RT Article
T1 Convergent and predictive properties of three risk assessment instruments in a Canadian forensic mental health sample
JF The journal of forensic psychiatry & psychology
VO 36
IS 4
SP 526
OP 549
A1 Cheng, Jeremy Chi Yeung
A1 Olver, Mark E.
A1 Haag, Andrew M.
A1 Wormith, J. S.
A2 Olver, Mark E.
A2 Haag, Andrew M.
A2 Wormith, J. S.
LA English
YR 2025
UL https://krimdok.uni-tuebingen.de/Record/1932042946
AB The use of risk assessment instruments is essential for the assessment, treatment, and management of violence risk; it is thus critical to examine their properties when implemented in novel settings with diverse forensic subpopulations. This study evaluated the convergent and predictive properties of three risk assessment instruments in a sample of 109 forensic patients found Not Criminally Responsible on Account of Mental Disorder (NCR). A retrospective longitudinal cohort design was employed to examine the Historical Clinical Risk Management-20 Version 3 (HCR-20V3), Revised Violence Risk Appraisal Guide (VRAG-R), and Level of Service/Case Management Inventory (LS/CMI) rated from archived hospital records. LS/CMI risk scores and risk bands predicted general (area under the curve [AUC] = .70–.73) and violent (AUC = .76–.91) recidivism with moderate to large effects and performed similarly to the VRAG-R and HCR-20V3. Calibration analyses demonstrated that LS/CMI scores overpredicted the risk of general recidivism in Moderate to Very High risk bands. Results supported the convergent validity and discrimination properties of study measures; however, mixed evidence was found for the calibration properties of the LS/CMI. The potential utility of risk instruments in the appraisal and management of offending behavior among forensic mental health patients is discussed.
K1 LS/CMI
K1 convergent validity
K1 Predictive Validity
K1 Forensic Mental Health
K1 forensic assessment
DO 10.1080/14789949.2025.2480548