Factors associated with returning to work after long-term absence due to mental disorders

Mental disorders are prevalent in OECD countries and constitute a major cause of disability pensions, with significant societal costs. In Finland, disability pensions (DP) can be temporary when recovery and restoration of work ability are possible. Prior research has shown differences in return to work (RTW) by gender, age, marital status, socio-economic status (SES), and employment situation, but evidence after long-term absence (over one year) is limited. This study asks: Which personal, socio-economic, and healthcare-related factors are associated with RTW (full or partial) after a temporary DP due to mental disorders? The purpose is to identify determinants and patterns of RTW using national register data and to assess the role of prior psychotherapy and vocational rehabilitation, with attention to gender-specific effects. Understanding these factors can inform policies and interventions to prevent permanent labor market exit and promote sustained re-employment.

Prior studies indicate gender differences in mental disorders and RTW: men more often have substance use and antisocial disorders, women more often depression/anxiety. RTW findings vary by gender, with several studies reporting lower RTW for men, though some contexts (e.g., Japan) found the opposite. Younger age generally predicts higher RTW and older age (>50) lower RTW; being married/cohabiting is associated with faster RTW. Higher SES (education/occupational status) is typically linked to better RTW in Finnish cohorts, though one Dutch study found higher education predicted longer time to RTW. Unemployment before disability is a risk for delayed/no RTW. Evidence on the effectiveness of treatments, psychotherapy, and work-oriented interventions is mixed: some Finnish studies suggest cognitive behavioral therapy and vocational rehabilitation during temporary DP improve RTW, while other reviews found no clear correlation for counseling, exercise, or RTW programs. Studies vary widely in definitions and durations of sickness absence, limiting comparability. The literature lacks comprehensive analyses of RTW after long-term (>1 year) absence and of distinct RTW trajectories, motivating this study.

Design and data sources: A nationwide register-based cohort study within the RETIRE project. Data were linked at the individual level from the Finnish Social Insurance Institution (SII), the Finnish Centre for Pensions (FCP), and Statistics Finland. Study population: All individuals in Finland granted a temporary DP due to a mental disorder (ICD-10: F10–F69, F80–F99) for the first time between 2010 and 2012 (N=8615). Exclusions: prior permanent DP for mental disorder before 1 Jan 2010; temporary DP within one year before index. Only individuals with complete 3-year follow-up from DP start were included. Variables: Personal/SES: age categories (18–24, 25–34, 35–44, 45–54, 55–63), family status (single; couple; couple with children; single with children), household disposable income (OECD equivalence) in quintiles (lowest to highest), education (basic; upper secondary; tertiary; higher degree), occupational status (upper-grade employees; lower-grade employees; manual workers; students; unemployed; self-employed; unknown). Health: primary DP diagnosis grouped as F2* psychotic disorders (F20–F29), F3* affective disorders (F30–F39), and other (F1*, F4*–F6*, F8*–F9*). Healthcare-related: receipt of rehabilitative psychotherapy and receipt of vocational/other rehabilitation during the 5 years preceding DP (yes/no; from SII/FCP registers). Outcomes and RTW patterns: Employment contracts and working days over 3 years post-DP were obtained from the FCP employment register. Log-transformed working days across 3 years were clustered using CLARA (k-medoids; Euclidean distance) to identify six clusters (a–f). Clusters were grouped into: RTW (a: returned during year 1 and stayed; b: returned during year 2 and stayed), partial RTW (c–e: some RTW but few days and did not remain working), and no RTW (f: none or only a few days). Additionally, status distributions at exactly 1–3 years after DP were examined. Statistical analysis: Group differences by RTW category and gender were tested with chi-squared tests for categorical variables. Multinomial logistic regression modeled associations of predictors with partial RTW and RTW (reference: no RTW). Univariable models were followed by adjusted multivariable models including variables with p<0.05 for either gender; model fit evaluated with Nagelkerke pseudo-R2 and deviance tests. Software: SPSS v25; R (cluster package). Follow-up period: 3 years post-DP start.

• RTW patterns: Six clusters of working-day trajectories over 3 years were identified and grouped into RTW, partial RTW, and no RTW. RTW groups included those returning in year 1 or year 2 and remaining at work; partial RTW groups had limited working days and did not sustain employment; the no RTW group had none or very few working days.
• Overall outcomes at 3 years: A majority remained on disability pension (DP) after 3 years (men 69% [n=2495]; women 64% [n=3179]). Unemployment at 3 years: men 12%, women 11%. At work at 3 years: men 11%, women 16%.
• Gender differences in RTW rate: Women returned to work more often than men (women 29.3% [n=1468] vs men 17.8% [n=544]). Nearly two-thirds remained on DP during follow-up (men 65%, women 62%).
• Diagnosis and RTW: Affective disorders (F3*) were the most common reason for temporary DP (women 71%, men 54%) and were associated with higher RTW than psychotic disorders (F2*). RTW among those with F3* was higher (women 82%, men 71%) than among those with F2* (women 9%, men 19%).
• Socio-economic differences: Individuals who RTW more often had higher education and employee status. For example, lower-grade employees constituted 21% of men and 44% of women in the RTW group versus 9% of men and 13% of women in the no RTW group.
• Multinomial regression (adjusted) highlights: • Occupational status vs unemployed strongly promoted RTW:
  • Women: lower-grade employees OR 7.85 (95% CI 5.35–11.51); manual workers OR 4.54 (3.06–6.73); upper-grade OR 4.28 (2.74–6.67); self-employed OR 3.24 (2.03–5.19).
  • Men: manual workers OR 5.47 (3.40–8.78); lower-grade OR 5.05 (3.07–8.30); upper-grade OR 4.28 (2.44–7.51); self-employed OR 4.18 (2.40–7.29). • Partial RTW vs unemployed also favored employee/manual roles:
  • Men: manual workers OR 3.51 (2.19–5.63); lower-grade OR 2.73 (1.63–4.56); upper-grade OR 2.31 (1.24–4.31); self-employed OR 3.58 (2.01–6.36).
  • Women: manual workers OR 2.44 (1.66–3.59); lower-grade OR 2.43 (1.65–3.57). • Education increased odds of RTW and partial RTW (vs basic):
  • Men RTW: secondary OR 1.50 (1.18–1.89); tertiary OR 2.73 (1.91–3.90); high OR 2.17 (1.27–3.69). Women RTW: secondary OR 2.11 (1.73–2.57); tertiary OR 2.51 (1.95–3.24); high OR 4.94 (3.36–7.27). • Age: RTW probability highest in middle-to-older groups. Men RTW: 45–49 years OR 3.27 (2.27–4.72); 35–44 OR 2.46 (1.73–3.50). Women RTW: 45–49 OR 1.51 (1.15–1.97); >55 OR 1.58 (1.03–2.41) vs <25. • Income: higher household income quantiles associated with higher odds of RTW (e.g., women middle-higher OR 2.44 [1.85–3.23] vs lowest). • Healthcare-related factors: psychotherapy and rehabilitation each increased odds of RTW and partial RTW; effects attenuated but remained significant in adjusted models. Combined psychotherapy+rehabilitation yielded higher RTW odds than either alone (men OR 2.15, 95% CI 1.56–2.97; women OR 2.22, 95% CI 1.51–1.99; p<0.001).
• Model fit: Nagelkerke pseudo-R2 ~0.23 (men) and ~0.26 (women).

The study demonstrates that after long-term work disability due to mental disorders, RTW is relatively uncommon within three years, and sustained RTW is even rarer. Nevertheless, distinct patterns of RTW exist, and both personal/SES and healthcare-related factors are associated with these trajectories. Higher educational attainment, employee occupational status (particularly lower-grade for women and manual roles for men), and greater household income were associated with higher odds of both partial and sustained RTW. Prior psychotherapy and vocational rehabilitation were positively associated with RTW, and their combination conferred the greatest benefit, suggesting that integrated therapeutic and rehabilitative approaches may enhance work ability recovery. Gender differences were evident: women overall had higher RTW rates, differing SES profiles, and diagnosis distributions (more affective disorders), while men more often had psychotic disorders and unemployment prior to DP. Contrary to some earlier literature, RTW was more common among middle-aged and older groups than the youngest, likely reflecting diagnosis severity (higher psychotic disorder prevalence among younger beneficiaries) and labor market attachment differences. The findings indicate that unemployment prior to DP and low SES strongly hinder RTW, underscoring the need for early, targeted employment support, workplace modifications, and coordinated care pathways that consider gender and SES contexts. Recognizing RTW risk profiles can help tailor interventions to sustain employment after initial return and prevent recurrent disability or prolonged unemployment.

This nationwide register study identifies gender- and SES-related determinants of RTW after temporary disability pension due to mental disorders and delineates distinct RTW trajectories. Higher education, favorable occupational status, and receipt of psychotherapy and vocational rehabilitation are associated with increased odds of RTW, with combined therapy and rehabilitation showing additive benefits. Given that most beneficiaries remained on DP or experienced only partial RTW, programs to promote RTW should be tailored to gender and SES differences and should include early, coordinated therapeutic and vocational supports, workplace adaptations, and sustained follow-up to prevent relapse into disability or unemployment. Future research should refine targeting and timing of interventions, evaluate the quality and intensity of rehabilitation, examine workplace factors that sustain RTW, and assess the appropriateness of temporary versus permanent DP decisions. Development of real-time indicators capturing personal, social, educational, occupational, and labor market dimensions could improve RTW prediction and intervention design.

• Follow-up duration was limited to 3 years from DP initiation, which may be insufficient to capture longer-term RTW and sustainability.
• Rehabilitation was measured as a binary indicator without details on type, timing, duration, intensity, or quality, limiting inference on specific interventions.
• Potential selection into psychotherapy and rehabilitation (e.g., by SES, education, employment status, insurer decisions) may bias associations despite adjustment.
• Cross-country generalizability is constrained by differences in disability benefit systems, eligibility, and labor market contexts; findings are most applicable to settings similar to Finland.