Unemployment is a risk factor for later development of mental health problems, but characterisation of this in real world clinical data is limited. This study aimed to investigate the association between employment status and time-to-first-contact with mental health services using survey data linked to electronic health records(EHR).

SELCoH (n = 1698, 2008–2010) was a representative population survey of South East London, with a 71.9% household participation rate. Anonymised survey data for participants was linked with EHR, generating survival data for time-to-first-contact. Cox regression was used to assess associations between unemployment and time to first contact with mental health services.

The rate in the unemployed was 22.84 contacts per 1000 person-years, and in those not unemployed, it was 10 contacts per 1000 person-years. The crude (age-adjusted) hazard ratio (HR) for unemployment was 3.09 (95% CI: 1.66–5.75). The HR for contact for unemployment, after adjusting for age, gender, ethnicity and education, was 2.8 (95% CI: 1.44–5.47). On addition of symptoms of common mental disorder, post-traumatic stress, psychosis and suicide attempts, to the model, unemployed participants remained at elevated risk (HR:2.65, 95% CI: 1.33–5.27). Finally, illicit drugs and alcohol had minimal influence on estimates, giving a fully-adjusted estimate for the association between unemployment and rate of contact of 2.6 (95% CI: 1.31–5.14).

Unemployment was associated with a greater than two-fold increase in risk of accessing mental health care for the first time within the observation time, after adjustment for sociodemographic confounders, psychopathology, and substance use. Explanations for this association could include unobserved confounding, health behaviours associated with unemployment or effects of unemployment on stress processing.

The authors have not supplied their declaration of competing interest.