Can we assess risk for violence in women? Predictive accuracy of the HCR-20

Violence risk assessments are used routinely in forensic settings, in decisions about psychiatric admission and release, as well as choices about mental health treatment. Structured professional judgment (SPJ) techniques such as the HCR-20 Violence Risk Assessment Scheme (HCR-20, Webster, Douglas, Eaves, & Hart, 1997) represent improvement over unstructured methods in a range of settings and populations. However, whether the HCR-20, which was primarily validated in men, is valid in women is unclear from the existing research. This study utilized data from the MacArthur Study of Mental Disorder and Violence to score the HCR-20 for each participant (N = 836). Accuracy of each of the scales (i.e., H, C, and HC) in predicting violent outcome was evaluated for the entire sample at two time points (20 weeks and 50 weeks). This study compared predictive accuracy for men and women for each of the HC scales and each HC item, at each time point. Additionally, mean scores on each of the HC scale and HC item scores were compared for men and women. The combined HC score yielded moderate predictive accuracy in both the short and long-term in the sample as a whole. The HC score did not differ in predictive accuracy for men and women in either the short or the long-term; however, the predictive accuracy in women in the short-term was poor. The predictive accuracy of the H scale was significantly better for men as compared to women in the short-term, and approached significance in the long-term. None of the individual HC items differed in predictive accuracy between men and women; however, there were significant score differences between men and women on the majority of the items. This study offers limited support for the HCR-20 in a civil psychiatric sample. Results suggest that the HCR-20 is slightly better for evaluating risk for future violence in men than in women. Results do not suggest that the measure needs to be altered for women, or should not be used on women. However, results of the HCR-20 need be interpreted with more caution in women than in men, even if still valid