The main finding in this study was that working women (WW) with FM displayed better ratings than nonworking women (NWW) with FM in terms of pain, fatigue, stiffness, depression, disease specific health status and physical health related quality of life, which represent body functions and overall health status.
Physical capacity did not differ significantly between WW and NWW in terms of performance-based tests (see Table 2) where both groups presented lower capacity than the average population [20, 21]. This supports earlier studies showing impaired body function in women with FM [4, 29]. However, the physical work demands might influence the work ability in persons who have an impaired physical capacity. Earlier studies have reported the importance of the work environment in women with FM [8, 11, 13, 30–32] and in other rheumatic diseases .
The number of pain localizations was significantly lower in WW than in NWW and global pain (FIQ pain) was significantly milder in WW than in NWW (see Table 2). The mean pain (FIQ pain) of WW was well above 50 (0–100), which corresponds to the average pain level in previous studies of FM . Mean pain was above 75 (0–100) in NWW, which corresponds to the ratings of severely afflicted patients with FM . FIQ pain was found to be the only independent explanatory factor for work in this study. Pain has previously been found to be a critical factor for work in rheumatic diseases [10, 35]. Our results indicate that women with FM having moderate pain generally could be expected to work. Some women appear to be able to work despite severe pain, which raises the question if there are workplace related factors that support their ability to work [32, 36]. The influence of work related factors on work ability in FM need to be further studied.
Global fatigue (FIQ fatigue) was found to be significantly lower in WW than in NWW as well as physical fatigue (MFI-20), reduced activity (MFI-20), and mental fatigue (MFI-20) (see Table 2). Fatigue has previously been found to be an important factor for work disability in rheumatic diseases . However, our results showed severe global fatigue (FIQ fatigue) with mean ratings of over 70 (0–100)  also in WW, indicating that fatigue might not be a critical factor for work disability.
Depression was rated significantly lower in WW than in NWW in the HADS, assessing depression. This supports the results of an earlier study on work disability in FM reporting the negative impact of depression symptoms on work ability .
WW displayed a significantly better disease specific health status (FIQ total, eight-item) than NWW (see Table 2). This supports the results from an earlier study on work disability in FM where the FIQ total score was found to predict work disability . Physical health-related quality of life (SF-36 PCS) was significantly higher in WW than in NWW (see Table 2), which is in line with a previous study of FM . However, the quality of life of workers in our population, assessed by SF-36, was very low as compared to a national sample . Impaired health status assessed by SF-36 has earlier been associated with work disability in rheumatoid arthritis (RA) , systemic lupus erythematosus (SLE)  and musculoskeletal pain .
The theory of the healthy worker effect suggests that healthier individuals are more likely to remain in the workforce . On one hand, this agrees well with the results of the present study. On the other hand, work is an important factor for health status in women in general [43, 44] and in women with FM . Further studies are needed to explore if working women with FM maintain their health status, or if it deteriorates over time.
The main strength of the present study is the integration of physical, social and psychological assessments including subjective ratings as well as clinical assessments and performance-based tests of physical capacity. About 40% of the patients in this study worked part-time or full-time which is in line with international reports of work ability in FM . No significant differences were found in age, symptom duration, cohabitation, ethnicity, education, pharmacological treatment, mean income in the area of residence and social support, i.e. personal and environmental barriers or facilitators for health . A limitation of this study is the cross sectional design which does not allow analyses of cause and effect. Also, the specific demands in work were not reported in the study and need further investigation.