Routine violence risk assessment in community forensic mental healthcare

Abstract

We developed a method for periodic monitoring of violence risk, as part of routine community forensic mental healthcare. The feasibility of the method was tested, as well as its predictive validity for violent and risk enhancing behavior in the subsequent months. Participants were 83 clients who received forensic psychiatric home treatment, and six case managers. The method proved feasible and informative. Violent and risk enhancing behavior could be predicted to a reasonable extent (AUC = .77, 95% CI = .70-.85; respectively .76, .70-.82). Dynamic risk factors had an incremental predictive value over static factors in the prediction of violent behavior (OR = 4.30, 1.72-10.73). The professional judgment of the case managers added further predictive power (OR = 2.16, 1.40-3.33), corroborating the structured professional judgment approach. Finally, unmet needs for care of the client were associated with a reduced risk for violent and risk enhancing behavior (OR = .80, 0.69-0.93, and 0.84, 0.72-0.97). This latter finding suggests that in cases with unmet needs the case manager saw opportunities to do something about the risk. Currently we are testing whether using the method actually prevents violence.