Risk Assessment and Provisional Discharge: Predictive Utility of the HCR-20

The study examines the predictive utility of the HCR-20 in risk assessments for supervised release to the community as well as returns to the forensic hospital for individuals under indeterminate civil commitment in Minnesota. This archival review included 331 patients who were evaluated for provisional discharge to the community using either the HCR-20V2 or HCR-20V3, and 135 patients were released. A series of logistic regressions and receiver operating characteristic (ROC) analyses indicated lower scores on the HCR-20V2 Clinical and Risk Management scales and on the HCR-20V3 Risk Management scale were predictive of review board decisions regarding release to the community. Additionally, individuals housed in the transition unit had odds of being provisionally discharged that were 6 (HCR-20V3) to 23 (HCR-20V2) times higher than individuals placed in the security hospital. Over a 15-year period, 30 (22%) patients on provisional discharge returned to the hospital within an average of just under 3 years and information was obtained regarding reasons for return. Only the HCR-20V3 Clinical scale was predictive of returns to the hospital using logistic regression; none of the scales showed predictive validity using ROC analyses. Success on provisional discharge underscores the ability of community services to perform well in maintaining mental health stability to high-risk individuals. As expected, but with some caveats, the "C" and "RM" scales from the HCR-20 demonstrated predictive power in determining who was discharged and who returned. This paper calls for additional research and resources on how to improve chances of remaining in the community after discharge.