- Top of page
- MATERIALS AND METHODS
- APPENDIX A
- APPENDIX B
The surveillance method was based on a large regional sentinel network of OPs, which allowed the inclusion of a large sample of workers (14). The 80 OPs who participated in the study had professional characteristics similar to the 370 who did not participate. One of the original aspects of this study is the size and representativeness of the sample with regard to the region's workforce. The random selection of workers during their compulsory annual occupational health examination was designed to ensure a representative sample of the region's workforce. This was achieved, with the exception that women were slightly underrepresented, and skilled and unskilled workers were somewhat overrepresented. The underrepresentation of female workers could be explained by the lack of OPs in 2 highly female economic sectors, i.e., education and health. Despite these limits, the relatively good representativeness of the sample allows a better estimation of prevalence rates of MSDs than recent large French surveys conducted in high-risk sectors (15, 16).
The health assessment procedure combining a self-administered questionnaire with a standardized physical examination is another originality of the study. The Nordic questionnaire permits a sensitive and reproducible assessment of prevalence rates of MSD symptoms (5, 17). To our knowledge, this is the first study to apply the recommendations of the criteria document (4, 6) to a large working population. This document proposes one of the most exhaustive classifications existing in the literature (4), but to date the diagnostic value of the consensus has not been assessed. However, the disorder and physical examination signs definitions are close to those of the Health and Safety Executive consensus group (2) and of the Southampton examination protocol diagnostic criteria, which demonstrated good diagnostic properties (5, 18). The physical examination procedures, particularly the diagnostic criteria charts and the clinical guide using photographs of clinical tests, were judged positively by the OPs and were easy to apply in an occupational health setting to standardize the physical examination of any workers included in MSD surveillance programs. Physicians were trained to increase the homogeneity of the examination procedure and to enhance the quality of data collection (19), but at this time quality control health assessment measures have not been incorporated in the surveillance system. The intraobserver variability was not systematically studied, but we observed no significant year-to-year variation in prevalence rates of MSDs among OPs, which is reassuring. Interobserver variability in the performance of the physical examination may partly explain the variation in the prevalence of MSDs from one OP to another. However, prevalence data collected by each OP fluctuated by a narrow interval (0–30%), following a normal distribution. Consequently, it could be hypothesized that, taking into account the large number of OPs, the global effect of the interobserver variability on the precision of estimates of the prevalence of MSDs was low, because some OPs probably overestimated the prevalence of the MSDs whereas others underestimated the prevalence. It should also be noted that interobserver variability could partly be explained by the differences in the characteristics (age, economic sector, occupations, etc.) of workers surveyed by each OP.
The surveillance system implemented in the Pays de la Loire region reported prevalence data for NSRMS symptoms and clinically diagnosed MSDs in a large sample of workers. Few workers failed to participate, but due to the cross-sectional design of the study only individuals who were healthy enough to work were included. Therefore, as in other investigations based on working populations, a healthy worker effect probably occurred. Nevertheless, if some workers with severe disorders do have to leave the workforce, the selection phenomenon was probably low, because the prevalence rates we report are in line with the estimate that 5–20% of individuals of working age are affected by upper-limb disorders (1). Diagnoses were based on a rigorous clinical procedure using a set of physical examination signs, which are recommended to enhance the validity of epidemiologic surveillance of MSDs (19). Our surveillance strategy optimized the precision of estimates of the prevalence of MSDs in the working population compared with surveillance systems based on WC claims data (20–22), because several surveys reported a substantial degree of underclaiming for WC benefits, which contrasts with the fraudulent overclaiming frequently brought to public attention (23). The methodology we used allows better estimates of MSD prevalence rates than surveys based on self-administered questionnaires on musculoskeletal symptoms (24, 25). Our results confirm the overestimation of prevalence rates of MSDs when outcomes are defined as symptoms only rather than physician-diagnosed disorders (19, 26–28), because an MSD was diagnosed clinically in approximately half of the workers with NSRMS during the preceding week. Implementation of such surveillance systems across European countries will permit more precise comparisons of musculoskeletal health in the European workforce (2, 4, 19) than the periodic European survey based on self-assessment of symptoms (25).
Work exposure assessments combined with health assessments are recommended for the surveillance of MSDs to determine the level of risk of MSDs with accuracy (1, 29). To our knowledge, this is the first European study describing exposure to risk factors of MSDs in a large and representative working population using the criteria document. Occupational risk factors were assessed for each worker through a self-administered questionnaire, which permitted a more reliable assessment than the job category often used for the epidemiologic surveillance of MSDs (30). Nevertheless, self-assessments of work exposures are less accurate than work analysis by direct observation or video and can lead to an overestimation of work constraints (31). The risk factors described in the present study are those defined by the criteria document (4) and include the main occupational risk factors described in the literature on MSDs (1, 32, 33). Several other risk factors of MSDs involving work organization (e.g., machine paced work and assembly line work) were also included in our study and will be studied in the future. We strictly applied the definitions proposed by the criteria document, except for force, which was assessed according to effort frequency and not its daily duration. The questionnaire presented awkward postures in picture form to facilitate workers' understanding and thus increase the validity of posture self-assessment (34). Work exposures were self-reported and workers who experienced pain may have overrated their exposure levels. Underrating was also possible, especially for workers who moved to lighter work because of recurrent symptoms. As much as possible, we used standardized and validated instruments for assessment of both exposure and outcomes, which reduces potential bias (4). Occupational physical and nonphysical risk factors were ranked according to a 3-level model proposed by the criteria document to standardize risk assessment (4). This approach, which is similar to the procedure used in the European Committee for Standardization Standard European Norm 614-1 and the proposal of the 3-zone model for action by Buckle and Devereux (35), is useful in comparing risk levels across economic sectors and occupations (4). However, the method is too insensitive to assess the risk of MSDs precisely (36) and complementary analyses are needed.
This study provides new information in comparison with several recent studies based on WC claims (20–22) and health self-administered questionnaires (24, 25). The high prevalence of clinically diagnosed MSDs (∼13% of workers) contrasts with the relatively low level of WC claims for upper-limb MSDs in the Pays de la Loire region (in 2003, ∼3.7 WCs for 1,000 workers). This confirms that in France (29), as in other countries (4, 22–24), using WCs as the unique source of information about MSDs leads to the underestimation of the prevalence in the working population. Prevalence rates of NSRMS of the upper limbs during the preceding year and the preceding week (approximately one-half and one-third of workers, respectively) were higher than those reported by the third European survey conducted in 2000 (25) and by a large survey conducted in The Netherlands (37). This could be partly due to methodologic differences, particularly a longer reference period defining symptoms in the Pays de la Loire study. Symptoms were predominant in the neck and shoulders, and prevalence rates of shoulder symptoms were higher than estimates (20–50%) reported in the literature (16, 38). The frequency of wrist symptoms was close to that observed in a previous large survey of French workers exposed to repetitive work (15). Prevalence rates of symptoms did not significantly differ between men and women, contrary to several surveys conducted in general populations (13, 24, 37, 39) and in French workers exposed to repetitive work (15).
This study reports high prevalence rates of clinically diagnosed MSDs in this representative working population, with 13% of workers experiencing 1 of the 6 MSDs considered and ∼3% experiencing at least 2 disorders. Contrary to French and regional WC claims data, the most prevalent disorder was not CTS but rotator cuff syndrome, the prevalence of which was approximately twice that of CTS. This finding agrees with the predominance of NSRMS in the neck and shoulder region. Estimates of the prevalence rates of rotator cuff syndrome reported in the literature vary widely across studies and study populations (1). The prevalence rate of rotator cuff syndrome in Pays de la Loire workers (9% in women and 7% in men) is nevertheless lower than that previously observed in French workers exposed to repetitive work: 29% in highly exposed workers and 16% in weakly exposed workers (15). This difference could be explained by a more restricted definition of the disorder in the present study and a more varied and representative sample of workers, including economic sectors and occupations characterized by a low risk of MSD. In contrast, the prevalence rate of rotator cuff syndrome in the present study was higher than in a recent study conducted in a large British general population using a similar case definition: 6.1% in women and 4.5% in men for physician-diagnosed shoulder tendinitis (28). The high prevalence rate of rotator cuff syndrome is troublesome because of the poor medical and social prognosis of this disorder (40).
There are few studies of lateral epicondylitis conducted in large working populations (1, 41). The prevalence rate estimated in the Pays de la Loire study (2% in women and 2% in men) was higher than the prevalence rate reported in the general population of Southampton (1.3% in women and 1.1% in men) (28), but was similar to that observed in a Swedish general population (1–3%) (42). As observed for rotator cuff syndrome, and probably for the same reasons, the prevalence rate of lateral epicondylitis was lower than in French workers exposed to repetitive work (12% in highly exposed workers and 8% in weakly exposed workers) (15).
Estimates of the prevalence of CTS vary widely between studies conducted in the working population (1), depending on whether or not case definition required symptoms alone, clinical findings, and electrophysiologic testing (26, 27). The prevalence rate estimated in Pays de la Loire workers (4% in women and 3% in men) was close to that observed in a Swedish general population (3.8%) using a less restricted clinical definition (26). However, our prevalence rate estimate was higher than in the general population of Southampton using a similar case definition (0.9% in women and 1.2% in men) (28) and was higher than in a large American survey in 1988 (1.5%) (28). As observed for other disorders, the prevalence rate of CTS was lower than in French workers exposed to repetitive work (19% in highly exposed workers and 7% in weakly exposed workers) (15).
Few epidemiologic studies assessed the prevalence of ulnar tunnel syndrome in the working population. Prevalence rates of ulnar tunnel syndrome in Pays de la Loire workers (0.8% in women and 0.6% in men) were much lower than for CTS, which confirms previous results in French workers (43). Moreover, the prevalence of ulnar tunnel syndrome was lower than in French workers exposed to repetitive work (2% in highly exposed workers and 1.2% in weakly exposed workers) (15).
The prevalence rate of de Quervain's disease was lower than that of CTS in men (0.7%) and to a lesser extent in women (2%). The prevalence rate of flexor-extensor tendinitis (0.5% in women and 0.8% in men) was lower than that of CTS for both sexes, as reported by other studies (15, 44). The prevalence rate of the 3 types of wrist tendinitis altogether was higher in the Pays de la Loire study than in the large American national health interview survey: 0.5% versus 1.9% (44). As for de Quervain's disease, wrist flexor, and wrist extensor tendinitis, the prevalence rates of these wrist disorders were lower than in French workers exposed to repetitive work (3.9%, 3.6%, and 4.2% in highly exposed workers and 2.2%, 2.5%, and 1.9% in weakly exposed workers, respectively) (15).
The prevalence of clinically diagnosed MSDs increased with age even after adjusting for job seniority, which agrees with some studies of working (1, 16, 32) and general populations (28). However, it is always difficult to disentangle the role of age from the effects of cumulative exposure to occupational hazards (45), and additional analyses adjusting for work-related risk factors are needed. One important finding for the prevention of MSDs is the very high prevalence rate after age 50. Although these aging workers were all employed at the time of the survey, the majority experienced recurrent nonspecific regional musculoskeletal pain of the upper limbs, of which 25% resulted from 1 clinically diagnosed MSD, particularly rotator cuff syndrome, and 6% resulted from ≥2 disorders. This accumulation of MSDs in aging workers probably lowers functional capacities and increases the risk of disability and job dismissal (40).
The study demonstrates wide variations in the prevalence rate of MSDs across economic sectors (1, 32). The sectors most affected were the manufacturing industries and public administration. Broadly, this confirms the results of large North American surveys conducted in the working population (24, 44) or based on WC claims (20, 21, 46). Similarly, our study demonstrated wide variations in the prevalence of MSDs according to occupation. Occupations with the highest prevalence rates were those employing unskilled industrial workers and agriculture workers of both sexes, as well as material handlers, drivers, and employees of public services for men, and personal care employees for women. These occupations are similar to those identified as being at high risk of WC claims in the US (20, 21). Contrary to findings from comparable surveys from North America, very few cases of MSDs were observed in administrative employees and clerks.
Overall, the study shows a high level of exposure to risk factors of MSDs for most workers of both sexes. Contrary to several studies conducted in high-risk sectors, no difference in work exposures was observed between men and women (1, 13), except for the neck region. Exposure to risk factors of MSDs did not decrease with age, which demonstrates the need for ergonomic interventions that reduce the level of exposure to occupational risk factors because of the high prevalence of disorders after age 50. Nevertheless, complementary analyses are needed to precisely assess work exposures in this workforce accurately.
For descriptive purposes, we used a modified version of the decision process proposed by the criteria document to assess the probability of disorders' work relatedness. Exposure bias was limited by the low job turnover in the population studied (87% had been in the same job for over a year), and under exposure to physical work factors, latency periods of MSDs are probably short (i.e., several weeks). It is therefore likely that the risk of MSDs was mainly associated with recent work exposure (47, 48). Only workers with length of service >12 months were studied because health and occupational risk factor assessments concerned mainly the last 12 months. This restrictive rule was adopted to respect the first step of the decision process, which is to ensure that all cases of MSDs began, worsened, or recurred after the start of the current job. Our methodology did not allow us to assess nonoccupational risk factors during home duties or leisure. Although the effect of some medical conditions, such as obesity, diabetes mellitus, and thyroid disorders, on the risk of MSDs is controversial (1), these individual characteristics were taken into account in the classification of the MSDs to follow the third step of the decision process proposed by the criteria document. Overall, the prevalence rate of self-reported diabetes mellitus was comparable with that of the general population of the Pays de la Loire (2.2% versus 2.0%) (49). The prevalence of obesity (9% for both sexes) was slightly higher than in the general population (7.2%) (50), and no information was available on the prevalence of thyroid disorders in the general population. Our results demonstrate that the majority of workers diagnosed by OPs were highly exposed to the main work-related risk factors of MSDs, without association with any major potential individual risk factor. This was the case for 53% of men and 63% of women age <50 years and 55% of men and 56% of women age >50 years. No significant differences in work relatedness were observed between men and women, regardless of age. When the decision rules of the criteria document were applied, the majority of cases could be considered as work related. Nevertheless, the classification used is relatively brief and strong residual confounding effect by unmeasured past and current work-related risk factors may remain. Complementary analyses assessing more precisely the relationships between MSDs, occupational risk factors, age, sex, and some medical conditions that could increase the risk of MSDs are necessary. Despite these limitations, the study shows that a large number of MSD cases were probably attributable to work, which is consistent with results from studies of CTS in the general population (50, 51) and in the working population (15).
Our study shows the results of the first 2 years of a new sentinel network designed for the epidemiologic surveillance of MSDs in a working population. This type of system can play a significant role in informing government agencies and companies on the state of the current epidemic of MSDs.
Overall, approximately half of the workers experienced nonspecific musculoskeletal symptoms of the upper limbs during the previous year and approximately one-third exerienced these symptoms during the previous week. The physical examination performed by OPs confirmed that upper-limb MSDs were common in the working population. According to criteria similar to those currently applied in clinical practice in France, 13% of workers experienced upper-extremity MSDs. The level of work exposures was high for more than half of the workers, which demonstrates the need for prevention programs aimed at reducing the prevalence of MSDs and reducing the associated socioeconomic costs in most economic sectors.
- Top of page
- MATERIALS AND METHODS
- APPENDIX A
- APPENDIX B
We are grateful to the occupational physicians involved in the sentinel network: Drs. Abonnat, Banon, Bardet, Benetti, Becquemie, Bertin, Bertrand, Bidron, Biton, Bizouarne, Boisse, Bonamy, Bonneau, Bouguer, Bouguer-Diquelou, Bourut-Lacouture, Breton, Caillon, Cesbron, Chisacoff, Chotard, Compain, Coquin-Geogeac, Cordes, Couet, Coutand, Daniellou, Darcy, Davenas, De Lescure, Delansalut, Dupas, Evano, Fontaine, Frampas-Chotard, Guiller, Guillimin, Harinte, Harrigan, Hervio, Hirigoyen, Jahan, Joliveau, Jube, Kalfon, Laine-Colin, Laventure, Le Dizet, Lechevalier, Leclerc, Ledenvic, Leroux, Leroy-Maguer, Levrard, Levy, Logeay, Lucas, Mallet, Martin, Mazoyer, Meritet, Michel, Migne-Cousseau, Moisan, Page, Patillot, Pinaud, Pineau, Pizzala, Plessis, Plouhinec, Raffray, Roussel, Russu, Saboureault, Schlindwein, Soulard, Thomson, Treillard, and Tripodi.