Gender differences in structured risk assessment: Comparing the accuracy of five instruments.

Abstract

Structured risk assessment should guide clinical risk management, but it is uncertain which instrument has the highest predictive accuracy among men and women. In the present study, the authors compared the Psychopathy Checklist-Revised (PCL-R; R. D. Hare, 1991, 2003); the Historical, Clinical, Risk Management-20 (HCR-20; C. D. Webster, K. S. Douglas, D. Eaves, & S. D. Hart, 1997); the Risk Matrix 2000-Violence (RM2000[V]; D. Thornton et al., 2003); the Violence Risk Appraisal Guide (VRAG; V. L. Quinsey, G. T. Harris, M. E. Rice, & C. A. Cormier, 1998); the Offenders Group Reconviction Scale (OGRS; J. B. Copas & P. Marshall, 1998; R. Taylor, 1999); and the total previous convictions among prisoners, prospectively assessed prerelease. The authors compared predischarge measures with subsequent offending and instruments ranked using multivariate regression. Most instruments demonstrated significant but moderate predictive ability. The OGRS ranked highest for violence among men, and the PCL-R and HCR-20 H subscale ranked highest for violence among women. The OGRS and total previous acquisitive convictions demonstrated greatest accuracy in predicting acquisitive offending among men and women. Actuarial instruments requiring no training to administer performed as well as personality assessment and structured risk assessment and were superior among men for violence.

The predictive validity of risk assessment tools for female offenders: A systematic review

“…None of the LSI instruments were evaluated with psychiatric samples. The PCL-R was able to predict recidivism with moderate accuracy in a correctional sample (AUC = .73; Coid et al, 2009), but the same was not found for a psychiatric sample (AUC = .57; Schaap et al, 2009). When the primary author compared effect sizes between the correctional and psychiatric sample using independent t-tests, no significant differences were found (p N .05).…”

“…Duplicates identified n = 62 (AUC = .70) in one correctional sample (Coid et al, 2009), but not in another (Warren et al, 2005). However, the 95% confidence interval for the AUC in Coid et al's (2009) study was wide suggesting that obtained AUC values cannot be interpreted with confidence.…”

“…However, the 95% confidence interval for the AUC in Coid et al's (2009) study was wide suggesting that obtained AUC values cannot be interpreted with confidence. Other studies found the HCR-20 to perform no better than chance at distinguishing recidivists from non-recidivists (Schaap, Lammers, & de Vogel, 2009).…”

Predicting community violence from patients discharged from acute mental health units in England

“…Step 2 covariates age and gender added best performing scales was evident at the median split; 65% for the HCR-20 and 66% for the NAS, and the relative predictive accuracy of the all the risk measures is moderate when compared with the larger effects found in other similar studies in forensic, mixed, prison and learning disability samples [4,8,9]. For the purposes of this study the sample comprised of discharged acute psychiatric patients.…”

“…Coid et al [7,8] investigated risk factors and five structured risk assessment instruments associated with offending post-discharge from medium secure forensic units in England. Predictors of future offending included male gender, young age, early onset offending and a diagnosis of personality disorder and all risk instruments demonstrated moderate predictive validity.…”