Being found Not Criminally Responsible on Account of Mental Disorder in Alberta: Population gender differences, and violence prediction with the VRAG-R

Abstract

Objective: This study had two main goals. The first was to examine those individuals who had been found Not Criminally Responsible on Account of Mental Disorder (NCRMD) in Alberta, Canada for gender differences. The second was to examine the predictive accuracy of the Violence Risk Appraisal Guide Revised (VRAG-R) within this population, specifically the discrimination and calibration properties. Method: The study was archival and retrospective in nature. There were 574 individuals identified via The Alberta NCR Project database for inclusion. Information was taken from various sources including hospital and disposition records. A VRAG-R was scored on every file that contained appropriate information. Results: Gender differences were identified that indicated a unique sociodemographic, clinical, and criminological profile for both genders. The VRAG-R demonstrated strong discrimination properties for both total score and bin number for both general and violent recidivism over 5-year, 10-year and global follow up. The calibration properties however indicated that the VRAG-R substantially over estimated risk and that there was poor agreement between expected and observed recidivism rates for the overall population. When examined by gender, these issues remained but to a lesser degree. Examination of both discrimination and calibration for females was not possible. Discussion: Results supported previous research indicating that unique sociodemographic, clinical, and criminological profiles exist for the genders. Results indicated strong discrimination and poor calibration properties of the VRAG-R. Though caution should be noted in interpreting these results for a variety of reasons. Overall, the results support the use of the VRAG-R within a population of persons found NCRMD when embedded within a comprehensive risk assessment battery of tools.