatment for 8. Goleman D, Gurin D, eds: Mind body medicine: how to use your mind for better upper extremity pain due to lack of studies.32 New studies should health. New York, 1993, Consumer Reports Books. consider other MSDs as well. Future research would benefit also 9. Haufler AJ, Feuerstein M, Huang GD: Job stress, upper extremity pain and functional from a comprehensive multivariate causal model that would limitations in symptomatic computer users. Am J Ind Med 38(5):507-515, 2000. allow the assessment of interactions among psychologic, social, 10. Hoogendoorn WE, van Poppel MN, Bongers PM, Koes BW, Bouter LM: Systematic and physical variables. Adjustments for workload could be made review of psychosocial factors at work and private life as risk factors for back pain. when studying the effects of other factors; such a model would Spine 25(16):2114-2125, 2000. also permit an understanding of how these risk factors emerge 11. Ijzelenberg W, Molenaar D, Burdorf A: Different risk factors for musculoskeletal during the transition from acute to chronic problems. complaints and musculoskeletal sickness absence. Scand J Work Environ Health 30(1):56-63, 2004. SUMMARY 12. Karjalainen K, Malmivaara A, van Tulder M, et al: Multidisciplinary biopsychosocial rehabilitation for subacute low back pain among working age adults. Cochrane Traditional treatments for occupational MSDs have fallen short Database Syst Rev (3):CD002193, 2000. of expectations. The biopsychosocial model goes beyond physi- 13. Koes BW, van Tulder MW, Ostelo R, Burton K, Waddell G: Clinical guidelines for the cal factors to include psychologic and psychosocial elements that management of low back pain in primary care: international comparison. Spine affect the worker before and after an injury. Investigations into 26(22):2504-2513, 2001. the practical application of this model are most convincing. 14. Koopman FS, Edelaar M, Slikker R, Reynders K, van der Woude LH, Hoozemans MJ: Studies have shown that psychosocial factors are at least as impor- Effectiveness of a multidisciplinary occupational training program for chronic low tant as and often more important than physical factors in deter- back pain: a prospective cohort study. Am J Phys Med Rehabil 83(2):94-103, 2004. mining disability. Programs that have attempted to prevent 15. Linton SJ: A review of psychological risk factors in back and neck pain. Spine injury or chronicity associated with occupational low back pain 25(9):1148-1156, 2001. have flourished when they include a biopsychosocial framework. 16. Luo X, Edwards CL, Richardson W, Hey L: Relationships of clinical, psychologic, and individual factors with the functional status of neck pain patients. Value Health Though more research is needed, it is suggested that our best 7(1):61-69, 2004. option is for primary health care providers and all health care 17. Main CJ, Wood PL, Hollis S, Spanswick CC, Waddell G: The distress and risk practitioners to apply a biopsychosocial model in the same fash- assessment method: a simple patient classification to identify distress and evaluate ion for all MSDs that do not respond to traditional care (Fig. 2.2). the risk of poor outcome. Spine 17(1):42-52, 1992. Because we all share the burden of disability, we must all share in 18. Marras WS, Davis KG, Heaney CA, Maronitis AB, Allread WG: The influence of the solution. Much can be done by all medical professionals to psychosocial stress, gender, and personality on mechanical loading of the lumbar make this process easier and more successful for everyone. For spine. Spine 25(23):3045-3054, 2000. physicians, establishing a good rapport with the patient, giving 19. Melzack R, Wall P: Pain mechanisms: a new theory. Science 150:971-979, 1965. clear and intelligible information, encouraging an active approach 20. National Research Council and the Institute of Medicine: Musculoskeletal disorders to treatment, and making appropriate referrals at the first sign of and the workplace: low back and upper extremities. Panel on Musculoskeletal delayed recovery can go a long way toward reducing occupa- Disorders and the Workplace. Commission on Behavioral and Social Sciences and tional MSDs. Physical therapists can follow a behavior model of Education. Washington, DC, 2001, National Academy Press. active care that encourages patient responsibility for outcome 21. New Zealand Guidelines Group: New Zealand acute low back pain guide. Wellington, and reinforces function over pain. NZ, October 2004, Accident Compensation Corporation. 22. Patrick LE, Altmaier EM, Found EM: Long-term outcomes in multidisciplinary REFERENCES treatment of chronic low back pain: results of a 13-year follow-up. Spine 29(8):850-855, 2004. 1. Campello M: Physical and psychosocial predictors of work retention after a 23. Pincus T, Burton AK, Vogel S, Field AP: A systematic review of psychological factors as multidisciplinary rehabilitation program for non-specific low back pain patients. predictors of chronicity/disability in prospective cohorts of low back pain. Spine UMI Dissertation Services, 2002. 27(5):E109-E120, 2002. 24. Selye H: The stress of life. New York, 1956, McGraw-Hill. 2. Canon W: Bodily changes in pain, hunger, fear and rage, ed 2. New York, 1936, 25. Staal JB, Hlobil H, van Tulder MW, et al: Occupational health guidelines for the Appleton-Century-Crofts. management of low back pain: an international comparison. Occup Environ Med 60(9):617, 2003. 3. Carroll LJ, Cassidy JD, Cote P: Depression as a risk factor for onset of an episode of 26. van den Hout JH, Vlaeyen JW, Heuts PH, Zijlema JH, Wijnen JA: Secondary prevention troublesome neck and low back pain. Pain 107(1-2):134-139, 2004. of work-related disability in nonspecific low back pain: does problem-solving therapy help? A randomized clinical trial. Clin J Pain 19(2):87-96, 2003. 27. van Tulder MW, Ostelo RWJG, Vlaeyen JWS, Linton SJ, Morley SJ, Assendelft WJJ: Behavioural treatment for chronic low-back pain (Cochrane Review). In The Cochrane Library, Issue 3. Chichester, UK, 2004, John Wiley & Sons, Ltd. 28. Waddell G: Biopsychosocial analysis of low back pain. In M Nordin, TL Vischer, eds: Common low back pain: prevention of chronicity. London, 1992, Bailliere Tindall. 29. Waddell G, Burton AK: Occupational health guidelines for the management of low back pain at work: evidence review. Occup Med (2):124-135, 2001. 30. Waddell G, McIntosh A, Hutchinson A, Feder G, Lewis M: The clinical guidelines for the management of acute low back pain. In Low back pain evidence review. London, December 2001, Royal College of General Practitioners. 31. Weiser S, Cedraschi C: Psychosocial issues in the prevention of chronic low back pain: a literature review. In M Nordin, TL Vischer, eds: Common low back pain: prevention of chronicity. London, 1992, Bailliere Tindall. 32. Work Loss Data Institute: Disorders of the neck and upper back. Occup Environ Med 60:618-626, 2003.

3C H A P T E R Neck

3aC H A P T E R during a specific period of time; its measurement is useful to quantify the burden of a condition in a defined population. A Review of the 2. What demographic, socioeconomic, health- and work- Epidemiology of Neck Pain related, psychosocial, and societal factors are associated with in Workers: Prevalence, prevalent neck pain in workers? We refer to “associated fac- Incidence, and Risk Factors tors” to describe variables that are correlated with neck pain in cross-sectional studies. Although these cannot be used to Pierre Côté, Linda Carroll, J. David Cassidy, Mana Rezai, Vicki Kristman, draw causal inferences about its etiology, they are useful to and the Scientific Secretariat of the 2000-2010 Bone and Joint Decade describe the potential risk factors for neck pain. Task Force on Neck Pain and Its Associated Disorders* 3. What is the incidence of neck pain in workers? We use the term incidence to describe the average risk of developing In the past half century, the nature of work has rapidly changed. neck pain during a time period. A measure of incidence is Once dependent mainly on manufacturing and resource extrac- not only useful to describe the proportion of the workers tion, the economy of industrialized countries increasingly relies who will develop a new episode of neck pain during a on the service sector. This new reality has transformed the nature specific time period but also necessary to quantify the of work injuries and disability. The shift from manufacturing and etiologic contribution of risk factors. Incidence is meas- resource-based jobs to the service industry led to a sharp reduc- ured from cohort studies. tion in acute and fatal injuries but contributed to the current 4. What are the risk factors for neck pain in workers? Risk factors epidemic of musculoskeletal disorders (MSDs).45 Today MSDs are things that may contribute to on help predict neck pain. are the most burdensome ailments that affect the health and We have collected information on risk factors—variables that reduce the productivity of workers.56 increase or reduce risk—from case-control and cohort studies. Neck pain is one of the most prevalent MSDs in the working- METHODS age population48 and a leading source of disability and health care utilization among adults.18,19 During any 6-month period, 54% of We wrote this chapter in collaboration with the members of the adults suffer from neck pain and 4.6% experience significant scientific secretariat of the 2000-2010 Bone and Joint Decade activity limitations because of neck problems.18 Contrary to Task Force on Neck Pain and Its Associated Disorders. The Task popular belief, most adults with neck pain do not experience Force includes a multidisciplinary scientific secretariat of 12 clini- complete permanent resolution of their discomfort; in fact, it cians and methodologists representing rheumatology, orthopedic follows an episodic course marked by periods of remission and surgery, physical and rehabilitative medicine, neurology, physio- exacerbation.20 Each year between 14.6% and 18.0% of adults therapy, health psychology, chiropractic, epidemiology, clinical develop a new episode of neck pain,20,21,48 which causes a new epidemiology, and biostatistics. A main objective was to complete episode of disability in 600/100,000 adults annually.20 a systematic search and critical review of the scientific literature on neck pain and its associated disorders. The literature reviewed in The etiology of neck pain in workers is complex and multifac- this chapter was critically appraised by members of the Task Force torial. Its hypothesized risk factors include demographics, psy- using a standardized critical appraisal methodology. chosocial and work-related issues, and poor general health.5,7,16,17,20 Although the etiologic contribution of work receives increasing Definition of neck pain attention, most studies remain cross-sectional and can be used only to hypothesize about the causes of neck pain.5,7 Understanding We define neck pain as soft tissue disorders of the cervical its etiology in workers is necessary to develop, test, and imple- spine/shoulder area, including disk lesions and radiculopathy ment prevention strategies to reduce its burden. that are experienced by workers. Our definition does not assume a causal link between work and neck pain. It includes neck PURPOSE pain that is caused or aggravated by work, that which interferes with work, and symptomatic neck pain that does not result in disability. The purpose of this chapter is to review and appraise the recent lit- Literature search erature on the epidemiology of neck pain in workers. Specifically, we review the literature to answer the following questions: Having developed our literature search strategy in collaboration with a library scientist (see Acknowledgments), we applied it to 1. What is the prevalence of neck pain in workers? We define two electronic databases: MEDLINE and Embase ( January 1997 prevalence as the proportion of workers with neck pain to October 2003). For the purpose of this chapter, we used a restricted portion of the exhaustive search undertaken by the *Eugene J. Carragee, Jaime Guzman, Scott Haldeman, Sheilah Hogg-Johnson, Lena Task Force. Our search was based on specific key words (neck pain, Holm, Eric Hurwitz, Margareta Nordin, Paul M. Peloso, and Gabrielle van der Velde. neck injury, pain, injury, epidemiology, incidence, prevalence,

26 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers workplace, worker, work, meta-analysis) and text words (neck to assist reviewers in understanding the impact of bias on the pain, cervical pain, neck ache, whiplash, pain, ache, sore, stiff, study results and to abstract information. discomfort, neck, occupational, work-related, review, literature synthesis, literature review). All search terms were limited to the Rotating pairs of scientific secretariat members independently English and French languages and adult subjects. We excluded conducted these in-depth reviews of each article, and the remain- duplicate articles. ing members read the studies. The two reviewers presented their critical appraisal to a larger group of members for discussion of Article selection the scientific merit of the study. The members decided by con- sensus whether a study was scientifically admissible or inadmis- Each citation found in the search was screened by rotating sible, the latter being one in which methodologic flaws were groups of three Task Force scientific secretariat members to assess judged to have fatally biased the results and therefore made the its relevance to the epidemiology of neck pain in workers. For paper’s conclusions scientifically invalid. the purpose of this review, articles were deemed relevant if they passed the following inclusion and exclusion criteria: Data synthesis Inclusion criteria Each research question was answered by a best-evidence synthesis,52,53 which consists of assembling and qualitatively 1. Studies that report results relevant to neck pain in workers synthesizing the best evidence relevant to it. We described the with or without its associated disorders, which might include strength of evidence for risk factors by adapting a methodology radiating arm pain, upper thoracic pain, headache, and tem- that has been used to rank studies on breast cancer, whiplash poromandibular joint disorder pain; injuries, and mild traumatic brain injuries.4,13,19 The methodol- ogy distinguishes among three phases of studies. Phase I studies 2. Studies that examine the prevalence, incidence, risk, are hypothesis-generating investigations that descriptively and/or prevention of neck pain with or without its associ- explore associations between potential risk factors and disease ated disorders; outcomes. Phase II studies are extensive exploratory analyses that focus on particular sets of risk factors or attempt to discover 3. Studies of the risk or prevention of neck pain that include which ones predict the development of neck pain. Phase III stud- at least 20 human subjects at risk for neck pain with or ies are large confirmatory studies of explicit prestated hypotheses without its associated disorders. that allow for focused examinations of the strength, direction, and independence of the proposed relationship between a risk Exclusion criteria factor and the development of neck pain. 1. Studies that do not include human subjects with neck pain We present the prevalence and incidence of neck pain with with or without its associated disorders; 95% confidence intervals. Where these figures were not reported in the article but sufficient information was provided, we com- 2. Studies of neck pain due to fracture, dislocation, tumor, puted these estimates and report our computations. We describe skin lesions, throat disorders, inflammatory disorders, the cross-sectional associations between various risk factors and cervical myelopathy, and spinal cord injury; neck pain as positive or negative. We report odds ratios, relative risks or hazard rate ratios, and 95% confidence intervals to 3. Studies concerning radiating arm pain, upper thoracic describe associations between risk factors and neck pain obtained pain, headaches, temporomandibular joint disorder, and from case-control or cohort studies. When these estimates did thoracic outlet syndrome not associated with neck pain; not appear in the paper, they were computed from sufficient data. Confidence intervals for prevalence and incidence rates 4. Opinion papers, letters to the editor, traditional narrative were calculated using the standard confidence interval formula reviews, and papers without scientific data; around proportions described by Kuzma.37 All other analyses were performed using SAS.49 5. Single-case reports of workers with neck pain with or with- out its associated disorders; RESULTS 6. Studies using cadavers or nonhuman subjects, such as crash test dummies, animals, or laboratory simulations; 7. Biomechanical studies without human subjects with neck pain. Critical review of the literature Selection and critical appraisal of articles All articles meeting the inclusion and exclusion criteria were The results from the literature search and selection of articles are critically reviewed using standard criteria and computerized crit- summarized in Figure 3a.1. Overall, we retrieved 162 publica- ical review forms. Modified for the purpose of the 2000-2010 tions, including 152 articles from the literature search and 10 Task Force, the forms were similar to the ones used by the from our files. Of the 159 nonduplicate articles, 45 met the Quebec Task Force on Whiplash-Associated Disorders and the inclusion criteria and were reviewed. Fifteen articles were Best Evidence Synthesis on Traumatic Mild Injury (Appendix excluded by the scientific secretariat after they were appraised as 3a.1).12,55 These criteria prompt the reviewers to appraise the scientifically inadmissible.10,11,15,26,31-34,36,40,41,43,61,63,64 methodologic merit of a study by focusing on selection bias, information bias, and confounding. Rather than provide a rating Of the 30 accepted articles, 18 were cross-sectional studies scale for determining scientific acceptability, the forms are used that reported on the prevalence and factors associated with neck pain,1-3,22-25,28-30,39,42,44,46,47,58-60 and 10 reported on its incidence

Chapter 3a ● Results 27 1. Literature search Medline Embase Personal files Librarian 109 43 10 162 articles 2. Screening 159 non-duplicate articles 45 relevant articles Scientific Inclusion criteria 114 irrelevant secretariat 3. Review 15 rejected Scientific Critical appraisal secretariat 30 articles accepted and reviewed Figure 3a.1 Results of literature search and selection of articles included in the review. and risk factors.6,8,27,35,38,50,51,54,57,62 Two studies reported cross- Factors associated with neck pain sectional and prospective results.9,14 All four studies reported that the prevalence of neck pain increases Neck pain in general working populations with age and is more common in women (Table 3a.1).9,14,23,46 In Sweden, self-employed males had a higher prevalence of Prevalence neck/shoulder symptoms in the past month than other workers.9 We accepted four cross-sectional studies that described the In the United Kingdom, workers with musculoskeletal prevalence of neck pain in samples of workers from the general comorbidities were more likely to report neck pain. Palmer et al46 population, in this case in Europe.9,14,23,46 The source population reported that the prevalence of neck pain in the preceding week for these studies varied from registers of general practitioners was higher in those with shoulder, elbow, wrist-hand, hip, and in the United Kingdom to age cohorts of workers under the knee pain. Similarly, neck pain was more common in workers supervision of occupational physicians in France, workers who who had experienced troublesome low back pain in the preced- participated in periodical health surveys in the Netherlands, and ing year and in those who suffered from frequent headaches, 34-year old workers in Sweden. Overall, these studies suggest fatigue, or stress.9,46 Barnekow-Bergkvist et al9 found that work- that neck pain is highly prevalent and a common source of ers with better flexibility of the cervical spine (rotation) reported activity limitations. fewer neck/shoulder symptoms in the preceding month. In the United Kingdom, 34% of workers reported that they Three studies reported that ergonomic, physical, and psychoso- experienced neck pain in the preceding year (Table 3a.1).46 cial factors at work are associated with neck pain (Table 3a.1). Two During the same period, 11% of workers limited their activities studies reported on its association with lifting. In Sweden, the because of neck pain. Overall, 20% of surveyed workers had prevalence of neck symptoms in the preceding month was lower experienced neck pain in the preceding week. In Sweden, the in workers who performed repetitive or heavy lifting,9 whereas in 1-year prevalence of neck pain was 44% and 61% in 34-year-old contrast, Palmer et al46 reported that lifting increased it. Among men and women, respectively (Table 3a.1).9 In this cohort, 26% U.K. workers, the prevalence of neck pain was found to be higher of men and 43% of women reported symptoms that lasted more also in workers who reached overhead for more than 1 hour per than 1 month. In France during 1990, the prevalence of chronic day and in those exposed to hand and arm vibration.46 Using neck/shoulder pain lasting for at least 6 months and associated a keyboard for 4 hours or more in an average working day was with functional limitations was 7.8% in men and 14.8% in positively associated with neck pain in women but not in men.47 women (Table 3a.1).14 In 1995 this same study reported that the A positive association between monotonous work, high decision prevalence of chronic neck/shoulder pain had increased slightly latitude, and prevalent neck/shoulder symptoms was found to 9.5% in men and 17.6% in women. among female workers in the Swedish sample.9 Finally, Dutch adult workers who reported a combination of mentally and physically demanding work and those performing heavily

Table 3a.1 Studies of prevalence and factors associated with neck pain in workers 28 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Study Source Inclusion/exclusion Case Prevalence (95% CI) Associated factors (95% CI) population criteria definitions Barnekow- Swedish student Inclusion: 34-year-old subjects Neck-shoulder pain 1-year prevalence: Men: aged 16 years in 1992 who worked at least measured with Men: 44% (36-52) Lifting (−); self-employed (+); worrying (+); Bergkvist in 1974 16 hours/week at time of survey Standardized Nordic Women: 61% (51-71) et al, 19989 Questionnaire Prevalence of symptoms neck lateral rotation (−) Exclusion: maternity/paternity leave, Women: Cassou et al, Workers under the students, long-term sick leave Chronic neck pain: > 1/month: Monotonous work posture (+); lifting (−); 200214 supervision of occupational self-report of neck/ Men: 26% (19-33) physicians from France Sample size = 238 shoulder pain lasting Women: 43% (33-54) high decision latitude (+); hand grip in 1990 and 1995 at least 6 months Point prevalence: performance (−); neck lateral rotation (−) Inclusion: workers born in 1938, with functional Men: Age (+); women (+) Palmer et al, Registers of 163 general 1943, 1948, or 1953 who limitations confirmed 1990: 7.8% (7.3-8.3) 200146 practitioners in 34 general underwent an annual medical by clinical examination 1995: 9.5% (8.9-10.1) Men: practices in England, examination Women: Age (+); shoulder pain (+); elbow pain (+); Scotland, and Wales and Neck pain measured 1990: 14.8% (14.0-15.6) members of the British Exclusion: Individuals not with Standardized 1995: 17.6% (16.7-18.5) wrist-hand pain (+); hip pain (+); knee pain (+); armed services in employed in 1995 Nordic Questionnaire 1-year prevalence: 34% (33-35) low back pain (+); headaches (+); tiredness/ 1997-1998 1-year prevalence of neck stress (+); working overhead (+); heavy Sample size = 16,950 Regular pain or lifting (+); use of keyboard for > 4 hours/day De Zwart et al, Dutch workers who Inclusion: individuals stiffness in the neck pain interfering with normal (−); hand-arm vibration (+) 199723 participated in the activities: 11% (10-12) Women: periodical occupational aged 16-64 years 1-week prevalence: 20% (19-21) Age (+); shoulder pain (+); elbow pain (+); health surveys between Sample size = 12,907 wrist-hand pain (+); hip pain (+); knee 1982 and 1993 Not applicable pain (+); low back pain (+); headaches (+); Inclusion: active workers tiredness/stress (+); working overhead (+); aged 16-64 years use of keyboard for > 4 hours/day (−); heavy lifting (+); hand-arm vibration (+) Sample size = 44,486 Prevalence varies with occupation Prevalence is higher in women than men Men: Age (+); mentally and physically demanding work (+); heavy physically demanding work (+) Women: Age (+); mentally and physically demanding work (+); heavy physically demanding work (+) (−), negative association; (+), positive association; CI, confidence interval.

Chapter 3a ● Results 29 demanding work were more likely to report regular pain or stiff- psychosocial factors, comorbidities, and health risk behaviors ness in the neck.23 (Table 3a.2). In their large cohort, Cassou et al14 found that increasing age, repetitive work, and high job demands were asso- Incidence ciated with chronic neck pain. Moreover, workers with depressive symptoms and those suffering from MSDs in the preceding year Three cohort studies, two conducted in France and one in the were more likely to develop chronic neck pain. United States, provided information about the incidence of neck pain in samples of workers from the general population. The Neck pain in specific occupational groups studied populations included workers under supervision of French occupational physicians/medical departments in the early Prevalence 1990s14 and those covered by the Washington State Department of Labor and Industries state fund (including employees of self- Twelve cross-sectional studies described the prevalence and insured employers) between 1990 and 1998.51 These studies sug- factors associated with neck pain in samples of health care gested that a significant proportion of the working population workers, including dental personnel, physical therapists and experiences a new episode of neck pain each year. nurses,1-3,22,24,39,58-60 music and nursery school teachers,25,44 and blue collar workers (sewage workers, spinning industry, forestry, These cohort studies offered insights about the “high-risk” and scaffolding).28-30,42 The available evidence suggests that the groups for the development of an episode of neck pain. In prevalence of neck pain varies across occupations and tends to France the 6-month incidence of neck pain among workers be higher for health care workers (Table 3a.3). with no prior history was 17.4%, whereas it was 44.4% in those who reported a past history of neck disorders.38 Between 1990 Among these workers, the 1-year prevalence of neck pain varies and 1995, 12.5% female French workers and 7.3% of male French from 45.8% in nurses to 47.6% in physical therapists, 64% in workers developed chronic neck pain (Table 3a.2).14 In dentists, and 72% in dental hygienists (Table 3a.3).1,2,22,39,58-60 In Washington State, 40.1 per 10,000 full-time equivalent workers Sweden, it is estimated that 50% of dental hygienists experience covered by the state fund developed neck pain.51 However, the neck pain in the preceding week, whereas in Norway, 22.6% of reported incidence was only 5.1 per 10,000 full-time equivalent nursing aides report that they experienced intense neck pain in employees in self-insured firms. the preceding 2 weeks.2,24 The prevalence of chronic neck pain is 9% among the nursing staff of Greek hospitals.3 In Sweden, most Risk factors for neck pain music teachers (59%) experience neck pain every year, and 30% experience an episode each week.25 Similarly, one fourth to one Four phase II studies reported on factors associated with the third of Japanese nursery school teachers report neck pain each development of neck pain in general working populations. Three month.44 Interestingly, except for sewage workers the prevalence Swedish studies described risk factors for developing neck pain of neck pain is lower in industrial, forestry, and construction in the past year, and one French study reported on the risk fac- workers than in health care workers or teachers. In Austrian tors for chronic neck pain. sewage workers, the annual prevalence of neck pain is 52.4%, and 20.9% of workers suffer from neck pain each day.28 In a Swedish cohort of 16-year-old student-workers, bench press performance was negatively associated with the development of Factors associated with neck pain neck/shoulder pain 18 years later (Table 3a.2).9 In other words, workers who performed better while bench pressing when 16 years In health care workers, the presence of neck pain was associated of age in 1974 were protected against developing neck/shoulder with demographic, ergonomic, and workplace factors. Among pain in 1992. In another study, Fredriksson et al27 found that the nurses, neck pain was more common in older workers and in those development of neck/shoulder pain was associated with physical with moderate/high physical or psychologic demands.39,58-60 Nurses and psychosocial factors at work, specifically high physical work and nursing aides who worked long hours reported neck pain load for men and frequent hand movement for women. The more frequently.24,39,58-60 Although neck pain was more common authors found, moreover, that physical and psychosocial risk fac- in those who worked under strenuous back postures, it was less tors frequently interact and modify their effects on the incidence so in those who used mechanical lifts or received ergonomic of neck/shoulder pain in workers. A third Swedish study found training.3,24,39,58-60 In physical therapists, neck pain was positively that both physical and psychosocial risk factors contribute to associated with the pace of work and the type of practice.22 workers seeking health care for neck/shoulder pain.57 A previous episode of neck pain that persisted for more than 3 months and The association between neck pain and physical/psychologic a nonfixed salary strongly predicted seeking health care for this variables was observed also in Japanese nursery school teachers, symptom. The risk factors as well as the strength and the direction among whom it was more common in those with poor supervi- of association varied between men and women, suggesting that sor support and in those caring for young children.44 Similarly, the etiology has a gender-specific component.57 For men, exposure male Swedish music teachers with low social support and women to psychosocial factors such as high job demands or high time teachers with high psychologic demands were more likely to pressure reduced the risk of seeking care (Table 3a.2). In women, report neck pain.25 Neck pain in blue collar workers was associ- however, exposure to other psychosocial risk factors such as high ated with age, physical work, and psychologic demands.28,30 degree of hindrance with work increased the risk of seeking care for neck/shoulder pain. Tornqvist et al57 also reported that Incidence of neck pain workers with multiple risk factors were more likely to seek care. Five cohort studies describe the incidence of neck pain in specific In France, the risk of developing chronic neck pain and dis- occupational groups, including industrial/service companies,6,8 ability was associated with sociodemographic variables, workplace

Table 3a.2 Studies of incidence and risk factors associated with neck pain in workers 30 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Study, design, Source Inclusion/exclusion Case definitions Incidence (95% CI) Risk factors (95% CI) and phase population criteria Barnekow- 16-year-old Inclusion: 34-year-old subjects in Neck-shoulder pain Men: Bergkvist et al, measured with Bench press performance at 16 years old 19989 Swedish student- 1992 who worked at least Standardized Nordic Questionnaire (OR = 0.30 (0.09-0.85)) Phase II workers in 1974 16 hours/week at time of survey Chronic neck pain: Men: Cassou et al, Exclusion: maternity/paternity leave, self-report of neck/ Year of birth 1938 vs. 1953 OR = 2.0 (1.6, 2.6); year of 200214 shoulder pain lasting students, long-term sick leave at least 6 months with birth 1945 vs. 1953 OR = 1.5 (1.2, 1.9); repetitive work Cohort study functional limitations before 1990 OR = 1.3 (1.0, 1.7); awkward work OR = 1.3 Phase II Sample size = 238 confirmed by clinical (1.1, 1.7); high job demand OR = 1.2 (1.0, 1.4); examination depressive symptoms OR = 1.3 (1.0, 1.8); MSD in past Workers under Inclusion: workers born in 1938, 1943, 5-year incidence: year OR = 1.5 (1.3, 1.8); sporting activities OR = 0.8 Men: 7.3% (6.8, 7.8) (0.7, 0.9); smoking OR = 1.2 (1.0, 1.4) the supervision 1948, or 1953 who underwent Women: 12.5% Women: Year of birth 1938 vs. 1953 OR = 1.5 (1.2, 2.0); year of of occupational an annual medical examination (11.7, 13.3) birth 1945 vs. 1953 OR = 1.6 (1.3, 2.0); year of birth 1948 vs. 1953 OR = 1.2 (1.0, 1.5); repetitive work in physicians from Exclusion: individuals not employed 1990 OR = 1.3 (1.0, 1.6); repetitive work before 1990 OR = 1.2 (1.0, 1.5); high job demand OR = 1.2 France in 1990 in 1995 (1.0, 1.4); depressive symptoms OR = 1.5 (1.2, 1.9); MSD in past year OR = 1.7 (1.5, 2.0) and 1995 Sample size = 16,950 Risk factors in the past 5 years. Men: High physical workload index OR = 2.2 (1.1, 4.6); Fredriksson Individuals from Inclusion: individuals living in Cases: neck/shoulder Not applicable frequent hand/finger work during et al, 200027 Stockholm, Sweden in 1993, without a leisure time OR = 2.1 (1.0, 4.5) Sweden who diagnosis of MSD in 1969 pain defined by a Significant interactions: Nested case- participated in and below age 59 in 1993 Work with vibrating tools and frequent hand movements; control study the REBUS-69 questionnaire reporting frequent hand work during leisure time and high study in 1969 Sample size = 484 perceived workload; frequent hand work during leisure Phase II consultation with a time and work with vibrating tools; precision work during leisure time and high physical workload physician, physiotherapist, Women: Frequent hand movements at work OR = 1.5 (1.0, 2.3) Significant interactions: chiropractor, osteopath, or High perceived workload and low influence over work condition; high perceived workload outside work and low influence over other medical professional work conditions; frequent hand work during leisure time and high perceived workload; frequent hand work during leisure or sick leave for more time and high physical workload; precision work during leisure time and few possibilities of development; precision work during than 7 consecutive days leisure time and low influence over work conditions between 1970 and 1992 Neck/shoulder pain defined as pain lasting at least 7 consecutive days during the last year Controls: Individual matched on age, gender, and index year

Leclerc Hospital, Inclusion: at least 1 year Neck disorder (ache, pain, Among 311 workers Not applicable et al, 199938 warehouse, office, of employment discomfort) in the past and airport 6 months defined by without prior neck pain: Cohort study workers recruited Exclusion: workers with sick leave questionnaire and body from medical for more than 3 months in the diagram 6-month incidence: 17.4% Silverstein departments in previous year, pregnancy, temporary et al, 200251 France in 1991 work contract, or retirement within Neck disorder for (13.3-22.0) the following 12 months more than 30 days Cohort study Workers covered by 6-month incidence of disorders the Washington Sample size = 511 Accepted claims for Tornqvist et al, State Department nontraumatic neck that lasted > 30 days: 1.3% 200157 of Labor and Inclusion: accepted State Fund soft tissue disorders Industries State claims for specific diagnostic or with the neck as the (0.0-3.3) Case-control Fund and treatment codes. Of the self-insured primary site of injury study employees of claims, only those with more than Among 27 workers with a prior self-insured 3 days of lost time were included. Phase II employers history neck pain: between 1990 Sample size = 392,925 and 1998 6-month incidence: 44.4% Inclusion: workers aged 20-59 years Workers in the who worked more than 17 hours/ (25.5-64.7) municipality of week and had worked at least Norrtälje, 2 months during the past year 6-month incidence of disorders Sweden from Sept 1, 1994 to Exclusion: seeking health care that lasted > 30 days: 11.1% June 30, 1997 for neck, shoulder, or low back disorders during the 6 months (2.4-29.2) before enrollment Cumulative incidence Not applicable Sample size = 1903 (per 10,000 FTE): State fund: 40.1 (39.6, 40.6) Self-insured employers: 5.1 (4.9, 5.3) Cases were those who Not applicable Men: Suffered from neck or shoulder symptoms > 3 months sought treatment for earlier in life: RR = 4.4 (2.7-7.1); work with vibrating neck-shoulder disorders tools: RR = 1.6 (1.0-2.3); nonfixed salary RR = 1.9 (1.1-3.1); high demands RR = 0.7 (0.4-1.0); high time Controls were randomly pressure RR = 0.5 (0.3-1.0); high creativity/low routine work profile RR = 0.6 (0.4-1.0); high quantitative demands from the population register RR = 0.2 (0.1-0.9); low demands in relation to competence RR = 1.5 (1.0-2.4) Control group = 1511 Women: Suffered from neck or shoulder symptoms > 3 months earlier in life: RR = 4.1 (3.0-5.7); repetitive hand/finger Chapter 3a ● Results movements RR = 1.6 (1.2-2.2); nonfixed salary RR = 2.0 (1.0-4.2); nightshift/shift work including night work RR = 1.3 (1.0-1.8); long working hours RR = 0.7 (0.5-0.9); solitary work RR = 1.8 (1.2-2.9); job strain RR = 1.4 (1.1-2.0); high degree of hindrances at work RR = 1.4 (1.0-1.9) CI, confidence interval; MSD, musculoskeletal disorder; OR, odds ration; RR, relative risk. 31

Table 3a.3 Studies of prevalence and factors associated with neck pain in specific occupations 32 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Inclusion/exclusion Study Source population criteria Case definitions Prevalence (95% CI) Associated factors (95% CI) ˚Akesson et al, Dental personnel (dentists, Exclusion: personnel Neck pain measured 1-week prevalence: Not applicable 19991 dental hygienists, and on leave (other than with the Standardized Hygienists: 50% (33, 67); male dentists: 31% dental assistants) from sick leave) and Nordic Questionnaire Strenuous back posture (+); moderate/bad ˚Akesson et al, the public health specialist dentists (16, 45); female dentists: 27% (12, 42); perceived general health (+) 20002 services in the county general practice dental assistants: 17% (10, 24); of Blekinge, Sweden Sample size = 244 specialist dental assistants: 16% (4, 27) Male (+), private practice (+), sports Alexopoulos 1-year prevalence: physical therapy (+), manual techniques (+), et al, 20033 Nursing staff from six Inclusion: at least one Neck pain measured Hygienists: 72% (56, 87); male dentists: 64% sedentary work (+), repetitive tasks (+), large hospitals in Athens, year of experience in with the Standardized (49, 79); female dentists: 64% (47, 80); number of patients/day (+), few rest Cromie et al, Greece between 2000 current job Nordic Questionnaire general practice dental assistants: 46% (37, 55); periods (+), rest periods (+), work when 200022 and 2001 Chronic neck pain is specialist dental assistants: 34% (19, 49) injured (+), mobilization/manipulation (+), Sample size = 351 pain lasting > 3 months Prevalence at 5-year follow-up for females. other manual treatment (+) Eriksen, 1-week prevalence: 200324 Physical therapists Exclusion: therapists Neck pain measured Hygienists: 41%; dentists: 38%; dental Female (+), marital status/single (+), registered in the state not living in Australia with the Standardized assistants 31% working >36 hours/week (+), working in old of Victoria, Australia Nordic Questionnaire 1-year prevalence: people’s home (+) Sample size = 536 Hygienists: 73%; dentists: 54%; dental assistants 62% Norwegian nurse’s Inclusion: vocationally active Neck pain measured 1-year prevalence: 47% (42-52) aides in 1999 Prevalence of chronic neck pain: 9% (6-12) Exclusion: sick leave because with the Standardized Prevalence of sick leave: 5% (3-7) of illness or pregnancy Nordic Questionnaire 1-year prevalence: 47.6% (43.3-51.8) 2-week prevalence: Any pain: 53.5 % (52.3-54.7) Intense pain: 22.6% (21.6-23.6) Sample size = 6485

Fjellman- Music teachers in two Inclusion: all music teachers Neck pain measured 1-year prevalence: Men: Cumulative: 59% (52-66) Low social support (+); lifting (+); Wiklund counties of Sweden in Sample size = 208 with the Standardized Men: 52% (43-61) Women: 67% (57-77) playing guitar (+) et al, 200325 2000 Nordic Questionnaire 1-week prevalence: Women: Cumulative: 30% (24-36) High psychologic demands (+); teaching Friedrich Sewage workers from Inclusion: job involving Neck pain measured Men: 26% (18-34) heavy lifting or working with the Standardized Women: 36% (26-46) at 5-12 schools/week (+) et al, 200028 Vienna, Austria while bent over Nordic Questionnaire Lifetime prevalence: 67.5% (61.6-73.3) 1-year prevalence: 52.4% (46.0-58.5) Age (+) Gamperiene Workers from the Sample size = 255 Neck pain measured 1-week prevalence: 25.7% (20.3-31.3) and Stigum, spinning industry in Inclusion: production with the Standardized Point prevalence: 20.9 (15.8-26.0) Not applicable 199929 Lithuania in 1996 Nordic Questionnaire 1-year prevalence: 16.5% (12.4, 20.8) workers Age (+); physical work (+); low intellectual Hagen et al, Forestry workers Sample size = 363 Neck/shoulder pain for 1-year prevalence: 27.7% (21.9, 33.5) discretion (+); high psychologic demands (+) 199830 from Norway Inclusion: manual workers, at least 30 days in previous 1 year 1-year prevalence: 27% (22, 32) Not applicable Molano Workers from a scaffolding machine operators, and measured with the et al, 200142 company in the administrative workers Standardized Nordic 10 years of experience (+); caring for children Netherlands from June who worked for at least Questionnaire aged 0 (+); holding/lifting a child/material Ono et al, to September, 1998 10 months/year during exceeding 20 kg (+); poor supervisor 200244 the previous 5 years Neck pain measured support (+) Nursery school teachers Sample size = 835 with the Standardized Continued in a Japanese city, Inclusion: all workers Nordic Questionnaire in October, 1995 Sample size = 323 Exclusion: male workers, Neck/shoulder pain in Not applicable nurses, supervisors the previous month greater than grade 3/5 Sample size = 959 on a self-reported scale Chapter 3a ● Results 33

Table 3a.3 Studies of prevalence and factors associated with neck pain in specific occupations—cont’d 34 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Inclusion/exclusion Study Source population criteria Case definitions Prevalence (95% CI) Associated factors (95% CI) Palmer et al, Registers of 163 general Inclusion: workers aged Neck pain measured 1-week prevalence: Not applicable 200146 practitioners in 34 general 16-64 years employed in with Standardized Men: practices across nonmanual occupations Nordic Questionnaire Keyboard: 14.8% (12.3-17.4) Age (+); mechanical lifting devices (−); training Britain in 1997-1998 No keyboard: 13.8% (12.0-15.6) on adjustment of work station (−); training Exclusion: manual Neck pain measured Women: on recognizing workplace hazards (−); occupations with the Standardized Keyboard: 22.9% (20.3-25.4) working long hours (+); moderate and high Nordic Questionnaire No keyboard: 18.3% (16.3-20.3) physical demands (+); psychologic demands; Sample size = 4889 1-week prevalence of disability among dependents (+); working long hours (+) Case: neck pain lasting Trinkoff et al, Licensed registered Inclusion: at current job ≥ 1 week and ≥ 3/5 workers with neck pain: 200259 nurses from Illinois and for at least 1 year in intensity Men: New York state in Keyboard: 26% (18-34) Lipscomb 1999-2000 Exclusion: nurses with Symptoms: neck pain No keyboard: 33% (26-39) et al, 200239 non-work related injuries without full definition Women: within a 3-month period of a case Keyboard: 33% (27-39) Trinkoff et al, before onset of work- No keyboard: 37% (31-42) 200358 related neck pain 1-year prevalence: Cumulative: 45.8% (43.0-48.7) Trinkoff et al, Sample size = 1163 Case: 20.4% (18.1-22.8) 200360 Symptoms: 25.4% (22.9-27.9) (−), negative association; (+), positive association; CI, confidence interval.

Chapter 3a ● Relevance 35 municipal office workers,35 teaching hospital employees,50 female neck pain; specifically, we found its incidence to be higher for nurses from acute care hospitals,54 and workers from a large nurses and office workers.35,50,51 Studies that quantified the role forestry industry.62 Overall, the evidence suggests that the inci- of risk factors supported that neck pain is more common in dence of neck pain varies across occupational group, tending to workers with high quantitative job demands, low coworker sup- be higher for those working in hospitals and offices (Table 3a.4). port, repetitive work, nonfixed salary, increased sitting time, poor ergonomics, previous musculoskeletal pain, and depressive In England, female nurses working in acute care hospital set- symptoms.5,8,14,27,35,54,57 tings developed neck pain at a rate of 17% per year.54 In Canada, 40.5% of employees at a teaching hospital who were asympto- Although we identified several risk factors for neck pain, it is matic in 1996 reported to have experienced an episode of neck important to note that only one phase III study was designed pain in the preceding week when followed up a year later.50 In specifically to test their independence.5,8 This finding indicates Finnish municipal office workers, the annual incidence of neck that very few risk factors can be considered well established. pain with or without radiations was 34.4%. Approximately 13% Efforts should be devoted to designing phase III studies to had local neck pain only, and 14.4% had radiating pain without confirm the results of cross-sectional observations and phase II neck pain. The annual incidence of workers reporting both local (exploratory) studies. Moreover, future studies should pay and radiating neck pain was 6.7%.35 attention to interactions among individual, psychosocial, and workplace risk factors. The growing body of evidence supporting The incidence of neck pain was lower among workers in the role of psychosocial and workplace factors in the etiology industrial/service and forestry sectors. In workers from 34 indus- of neck pain must not be considered in isolation. Because it is trial/service companies located throughout the Netherlands, the very likely that the risk factors for the development of neck pain 3-year incidence of neck pain was 14.4%.6,8 The overall 3-year and disability vary across occupations, as demonstrated in our incidence among Finnish forestry workers was a similar 15.6%62: review, future studies should ensure that the risk profiles of vari- 9.2% developed mild neck pain, and 6.4% developed severe pain. ous occupations are explored. Risk factors for neck pain RELEVANCE Two phase II and one phase I study examined the risk factors Our review has important implications for prevention. First, for neck pain in specific occupational groups (Table 3a.4). intervention must target clearly modifiable risk factors. Second, Among Finnish municipal workers, females and those with the multifaceted etiology of neck pain highlights the importance poor keyboard position were more likely to develop neck pain.35 of designing preventive interventions focused on multiple rather In English female nurses, a new episode of neck pain was associ- than individual risk factors. Current research has not yet identi- ated with previous neck and low back pain and with various fied the “necessary” causes for neck pain and disability, and very physical aspects of care such as moving and transferring little is known about the interrelationships among risk factors. patients.54 Multimodal interventions targeting multiple modifiable risks such as workplace and psychosocial factors may thus prove more Only one phase III study supported the observation that promising than one-dimensional approaches targeting a specific physical and psychosocial factors contribute to the development one. Finally, the complex etiology of neck pain suggests that of neck pain.5,8 In their cohort study of industrial and service the roles of economic and legal factors, work organization, and workers from the Netherlands, Ariens et al5,8 demonstrated that health care access and delivery are important and must be high quantitative job demands, low coworker support, and pro- studied as well.14 longed sitting are independent risk factors for a new episode of neck pain. SUMMARY Neck pain is endemic in workers. Our best evidence synthesis ACKNOWLEDGMENTS demonstrates that it is highly prevalent and a common source of disability. By far, most neck pain in workers is nontraumatic. Its We are indebted to Emma Irvin and Stephen Greenhalgh, research librar- etiology is consequently multifaceted, and although occupa- ians, for their expertise and guidance with the literature search. tional factors may be important contributors to its development, The Bone and Joint Decade Task Force on Neck Pain and Its Associated they are neither sufficient nor necessary. This reality clearly Disorders is supported by a grant to the University of Alberta from the emphasizes that the development of an episode of neck pain National Chiropractic Mutual Insurance Company and the Canadian cannot be attributed entirely to the physical and psychosocial Chiropractic Protective Association, Jalan Pacific Inc., Länsförsäkringar environment of a workplace. Wasa, and the Insurance Bureau of Canada. This article was made possi- ble also through the financial support of the Workplace Safety and Our review of cross-sectional studies supports the view that neck Insurance Board of Ontario. Dr. Côté is supported by the Canadian pain is more common in various subgroups of workers. It is more Institutes of Health Research through a New Investigator Award and by prevalent among older workers, women, and those with muscu- the Institute for Work & Health by the Workplace Safety and Insurance loskeletal comorbidities.1,2,14,23,24,28,30,39,46,54,57-60 Moreover, it varies Board of Ontario. Dr. Carroll is supported by a Health Scholar Award significantly across occupations and is associated with ergonomic, from the Alberta Heritage Foundation for Medical Research. Dr. Cassidy physical, and psychosocial factors.3,5,8,9,14,22,23,25,27,30,35,39,44,46,54,57-60 is supported by an endowed research chair from the University Health Evidence from case-control and cohort studies confirms that cer- Network. Vicki Kristman is supported by a Doctoral Training Award from tain occupations pose higher risk of developing an episode of the Canadian Institutes of Health Research in partnership with the

Table 3a.4 Studies of incidence and risk factors associated with neck pain in specific occupations 36 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Study Source population Inclusion/exclusion criteria Case definitions Incidence (95% CI) Risk Factors (95% CI) Ariens et al, Workers from 34 Inclusion: workers who worked Neck pain measured by a modified 1-year incidence: High quantitative job demands RR = 2.14 20018 industrial/service at least 20 hours per week Standardized Nordic Questionnaire 5.7% (4.3, 7.2) (1.28, 3.58); low coworker support companies located and worked at their current RR = 2.43 (1.11, 5.29); sitting > 95% Ariens et al, throughout the job for at least 1 year Cases were defined as regular or 3-year incidence: of the time RR = 2.34 (1.05-5.21) 20015 Netherlands from prolonged neck pain, with episodes that 14.4% (12.2, 16.6) 1994 to 1997 Exclusion: workers who had another lasted for at least 1 day during the Female OR = 2.9 (1.3-2.7); poor keyboard Cohort study paid job for any substantial previous 12 months on at least 1-year incidence: position OR = 2.1 (1.0-4.5) Phase III Finnish municipal amount of time, received a work one of 3 follow-up measurements Local or radiating neck office workers in disability payment for neck pain Not applicable Korhonen 1998-1999 in the preceding year, and Incident cases were healthy subjects at pain: 34.4% et al, 200335 had prolonged neck pain during baseline (those who reported local or (12.5-41.3) Employees of a teaching the year before baseline radiating neck pain for less than 8 days Local and radiating neck Cohort study hospital in Hamilton, during the preceding 12 months) who pain: 6.7% (3.0-10.3) Phase II Ontario, Canada from Sample size = 977 reported at follow-up local or radiating 1995 to 1997 Inclusion: full-time neck pain for at least 8 days during the Incidence of an episode Shannon preceding 12 months. Question for local of neck pain in the et al, 200150 employees working on neck pain was: “Estimate the total number previous week: 40.5% video display units for of days you have had local neck pain (33.1, 47.8) Cohort study more than 4 hours/week (not radiation) during the preceding Sample size = 180 12 months: 0 days, 1-7 days, 8-30 days, >30 days but not daily, and daily.” Healthy Inclusion: workers still was defined as 0-7 days, and 8 or employed by the hospital more days defined incident neck pain. at the end of the study Neck pain measured by question: “In the Exclusion: physicians past week, how often have you suffered Sample size = 173 from neck and/or shoulder pain?”

Smedley Female nurses Inclusion: working at same job Neck pain measured with the Standardized 2-year incidence of neck Previous neck pain: et al, 200354 from two acute throughout the study Nordic Questionnaire pain: 34% (30-38) > 1 year before baseline HR = 1.6 (1.1-2.3); hospitals in the < 1 year before baseline HR = 2.8 (2.0-3.9) Cohort study south of England Exclusion: agency staff, student Radiating neck pain measured by a Cumulative 3-year Phase II nurses, and community nurses modified version of the Standardized incidence: Duration of previous neck pain: Nordic Questionnaire < 1 week HR = 1.7 (1.1-2.5); Sample size = 587 Mild pain: 9.2% 1-4 weeks HR = 2.3 (1.5-3.3); Mild pain: 8-30 days in previous year (7.9, 10.5) > 4 weeks HR = 2.6 (1.7-4.0) Viikari-Juntura Workers from a large Exclusion: workers with Severe pain: >20 days in previous year et al, 200162 forest industry rheumatoid arthritis and Severe pain: 6.4% Previous low back pain: enterprise in Finland part-time workers (5.3, 7.5) >1 year before baseline HR = 1.8 (1.2-2.7); Cohort study from 1992 to 1995 < 1 year before baseline HR = 1.9 (1.4-2.7). Phase II Sample size = 5180 Duration of previous low back pain: 1-4 weeks HR = 1.8 (1.2-2.7); > 4 weeks HR = 2.3 (1.6-3.3) Assisting patients to mobilize > 4 times/shift HR = 1.6 (1.1-2.3); moving patients around > 4 times/shift HR = 1.6 (1.1-2.4); transferring patients in/out bath HR = 1.4 (1.0-2.0); washing or dressing patients while on chair > 4 times/shift HR = 1.7 (1.1-2.8); washing or dressing patients while on bed > 4 times/shift HR = 1.6 (1.0-2.5) Not applicable CI, confidence interval; HR, hazard ratio; OR, odds ratio; RR, relative risk. Chapter 3a ● Relevance 37

38 Chapter 3a ● Prevalence, incidence, and risk factors of neck pain in workers Canadian Institute for the Relief of Pain and Disability (formally known 25. Fjellman-Wiklund A, Brulin C, Sundelin G: Physical and psychosocial work-related as the Physical Medicine Research Foundation) Woodbridge Grants and risk factors associated with neck-shoulder discomfort in male and female music Awards Program and by the Institute for Work & Health by the Workplace teachers. Med Prob Perform Art 18(1):33-41, 2003. Safety and Insurance Board of Ontario. 26. Fjellman-Wiklund A, Sundelin G: Musculoskeletal discomfort of music teachers: an REFERENCES eight-year perspective and psychosocial work factors. Intl J Occup Environ Health 4(2):89-98, 1998. 1. Åkesson I, Johnsson B, Rylander L, Moritz U, Skerfving S: Musculoskeletal disorders among female dental personnel—clinical examination and a 5-year follow-up study 27. Fredriksson K, Alfredsson L, Thorbjornsson CB, et al: Risk factors for neck and of symptoms. Intl Arch Occup Environ Health 72(6):395-403, 1999. shoulder disorders: a nested case-control study covering a 24-year period. Am J Indust Med 38(5):516-528, 2000. 2. Åkesson I, Schutz A, Horstmann V, Skerfving S, Moritz U: Musculoskeletal symptoms among dental personnel: lack of association with mercury and selenium status, 28. Friedrich M, Cermak T, Heiller I: Spinal troubles in sewage workers: epidemiological overweight and smoking. Swed Dent J 24(1-2):23-38, 2000. data and work disability due to low back pain. Intl Arch Occup Environ Health 73(4):245-254, 2000. 3. Alexopoulos EC, Burdorf A, Kalokerinou A: Risk factors for musculoskeletal disorders among nursing personnel in Greek hospitals. Intl Arch Occup Environ Health 29. Gamperiene M, Stigum H: Work related risk factors for musculoskeletal complaints in 76(4):289-294, 2003. the spinning industry in Lithuania. Occup Environ Med 56(6):411-416, 1999. 4. Altman DG, Lyman GH: Methodological challenges in the evaluation of prognostic 30. Hagen KB, Magnus P, Vetlesen K: Neck/shoulder and low-back disorders in the factors in breast cancer. Breast Cancer Res Treat 52(1-3):289-303, 1998. forestry industry: relationship to work tasks and perceived psychosocial job stress. Ergonomics 41(10):1510-1518, 1998. 5. Ariens GAM, Bongers PM, Douwes M, et al: Are neck flexion, neck rotation, and sitting at work risk factors for neck pain? Results of a prospective cohort study. 31. Hildebrandt VH, Bongers PM, van Dijk FJ, Kemper HC, Dul J: The influence of climatic Occup Environ Med 58(3):200-207, 2001. factors on non-specific back and neck-shoulder disease. Ergonomics 45(1):32-48, 2002. 6. Ariens GA, Bongers PM, Hoogendoorn WE, Houtman IL, van der Wal WG, van Mechelen W: High quantitative job demands and low coworker support as 32. Holte KA, Westgaard RH: Daytime trapezius muscle activity and shoulder-neck pain risk factors for neck pain: results of a prospective cohort study. Spine 26(17): of service workers with work stress and low biomechanical exposure. Am J Indust 1896-1901, 2001. Med 41(5):393-405, 2002. 7. Ariens GA, van Mechelen W, Bongers PM, Bouter LM, van der Wal WG: Physical risk 33. Inoue M, Harada N: Habitual smoking and musculoskeletal symptoms in Japanese factors for neck pain. Scand J Work Environ Health 26(1):7-19, 2000. blue-collar workers. J Occup Health 44(5):315-320, 2002. 8. Ariens GA, van Mechelen W, Bongers PM, Bouter LM, van der Wal WG: Psychosocial 34. Joshi TK, Menon KK, Kishore J: Musculoskeletal disorders in industrial workers of risk factors for neck pain: a systematic review. Am J Indust Med 39(2):180-193, Delhi. Intl J Occup Environ Health 7(3):217-221, 2001. 2001. 35. Korhonen T, Ketola R, Toivonen R, Luukkonen R, Hakkanen M, Viikari-Juntura E: Work 9. Barnekow-Bergkvist M, Hedberg GE, Janlert U, Jansson E: Determinants of related and individual predictors for incident neck pain among office employees self-reported neck-shoulder and low back symptoms in a general population. working with video display units. Occup Environ Med 60(7):475-482, 2003. Spine 23(2):235-243, 1998. 36. Krause N, Ragland DR, Fisher JM, Syme SL: Psychosocial job factors, physical 10. Best M, French G, Ciantar J, et al: Work-related musculoskeletal disorders in workload, and incidence of work-related spinal injury: a 5-year prospective study hairdressers. J Occup Health Safety Austr New Zeal 18(1):67-76, 2002. of urban transit operators. Spine 23(23):2507-2516, 1998. 11. Brulin C, Gerdle B, Granlund B, Hoog J, Knutson A, Sundelin G: Physical and 37. Kuzma JW. Basic statistics for the health sciences, ed 3. Mountain View, California, psychosocial work-related risk factors associated with musculoskeletal symptoms 1988, Mayfield Publishing Company. among home care personnel. Scand J Caring Sci 12(2):104-110, 1998. 38. Leclerc A, Niedhammer I, Landre MF, Ozguler A, Etore P, Pietri-Taleb F: One-year 12. Carroll LJ, Cassidy JD, Peloso PM, Garritty C, Giles Smith L, WHO Collaborating Centre predictive factors for various aspects of neck disorders. Spine 24(14):1455-1462, Task Force on Mild Traumatic Brain Injury: Systematic search and review procedures: 1999. results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med (43 Suppl):11-4, 2004. 39. Lipscomb JA, Trinkoff AM, Geiger Brown J, Brady B: Work-schedule characteristics and reported musculoskeletal disorders of registered nurses. Scand J Work Environ 13. Cassidy JD, Carroll LJ, Peloso PM, et al: Incidence, risk factors and prevention of mild Health 28(6):394-401, 2002. traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med (43 Suppl):28-60, 2004. 40. Massaccesi M, Pagnotta A, Soccetti A, Masali M, Masiero C, Greco F: Investigation of work-related disorders in truck drivers using RULA method. Appl Ergon 14. Cassou B, Derriennic F, Monfort C, Norton J, Touranchet A: Chronic neck and shoulder 34(4):303-307, 2003. pain, age, and working conditions: longitudinal results from a large random sample in France. Occup Environ Med 59(8):537-544, 2002. 41. Miranda H, Viikari-Juntura E, Martikainen R, Takala EP, Riihimaki H: Physical exercise and musculoskeletal pain among forest industry workers. Scand J Med Sci Sports 15. Chan OY, Ho SF: Study on musculoskeletal complaints involving the back, neck and 11(4):239-246, 2001. upper limbs. Singapore Med J 39(8):363-367, 1998. 42. Molano SM, Burdorf A, Elders LA: Factors associated with medical care-seeking due 16. Côté P: The epidemiology of neck pain. Adv Chiropr 4, 1997. to low-back pain in scaffolders. Am J Indust Med 40(3):275-281, 2001. 17. Côté P, Cassidy JD, Carroll L: The factors associated with neck pain and its related 43. Nelson NA, Silverstein BA: Workplace changes associated with a reduction in disability in the Saskatchewan population. Spine 25(9):1109-1117, 2000. musculoskeletal symptoms in office workers. Hum Fact 40(2):337-350, 1998. 18. Côté P, Cassidy JD, Carroll L: The Saskatchewan Health and Back Pain Survey. The 44. Ono Y, Imaeda T, Shimaoka M, et al: Associations of length of employment and prevalence of neck pain and related disability in Saskatchewan adults. Spine working conditions with neck, shoulder and arm pain among nursery school 23(15):1689-1698, 1998. teachers. Indust Health 40(2):149-158, 2002. 19. Côté P, Cassidy JD, Carroll L: The treatment of neck and low back pain: who seeks care? who goes where? Med Care 39(9):956-967, 2001. 45. Ostry A: From chainsaws to keyboards: injury and industrial disease in British 20. Côté P, Cassidy JD, Carroll LJ, Kristman V: The annual incidence and course of Columbia. In T Sullivan, ed: Injury and the new world of work. Vancouver, 2000, neck pain in the general population: a population-based cohort study. Pain UBC Press, pp. 27-45. 112(3):267-273, 2004. 21. Croft PR, Lewis M, Papageorgiou AC, et al: Risk factors for neck pain: A longitudinal 46. Palmer KT, Cooper C, Walker-Bone K, Syddall H, Coggon D: Use of keyboards study in the general population. Pain 93(3):317-325, 2001. and symptoms in the neck and arm: evidence from a national survey [comment]. 22. Cromie JE, Robertson VJ, Best MO: Work-related musculoskeletal disorders in Occup Med 51(6):392-395, 2001. physical therapists: prevalence, severity, risks, and responses. Phys Ther 80(4):336-351, 2000. 47. Palmer KT, Walker-Bone K, Griffin MJ, et al: Prevalence and occupational associations 23. De Zwart BCH, Broersen JPJ, Frings-Dresen MHW, Van Dijk FJH: Musculoskeletal of neck pain in the British population. Scand J Work Environ Health 27(1):49-56, complaints in the Netherlands in relation to age, gender and physically demanding 2001. work. Intl Arch Occup Environ Health 70(5):352-360, 1997. 24. Eriksen W: The prevalence of musculoskeletal pain in Norwegian nurses’ aides. 48. Picavet HS, Schouten JS: Musculoskeletal pain in the Netherlands: prevalences, Intl Arch Occup Environ Health 76(8):625-630, 2003. consequences and risk groups, the DMC(3)-study. Pain 102(1-2):167-178, 2003. 49. SAS/STAT Software: Changes and enhancements, Release 8.1. Cary, NC, 2000, SAS Publishing. 50. Shannon HS, Woodward CA, Cunningham CE, et al: Changes in general health and musculoskeletal outcomes in the workforce of a hospital undergoing rapid change: a longitudinal study. J Occup Health Psychol 6(1):3-14, 2001. 51. Silverstein B, Viikari-Juntura E, Kalat J: Use of a prevention index to identify industries at high risk for work-related musculoskeletal disorders of the neck, back, and upper extremity in Washington state, 1990-1998. Am J Indust Med 41(3):149-169, 2002. 52. Slavin RE: Best evidence synthesis: an alternative to meta-analytic and traditional reviews. Educat Res 15:5-11, 1986.

Chapter 3a ● References 39 53. Slavin RE: Best evidence synthesis: an intelligent alternative to meta-analysis. 59. Trinkoff AM, Lipscomb JA, Geiger Brown J, Brady B: Musculoskeletal problems of the J Clin Epidemiol 48(1):9-18, 1995. neck, shoulder, and back and functional consequences in nurses. Am J Indust Me 41(3):170-178, 2002. 54. Smedley J, Inskip H, Trevelyan F, Buckle P, Cooper C, Coggon D: Risk factors for incident neck and shoulder pain in hospital nurses. Occup Environ Med 60(11):864-869, 2003. 60. Trinkoff AM, Lipscomb JA, Geiger Brown J, Storr CL, Brady BA: Perceived physical demands and reported musculoskeletal problems in registered nurses. Am J Prevent 55. Spitzer WO, Skovron ML, Salmi LR, et al: Scientific monograph of the Quebec Task Med 24(3):270-275, 2003. Force on Whiplash-Associated Disorders: redefining “whiplash” and its management [see comment]. [Erratum appears in Spine 20(21):2372, 1995.] Spine (8 Suppl): 61. Vasseljen O, Holte KA, Westgaard RH: Shoulder and neck complaints in customer 1S-73S, 1990. relations: individual risk factors and perceived exposures at work. Ergonomics 44(4):355-372, 2001. 56. Stewart WF, Ricci JA, Chee E, Morganstein D, Lipton R: Lost productive time and cost due to common pain conditions in the US workforce [see comment]. JAMA 62. Viikari-Juntura E, Martikainen R, Luukkonen R, Mutanen P, Takala EP, Riihimaki H: 290(18):2443-2454, 2003. Longitudinal study on work related and individual risk factors affecting radiating neck pain. Occup Environ Med 58(5):345-352, 2001. 57. Tornqvist EW, Kilbom A, Vingard E, et al: The influence on seeking care because of neck and shoulder disorders from work-related exposures. Epidemiology 63. West DJ, Gardner D: Occupational injuries of physiotherapists in North and Central 12(5):537-545, 2001. Queensland. Austr J Physiother 47(3):179-186, 2001. 58. Trinkoff AM, Brady B, Nielsen K: Workplace prevention and musculoskeletal injuries 64. Wihlidal LM, Kumar S: An injury profile of practicing diagnostic medical in nurses. J Nurs Admin 33(3):153-158, 2003. sonographers in Alberta. Intl J Indust Ergon 19(3):205-216, 1997.

Appendix 3a.1 GENERAL METHODOLOGIC ISSUES 3. Same data collection for all cohorts USED FOR ALL STUDY DESIGN 4. Important baseline variables measured, valid, and reliable 5. All aspects of exposure measured (dose, level, duration) 1. Research question is clearly stated 6. Exposure adequately measured (previous, at entry, during 2. Source and target population identified and described 3. Inclusion criteria described and appropriate study) 4. Exclusion criteria described and appropriate 7. Regular follow-up periods 5. Number of excluded or refusals (before study) reported 8. Coexposures monitored 6. Withdrawals (during study) reported, explained, and 9. Duration of follow-up adequate 10. Outcome(s) defined and measurable reasonable 11. Outcome(s) valid 7. Withdrawals equal in groups 12. Blind assessment of outcome(s) 8. Statistical analyses appropriate 13. Analyses controls for confounding 9. Adjustment for important variables measured at entry CRITERIA FOR THE APPRAISAL into study OF THE METHODOLOGIC QUALITY 10. Results verifiable from raw data OF CASE-CONTROL STUDIES CRITERIA FOR THE APPRAISAL 1. Similar population sources for cases and controls (time, OF THE METHODOLOGIC QUALITY place, potential for exposure) OF CROSS-SECTIONAL STUDIES 2. Referral and sampling independent of exposure 1. Similar sampling procedures for all subjects 3. Random selection of controls 2. Similar ascertainment of exposure for all subjects 4. Diagnostic criteria for cases clear, precise, and valid 3. Similar referral and diagnostic procedures for all subjects 5. Date of diagnosis for case operationally defined 4. Diagnostic criteria for disease (clear, reliable, and valid) 6. Ascertainment of disease adequate for cases and controls 5. Characteristics of subjects at enrollment reported 7. Comparison of cases and controls at enrollment reported 6. All aspects of exposure measured (level, dose, duration, etc.) 7. Coexposures measured (including confounding variables) 8. Recall bias controlled 8. All aspects of exposure measured (level, dose, duration, etc.) 9. Data collection valid and reliable 9. Exposure adequately measured (same in all groups, 10. Selection bias considered 11. Analyses controls for confounding factor blinded) 10. Coexposures measured CRITERIA FOR THE APPRAISAL 11. Recall bias controlled OF THE METHODOLOGIC QUALITY 12. Data collection valid and reliable OF COHORT STUDIES 13. Analysis according to level of exposure 14. Effect of matching assessed 1. Zero time identified 2. Baseline comparability reported (including confounding variables)

3bC H A P T E R makes it almost impossible for a patient suffering from neck pain to localize the exact origin of pain,11,12,26 the clinician must Biomechanics of the base the diagnosis on functional and palpatory examinations Cervical and Thoracic Spine of the cervical spine. To do this correctly requires analysis and interpretation of normal and disturbed motion patterns based Jiri Dvorak and Malcolm Pope on a knowledge of clinical biomechanics and an understanding of developmental anatomy. BIOMECHANICS OF THE NECK The natural aging process results in many changes in the The biomechanics of the cervical spine is determined by the shape cervical spine that must be taken into account in the clinical of the vertebral bodies and the orientation of the zygapophyseal assessment, especially as related to range of motion. It is well joints and can accordingly be divided into three sections: established that range of motion decreases with age, mainly in ● Upper cervical spine: occiput, atlas, axis; the middle and lower segments of the cervical spine.4 This is due ● Lower cervical spine: C2-C3 to C7; to the ongoing transformation process of the intervertebral disk ● Cervical-thoracic junction: C7 to the third thoracic vertebra and the development of uncovertebral joints14,29; later it may also be due to development of osteoarthritis of the zygapophy- (Fig. 3b.1). seal joints. The primary aim of the clinician treating a patient with neck In the first two decades of life, the uncovertebral spaces of pain is to find the region or even segment responsible for pain the lower part of the cervical spine begin to undergo a lifelong symptoms. Because the intersegmental nerve root anastomosis transformation into uncovertebral joints. This transformation is a response to the compressive force of the weight of the head, Upper cervical spine which the upright posture of the body requires the cervical spine to support (Fig. 3b.2).14,28,29 The result is the formation of lateral C2 tears of the disk annulus beginning in the second decade of C3 life (Fig. 3b.3). These lateral tears continue to develop into the medial center part of the disk until, in the third decade, com- plete transverse tears commonly occur (Fig. 3b.4).28 The resulting space in the middle of the intervertebral disk partially takes over the function of the zygapophyseal joints during the second and third decades. (At this stage, the nucleus pulposus dries out, appearing on radiographs as a narrowing between the vertebrae of the lower cervical spine, which provides a convenient way for the clinician to monitor these changes.) This new space within the disk significantly reduces the load- absorbing function of the intervertebral disk in the cervical spine. To support the load of the head, a transformation of the uncovertebral joints starts to occur.28 Instead of the original pointed shape, the uncovertebral processes now become flat C4 Lower C5 cervical spine C6 C7 Cervical-thoracic Th1 junction Th2 Th3 Figure 3b.1 Sections of the cervical spine. Figure 3b.2 Drawing of the uncovertebral joints. (From Luschka H: Die Halbgelenke des menschlichen Körpers. Berlin, 1858, Reimers.)

42 Chapter 3b ● Biomechanics of the cervical and thoracic spine Figure 3b.3 Frontal section of the cervical spine of a 9-year-old child. Figure 3b.4 Frontal section of the cervical spine of a 33-year-old Remnants of cartilage are still present. The arrow points to a space in the lateral part of the intervertebral disk at level C3-C4. man. In the three lowest segments, each intervertebral disk shows with a shape like that of a cow horn (Fig. 3b.5) and take over a complete transverse tear; however, the upper two levels have only the load of the cranial vertebrae. The overall result is a natural transformation of the structure and shape of the uncovertebral lateral tears. (From Töndury G, Theiler K: Entwicklungsgeschichte processes that is probably responsible for much of the decreased range of motion that accompanies aging and must be taken into und Fehlbildungen der Wirbelsäule, ed 2, Stuttgart, Germany, 1990, account in any clinical assessment of the cervical spine. Hippokrates-Verlag.) In the first two decades of life, the surfaces of the articular processes are covered by a thin layer of cartilage, with the uneven capsule that limits the movements possible. The dominant surfaces filled in by a synovial fold in the joint capsule. This movement in the atlantooccipital joint is flexion and extension synovial fold has been described by Penning and Töndury21 as of approximately 22 to 24 degrees, according to the author. meniscoid. It is found within the entire cervical spine (Fig. 3b.6) Lateral bending, again according to the author, is between 5 and and again degenerates or atrophies with increasing age. 10 degrees. The idea of axial rotation in this joint has long been rejected; however, newer investigations by Dvorak et al7 UPPER CERVICAL SPINE: OCCIPUT/C2 showed axial rotation in both in vitro and in vivo studies. A summary of movements possible in the atlantooccipital joint The upper cervical spine, which consists of the occiput, atlas, and is shown in Table 3b.1. axis, is responsible for most of the axial rotation and some of the flexion-extension and lateral bending of the head. In addition The atlantoaxial joint consists of four joint spaces: the two to allowing large rotations, it must be stable enough to support atlantoaxial lateral joints, the atlantoaxial median joint (between the weight of the head and protect the delicate spinal cord the anterior arch of the atlas and the dens axis), and a joint and intervertebral arteries from injury and is therefore quite a between the posterior surface of the dens and the transverse complicated structure. Possible motion in the atlantooccipital ligament, which is connected to the anterior joint space. From and atlantoaxial joints is determined by the orientation of the the medial part arises a large synovial fold in the lateral atlantoax- articular processes. The occipital condyles are an oval-shaped ial joint (Fig. 3b.10).4 This joint capsule is loose, which allows for bean-like form with a sagittal orientation of the joint axis of a great deal of motion. It is here that most of the axial rotation 28 degrees on average (Fig. 3b.7).8,10 The frontal orientation occurs. The movements possible in the atlantoaxial joints are of the joint axis (Fig. 3b.8) averages 124 degrees in males and summarized in Table 3b.2. 127 degrees in females.9,27 Motion within the upper cervical spine, especially in the The motion axis of the atlantooccipital joints was described atlantoaxial joint, is mainly limited by ligaments that, with by Knese9 (Fig. 3b.9). The atlantooccipital joints are described exception of the tectorial membrane, consist of nonstretchable as a spheroid articulation. They are connected with a tight joint collagen fibers.24,25 The tectorial membrane, which consists of elastic fibers, inserts at the great occipital foramen and distally joins the posterior longitudinal ligament. The biomechanical properties of the tectorial membrane have been studied by Oda et al,17 who documented their large elasticity.

Chapter 3b ● Upper cervical spine: occiput/C2 43 Figure 3b.5 Cow horn-like shapes (arrows). Changes of the uncovertebral disks at the C4-C5 and C5-C6 levels of the same subject at the age of 47 years (A) and 49 years (B). At the same level on the lateral views, anterior and posterior osteophytes are starting to form (C and D) and narrow the intervertebral foramen at the C5-C6 level (E and F). AB CD EF

44 Chapter 3b ● Biomechanics of the cervical and thoracic spine Frontal orientation of joint axis 124° Figure 3b.8 Frontal orientation of the occipital condyles according to Stoff. (From Stoff E: Verh Anat Gesch Jena 70:575, 1976.) Figure 3b.6 Parasagittal section of the intervertebral joints Z-axis Y-axis (zygapophyseal joints) at the level of C4 and C6. The articular X-axis processes show an inclination of approximately 45 degrees. The arrow points to the synovial folds in between the intervertebral joint surfaces, which have been described by Penning and Töndury as meniscoid. (Courtesy Professor Doctor W. Rauschning, Uppsala, Sweden.) Figure 3b.9 Possible joint axis of the upper cervical spine according to Knese. (From Knese KH: Z Anat Entwickl 114:67-107, 1949.) Table 3b.1 Possible movements in the atlantooccipital joint according to different authors Flexion/extension Occipto-C1 (total) rotation Side bending joint (one side) (one side) Axial Fick (1904) 50 30-40 0 Poirer and Charpy 50 14-40 0 (1926) 13 8 0 Werne (1957)30 35 0 Penning (1978)20 — 10 5.2 Dvorak et al (1985) 22.7 — 4.8 Clark et al (1986) — — 4 Dvorak et al (1987)3 — Penning and Wilmink — 1 24.5 — 7.2 (1987)22 5.5 Panjabi et al (1988)19 Figure 3b.7 Sagittal orientation of the occipital condyles is 28 degrees on average. (From Ingelmark BE: Acta Anat (Basel) 6:1-48, 1947.)

Chapter 3b ● Upper cervical spine: occiput/C2 45 AB Figure 3b.10 Dissection of normal left atlantoaxial meniscoid from a fresh cadaveric specimen (A); the surface is covered with meniscoid (B). The cruciate ligament (Fig. 3b.11) has the important function the transverse ligament,5 which allows the ligament to move of restricting potentially dangerous anterior gliding of the more freely and protects it from damage caused by friction. atlas during flexion movement of the head while still allowing The transverse ligament consists exclusively of collagen fibers the atlas to turn freely around the dens during axial rotation. with an interesting fiber orientation similar to a folding lattice It consists of two main parts: a horizontally oriented transverse (Fig. 3b.13). This allows extensive stretching of the ligament ligament and vertically oriented longitudinal fibers. The trans- during axial rotation without damage to the fibers. In vitro exper- verse ligament inserts at the medial portion of the lateral mass iments show failure of the transverse ligament to occur between of the atlas. The caudal fibers are occasionally fixed at the base 170 and 700 N.5 of the dens and may additionally stabilize the dens (Fig. 3b.12). At the level of the dens is a thin layer of cartilage covering The main limiting structures for the upper cervical spine are the alar ligaments. Consisting exclusively of nonstretchable Table 3b.2 Summary of possible motions at the collagen fibers, the alar ligaments connect the dens axis with the atlantoaxial joint according to various authors occipital condyles and the anterior arch of the atlas (Fig. 3b.12).1,3,13 Occasionally, a loose connection is also found between the basis Flexion/ Axial of the dens axis and the anterior arch of the atlas3; this has been extension described by Von Barrow as the atlantodental anterior ligament. (total) Side bending rotation According to Werne,30 alar ligaments are of great importance in limiting axial rotation (Fig. 3b.14), a belief that has been C1-C2 joint (one side) (one side) confirmed by newer investigations.5,18 In conjunction with the tectorial membrane, the alar ligaments also limit flexion of the Fick (1904) 00 60 occiput. During lateral bending (Fig. 3b.15), the alar ligament is responsible for forced rotation of the second vertebra.5,30 Poirer and Charpy The apical ligament has no functional meaning and is actually a rudiment of the chorda dorsalis.10 (1926) 11 — 30-80 Clinical analysis of upper cervical spine motion can be done Werne (1957)30 10 0 47 through the use of functional radiographs. In the anteroposterior view, lateral bending can be assessed.23 Physiologic movements, Penning (1978)20 30 10 70 gliding of the atlas in bending direction, and forced rotation of the axis are presented in Figure 3b.16 as seen on functional radio- Dvorak et al (1985) — — 32.2 graphs. Axial rotation is currently assessed through measurements of functional computed tomography and may in the future be Clark et al (1986) 10 — 14.5 tested by functional magnetic resonance imaging (Fig. 3b.17).2,6 Dvorak et al (1987)3 — — 43.1 Penning and Wilmink (1987)22 —— 40.5 Panjabi et al (1988)19 22.4 6.7 38.9

46 Chapter 3b ● Biomechanics of the cervical and thoracic spine Dura Jugular foramen Clivus Hypoglossal canal Tectorial membrane Transverse occipital ligament Atlantooccipital articulation Alar ligament Cruciate ligament Atlantoaxial articulation Posterior longitudinal Transverse ligament ligament Figure 3b.11 Ligaments of the upper cervical spine, posterior view. (From Lang J: Klinische Anatomie der Halswirbelsäule. New York, 1991, Georg Thieme Verlag.) LOWER CERVICAL SPINE The motion segments are connected and stabilized by liga- ments, anteriorly by the anterior longitudinal ligament (Fig. 3b.19) The anatomic structures of the motion segments of the lower and dorsally by the posterior longitudinal ligament. The density cervical spine are different from those in the upper cervical of nociceptive and mechanoreceptive innervation of the poste- spine. Their particularities include the uncovertebral joints, rior longitudinal ligament is high in comparison with other which support part of the axial load once the intervertebral disk cervical spine ligaments and the disk.16 This results in a very sen- loses its elasticity due to age-related transformations.14,28,29 The sitive ligament that indirectly controls the innervation of neck articular processes of the cervical spine are inclined approxi- muscles through nociceptive and mechanoreceptive reflexes.32 mately 45 degrees from the horizontal plane (Fig. 3b.18), with The laminae are connected by the strong ligamentum flavum, steeper inclinations in the lower segments. The transverse processes which consists almost exclusively of elastic fibers and is a major hide and protect the spinal nerve and vertebral artery. limiting structure in flexion movement. Anterior atlantodental ligament The dominant motion in the lower cervical spine is flexion- extension. Different parameters can be measured with flexion- Alar ligament extension x-ray views, including segmental rotation, translatory (atlantal portion) movement, and the location of the center of rotation (Fig. 3b.20).7 Because a significant motion difference exists between actively Alar ligament and passively performed movements, the use of passively per- (occipital portion) formed radiographs has been recommended in diagnosing segmental instability, such as can occur after trauma (Fig. 3b.21).6,7 Transverse The first description of the center of rotation in healthy adults ligament came from Penning and Töndury’s measurements21 of flexion and extension radiographs. The center of rotation has been Figure 3b.12 Drawing of the ligaments of the upper cervical determined with computer-assisted methods7 and has confirmed spine (axial dissection). (From Dvorak J, Froehlich D, Penning L, Penning and Töndury’s findings (Fig. 3b.22). Table 3b.3 shows Baumgartner H, Panjabi MM: Spine 13(7):748-755, 1988.) the relevant data on flexion-extension movements of healthy adults for in vitro and in vivo examinations. Table 3b.4 presents data on rotation, translation, and center of rotation as measured by computer-assisted methods on a healthy population.7 Lysell15 described the so-called top angle, or segmental arch, as being flat at the level of C2 and steep at the lower cervical spine. Motion of the upper segments during flexion-extension is there- fore fairly horizontal, whereas motion of the lower segments is more like that of an arc (Fig. 3b.23).

Chapter 3b ● Lower cervical spine 47 Figure 3b.13 The orientation of the collagen fibers of the transverse ligament is similar to a folding lattice and allows extensive stretching of the ligament during flexion and axial rotation without damage to the fibers. Figure 3b.14 Drawing of possible movements in the atlantoaxial joint during axial rotation of the head according to Werne. (From Werne S: Acta Orthop Scand Suppl 23:80, 1957.)

48 Chapter 3b ● Biomechanics of the cervical and thoracic spine Left lateral bending Right lateral bending Figure 3b.15 Function of the alar ligaments during –∅z y side bending of the head. (From Dvorak J, Froehlich D, y y +∅z Penning L, Baumgartner H, Panjabi MM: Spine y1 C0 y1 13(7):748-755, 1988.) –x C1 +∅y +x +∅y C2 Figure 3b.16 Forced rotation of the axis and gliding of the atlas in the direction of bending as seen from functional radiographs in the anteroposterior view. (From Reich C, Dvorak J: Manual Med 2, 1986.) Figure 3b.17 Functional computed tomography of the upper cervical spine during axial rotation within the atlantoaxial and atlantooccipital joints.

Chapter 3b ● Lower cervical spine 49 Figure 3b.17 Cont’d Figure 3b.18 Orientation of the articular processes of the lower cervical spine in the frontal plane. (From White AA, Panjabi MM: Clinical biomechanics of the spine, ed 2. Philadelphia, 1990, JB Lippincott.)

50 Chapter 3b ● Biomechanics of the cervical and thoracic spine Figure 3b.19 Ligaments of the anterior and posterior parts at the lower cervical spine. (From White AA, Panjabi MM: G D Clinical biomechanics of the spine, ed 2. Philadelphia, 1990, F A JB Lippincott.) C HB IE D G. Facets E. Intertransverse ligament F. Capsular A. Anterior ligament longitudinal ligament I. Interspinous and B. Anterior one-half superspinous annulus fibrosus ligaments C. Posterior one-half H. Ligamentum annulus fibrosus flavum D. Posterior longitudinal ligament Y RX BY BZ +Y AY +RY AZ +RX +X +Z +RZ CRY Z X CRZ Figure 3b.20 Parameters for measurement of segmental motion using computer-assisted methods. RX is the rotation about the x-axis; CRY and CRZ are the centers of rotation of the Y and Z rotations, respectively; AY, AZ, BY, and BZ are the translations of point A in the Y and Z directions and point B in the Y and Z directions, respectively.

Chapter 3b ● Lower cervical spine 51 AB D C Figure 3b.21 Female, aged 51, 1 year after injury to the cervical spine. The examination has been performed during active and passive motion (A-D) and measured according to Penning’s method (E and F).

52 Chapter 3b ● Biomechanics of the cervical and thoracic spine 35 Hypermobile Figure 3b.21 Cont’d According to the functional 30 diagram, there is a significant difference in segmental 25 Passive C3/C4 C4/C5 C5/C6 motion, especially as related to segments C3-C4 and 20 C2/C3 C6-C7. Drawing the vertebral bodies on a transparent 15 paper (G) makes the difference obvious. (From Dvorak J, 10 Average + 2 STDV Froehlich D, Penning L, Baumgartner H, Panjabi MM: Average + 1 STDV Spine 13(7):748-755, 1988.) 5 Average 0 Average – 1 STDV C6/C7 C1/C2 Average – 2 STDV Z.A., F, 51y C6/C7a E 35 Active Hypomobile 30 C2/C3a 25 C3/C4a C4/C5a C5/C6a 20 15 Average + 2 STDV 10 Average + 1 STDV Average 5 Average – 1 STDV 0 Average – 2 STDV C1/C2a Z.A., F, 51y F Active Active Passive Passive Hypermobile G Z.A., f, 51Y