Indicators relevant to the epidemiology of MSD are scarce [1]. However, studies of sickness absence (SA) and permanent work-disability might outline the functional consequences of MSDs when cause-specific information is present [16]. Within universal social insurance systems, like in Sweden and Norway, population-based studies are feasible, since most women have paid work [7,17-20].

Gender differences are a key feature of the MSD epidemiology. A female excess in SA is found in countries with different social insurance systems and different levels of sickness absenteeism [7,17-19], even though some studies differ [21]. This corresponds to findings from epidemiological surveys of MSDs, in the general population [22,23] or in occupational samples [24,25] which have consistently found a higher prevalence among women [26]. A gender and site specific analysis of data from the Dutch cross sectional DMC3 study [22] showed a "slightly different" pattern in women and men with regard to the association of musculoskeletal pain with age, education, use of health care, disability and work-leave variables [27]. Behavioural factors and hazardous workplaces lead to more fractures and other injuries among men. Several studies on predictors of chronicity in MSD have shown that women with MSDs might have an increased risk of chronicity [13,15,17]. However, this has not been a subject for in-depth assessment.

Recently, investigators from different countries have discussed the need for gender specific studies of work related MSDs, without specifically mentioning the aspect of chronicity [28-30]. A main issue here is how statistical analyses should be performed, with adjustment or stratification for gender. Data from two large epidemiological studies from the US [29] and Canada [30] were analyzed both with adjustment for gender and separately for women and men in order to identify the full range of associations. Thus, in this study we decided to assess the gender difference with both approaches.

MSDs dominate social insurance expenditure in the Nordic countries, with respect to both short-term and permanent work disability. In the period 1994-2003, 150,374 women and 129,612 men obtained disability pension (DP) in Norway. Forty-one percent of these women and 28% of the men were certified with a MSD [31] whereas the incidence of DP caused by non-MSD was nearly similar. Thus, the higher prevalence of MSD among women has led to a "gender-gap" in the use of social insurance.

MSD might have different meanings and consequences for women and men and there are several explanations of the gender difference in MSD, with varying degrees of evidence [32] which may also have relevance with respect to chronicity: The exposure-theory claims that women generally have poorer working conditions, whereas the vulnerability theory claims that women are more strongly affected by pathogenic factors. Biological factors might explain the increased vulnerability and a higher prevalence of MSDs like osteoporosis, rheumatoid arthritis and fibromyalgia in women [33,34]. Mental/physical co-morbidity [35] and the interplay between pain and depression contribute to the functional consequences of MSDs, since depression is more common in women [36] and may affect pain differently [37] There are also differences in health beliefs and help seeking [27,38,39]. Gender differences in self rated health has been shown to vary in countries with different welfare state models, with the largest female/male difference in the Nordic countries [40]. The double-burden theory implies that the combined exposure to stressors in the family and at work might lead to musculoskeletal problems and SA [41,42] which might have different consequences depending on the extent of gender inequality [43,44]. Finally, socioeconomic variables play an important role in musculoskeletal morbidity [17,18,23,45,46] and different positions in society and workplace hierarchies might therefore also contribute to gender differences. Low socioeconomic status, education and income are both exposures and might lead to increased vulnerability. Adjusting for socioeconomic variables might reduce the gender differences significantly [17,18]. When also including workplace factors, the gender difference might "disappear" completely [21].

When the full sample was analyzed together the risk or HR for obtaining DP related to female gender was 1.34 (95% CI 1.28-1.40) when adjusted only for age. When income and education were included, the HR decreased to 1.21 (1.15-1.27), and significantly more when adjusted for part-time work and family obligations 1.09 (1.03-1.14). However, only after including the diagnostic distribution the gender difference was completely "removed". In this final model the risk of DP according to diagnostic subgroups is also presented after adjustment for all socio-demographic variables (table ​(table33).

The gender-stratified analysis showed that age above 50 and very low income, were more strongly linked to the risk of DP among the men, when adjusted for all other variables. Higher education had a similar protective effect for both genders. In both genders, those working part time and those without children had an increased risk of DP. Among women, the prevalence of myalgia/fibromyalgia was 2.5 times higher than among men and the risk of DP for women with this diagnosis was even higher than among those with osteoarthrosis. Among men, neck problems carried a higher DP risk than back problems. Cases with fractures or other injuries, 22% of the male and 10% of the female sample had a significantly lower risk than all other groups for both genders (table ​(table44).