Prediction of Violence, Suicide Behaviors and Suicide Ideation in a Sample of Institutionalized Offenders With Schizophrenia and Other Psychosis

Miriam Sánchez SanSegundo, Rosario Ferrer-Cascales, Jesús H. Bellido, Mar P. Bravo, Javier Oltra-Cucarella, Harry G. Kennedy

This study examined the predictive validity of the Spanish version of the Suicide Risk Assessment Manual (S-RAMM) and the Historical-Clinical-Risk Management-20 (HCR-20) in a sample of violent offenders with schizophrenia and other psychosis, who had committed violent crimes and had been sentenced to compulsory psychiatric treatment by the criminal justice system. Patients were prospectively monitored within the institution for 18 months. During the follow-up period, 25% of offenders were involved in any suicidal behavior including acts of self-harm, suicidal ideation and suicide attempts and 34% were physically or verbally violent. The S-RAMM and HCR-20 risk assessment tools were strongly correlated and were able to predict suicidal behavior and violence with a moderate-large effect size (AUCs = 0.81-0.85; AUCs = 0.78-0.80 respectively). Patients scoring above the mean on the S-RAMM (>20-point cut-off) had a five times increased risk of suicide related events (OR = 5.05, 95% CI = 2.6-9.7) and sevenfold risk of violence in the HCR-20 (>21-point cut-off) (OR = 7.13, 95% CI = 2.0-21.2) than those scoring below the mean. Offenders at high risk for suicide and violence had significantly more suicide attempts (p < 0.001) and more prior sentences for violent crimes (p < 0.001). These results support the use of the S-RAMM and HCR-20 for clinical practice by providing evidence of the utility of these measures for predicting risk for suicidal and violent behavior in mentally disordered offenders.