Association of major depression with return to work and changes in employment rate across sociodemographic municipality characteristics: A register-based cohort study

We examined whether the association between depression and return to work (RTW) varied by municipality characteristics.

In this nationwide register-based cohort study (2012-2018), adults aged 25-55 years with a first-time in- or outpatient contact for depression were matched 1:3 by birth year and sex to the general population. For comparison, parallel cohorts were established for first-time antidepressant users and individuals with cancer. Labour market affiliation was tracked from 1 year before to 3 years after diagnosis. Municipalities were characterized by urbanicity, unemployment level, and tax revenue. Among individuals outside the workforce, Cox proportional hazard regression estimated hazard ratios (HR) for RTW (early and late), comparing depression and reference groups, with interaction by municipality characteristics. Difference-in-difference analyses assessed absolute changes in the proportion in work, with parallel analyses for antidepressant use and cancer.

In the main analyses, HRs of RTW varied by urbanicity. The smallest difference (depression vs. reference group) was observed in capital municipalities (early HR 0.54, 95% confidence interval (CI) (0.51, 0.57)), late HR 0.72 (0.68, 0.75)) while larger differences were seen in upland (early HR 0.38 (0.35, 0.42)) or rural municipalities (late HR 0.59, 95% CI (0.55, 0.63)). Absolute workforce loss ranged from -9.6% (95% CI ( -10.7%, -8.4%)) in capital to -13.6% (-15.3%, -11.9%) in big city municipalities. A similar variation by urbanicity, though of smaller magnitude, was observed for antidepressant use but not for cancer.

The association between depression and RTW varied by municipality urbanicity, whereas no clear differences were observed for cancer.