To describe the 5 year work status in patients referred for suspected work-related common mental disorders. To develop a prognostic model.

Register-based nationwide longitudinal follow-up study.

All departments of occupational medicine in Denmark.

17 822 patients aged 18–67 years, seen for the first time at a Department of Occupational Medicine in Denmark from 2000 to 2013 and diagnosed with stress, depression, post-traumatic stress disorder, anxiety or other mental disorders.

All patients were seen for diagnostic assessment and causal evaluation of the work-relatedness of their disorders. Some departments offered patients with stress disorders psychological treatment, which, however, was not organised according to patient selection or type of treatment.

Register data were collected for 5 year periods before and after the patients’ first assessment at a department. Weekly percentages of patients are presented according to work status. The outcome in the prognostic model was a high Work Participation Score (ie, working>75% of potential work weeks/year) at 5 year follow-up.

For all subgroups of patients, a high proportion were working (>75%) 1–5 years before assessment, and all experienced a large reduction in work status at time of assessment. At 1 year follow-up, almost 60% of patients with stress were working, whereas in the other patient subgroups, less than 40% were working. In the following years, practically no increase was observed in the percentage of patients working in any of the subgroups. Based on these 5 year follow-up data, we developed a work participation model with only moderate discrimination and calibration.

In Denmark, not all return to previous work status 5 years after a referral due to a suspected work-related common mental disorder. We developed a prognostic model with only moderate discrimination and calibration for long-term work participation after suggested work-related common mental disorders.