Predicting the Onset of Forearm Pain: A Prospective Study Across 12 Occupational Groups

Arthritis & Rheumatism (Arthritis Care & Research) Vol. 49, No. 4, August 15, 2003, pp 519 525 DOI 10.1002/art.11202 2003, American College of Rheumatology ORIGINAL ARTICLE Predicting the Onset of Forearm Pain: A Prospective Study Across 12 Occupational Groups ELIZABETH S. NAHIT, STEWART TAYLOR, ISABELLE M. HUNT, ALAN J. SILMAN, AND GARY J. MACFARLANE Objective. To determine, among workers free of forearm pain, the role of mechanical and psychosocial factors in predicting future onset. Methods. A prospective cohort study was conducted among 782 newly employed workers from 12 occupational groups. At baseline, a cohort of 782 workers free of forearm pain was identified and measurement was made about physical and psychosocial aspects of their job and working environment. Subjects were recontacted after 1 year to determine new onsets of forearm pain. A sample of those reporting new onset forearm pain underwent a structured examination of the upper limb. Results. One year after baseline, 666 (85%) subjects were followed up. The overall prevalence of new onset forearm pain was 8.3% (n 55). The strongest mechanical risk factor was frequent repetitive movements of the arm or wrist (odds ratio [OR] 2.9, 95% confidence interval [95% CI] 1.6 5.2). The strongest psychosocial risk factors were work considered monotonous at least half of the time (OR 3.0, 95% CI 1.6 5.7) or work with little autonomy (OR 2.6, 95% CI 1.1 6.1). Three specific independent risk factors (monotonous work, repetitive wrist movement, working with hands above shoulder level) could distinguish groups of subjects at substantially different risks of onset. Conclusions. Along with repetitive movements of the arms and wrists, mechanical postural factors and psychosocial factors also are important risk factors for onset of forearm pain. Our study emphasizes the multifactorial nature of risks for onset of forearm pain, and provides leads as to possible mechanisms for prevention. KEY WORDS. Regional pain syndrome; Forearm; Occupational risk; Epidemiology. INTRODUCTION The etiology of diffuse, nonspecific forearm pain has been the subject of intense controversy. Some believe it to be the result of injury from mechanical exposures to the wrist or forearm, such as repetitive movement. Others believe that it is commonly a manifestation of a fibromyalgia-type syndrome in which adverse psychological and psychosocial factors are prominent. Epidemics of conditions, with forearm pain as a prominent feature, have been common in occupational settings and are not confined to recent times. Supported by grants from the British Occupational Health Research Foundation and The Arthritis Research Campaign, United Kingdom. Elizabeth S. Nahit, PhD, Stewart Taylor, BSc, Isabelle M. Hunt, BSc, Alan J. Silman, MD, Gary J. Macfarlane, PhD: University of Manchester, Manchester, United Kingdom. Address correspondence to Professor Gary J. Macfarlane, The Unit of Chronic Disease Epidemiology, School of Epidemiology and Health Sciences, The Medical School, University of Manchester, Oxford Road, Manchester M13 9PT, UK. E-mail: G.Macfarlane@man.ac.uk. Submitted for publication November 8, 2001; accepted in revised form September 5, 2002. In the early 18th century, Ramazzini described pain in the wrist and forearm among bakers, pain that he considered a result of the constant pressure of kneading dough. Forearm pain was also a common complaint among writers and scribes (writers cramp). In the late 19th century, the introduction of the telegraph led to pain, numbness, and weakness being commonly reported among telegraph operators (telegraphists cramp) (1). More recently, in the 1980s, cumulative trauma disorders and repetitive strain injuries have been commonly reported across a variety of industries with such terms implying symptoms consequent upon mechanical tasks (principally in the workplace). Nevertheless, the hypothesis that the onset and reporting of these recent conditions are also influenced by psychosocial factors is not confined to recent times (2). Gowers (3), in 1888, on writers cramp described sufferers as of distinctly nervous temperament the condition being easily imagined by those who have witnessed the disorder. The relative contribution of mechanical exposures and psychosocial factors to this disorder remain to be determined. The majority of previously conducted studies have primarily been cross-sectional or retrospective. These are 519

520 Nahit et al further information (baseline). This was because at the time of recruitment, many workers would not know the physical characteristics of their usual work and there would have been insufficient time for adverse psychosocial factors to develop. At baseline, subjects were asked Have you had pain in the shaded area lasting at least 24 hours during the past month? Subjects were provided with an upper-body manikin with the area from the wrist to just below the elbow shaded (see reference 5). Because the aim of the study was to determine factors associated with pain onset, only those subjects who were free of forearm pain at baseline were eligible for followup. Figure 1. Study design. particularly susceptible to recall bias about past and current experiences between those with and without symptoms. Furthermore, the diagnostic terms used in defining the conditions of interest in the forearm or wrist have often implied not only symptoms but also etiology (e.g., repetitive strain injury), making an evaluation of possible etiology impossible. The aim of the current study was therefore to determine, among workers free of forearm pain, the role of mechanical and psychosocial factors in predicting future onset. MATERIALS AND METHODS Design. The design of the study was a prospective cohort study of newly employed workers (Figure 1). Subjects were recruited at or close to the time of taking up employment (recruitment phase). One year later, mechanical exposures and psychosocial factors in the workplace were measured on those who were free of forearm pain (baseline phase). This pain-free cohort was followed for an additional year to determine who had developed forearm pain (followup phase). Recruitment. Subjects were recruited from 12 occupational groups. These occupations were selected such that most workers would be taking up their first full-time employment. They represented a range of industries or occupations where high rates of musculoskeletal disorders had previously been identified, with comparatively stable workforces. Only workers who were employed for at least 20 hours per week were eligible for recruitment and, for practical purposes, each workplace had to contribute a minimum of 40 newly employed workers. The sources for recruitment of subjects have been described previously (4) but full details are given in Appendix A. Identifying pain-free participants. All information was collected by means of a self-completion questionnaire. Because workers were recruited to the study very close to their taking up employment (or in the case of vocational courses and workers on training courses, before taking up employment), we waited for 12 months before collecting Measuring exposures in the workplace at baseline. Information was collected relating to mechanical factors and psychosocial factors in the workplace. Information on psychosocial factors was collected according to the model of Karasek (6). Participants were asked to assess job demands (i.e., the hectic nature of work, perceived stress), control (i.e., the ability to decide how to carry out one s work, whether one learned new things at work), and satisfaction with social support from colleagues. Subjects were also asked, overall, how satisfied they were with their job. Information on mechanical factors was collected, relating to the last shift/day worked, using a previously validated questionnaire (7). Details of posture, repetitive movements of the upper limb, and manual handling activities were collected. For each posture, information was asked on the duration of time spent in the posture during the individual s last working day. Subjects were required to select 1 of 4 categories: less than 15 minutes, 15 minutes to less than 2 hours, 2 hours to less than 4 hours, and 4 hours or more. Information was collected on repetitive movements (of the wrist and arms) in the same way. For manual handling activities, information was requested on the average weight involved in the particular activity, using a visual analog scale. Forearm pain at followup. One year later, subjects who were free of forearm pain at baseline were recontacted to establish whether they now had forearm pain. The question and manikin used were identical to baseline. In addition, subjects with new forearm pain were asked to complete a disability questionnaire, inquiring about 10 activities that could potentially be difficult to carry out because of forearm pain. Finally those subjects with forearm pain who worked near the study coordinating center were asked to undergo a structured clinical examination of the upper limb according to a previously validated schedule (8). From this, it could be determined whether these subjects satisfied criteria for a specific upper limb disorder or had nonspecific forearm pain (9). Data analysis. Workplace exposures were divided into those who were not exposed, those exposed below the median level, and those exposed above the median level of exposure. Posture and repetitive movements were dichotomized according to the distribution of responses. Psychosocial factors of demand, support, and control were dichotomized, respectively, by those exposed at least half of

Predicting the Onset of Forearm Pain 521 Table 1. Participation rates and prevalence of forearm pain at followup Occupation the time, those dissatisfied/very dissatisfied, and those with control seldom/very seldom. A case-control analysis (subjects with and without forearm pain, respectively) was conducted to determine the univariate association between each exposure and the onset of forearm pain. The associations were assessed using logistic regression analyses, adjusting for age, sex, and occupational group. Thereafter, a stepwise analysis was conducted to determine if a small group of independent factors could identify those at high risk of developing forearm pain. Relationships are presented as odds ratios (ORs) with 95% confidence intervals (95% CIs). Finally, the population attributable risk (in the working populations studied) of the factors found to be independently associated with onset was calculated (10). RESULTS Baseline total* Followup participation N (%) Forearm pain at followup N (%) All groups 782 666 (85) 55 (8.3) Firefighters 135 120 (89) 6 (5.0) Shipbuilders 92 79 (86) 6 (7.6) Retail workers 87 74 (85) 10 (13.6) Dentists 76 71 (93) 4 (5.6) Nurses 68 59 (87) 3 (5.1) Podiatrists 58 53 (91) 7 (13.2) Army clerks 57 53 (93) 6 (11.3) Army officers 56 46 (82) 2 (4.4) Postal workers 51 35 (69) 1 (2.9) Army infantry 43 25 (58) 7 (28.0) Police 38 34 (89) 0 (0.0) Forestry 21 17 (81) 3 (17.7) * Free of forearm pain. There were 835 participants at baseline of whom 53 reported forearm pain, thus leaving 782 subjects eligible for followup (Table 1). This group had a median age of 24 years (inter-quartile range 21 29 years) and 66% were male. At followup, 85% (n 666) of subjects were successfully recontacted and agreed to participate. Nonparticipants (compared with participants) were the same median age (24 years) but were more likely to be male (nonparticipants 78%, participants 64%; 2 P 0.004). The highest participation rates were achieved among dentists, podiatrists, army clerks, and the police; and the lowest among army infantry, who were based worldwide and consequently were particularly difficult to contact (Table 1). The overall prevalence of forearm pain at followup was 8.3% (n 55). Prevalence was higher in women (10.0%, n 24) than men (7.2%, n 31) ( 2 P 0.22) and was highest in the youngest age group (11.9% 20 years; 7.3% 21 23 years; 6.0% 24 27 years; 7.9% 28 years; 2 P 0.26). There was a wide variation between occupational groups. The highest prevalence was noted among army infantry (28%), forestry workers (18%), retail (supermarket) workers (14%), and podiatrists (13%), while low prevalence was noted among army officers (4%), postal workers (3%), and police officers (0%). Work-related mechanical factors. A number of the mechanical exposures were associated with an increased risk of pain. Frequent repetitive movements of the arm or wrist were the strongest predictors of future forearm pain (both activities OR 2.9, 95% CI 1.6 5.2; Table 2). Several mechanical load activities were associated, although not significantly, with increased risk; these included carrying weights on 1 shoulder (OR 2.1, 95% CI 0.9 4.5) and lifting above shoulder level (OR 2.1, 95% CI 0.9 4.7). Three postures were associated with onset: working with hands above shoulder level (OR 2.4, 95% CI 1.3 4.5), bending (OR 2.2, 95% CI 1.2 3.8), and squatting (OR 2.0, 95% CI 1.0 3.9). Work-related psychosocial factors. Some aspects of job demand and autonomy were also significantly associated with future risk of forearm pain onset. The strongest risk was among participants who reported that their work was monotonous at least half of the time or who very seldom were able to decide how to carry out their work (i.e., low job autonomy). Both factors were associated with an approximate 3-fold increased risk of developing forearm pain (Table 3). Multivariate model. To identify whether a small group of independent factors could predict onset, and if so, to identify how well these identified high risk groups, a further logistic regression analysis was conducted using a forward stepwise model. All previous significant predictors (or where the OR exceeded 1.5 irrespective of statistical significance) were offered as candidates for entry into the model. Age and sex were added into the final model. Three factors were included in the final model: monotonous work, repetitive wrist movement, and working with hands above shoulder level (Table 4). Each 2-way interaction was tested between these 3 variables. However, there was no evidence of deviation from the multiplicative interaction assumed by the logistic model. We then calculated for each subject, using this model, how many of these factors were reported at baseline. The prevalence of forearm pain at followup increased from 4% in those who reported none of these factors to 8% and 22% with 2 and 3 factors, respectively, and 36% in those reporting all 3 factors. Features of forearm pain. Of the 55 subjects with forearm pain, 36 completed the disability schedule by telephone. The majority (78%) reported at least 1 limitation as a result of the pain. The most common problems were difficulty in moving hands and fingers (47%) and difficulty writing or typing (44%). Many (42%) subjects reported spending less time on hobbies because of forearm pain. Sixteen percent of subjects reported having consulted their general practitioner about the problem during the past year, and 7% of those with forearm pain reported taking time off work during the past month.

522 Nahit et al Table 2. Odds of forearm pain at followup in relation to mechanical exposure at baseline Table 3. Odds of forearm pain at followup in relation to psychosocial exposure at baseline Mechanical exposure Forearm pain at followup* No Yes Odds ratio (95% CI) Psychosocial exposure Forearm pain at followup* No Yes Odds ratio (95% CI) Manual handling activity Lifting with 1 hand Never 257 25 1.0 16 lbs 180 17 1.0 (0.5 1.8) 16 lbs 163 12 0.8 (0.4 1.8) Lifting with 2 hands Never 281 19 1.0 25 lbs 168 18 1.7 (0.8 3.3) 25 lbs 151 16 1.9 (0.9 4.0) Carrying on 1 shoulder Never 459 40 1.0 30 lbs 83 6 0.8 (0.3 2.0) 30 lbs 57 9 2.1 (0.9 4.9) Lifting above shoulder level Never 453 41 1.0 20 lbs 74 6 0.9 (0.4 2.2) 20 lbs 71 8 1.5 (0.7 3.5) Pushing Never 406 39 1.0 69 lbs 96 6 0.6 (0.3 1.6) 69 lbs 97 10 1.2 (0.6 2.5) Pulling Never 466 44 1.0 58 lbs 67 3 0.5 (0.1 1.6) 58 lbs 63 7 1.3 (0.5 3.0) Posture Sitting 4 hours 429 38 1.0 4 hours 172 17 1.0 (0.5 1.9) Standing 4 hours 493 42 1.0 4 hours 107 13 1.4 (0.7 2.8) Driving 4 hours 553 51 1.0 4 hours 58 4 0.8 (0.3 2.4) Kneeling 15 mins 502 40 1.0 15 mins 99 15 1.8 (0.9 3.4) Squatting 15 mins 520 42 1.0 15 mins 79 13 2.0 (1.0 3.9) Bending 15 mins 456 33 1.0 15 mins 143 22 2.2 (1.2 3.8) Stretching below knee level 15 mins 532 46 1.0 15 mins 69 9 1.6 (0.7 3.3) Working with hands above shoulder 15 mins 501 37 1.0 15 mins 100 18 2.4 (1.3 4.5) Repetitive movement Repetitive wrist movements 2 hours 424 24 1.0 2 hours 178 31 2.9 (1.6 5.2) Repetitive arm movements 2 hours 477 31 1.0 2 hours 123 22 2.9 (1.6 5.2) * Total numbers vary due to missing items. Adjusted for age, sex, and occupational group. 95% CI 95% confidence interval. Job demand Stressful work Never/occasionally 499 43 1.0 At least half of the time 106 10 1.1 (0.5 2.2) Monotonous work Never/occasionally 505 33 1.0 At least half of the time 94 19 3.0 (1.6 5.7) Hectic work Never/occasionally 441 36 1.0 At least half of the time 164 17 1.2 (0.6 2.3) Job satisfaction Satisfaction with job Not dissatisfied 567 47 1.0 (Very)/dissatisfied 32 5 1.7 (0.6 4.7) Social support Satisfaction with support Not dissatisfied 578 50 1.0 (Very)/dissatisfied 26 3 1.4 (0.4 5.0) Control over work Able to decide how to carry out work At least sometimes 566 46 1.0 (Very)/seldom 38 7 2.6 (1.1 6.1) Learning new things at work At least sometimes 558 48 1.0 (Very)/seldom 47 5 1.3 (0.5 3.5) * Total numbers vary due to missing items. Adjusted for age, sex, and occupational group. 95% CI 95% confidence interval. Seventeen subjects with forearm pain who worked close to the study coordinating center were invited and agreed to a clinical examination. Three subjects satisfied criteria for tenosynovitis of the wrist, 1 satisfied criteria for carpal tunnel syndrome, and the remainder (n 13) did not Table 4. Combined regression model of risk factors for new onset of forearm pain Exposure Odds ratio (95% CI)* Repetitive movement of wrists 2 hours 1.0 2 hours 2.9 (1.5 5.3) Monotonous work Never/occasionally 1.0 At least half of the time 3.0 (1.5 5.8) Working with hands above shoulder 15 mins 1.0 15 mins 2.2 (1.1 4.3) * Adjusted for age, sex, and occupational group. 95% CI 95% confidence interval.

Predicting the Onset of Forearm Pain 523 satisfy criteria for any upper limb disorder (nonspecific forearm pain). Population attributable risk in the working populations studied. We calculated the population attributable risk (PAR) for the 3 factors independently associated with forearm pain onset to determine their contribution to the overall occurrence of forearm pain among the group of workplaces studied. The PAR associated with repetitive wrist movements was 36%; monotonous work, 23%; and working with hands above shoulder level, 18%. A total PAR using these 3 factors was 77%. DISCUSSION We have previously shown, from the cross-sectional phase of this study, that forearm pain was associated with the reporting of physical factors in the workplace (particularly repetitive wrist movements) and adverse psychosocial factors (monotony, lack of autonomy, and stress) (4,11). The prospective phase of the study has clarified the temporal relationships between forearm pain and these factors. Physical factors in the workplace (repetitive movements, posture) and psychosocial aspects of the workplace (aspects of job demands and control) influence the risk of future onset. They are therefore not merely a consequence of having symptoms and/or a reflection of altered perception of job tasks. Three specific factors (monotonous work, repetitive wrist movement, working with hands above shoulder level) could distinguish groups of subjects at substantially different risks of onset; among the highest risk group, 1 in 3 had developed forearm pain at the end of the followup year compared with 1 in 25 of the low risk group. There are several methodologic issues to consider in relation to the study. 1) It is prospective, among subjects initially free of forearm pain. Therefore, the exposures were evaluated prior to the onset of symptoms and the study is not affected by recall bias. 2) The cohort consisted of young workers who had recently taken up employment (the majority in their first full-time job). Among those not followed, there was none who had left employment for health-related reasons. Therefore, this study is less affected by the healthy-worker effect than studies on established working populations. Nevertheless to be included in the cohort, subjects were required to have been employed for 1 year. 3) Reports of workplace exposures are self reported and, although the mechanical exposure questionnaire was validated, there will inevitably be errors in reporting. However, because all the persons are reporting exposures when forearm pain free, the errors will be unrelated to future forearm pain unless, for example, subjects who overreport exposures at baseline are more prone on followup to do the same with respect to their symptoms. 4) Despite the fact that we had a large cohort, forearm pain is relatively uncommon (in relation to other regional pain syndromes). Given that we were only able to clinically examine subjects located close to the study coordinating center, and the fact that among these subjects specific upper limb conditions were uncommon, there were only 4 subjects who satisfied criteria for a specific disorder. We therefore were not able to address an issue of considerable interest. Does the etiology of specific conditions and nonspecific forearm pain differ? Evidence from a recently reported population cross-sectional survey of upper limb disorders suggests that etiology may be different (12). 5) The design of the study dictates that all forearm pain in this study was of new onset since baseline. Although the majority reported some consequence of symptoms, we have not determined impact further, e.g., were they minor discomforts or serious handicaps? Only 1 in 6 of those with new onset forearm pain reported having consulted their general practitioner about the problem, suggesting that it was a serious handicap in only a minority. Results from previous studies of persistence of regional pain syndromes suggest that adverse psychosocial factors may be particularly important in determining the likelihood of chronicity (13,14). 6) Although the participation rate at followup was high (85%), there is still concern that subjects lost to followup may result in a bias to the study results. We compared response rates according to the 3 independent predictors of outcome: there was neither any large nor significant differences (data not shown). 7) The current study has collected information on some psychosocial factors. In general, adverse factors were relatively uncommon, perhaps reflecting the newly employed status of the cohort. In addition, risk may well also be affected by illness attitudes and other premorbid individual factors that were not collected. Social factors (e.g., acceptability of specific symptoms, perceptions of their cause) may also be strong influences on risk, but cannot be studied in the short term. In attempting to compare the results of this study, there are no other prospective epidemiologic studies in the workplace examining factors influencing onset of forearm pain. We have previously conducted a study in the general population in which we collected information on some factors examined in this study (5). This was a study of 1,953 working subjects aged 18 65 years, free of forearm pain, randomly sampled from the general population in an area of northwest England. Similar, but not identical, instruments were used to gain information on mechanical and psychosocial factors at work. Interestingly, the workplace factors associated with an increased risk of onset were repetitive movements of the arms, repetitive movements of the wrists, lack of control over work conduct, monotonous work, dissatisfaction with support from colleagues/supervisor, and stressful work. These predictors of onset overlap considerably with those reported in the current study. Such consistency of information in different settings using different instruments provides stronger evidence that these factors are indeed important in influencing onset. Nevertheless, there are studies examining other regional pain syndromes or specific clinical conditions in the forearm whose results can be compared with the current study. Armstrong et al (15) examined the role of upper limb repetitiveness and forcefulness in hand and wrist tendinitis in a cross-sectional study among 652 workers. Exposure assessment included electromyography and videotape analysis. The prevalence increased from 0.6%

524 Nahit et al among persons with low force, low repetitiveness jobs rising to 10.8% among persons with high force, high repetitiveness jobs. A further cross-sectional study compared the prevalence of tenosynovitis in assembly line packers doing repetitive work (with movements up to 25,000/day) with shop assistants (described as having variable tasks). The prevalence of tenosynovitis/peritendinitis was 53% and 14%, respectively (16). In a review of evidence from occupational studies, The National Institute for Occupational Safety and Health in the United States concluded that there was evidence that repetitive tasks were associated with hand/wrist tendinitis and that the evidence of an association was particularly strong when combined with other factors, such as force and posture (e.g., forceful hand or wrist extensions) (17). The current study adds prospective evidence that repetitive tasks are strong risk factors for the onset of forearm pain. Is the finding that monotonous work is a predictor of onset a reflection that lack of demands is important or is it simply a marker for jobs that involve repetitive (mechanical) factors? Given that monotonous work has been found to be a risk factor after adjustment for repetitive upper limb movement (in the multivariate model), this implies that it is not simply a marker for such repetition but that the perception of work monotony is, independently, a risk for forearm pain. Other studies have similarly reported that the perception of work as monotonous increases the risk of reporting other upper limb pain syndromes. In a case control study of working persons consulting a community physician with neck/shoulder problems in Sweden (n 109) compared with pain-free community controls, cases were significantly more likely to report low work content (OR 23, 95% CI 5 107) (18). In a study of more than 22,180 workers undergoing a screening examination, those whose work was monotonous were more than twice as likely to have neck or shoulder pain (19). One of the issues in such cross-sectional studies is dissecting the temporal association. The current study adds evidence that the perception of work as monotonous leads to an increased risk of forearm pain onset, rather than simply being a consequence of having symptoms. This also agrees with results from studies that have found that monotonous work is a risk factor for other regional pain syndromes (20). However, for some associations found in this study on univariate analysis, such as bending and squatting, it is difficult to propose a biologic mechanism. Instead these apparent relationships are explained by confounding. Persons who reported these activities were also more likely to report that their job was monotonous or involved repetitive upper limb movement (data not shown). In summary, this study has shown that with respect to the onset of forearm pain, although repetitive movements of the arms and wrists are important, they are not the only (or only type of) risk factor. Mechanical postural factors also influence risk, as do psychosocial factors monotony and lack of control over the working environment. Our observations are consistent with some hypotheses about historical epidemics. These occurred generally in workers with exposure to tasks involving repetitive upper limb movement, but were specific with respect to country, occupational group, company, and time the epidemic occurred. It has been speculated that in addition to social influences (which could not be examined in the current study), workplace psychological and psychosocial factors were involved. Our study emphasizes the multifactorial nature of risks for the onset, and provides leads as to possible mechanisms for prevention. ACKNOWLEDGMENTS The authors are indebted to all the organizations that allowed us access to their workforces/students and our contacts in these organizations who were crucial in facilitating the practical aspects of conducting the study. Nicola M. Cherry was involved in aspects of study design and supervising conduct; Christina Pritchard was involved in recruitment of organizations and study subjects; Ann Papageorgiou conducted the clinical examinations with the help of Dr. Jaechul Song. We thank colleagues at the MRC Environmental Epidemiology Unit at the University of Southampton for help with respect to the upper limb clinical examination schedule. REFERENCES 1. Dembe AE. Occupation and disease: how social factors affect the conception of work-related disorders. New Haven: Yale University Press; 1996. 2. Ireland DCR. Psychological and physical aspects of occupational arm pain. J Hand Surg 1988;13B:5 10. 3. Gowers WR. A manual of diseases of the nervous system. Philadelphia: Blakiston; 1888. 4. Nahit ES, Macfarlane GJ, Pritchard CM, Cherry NM, Silman AJ. Short-term influence of mechanical factors on regional musculoskeletal pain: a study of newly employed workers. Occup Environ Med 2001;58:374 81. 5. Macfarlane GJ, Hunt IM, Silman AJ. Role of mechanical and psychosocial factors in the onset of forearm pain: prospective population based study. BMJ 2000;321:676 9. 6. Karasek RA. Job demands, job decision latitude and mental strain: implications for a job redesign. Admin Sc Q 1979;24: 285 311. 7. Pope DP, Croft PR, Pritchard CM, Silman AJ Macfarlane GJ. Occupational factors in relation to shoulder pain and disability. Occup Environ Med 1997;54:316 21. 8. Palmer K, Walker-Bone K, Linaker C, Reading I, Kellingray S, Coggon D, et al. The Southampton examination schedule for the diagnosis of musculoskeletal disorders of the upper limb. Ann Rheum Dis 2000;59:5 11. 9. Harrington JM, Carter JT, Birrell L, Gompertz D. Surveillance case definitions for work-related upper limb syndromes. Occup Environ Med 1998;55:264 71. 10. Silman AJ. Epidemiological studies: a practical guide. Cambridge: Cambridge University Press; 1995. 11. Nahit ES, Pritchard CM, Cherry NM, Silman AJ, Macfarlane GJ. The influence of work related psychosocial factors and psychological distress on regional musculoskeletal pain: a study of newly employed workers. J Rheumatol 2001;28: 1378 84. 12. Walker-Bone K, Palmer K, Reading I, Linaker C, Byng T, Ryall C, et al. Brawn of brain? Mechanical and psychosocial risk factors for neck and upper limb musculoskeletal disorders. Rheumatology 2001;40:S74. 13. Macfarlane GJ, Hunt IM, Silman AJ. Predictors of chronic shoulder pain: a population-based study. J Rheumatol 1998; 25:1612 5. 14. Thomas E, Silman AJ, Croft PR, Papageorgiou AC, Jayson MIV, Macfarlane GJ. Predicting who develops low back pain in primary care: a prospective study. BMJ 1999;318:1662 7. 15. Armstrong TJ, Fine LJ, Goldstein SA, Lifshitz YR, Silverstein

Predicting the Onset of Forearm Pain 525 BA. Ergonomic considerations in hand and wrist tendonitis. J Hand Surg 1987;12A:830 7. 16. Luopajarvi T, Kuorinka I, Virolainen M, Holmberg M. Prevalence of tenosynovitis and other injuries of the upper extremities in repetitive work. Scand J Work Environ Health 1979; 5:48 55. 17. National Institute for Occupational Safety and Health (NIOSH). Musculoskeletal disorders and workplace factors: a critical review of epidemiologic evidence for work-related musculoskeletal disorders of the neck, upper extremity and low back. Washington, DC: US Department of Health and Human Services; 1997. 18. Ekberg K, Björkqvist B, Malm P, Bjerre-Kiely B, Karlsson M, Axelson O. Case-control study of risk factors for disease in the neck and shoulder area. Occup Environ Med 1994;51:262 6. 19. Linton SJ. Risk factors for neck and back pain in a working population in Sweden. Work Stress 1990;4:41 9. 20. Bongers PM, de Winter CR, Kompier MAJ, Hildebrandt VH. Psychosocial factors at work and musculoskeletal disease. Scand J Work Environ Health 1993;19:297 312. APPENDIX A: SOURCES FOR RECRUITMENT OF STUDY COHORTS Companies opening a new branch in the northwest of England or employing an entire new workforce were contacted: 2 met our criteria for inclusion. The first, a supermarket, employed workers in a number of areas, including checkouts, service counters, nursery, general office, stock management, and shelf stacking. The second, a postal distribution center, employed workers who were, in the main, responsible for unloading and loading trolleys of mailbags on and off trains and lorries. Other workers were employed in administrative duties and catering. All workers from both sources, who were employed for the required number of hours (at least 20 per week), were invited to participate. Service organizations that regularly recruited cohorts of new workers were contacted. There were 3 recruited to the study. All full-time paid firefighters from 4 local counties who were in their initial training period were invited to participate. A total of 9 training intakes were included. One police force contributed all trainee police officers belonging to 3 training intakes. Three types of new army recruits were involved in the study: officers, infantry, and clerks. Thirty-four officers were selected at random from each of 3 companies. Fifty infantry soldiers were selected at random from each of 2 battalions, and all clerks enrolled on 3 training intakes were included. One established organization, a shipbuilding company, that was recruiting a large group of apprentices was contacted. These workers carried out construction and engineering tasks. All apprentices were invited to participate. Students in their final year of a vocational course were contacted. All nursing students from 1 academic institution, all dental students from 2 academic institutions, all podiatry students from a further 2 academic institutions, and all forestry students enrolled on 2 courses at a specialized college were invited to participate.