Antecedents of Long‑Term Work Disability After Mental Health-Related Sickness Absence: The Role of Occupational Health Care Pathways

Purpose The relationship between specific occupational health care (OHC) pathways and subsequent work disability following an initial mental health-related sickness absence (SA) is poorly understood. This study aimed to predict long-term SA outcomes by modeling the timing and composition of OHC pathways among employees absent due to mood disorders or anxiety. Methods We conducted a register‑based observational cohort study using structured occupational health records from a major Finnish OHC provider (2015–2023, N = 35 385). Following a 12‑month washout period, the first mental health-related SA (ICD‑10 F30–F48) occurring in 2016–2021 was defined as the index episode. The OHC pathways were characterized within a 3-month window. XGBoost was employed to predict 24‑month recurrence of mental health-related SA and cumulative SA days using care pathway features and socio‑demographic variables. Model interpretability was assessed using Shapley Additive Explanations (SHAP) values. Results SA patterns emerged as the most potent predictors of long‑term outcomes. The accumulation of absence days, particularly between days 61 and 90, and internal recurrence within the index period markedly increased the predicted risk of future mental health-related SA. In contrast, specific OHC pathway elements and the timing of clinical contacts contributed less to overall model performance. Conclusions Our findings suggest that long‑term work disability risk is most strongly predicted by indicators of condition severity rather than the specific composition of early OHC interventions. To improve outcomes, timely identification of workers whose recovery trajectories deviate from expected norms may allow more targeted and sustained support for work ability.