Objective: National efforts to improve responses to persons with mental illness involved with the criminal justice system have traditionally focused on providing mental health services under court supervision. However, a new policy emphasis has emerged that focuses on providing correctional treatment services consistent with the risk-need-responsivity (RNR) model to reduce recidivism. The objective of this review was to evaluate empirical support for following the RNR model (developed with general offenders) with this group and to pose major questions that the field needs to address.

Methods: A comprehensive search using PubMed and PsycINFO yielded 18 studies that addressed the applicability of the RNR model to the target population. The results of these studies were synthesized.

Results: There is strong support for using general risk assessment tools to assess this group’s risk of recidivism. Preliminary evidence indicates that cognitive-behavioral programs targeting general risk factors are more effective than psychiatric treatment alone. However, there is as yet no direct support for the applicability of the three core RNR principles to treat this population.

Conclusions: Although the new policy emphasis shows substantial promise, the field must avoid rushing to the next “evidence base” too rapidly and with too little data. There must be explicit recognition that RNR principles are being applied to a new population with unique characteristics (mental illness combined with justice system involvement), such that generalizability from general offender samples is uncertain. Moreover, public safety goals for the target population should not eclipse those related to public health. This group’s unique features may affect both the process and outcomes of treatment.