An Examination of Dynamic Risk, Protective Factors, and Treatment-Related Change in Violent Offenders
The present study was archival in nature and examined risk for recidivism, treatment-related changes in risk, protection against recidivism, treatment-related changes in protection, the relationship between risk and protective factors, and the prediction of positive community outcomes. A select set of risk- and protective-factor measures were used, including the Violence Risk Scale (VRS), the Historical Clinical Risk Management scheme-20 (HCR-20 version 2), the Structured Assessment of Protective Factors (SAPROF), and the PF List (an operationalized list of protective factors developed by the investigators). Participants included 178 federally incarcerated adult male violent offenders who participated in the Aggressive Behaviour Control treatment program at the Regional Psychiatric Centre (Saskatoon, SK) between 1998 and 2003. Participants were followed for an average of 9.7 years (SD 2.6) to assess community recidivism. Approximately 60% had at least one new violent conviction, 60% had at least one new nonsexual violent conviction, and 79% had at least one new conviction (i.e., any reconviction). Additionally, participants were followed for an average of 30.7 months (SD = 40.3) to assess institutional recidivism. Approximately 31% had at least one post-treatment major misconduct, 51% had at least one post-treatment minor misconduct, 12% had at least one post-treatment violent misconduct, and 56% had at least one post-treatment misconduct (i.e., any misconduct). Correlations between the risk measures scores support their convergent validity. Both the VRS and HCR-20 predicted all violent, nonsexual violent, and any recidivism. Dynamic variables on these tools generally added uniquely to the prediction of community recidivism over static variables. A similar but weaker pattern of results was observed for institutional recidivism. Additionally, treatment-related change scores on the risk measures added uniquely to the prediction of most recidivism outcomes, supporting the dynamism of these tools and the hypothesis that treatment-related changes translate to actual reductions in recidivism rates. Correlations between the protection measures’ scores support their convergent validity. The protective factor tools, the SAPROF and PF List, similarly predicted community recidivism and, to a lesser degree, institutional recidivism. Dynamism of the protective factor tools was supported and change scores on these tools added incrementally to the prediction of recidivism outcomes. Large correlations were observed between the risk and protection scores, suggesting that part of the predictive accuracy of the protection measures may relate to measuring the absence of risk rather than the presence of protection. Alternative hypotheses are discussed. Protection scores did not add incrementally to the prediction of recidivism over their respective risk scores. Risk, protection, and change scores were significant predictors of most positive community outcomes. Protection scores and risk change scores added incrementally to the prediction of positive community outcomes over their respective risk scores. As such, it appears that treatment-related changes may also represent increases in other positive community outcomes (beyond reduced reoffending) and that protection factors may have important benefits in risk assessment and treatment planning when other positive community outcomes are considered. Strengths, limitations, and implications are discussed.
Violence, Violent Offender, Risk, Recidivism, Treatment-related change in risk, Protective Factors, RNR model, Aggression
Doctor of Philosophy (Ph.D.)