Introduction Mental health problems among schoolteachers and related psychosocial factors at schools are important issues, yet there has not been sufficient study in this area. In the present study, we collected 6 years of annual stress survey data on individuals and linked these data with application for long-term sick leave due to mental illness during those periods. We then examined the relationship between long-term sick leave and workplace psychosocial factors.

Methods First, we created a database of 5631 male and 7760 female schoolteachers (total: 13,391) by linking data from the annual stress surveys from 2010 to 2015 for each individual. The survey used a questionnaire composed of the Brief Job Stress Questionnaire, K6 questionnaire, and two items on organisational justice, six items on work factors specific to schoolteachers, two items on family-related concerns and medical consultation, and questions on present illnesses under treatment. After excluding recurrence cases from subjects with long-term (>90 days) sick leave due to mental illness during 2010–2015, we identified 189 cases. We then linked the above two datasets for analysis.

Result From 3 to 2 years prior to sick leave, a remarkable increase in quantitative and qualitative job demands and considerable decline in job control and support from colleagues were observed.

During that period, however, depressive symptoms did not conspicuously increase. Transfer of workplace or occurrence of family-related concerns was observed immediately prior to starting sick leave.

Conclusion We examined the longitudinal relationship between long-term sick leave due to mental illness and psychosocial factors in the workplace. Increased job demands, declining job control, and deteriorating relations with coworkers during 3 to 2 years before the sick leave should be considered as background factors. Transfer of workplace or occurrence of family-related concerns may also serve as trigger factors of long-term sick leave.