Assessing the risk of violence in psychiatric outpatients: The predictive validity of the HCR-20 risk assessment scheme

Although a large proportion of psychiatric patients display violent behaviour after release from hospital, there currently are no validated violence risk assessment schemes which are applicable to civil psychiatric settings. Under most civil committment legislation, individuals only can be released from psychiatric institutions if they do not need to be detained for the protection of others. The present research sought to validate the HCR-20 Risk Assessment Scheme, which is a broad-band violence risk assessment tool with applicability to a variety of settings. This instrument possesses subscales which measure Historical, Clinical, and Risk Management correlates of violence. The HCR-20 was compared to the Psychopathy Checklist: Screening Version (PCL:SV), and a Violence Screening Tool (VST) designed originally to assess risk for psychiatric inpatient violence. Participants were 200 involuntarily committed patients who were followed into the community and whose violence was measured by criminal records and rehospitalizations involving violence. Using receiver operating characteristic analyses and the areas under their curves as indexes of predictive accuracy, and multiple regression analyses to determine the relative contributions of these scales to violence, it was found that the HCR-20 produced larger and more consistent relationships to violence than the PCL:SV. The VST was unrelated to violence. Subscales of the HCR-20 which measure Historical and Risk Management factors were the most consistent and strongest of all subscale predictors. It is concluded that there is support for the use of the HCR-20 in a civil psychiatric setting for assessing the risk of post-release violence.