Mehlum, Ingrid Sivesind PhD; Kristensen, Petter PhD; Veiersted, Kaj Bo PhD; Wærsted, Morten PhD; Punnett, Laura ScD
Socioeconomic and sex inequalities in health and specifically in work-related health are well documented. Musculoskeletal disorders (MSDs) in general and those that are work-related are more common among persons of lower than higher social class.1–3 They are generally also more prevalent among women than among men, especially in the neck and upper extremity, regardless whether symptoms or physical findings are assessed.4–7
Nevertheless, many of these data were generated using self-administered questionnaires. Self-reported musculoskeletal pain is a commonly used but somewhat contested outcome, considered to be more subjective and less valid than clinical diagnoses based on physical examinations.8 Self-report of work-related health problems is regularly used in national surveys in several countries, as well as the European Union, and in large population studies.5–7,9–11
If the threshold for reporting a health problem or the tendency to attribute it to the job is linked to socioeconomic position (SEP) or sex estimates of health inequality may be of questionable validity. Apparent differences could be artifacts of different perceptions or judgments. For example, if low-SEP subjects or women tend to overreport their symptoms or attribute work-relatedness to a greater extent than high-SEP subjects and men, respectively, the differences would be falsely inflated. Nevertheless, neither sex nor SEP differences seem to have been examined much in relation to these reporting issues.
Although physician assessment is no criterion standard, it may provide a reference against which self-assessment can be compared.12,13 Different proportions of agreement between physician assessment and self-assessment in population subgroups would suggest differential reporting tendencies.
A few studies have compared self-reported musculoskeletal symptoms with clinical diagnoses based on physical examinations.14 Some of these concern one sex primarily or exclusively or have not reported separate results for women and men. Others have made separate analyses but not presented the results in a way that facilitates direct comparisons. Comparable data have been presented in a handful of studies but without calculating sex differences.15–21 In post hoc analyses of one of these studies,15 Punnett and Bergqvist22 concluded that at the group level, women were considerably more likely than men to have upper extremity and neck discomforts “confirmed” by a diagnosis in the corresponding region. As far as we know, no one has compared self-reported musculoskeletal symptoms with clinical diagnoses between different SEP groups.
In a previous study,13 we compared self-reported work-relatedness of neck–shoulder and arm pain with physicians' assessments based on specific criteria and found that the physicians agreed somewhat more often with men than with women who believed their pain was work-related. To our knowledge, only two other studies comparing self-assessment and physician assessment of work-relatedness have been published, but sex or socioeconomic differences were not examined.10,23
In this study, we examined women and men of high and low SEP who reported pain in the neck and upper extremity. The overall objective was to examine the effect of sex and SEP on individuals' perceptions of pain and its work-relatedness. Self-reported pain was compared with clinical diagnoses assigned through physical examination, and workers' judgments of work-relatedness were compared with physicians' assessments. Physicians' clinical diagnoses and assessments of work-relatedness were based on specific criteria. This study was based on the data collected for another study.24
Participants and Data Collection
The study was part of the Oslo Health Study 2000–2001 (HUBRO), a cross-sectional population study conducted under the collaboration of the Norwegian Institute of Public Health (NIPH), the University of Oslo, and the Municipality of Oslo. All individuals in Oslo county, born in 1970, 1960, 1955, 1940 to 1941, and 1924 to 1925 were invited by letter to attend a health screening.25 Of the three youngest cohorts (n = 26,074; aged 30, 40, and 45 years), 10,712 attended the screening and 8594 (33%) returned an age-specific supplementary questionnaire, which included questions on work-related health problems.5
To evaluate the self-reported attribution of health problems to work, 668 subjects (aged 30, 40, and 45 years) were invited to further health examinations. All subjects reporting work-related eczema or respiratory symptoms in the past month (n = 508; 68% with pain in the neck–shoulder or more distal arm regions) were identified by the NIPH and asked for consent to be contacted by the National Institute of Occupational Health. In addition, 160 subjects were invited on the basis of their reported neck–shoulder pain or pain in the elbow–forearm–hand (arm pain). Of the 668 identified subjects, 302 gave their consent and were subsequently contacted; 268 of them were examined between September 2000 and September 2002. Thirty-four subjects were contacted but not examined mainly because they had less pain or did not want an examination. Before the health examination, the subjects completed questionnaires on musculoskeletal symptoms, their work-relatedness (“not caused by work” or “totally or partially caused by work”), occupation, and working conditions. This study is restricted to the 217 employed subjects reporting neck–shoulder pain or arm pain in the past month before the examination, 142 women and 75 men (Table 1).
Assessment of Diagnoses and Work-Relatedness
Criteria to establish clinical diagnoses and assess the work-relatedness of pain with respect to the subjects' present job were based on the work of Sluiter et al26 (hereafter referred to as the criteria document). Subjects reporting pain in the neck, shoulder, or arm at the examination or on at least 4 days during the past 7 days (“current pain”; n = 141) were classified as having a clinical diagnosis if the following criteria were met: (1) symptom criteria for the relevant clinical diagnoses according to the region of pain and (2) sign criteria on relevant provocative tests, thoroughly described, with photos, in the criteria document.26 Four of the most common clinical diagnoses in the neck and upper extremity were selected for the examination: (1) radiating neck complaints, (2) rotator cuff syndrome, (3) epicondylitis (lateral and medial), and (4) peritendinitis–tenosynovitis in the forearm–wrist, in four corresponding body regions: neck, shoulder, elbow, and forearm–wrist.
The physicians assessed work-relatedness in all the subjects reporting pain in the past month (“recent pain”; n = 217), whether or not they had a clinical diagnosis, based on the following criteria: (1) temporal relationship, whether the symptoms began, recurred, or worsened after the current job started; (2) occupational risk factors (physical, work organizational, and psychosocial factors) known to be specifically associated with MSD in the relevant body region (neck, shoulder–upper arm, elbow–forearm, or wrist–hand), listed in the criteria document and categorized into three risk zones (“unacceptable,” “not suitable,” and “acceptable”), and (3) whether nonoccupational causes were plausible. The occupational risk factors in the second criteria included computer or mouse work, repetitive work with too short recovery time, psychological demands, decision latitude, and social support, in addition to specific physical risk factors for each of the four upper extremity body regions. The physicians interviewed the participants about the temporal relationship and the risk factors and considered their answers carefully before making the final assessment on the basis of specific criteria for each of the four body regions. The assessment resulted in one of the three possible conclusions: “probably work-related,” “possibly work-related,” and “most likely not work-related.” The former two were collapsed, leaving two categories in the analyses, “work-related” and “not work-related,” to allow comparison with self-reported work-relatedness of pain, which was a dichotomous variable. For the same reason, the four body regions were collapsed into two regions. More details about the population and the evaluation procedures are given elsewhere.24
Socioeconomic position was based on the participants' present occupation, coded according to the Norwegian Standard Classification of Occupations, which uses the International Standard Classification of Occupations (ISCO-88). Occupations were classified according to the Erikson-Goldthorpe-Portocarero schema27 into eight categories, which were collapsed into high SEP (Erikson-Goldthorpe-Portocarero categories I, II, IIIa, and IV) and low SEP (Erikson-Goldthorpe-Portocarero categories IIIb, V, VI, and VII).
Data Analysis and Statistics
Among subjects with current pain, we compared the proportions with clinical diagnoses between high- and low-SEP subjects, and between women and men. High SEP and men, respectively, were the reference categories. Comparisons were made for each of the four body regions (neck, shoulder, elbow, and forearm–wrist), as well as for combined regions (neck–shoulder and elbow–forearm–wrist) and for all regions together. Proportions and differences in proportions were estimated and presented as percentages and differences in percentage points.
We compared self-reported work-relatedness of recent pain with physicians' assessments and calculated the proportion of cases judged as work-related by the participants with which the physicians agreed. Comparisons were made separately for neck–shoulder pain and arm pain. Differences in proportions were calculated between high- and low-SEP subjects and between women and men.
Analyses were performed using the statistical software packages StataSE 11 (www.stata.com, StataCorp LP, College Station, TX) and SPSS 18.0 for Windows (Statistical Products and Service Solutions, Inc, Chicago, IL). Age-adjusted percentage differences with their corresponding 95% confidence intervals were calculated using the BINREG procedure (binomial regression) in StataSE 11. Sex differences were also adjusted for SEP because the distribution of SEP was different in women and men. The study protocol was approved by the Norwegian Data Inspectorate and recommended by the regional committee for medical research ethics.
High and low SEP were similarly common in the total population, 53% and 47%, respectively; but low SEP was more frequent among men than among women, and there were small sex differences in the age distribution (Table 1).
Occupational risk factors, as classified in the criteria document,26 were generally more frequent among low-SEP subjects than high-SEP subjects, but those of high SEP had more computer or mouse work (Table 2). Physical risk factors and low social support were somewhat more frequent among men, while low job control (the lower quartile of the nine-item scale) seemed more common in women (neither adjusted for SEP).
Among the 141 subjects who reported current pain, a higher proportion was assigned clinical diagnoses in low-SEP workers than in high-SEP workers, for all body regions (Table 3). The adjusted difference was larger for pain in the elbow–forearm–wrist region (28 percentage points [pp]) than for pain in the neck–shoulder region (12 pp), largest for elbow pain (31 pp), and lowest for shoulder pain (4 pp). The socioeconomic pattern was similar for women and men for all regions combined, but the region-specific differences were somewhat larger in men. Not all results were statistically significant, but the social pattern was consistent.
The proportion with clinical diagnoses among the 141 subjects was generally higher in women than in men (Table 4), except among those with shoulder pain, for whom there was a small, not significant male excess (3 pp). The sex differences for the other three regions were quite large (17 to 23 pp) and statistically significant for two regions (neck and forearm–wrist).
Work-relatedness was assessed for all 217 participants. The neck–shoulder pain judged to be work-related by the subjects was more frequently assessed as work-related by the physicians in low-SEP workers (93%) than in high-SEP workers (51%) (Table 5). The corresponding values for arm pain were 89% and 70%, respectively. The adjusted socioeconomic differences were somewhat larger in men than in women. The social pattern was consistent and statistically significant for all outcomes.
Pain attributed to work by the subjects was somewhat more frequently assessed as work-related by the physicians in men than in women in the total population, but the adjusted sex differences did not quite reach statistical significance. Higher male proportions were, however, only found in the low-SEP group (97% vs 90% for neck–shoulder pain and 96% vs 82% for arm pain); among high-SEP individuals, the physicians agreed slightly more frequently with women's work attribution (48% in men vs 52% in women for neck–shoulder pain, and 69% vs 70% for arm pain).
In this general population sample of 217 employed middle-aged adults, the 141 with current musculoskeletal pain were more likely to receive clinical diagnoses if they were in low-SEP occupations than in high-SEP occupations. The proportion with clinical diagnoses was also generally higher in women than in men. Pain reported as work-related by the 217 subjects with recent pain was more frequently agreed to as work-related by physicians for low-SEP subjects than for high-SEP subjects, and somewhat more frequently in men than in women, however, only in the low-SEP group.
The low participation in the HUBRO in general and in this subset in particular may have involved self-selection of certain groups of individuals. This could not be examined directly because the original NIPH records were not linked to our questionnaires. There were twice as many women as men in this study, probably because women to a larger extent participated in the HUBRO5,28 and because neck–shoulder pain and arm pain are more prevalent among women. Low SEP was more frequent in this study than in the HUBRO for both sexes, particularly in men. This is likely explained, at least in part, by higher frequencies of musculoskeletal pain and thus invitation to this study among low-SEP subjects.3 Self-selection of individuals may cause less systematic error in the association measures of this comparative study than in the occurrence measures of a descriptive study, as has been discussed elsewhere.5,24,28 In addition, the low participation, particularly among men of high SEP, led to small numbers in the stratified analyses and thus uncertainty in some results.
We can only speculate whether the selective loss of high-SEP men could have influenced the estimated associations. If the high-SEP men who did not participate were different from the ones who did, they probably had lower proportions with clinical diagnoses, because many of those who were contacted but were not examined had less pain. If they had been included, the difference between high and low SEP in men and in the total population (Table 3) would probably have been larger, as would the difference between women and men (Table 4). Similarly, they probably had lower proportions assessed as work-related. The effect of this would be larger social differences among men and in the total population but a smaller (or reversed) difference between men and women (Table 5). This may indicate that, except for the sex comparisons of work-relatedness, our results are conservative.
The physicians were not blind to the subjects' assessments. Blinding would have been difficult in this study because the physicians had to take a symptom history to apply the diagnostic criteria. Nevertheless, employing standardized criteria for diagnoses and work-relatedness was intended to reduce subjectivity in the physicians' judgments and the interdependence of the assessments between subject and physician. Nonetheless, it cannot be ruled out that some differences in agreement for different subgroups of the population might have resulted from the physicians responding differently to some individuals than to others.
The outcome measures of this study were the proportion (percentage) with clinical diagnoses among subjects reporting current pain and the proportion with physician-assessed work-relatedness of pain among subjects reporting work-related recent pain. Although there is no criterion standard in these cases, physician assessment provides a standard against which other case definitions can be tested.12 Nevertheless, their limited sensitivity and specificity imply leeway for subjective interpretation by the examiner.29
The proportion of symptomatic individuals who receive clinical diagnoses will depend on the case definitions of both the symptoms (body region, recall time, and symptom frequency, intensity, and duration) and the diagnoses (which ones are included and the criteria for their assessment).8,30,31 In our study, only four clinical diagnoses were selected for the examination, and the criteria employed were rather strict, thus excluding common nonspecific disorders like tension neck syndrome and trapezius myalgia. This may have led to the underestimation of the proportion with clinical diagnoses. Women tend to have higher prevalences of nonspecific MSD than men;32 thus, sex differences might have been larger if certain nonspecific disorders were included. Whether there are differences between high- and low-SEP subjects regarding nonspecific MSD is less known and it might differ for different diagnoses and for different occupations within each SEP category.
When comparing symptoms and clinical diagnoses, we included only subjects reporting pain during the past week. Other investigators of this question have used various prevalence periods: the past week,8,30,33 the past month,17 or up to 12 months.30,33,34 Two studies found better agreement between symptoms and clinical diagnoses for current pain than for longer prevalence periods.30,33 Musculoskeletal symptoms are often intermittent and episodic, especially in the early stages.35 Symptom reports during a long period may, therefore, not correspond well to clinical diagnoses at one specific time point.17,30
Interpretation of Results: Agreement Between Symptoms and Examinations
Sex differences in pain perception or pain expression have been suggested as explanations of sex differences in pain prevalences.36,37 Nevertheless, biological and psychological differences would be less likely to explain socioeconomic differences in symptom prevalences. Our results indicate that the excess of symptoms found in many studies among women and low-SEP subjects, using self-reported data, is not due to overreporting.
Punnett and Bergqvist22 concluded that, as a group, women were less likely than men to report symptoms without physical examination findings. Ratios between diagnoses and symptoms calculated at the group level from data given in other studies of different occupational groups15–21 showed similar results (summary in Table 6). Ratios were generally higher in women than in men, except among fish processing workers20 and for shoulder pain among car assembly workers19 and air traffic controllers.16 Compiled epidemiological data from 43 different occupational groups, including two of the referred studies with higher male ratios,16,20 showed higher ratios in women, except for pain in elbow–hand in occupational groups with repetitive or constrained industrial work.21
Similarly, Gerr et al38 found stronger associations between women and musculoskeletal outcomes for diagnoses than for symptoms. Andersen et al39 found that women and men had an equal risk for onset neck, shoulder, or both pain symptoms, but women had a higher risk for being a future clinical case. Nevertheless, in all of these prior studies, analyses were done at a group level. The present study found that at an individual level, women who reported pain were more likely than men to have physical examination findings. This stronger agreement might imply either higher validity of women's self-reported pain or perhaps a qualitative difference in physician assessment of women versus men.40 Nevertheless, the latter will usually result in more frequent nonspecific symptom diagnoses in women than in men.41
In this study, clinical diagnoses were assigned more frequently in low-SEP occupations than in high-SEP occupations among subjects reporting pain. We have found only one study that has analyzed data on self-reported musculoskeletal symptoms and physical examination findings in different socioeconomic groups, but focusing more narrowly on arthritis.42 Findings and symptoms ratios calculated from the data given in that study revealed a clear socioeconomic gradient, with approximately 25% higher values in the lowest social class than in the highest social class, both according to years of education and annual family income.42
A few studies have compared symptoms and diagnoses in different occupational groups. Nordander et al found no differences in complaints between male fish processing workers and workers with more varied work (caretakers, municipal employees working in parks/gardens or repairing and maintaining equipment/machines), whereas clinical diagnoses were three times more common among the fish processing workers.20 In their study using compiled data from 43 different occupational groups, they also found a higher proportion with clinical diagnoses in occupations with repetitive or constrained industrial work than in groups with varied or mobile work (Table 6).21 They suggested that this could reflect differences in the severity of disorders.20 In women, the pattern was somewhat different, with higher prevalences of both symptoms and diagnoses in fish processing workers and occupations with repetitive or constrained industrial work, respectively, than in the other groups of workers.20,21 Nevertheless, the diagnosis-to-symptom ratios were still higher in occupations with the more industrial type of work. Whether the different groups of workers in these studies belonged to different socioeconomic strata is uncertain.
One reason for higher proportions with clinical diagnoses in low-SEP subjects and among women could be that they have more severe pain, as suggested by Nordander et al.20 Wang et al17 found that subjects with physical findings of MSDs had higher mean pain frequencies and intensities.17 Baron et al30 found clear trends between the reported severity on each of the three scales, pain frequency, duration, and intensity, and the prevalence of physical examination findings. Differences in job exposure could lead to differences in the severity of pain.20 Occupational exposures explain a substantial proportion of the prevalence differences in MSD between high- and low-SEP subjects1–3,43 and are related to both symptoms and clinical diagnoses.35,44 A different but possibly related question is whether subjects with higher exposures are more likely to develop the types of symptoms that correspond to existing physical examination strategies, while individuals with less severe symptoms might fit less neatly into preexisting diagnostic categories.45
Another possible reason for higher proportions of clinical diagnoses in low-SEP subjects and among women could be that these groups have a higher threshold for reporting pain. People may experience their symptoms relative to the level of symptoms among their workmates.20 If everyone has some pain, this becomes a normal state, not worth reporting unless it gets severe. They may also become used to symptoms of long duration, because they have to live with those46 and may not consider it worthy of reporting.17 Cultural differences could also be an explanatory factor. The ratio between physical examination findings and reported pain calculated at the group level from data in a study by Wang et al17 was considerably higher among Asian garment workers than among Hispanic garment workers for arm–forearm pain (0.35 vs 0.15) and hand–wrist pain (0.43 vs 0.19); the difference was smaller for neck–shoulder pain (0.25 vs 0.20). In summary, interpretation of the differences in agreement between symptoms and examination findings or clinical diagnoses is a complex problem.
Interpretation of Results: Assessment of Work-Relatedness
Pain judged to be work-related by the subjects was more frequently assessed as work-related by the physicians in low-SEP subjects than in high-SEP subjects. We used occupational class as an indicator of SEP. Other indicators (education, income) were not available. Besides, for socially patterned exposures that have etiological effects specific to particular stages of the life course, it has been recommended to use SEP indicators that are related to these stages.47 In our study, occupation would thus be a logical choice. In some studies, occupation is used as a proxy for exposure. Occupational risk factors tend to be more prevalent in low-SEP groups,1,2 as we also found in our study. Nevertheless, this would not necessarily explain the SEP differences in physician agreement among subjects reporting work-related pain. The difference was largest for neck–shoulder pain. According to the criteria document,26 computer work is considered a risk factor only for pain in the wrist–hand region, and physical risk factors are emphasized more than the psychosocial ones. We do not have information about what occupational factors the subjects were attributing their pain to. High-SEP subjects may have attributed their neck–shoulder pain to computer work or stress, and this may explain the larger SEP differences in physician agreement for neck–shoulder pain than for arm pain.
Physicians were more likely to agree with workers about work-relatedness in men than in women but only in the low-SEP group; thus, there was effect modification between sex and SEP. Although occupational physical risk factors, according to the criteria document,26 were more frequent among men, job exposures may still be an important explanatory factor for excess musculoskeletal morbidity among women and a reason for their attribution of pain to work. Force-demanding tasks may be relatively more strenuous for women than for men.19 Workplace design factors may lead to sex differences in working technique.48 The criteria do not differentiate according to sex and may underestimate the occupational risk in women. These factors may affect low-SEP women more than high-SEP women and might explain effect modification between sex and SEP regarding physicians' agreement on work-relatedness. A related issue is that women spend more time on household activities than men (double exposure) and have less time for personal recovery and physical exercise,49 which may make them more vulnerable to the same levels of physical work exposure. Household activities resemble more the work exposures of low-SEP women (heavy, monotonous, and repetitive work) than of high-SEP women, which may be another explanation for the sex–SEP effect modification. Others have called attention to potential sex bias in ergonomic exposure assessment,50 but that should not be a concern here because each person was interviewed individually about his or her job demands.
According to the criteria document,26 a physical risk factor is either “unacceptable” (above the risk level with respect to duration per workday) or “acceptable” (not present or below the risk level). In reality, it is not “either/or”; rather, the longer the duration, the greater the risk, both when it comes to hours per workday and to months or years of exposure (cumulative duration). Thus, exposure levels below the risk levels of the criteria may not be risk-free, especially if combined with other occupational risk factors or in individuals who are more vulnerable, for example, because of a preexisting condition or a high load outside the work situation.
Again, we cannot rule out (subconscious) bias among physicians in their recognition of occupational exposures in certain groups of patients. Workers might more readily recognize their own relevant exposures than do physicians. Nevertheless, the physicians who performed the medical examinations were experienced specialists in occupational medicine and used the same criteria for all. It seems a more likely explanation simply that there are more etiological exposures to be identified in low-SEP groups and especially that these workers have less varied tasks, so they are more likely to exceed the exposure duration criteria.
We found that among workers who reported musculoskeletal pain, the proportions assigned with clinical diagnoses were higher in low-SEP subjects than in high-SEP subjects, and generally higher in women than in men. Thus, the excess musculoskeletal morbidity found in low-SEP subjects and among women, using self-reported data, cannot be explained by a lower threshold for reporting pain. Quite the contrary, the threshold might be higher, and thus, the health inequalities might be larger than these studies usually show.
Furthermore, pain judged as work-related by participants was more frequently assessed as work-related by physicians for low-SEP subjects than for high-SEP subjects, and somewhat more frequently in men than in women, however, male excess was only present in the low-SEP group. Thus, socioeconomic differences in work-related pain might in reality be larger than shown in surveys using self-reported data. On the contrary, in surveys showing higher prevalences of work-related pain among women, the sex differences might actually be smaller or even reversed, but only in low-SEP groups.
Nevertheless, this conclusion is based on the assumptions that our criteria equally capture clinical diagnoses and occupational risk factors in the different groups and that the same criteria apply for women and men. Studies of other disease categories or other age or cultural groups may yield different results and need to be performed. Studies of these research questions in population-based samples of people with and without pain are also needed. The results of this study indicate that differential misclassification of pain and its work-relatedness may be a problem in studies and surveys using self-reported data, and this should be kept in mind when interpreting such data.
The primary data collection was conducted as part of the HUBRO in collaboration with the NIPH, which also invited the subjects selected for this study.
1. Melchior M, Roquelaure Y, Evanoff B, et al. Why are manual workers at high risk of upper limb disorders? The role of physical work factors in a random sample of workers in France (the Pays de la Loire study). Occup Environ Med. 2006;63:754–761.
2. Aittomäki A, Lahelma E, Rahkonen O, Leino-Arjas P, Martikainen P. The contribution of musculoskeletal disorders and physical workload to socioeconomic inequalities in health. Eur J Public Health. 2006;17:145–150.
3. Mehlum IS, Kristensen P, Kjuus H, Wergeland E. Are occupational factors important determinants of socioeconomic inequalities in musculoskeletal pain? Scand J Work Environ Health. 2008;34:250–259.
4. Punnett L, Herbert R. Work-related musculoskeletal disorders: is there a gender differential, and if so, what does it mean? In: Goldman MB, Hatch MC, eds. Women and Health. New York, NY: Academic Press; 2000:474–492.
5. Mehlum IS, Kjuus H, Veiersted KB, Wergeland E. Self-reported work-related health problems from the Oslo Health Study. Occup Med (Lond). 2006;56:371–379.
6. Räsänen K, Notkola V, Husman K. Perceived work conditions and work-related symptoms among employed Finns. Soc Sci Med. 1997;45:1099–1110.
8. Perreault N, Brisson C, Dionne C, Montreuil S, Punnett L. Agreement between a self-administered questionnaire on musculoskeletal disorders of the neck-shoulder region and a physical examination. BMC Musculoskeletal Disord. 2008;9:34.
9. Parent-Thirion A, Fernández Macías E, Hurley J, Vermeylen G. Fourth European Working Conditions Survey. Dublin, Ireland: European Foundation for the Improvement of Living and Working Conditions; 2007.
10. Jones JR, Hodgson JT, Clegg TA, Elliott RC. Self-Reported Work-Related Illness in 1995. Results From a Household Survey. Norwich, England: Her Majesty's Stationery Office; 1998.
11. Health and Safety Executive. Self-Reported Work-Related Illness and Workplace Injuries in 2008/09. Results From the Labour Force Survey. United Kingdom: National Statistics Publication; 2010.
12. Coggon D, Martyn C, Palmer KT, Evanoff B. Assessing case definitions in the absence of a diagnostic gold standard. Int J Epidemiol. 2005;34:949–952.
13. Reitsma JB, Rutjes AWS, Khan KS, Coomarasamy A, Bossuyt PM. A review of solutions for diagnostic accuracy studies with an imperfect or missing reference standard. J Clin Epidemiol. 2009;62:797–806.
14. Lenderink AF, Zoer I, van der Molen HF, Spreeuwers D, Frings-Dresen MH, Van Dijk FJ. Review on the validity of self-report to assess work-related diseases. Int Arch Occup Environ Health. 2012;85:229–251.
15. Bergqvist U, Wolgast E, Nilsson B, Voss M. Musculoskeletal disorders among visual display terminal workers: individual, ergonomic, and work organizational factors. Ergonomics. 1995;38:763–776.
16. Arvidsson I, Arvidsson M, Axmon A, Hansson GÅ, Johansson CR, Skerfving S. Musculoskeletal disorders among female and male air traffic controllers performing identical and demanding computer work. Ergonomics. 2006;49:1052–1067.
17. Wang PC, Rempel DM, Hurwitz EL, Harrison RJ, Janowitz I, Ritz BR. Self-reported pain and physical signs for musculoskeletal disorders in the upper body region among Los Angeles garment workers. Work. 2009;34:79–87.
18. Nordander C, Ohlsson K, Balogh I, et al. Gender differences in workers with identical repetitive industrial tasks: exposure and musculoskeletal disorders. Int Arch Occup Environ Health. 2008;81:939–947.
19. Zetterberg C, Forsberg A, Hansson E, et al. Neck and upper extremity problems in car assembly workers. A comparison of subjective complaints, work satisfaction, physical examination and gender. Int J Ind Ergon. 1997;19:277–289.
20. Nordander C, Ohlsson K, Balogh I, Rylander L, Pålsson B, Skerfving S. Fish processing work: the impact of two sex dependent exposure profiles on musculoskeletal health. Occup Environ Med. 1999;56:256–264.
21. Nordander C, Ohlsson K, Åkesson I, et al. Risk of musculoskeletal disorders among females and males in repetitive/constrained work. Ergonomics. 2009;52:1226–1239.
22. Punnett L, Bergqvist U. Musculoskeletal disorders in visual display unit work: gender and work demands. Occup Med. 1999;14:113–124, iv.
23. Plomp HN. Employees' and occupational physicians' different perceptions of the work-relatedness of health problems: a critical point in an effective consultation process. Occup Med (Lond). 1993;43(suppl 1):S18–S22.
24. Mehlum IS, Veiersted KB, Wærsted M, Wergeland E, Kjuus H. Self-reported versus expert-assessed work-relatedness of pain in the neck, shoulder, and arm. Scand J Work Environ Health. 2009;35:222–232.
25. Norwegian Institute of Public Health. The Oslo Health Study: About the Study. Oslo, Norway: Norwegian Institute of Public Health; 2008. Available at: http://www.fhi.no/hubro-en
. Accessed January 19, 2013.
26. Sluiter JK, Rest KM, Frings-Dresen MH. Criteria document for evaluating the work-relatedness of upper-extremity musculoskeletal disorders. Scand J Work Environ Health. 2001;27(suppl 1):1–102.
27. Erikson R, Goldthorpe JH. The Constant Flux. A Study of Class Mobility in Industrial Societies. Oxford, England: Clarendon Press; 1992.
28. Søgaard AJ, Selmer R, Bjertness E, Thelle D. The Oslo Health Study: the impact of self-selection in a large, population-based survey. Int J Equity Health. 2004;3:3.
29. Viikari-Juntura E. Interexaminer reliability of observations in physical examinations of the neck. Phys Ther. 1987;67:1526–1532.
30. Baron S, Hales T, Hurrell J. Evaluation of symptom surveys for occupational musculoskeletal disorders. Am J Ind Med. 1996;29:609–617.
31. Mehlum IS. Work-related health problems in the population. Impact of working conditions on health and on social inequalities in musculoskeletal pain among Oslo citizens aged 30–45 years [PhD thesis]. Oslo, Norway: University of Oslo; 2010.
32. Roquelaure Y, Ha C, Leclerc A, et al. Epidemiologic surveillance of upper-extremity musculoskeletal disorders in the working population. Arthritis Rheum. 2006;55:765–778.
33. Björkstén MG, Boquist B, Talbäck M, Edling C. The validity of reported musculoskeletal problems. A study of questionnaire answers in relation to diagnosed disorders and perception of pain. Appl Ergon. 1999;30:325–330.
34. Descatha A, Roquelaure Y, Chastang JF, et al. Validity of Nordic-style questionnaires in the surveillance of upper-limb work-related musculoskeletal disorders. Scand J Work Environ Health. 2007;33:58–65.
35. Punnett L, Wegman DH. Work-related musculoskeletal disorders: the epidemiologic evidence and the debate. J Electromyogr Kinesiol. 2004;14:13–23.
36. Hooftman W. Gender differences in work-related risk factors for musculoskeletal symptoms and absenteeism [PhD thesis]. Amsterdam, the Netherlands: VU University Medical Center; 2006.
37. van Wijk CMTG, Kolk AM. Sex differences in physical symptoms: the contribution of symptom perception theory. Soc Sci Med. 1997;45:231–246.
38. Gerr F, Marcus M, Ensor C, et al. A prospective study of computer users: I. Study design and incidence of musculoskeletal symptoms and disorders. Am J Ind Med. 2002;41:221–235.
39. Andersen JH, Kaergaard A, Mikkelsen S, et al. Risk factors in the onset of neck/shoulder pain in a prospective study of workers in industrial and service companies. Occup Environ Med. 2003;60:649–654.
40. Hølge-Hazelton B, Malterud K. Gender in medicine—does it matter? Scand J Public Health. 2009;37:139–145.
41. Hamberg K, Risberg G, Johansson EE, Westman G. Gender bias in physicians' management of neck pain: a study of the answers in a Swedish national examination. J Womens Health Gend Based Med. 2002;11:653–666.
42. Cunningham LS, Kelsey JL. Epidemiology of musculoskeletal impairments and associated disability. Am J Public Health. 1984;74:574–579.
43. Boyer J, Galizzi M, Cifuentes M, et al. Ergonomic and socioeconomic risk factors for hospital workers' compensation injury claims. Am J Ind Med. 2009;52:551–562.
44. Punnett L, Fine LJ, Keyserling WM, Herrin GD, Chaffin DB. Shoulder disorders and postural stress in automobile assembly work. Scand J Work Environ Health. 2000;26:283–291.
45. Gold JE, d'Errico A, Katz JN, Gore R, Punnett L. Specific and non-specific upper extremity musculoskeletal disorder syndromes in automobile manufacturing workers. Am J Ind Med. 2009;52:124–132.
46. 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 of symptoms. Int Arch Occup Environ Health. 1999;72:395–403.
47. Galobardes B, Shaw M, Lawlor DA, Lynch JW, Davey Smith G. Indicators of socioeconomic position (part 1). J Epidemiol Community Health. 2006;60:7–12.
48. Dahlberg R, Karlqvist L, Bildt C, Nykvist K. Do work technique and musculoskeletal symptoms differ between men and women performing the same type of work tasks? Appl Ergon. 2004;35:521–529.
49. Strazdins L, Bammer G. Women, work and musculoskeletal health. Soc Sci Med. 2004;58:997–1005.
50. Messing K, Punnett L, Bond M, et al. Be the fairest of them all: challenges and recommendations for the treatment of gender in occupational health research. Am J Ind Med. 2003;43:618–629.
Copyright © 2013 by the American College of Occupational and Environmental Medicine