The Short- to Medium-Term Predictive Accuracy of Static and Dynamic Risk Assessment Measures in a Secure Forensic Hospital

Although violence risk assessment knowledge and practice has advanced over the past few decades, it remains practically difficult to decide which measures clinicians should use to assess and make decisions about the violence potential of individuals on an ongoing basis, particularly in the short to medium term. Within this context, this study sought to compare the predictive accuracy of dynamic risk assessment measures for violence with static risk assessment measures over the short term (up to 1 month) and medium term (up to 6 months) in a forensic psychiatric inpatient setting. Results showed that dynamic measures were generally more accurate than static measures for short- to medium-term predictions of inpatient aggression. These findings highlight the necessity of using risk assessment measures that are sensitive to important clinical risk state variables to improve the short- to medium-term prediction of aggression within the forensic inpatient setting. Such knowledge can assist with the development of more accurate and efficient risk assessment procedures, including the selection of appropriate risk assessment instruments to manage and prevent the violence of offenders with mental illnesses during inpatient treatment.