A systematic review on the effectiveness of sex offender risk assessment tools in predicting sexual recidivism of adult male sex offenders

摘要

Purpose: This study aimed to systematically review the effectiveness of risk assessment tools in predicting sexual recidivism of adult male offenders. Background: Sex offender risk assessment aids risk management within the criminal justice system. Some tools follow an actuarial approach and some adopt structured professional judgement. There has not been a systematic review evaluating and comparing the effectiveness of those tools and appraising the overall quality of the primary research. Methods: Six electronic databases and reference lists of relevant meta-analyses were searched. Three experts were contacted to obtain relevant studies. Inclusion criteria were applied to the identified references and the included studies were quality assessed, using pre-defined criteria, prior to data extraction and synthesis. Results: Electronic search yielded 4949 hits. Of these, 1419 duplicates, 1 meta-analysis and 3382 irrelevant hits were excluded. 14 publications identified from previous meta-analyses were included. 2 non-English language duplications of publications were excluded. 89 publications that did not meet inclusion criteria and 15 inaccessible publications were excluded. 11 studies that did not meet minimum threshold criteria and 1 study that re-analysed an already included sample were then excluded, leaving 43 publications containing 43 studies. Conclusions: All included tools demonstrated at least moderate predictive accuracy, with two reporting a large effect size (VRS-SO and SRA), although these two came under much less empirical scrutiny than the others and may have been the subject of developer bias in the research that is available. The VRS-SO was found to have the highest mean quality score, this again being limited by the number of studies and developer bias. The quality of the primary research is variable. More independent high quality research is needed, particularly on structured professional judgement incorporating dynamic risk factors. ? 2012 Elsevier Ltd.