The utility of the Level of Service/Risk-Need-Responsivity (LS/RNR) instrument in predicting recidivism for adult mentally ill offenders involved in a mental health cour

This research examined the predictive validity the Level of Service/Risk-Need-Responsivity (LS/RNR) instrument for recidivism in 138 male and female mentally ill offenders involved with a mental health court (MHC). Overall, the sample was medium-risk (MLS/RNR = 17.06) and had a general recidivism rate of 47.8%. The General Risk/Need section of this instrument strongly predicted general (AUC = .77) and technical (AUC = .75) recidivism, while the Specific Risk/Need section most accurately predicted violent recidivism (AUC = .65). No LS/RNR section had incremental validity beyond the General Risk/Need section for general and technical recidivism, but only the Specific Risk/Need section significantly explained variance for violent outcomes. Criminogenic needs were differentially predictive across gender and diagnostic groups. Recommendations for future research include expansion of the Specific Risk/Need item content to include variables that inform short-term violence risk. It was also argued that integration of formal risk assessment procedures into MHC programs is critical to inform decision-making regarding the admission and case management of offenders with mental illness. This assessment should use the full LS/RNR rather than consideration of the General Risk/Need section only, especially for the prediction of violent recidivism.