Journal of Epidemiology (Jun 2022)

Overtime Work and the Incidence of Long-term Sickness Absence Due to Mental Disorders: A Prospective Cohort Study

  • Yosuke Inoue,
  • Shuichiro Yamamoto,
  • Andrew Stickley,
  • Keisuke Kuwahara,
  • Toshiaki Miyamoto,
  • Tohru Nakagawa,
  • Toru Honda,
  • Teppei Imai,
  • Akiko Nishihara,
  • Isamu Kabe,
  • Tetsuya Mizoue,
  • Seitaro Dohi

DOI
https://doi.org/10.2188/jea.JE20200382
Journal volume & issue
Vol. 32, no. 6
pp. 283 – 289

Abstract

Read online

Background: Although previous research has focused on the association between long working hours and several mental health outcomes, little is known about the association in relation to mental health-related sickness absence, which is a measure of productive loss. We aimed to investigate the association between overtime work and the incidence of long-term sickness absence (LTSA) due to mental disorders. Methods: Data came from the Japan Epidemiology Collaboration on Occupational Health Study (J-ECOH). A total of 47,422 subjects were followed-up in the period between April 2012 and March 2017. Information on LTSA was obtained via a study-specific registry. Baseline information was obtained at an annual health checkup in 2011; overtime working hours were categorized into <45; 45–79; 80–99; and ≥100 hours/month. Results: During a total follow-up period of 211,443 person-years, 536 people took LTSA due to mental disorders. A Cox proportional hazards model showed that compared to those with less than 45 hours/month of overtime work, those with 45–79 hours/month were at a lower risk of LTSA due to mental health problems (hazard ratio [HR] 0.63; 95% confidence interval [CI], 0.56–0.71) while those with overtime work of ≥100 hours/month had a 2.11 (95% CI, 1.12–3.98) times higher risk of LTSA due to mental health problems. Conclusion: Engaging in excessive overtime work was linked with a higher risk of LTSA due to mental health problems while the lower risk observed among individuals working 45–79 hours/month of overtime work might have been due to a healthy worker effect.

Keywords