This longitudinal study showed that sickness absence due to LBP—with an annual incidence of 14% and a recurrence of 41%—was the most frequent single cause of absence among shipyard workers. Individual characteristics and prior absenteeism influenced the decision to take sick leave due to LBP. Prior sick leave due to LBP partly captured the effects of work-related factors. RTW was largely determined by a history of diagnosed herniated disc and musculoskeletal co-morbidity.
Our study considered various factors that may influence sickness absence due to LBP. A particular strength of this study was that the information about sickness absence was reliable, because the diagnosis was taken on return to work. Another strength was that all subjects worked in the same company and were comparable for several factors, such as cultural and socioeconomic factors. On the other hand, this may have hampered the influence of some work-related risk factors since there appears to be a limited contrast in some of the risk factors. The exposure information was self-reported, although objective measurements on individual characteristics like BMI took place. Since aspects of physical load were measured crudely on a four-point scale, these variables will lack discriminatory power [13, 19]. A substantial part of the study population with a MSD had gone on sick leave in the 12 months prior to the study. When work-related physical and psychosocial factors determine the risk on sick leave, it is expected that the inclusion on prior sick leave in the analysis will compromise the power of the current study to demonstrate the effect of work-related factors. In our study, when we excluded from the analysis all persons with previous sick leave due to LBP, work-related factors exhibited an influence of borderline significance. It seems that prior sick leave due to LBP partly captures the effect of work-related physical and psychosocial factors. Another disadvantage of this study is that psychological factors were not addressed and thus their potential influence on absenteeism could not be established. Another feature was that we did not examine support at work in the analysis. A previous study in the same population showed that 90% reported a highly supportive work environment and, thus, this has limited discriminatory power .
Of workers with LBP at baseline, the 1 year rate of absence due to LBP (23.7%) in our study lies near of those reported in other cohort studies [7, 33]. However, these studies were based on self reports rather than sickness absence registries and some bias may have occurred.
In the proportional hazard analysis, several factors were evaluated for their influence on the probability of occurrence of sickness absence due to LBP. Prior absenteeism was the most important prognostic factor, which has also been shown in other studies [7, 29]. Previous sick leave due to LBP reflects actually recurrence while the influence of prior absenteeism due to other reasons most likely reflects coping behaviors. In total, 8.4% of those with no previous LBP took a sick leave due to LBP while recurrence was 23.7%, higher than the rate reported in another study . The few studies that have investigated risk factors for recurrence of sick leave due to LBP found various work related and psychological factors as predictors [24, 26, 32].
Low educational level was also an important predictor. It might reflect to a great extent (as surrogate) differences in work activities and working conditions between job titles. Those workers with the lowest education also reported a higher physical load and, hence, it was not possible to disentangle the separate effects of education and physical load. This may explain why physical load factors did not appear to have a significant influence on the occurrence of sick leave due to LBP. Furthermore, the impact of education might also reflect differences in coping strategies and work motivations .
Workers with night shift work had significantly less sickness absence due to LBP. The possible explanation for this finding is a self-selection process since those engaged in night shift were rather few, well paid, and not easily being replaced. This aspect of a healthy worker effect has been observed before among shift workers .
Living alone was found to predict future absenteeism, while another study showed contradictory results . In our study population this may be due to the fact that workers on sick leave only received 50% compensation of their wage and, hence, may put their families under financial strain. Thus, living alone not only reflects marital status but also a different social and economic environment. Perhaps a more supportive environment at home may be a reason for earlier RTW.
Other individual characteristics such as age, height, weight, smoking, and duration of employment, were not predictive for low back complaints leading to absence from work. In contrast with other studies [9, 31, 32], work-related factors did not exhibit a significant influence on absenteeism due to LBP in the whole study population, but when we excluded from the analysis subjects with a previous sick leave due to LBP, work-related physical, psychosocial factors, and need for recovery had an influence of borderline significance. It seems that prior sick leave due to LBP partly captured the effects of work-related physical and psychosocial factors. Health-related aspects have been reported to be more strongly associated with sick leave than work-related aspects, which was reflected in our findings that prior absenteeism (health problems and coping behaviors) predicted the occurrence sickness absence.
On average, about 90% returned to work within 2 weeks. This is a higher RTW rate than reported in the literature [12, 28]. A reason for this may be the low compensation (50%). It is known that the sickness absence rate in Greece is among the lowest in Europe. While this is partly explained by the low social insurance benefits and the relatively high unemployment rate, other factors like the high percentage of permanent full-time employees and the high prevalences of reported morbidity (stress, MSDs, etc.) do not explain this large difference with other European populations .
In this study it was obvious that workers were not always fully recovered at return to work, given the high recurrence of sick leave due to LBP (43% in 1 year). It has been reported before that most workers experience residual low back complaints after returning to work . This high return to work within the first weeks combined with the fact that in most cases employees suffered residual complaints when returned to work suggests that additional management after RTW may be of importance [24, 34].
Only a history of herniated disc and comorbidity with musculoskeletal complaints were significant prognostic factors for RTW. It is known that ‘‘specific’’ (lesions of vertebrae and discs) causes account for the most long-term absences compared to a ‘‘nonspecific’’ (pain, sprains, and strains) origin of complaints .
In the present study, having concurrent LBP and hand/wrist pain was associated with a higher rate of return after a first spell of sickness absence. The opposite hold for concurrent shoulder/neck pain but this did not reach a statistically significant level. In another study having concurrent LBP and shoulder/neck pain was associated with a higher risk for sickness absence and also long-term sickness absence . These findings suggest that in research on risk factors for sick leave and prognostic factors for RTW musculoskeletal comorbidity has to be taken into consideration.
Our study showed that individual and job characteristics (living alone, night shift, lower education, sick leave or care seeking during the last 12 months) influenced the decision to take sick leave due to LBP. An increased awareness of those frequently on sick leave and additional management after return to work may have a beneficial effect on the sickness absence pattern.