Evidence for the ecological validity of the Brief Jail Mental Health Screen: Positive predictive value among remanded men and women.

BACKGROUND: Individuals with severe mental illness are over-represented in correctional institutions. The scarcity of mental health services in prison settings has increased the demand for tools to screen effectively for mental health need. While the need for sensitivity is widely recognised, there has been less attention to specificity of screening tools. In addition, prior studies have focussed on research-informed diagnostic performance rather than real-world service provision. OBJECTIVE: To examine the performance of the Brief Jail Mental Health Screen (BJMHS) for indicating secondary mental health need in 'real world' conditions. METHODS: Retrospective data were collected from 2977 individuals (2256 male) remanded in either of two correctional facilities in Ontario, Canada, who had been screened on reception as having clinically significant mental health needs by correctional health staff using the BJMHS and examined by specialist mental health staff at triage. The positive predictive value (PPV) of the BJMHS was calculated, using actual secondary mental health service referral as the performance criterion. RESULTS: Overall, the positive predictive value of the BJMHS was 67.2%. It was significantly higher for men (69.5%) than women (60.1%). CONCLUSIONS: While these findings add support to the use of the BJMHS in screening mental health need among people under custodial remand, its false positive rate, particularly among women suggests a need to improve its performance. One potentially important avenue for future research would be whether repeating the screen after an interval prior to specialist referral would improve efficiency.