# Sick leave patterns before and after commencement of psychological therapy among individuals with common mental disorders: a register-based, quasi-experimental study

## TL;DR

This study examines whether psychological therapy in primary care reduces sick leave for people with mental health issues, finding mixed results that suggest therapy may not consistently lower sick leave.

## Contribution

The study evaluates real-world effectiveness of psychological therapy on sick leave reduction for common mental disorders using a quasi-experimental design.

## Key findings

- Crude models showed reduced sick leave likelihood in treated groups at 12, 18, and 24 months.
- Inverse probability weighted models showed increased sick leave likelihood in treated groups at 6 months.
- Higher therapy doses were associated with worse sick leave outcomes in some analyses.

## Abstract

Psychological therapies have been shown to reduce sickness absence (SA) among individuals with common mental disorders (CMDs) in clinical trials, but their real-world impact is unclear. To address these knowledge gaps, we compared the likelihood of receiving publicly-financed SA compensation among individuals with CMDs who received at least one session of systematic psychological therapy in primary care and individuals with CMDs who did not receive these treatments.

Primary healthcare registers were used to identify individuals with CMDs in Region Stockholm who had received psychological therapies (N = 12,167) and untreated controls (N = 40,517). SA was measured at six-monthly intervals in the two years before and after treatment commencement. Crude and inverse probability weighted (IPW) generalised estimating equation (GEE) models were used to compare the likelihood of having > 14 net SA days (primary outcome) and > 30 and > 90 days (secondary outcomes) at each six-month interval to the period preceding treatment commencement (t0). Crude and IPW models were performed in the treated and control groups separately, with effects at each time-point compared via the ratio of odds ratios (ROR).

In the crude model, the treated group showed significantly greater reductions in the likelihood of having > 14 net SA days at t12, t18, and t24 compared to the control group (ROR: 0.82, 0.80, and 0.79, respectively). However, in the IPW models, the likelihood of receiving > 14 net SA days at t6 relative to t0 was significantly higher in the treated group (ROR: 1.33, 95% CI: 1.25–1.40) with no group differences at t12, t18, or t24. Moreover, the treated group fared significantly worse than the control group in IPW models examining > 30 and > 90 net SA days. In sensitivity analyses, individuals receiving 1–2 sessions showed a greater reduction in the likelihood of having the primary outcome relative to the control group at t12, t18, and t24, whilst those receiving greater doses of psychological therapy (6–12 and > 12 sessions) had a significantly higher likelihood of SA at these timepoints.

Our findings tentatively suggest that psychological therapies delivered in primary care are not associated with a reduction in the likelihood of receiving sick leave in people with CMDs. However, due to the possibility of unmeasured confounders, our findings cannot support causal inferences. Further studies in real-world settings are needed to investigate whether treatment-related factors (e.g., therapeutic focus, quality, and fidelity) modify the effect of primary-care delivered psychological therapies on SA outcomes among people with CMDs.

Not applicable.

The online version contains supplementary material available at 10.1186/s12888-026-07818-3.