Predictors of repeated sick leave in the workplace because of mental disorders

INTRODUCTION Mental disorders create a considerable burden to society. Previous studies have shown that productivity loss constitutes the largest proportion of the total societal burden. For depression and anxiety disorders, in particular, more than half of the associated productivity loss occurs in the workplace. Many previous studies have clarified the risk factors for the relapse/recurrence of mental disorders in health care settings. However, the risk factors for repeated sick leave among mental disorders prevalent in the workplace have not yet been adequately evaluated. OBJECTIVE The objective of this study was to investigate which variables could predict repeated sick leave for workers with a history of sick leave because of mental disorders. METHODS Data regarding 194 subjects employed at a manufacturing company were obtained. Mental disorders were defined as disorders listed in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV). The duration between the return to work (RTW) and the repeated sick leave was regarded as a dependent variable. The subjects' age at the RTW, sex, age at the time of employment, job tenure, diagnosis, number of previous sick leave days, duration of most recent sick leave, and employee rank were examined as explanatory variables. Univariate analyses using a log-rank test and a multivariate analysis using the Cox proportional hazard model were conducted. RESULTS The results of the univariate analyses showed that the number of previous sick-leave episodes was a significant predictor of repeated sick leave. A multivariate analysis revealed that age at RTW and the number of previous sick-leave episodes were significant variables. CONCLUSION Age and the number of previous sick-leave episodes appeared to be predictors of repeated sick leave. Therefore, effective intervention to prevent repeated sick leave for those with high risk is quite crucial. Analyses including various work-related factors with subjects from multiple companies should be conducted in future studies.