Premium
Predictive validity of the HCR ‐20 for inpatient aggression: the effect of intellectual disability on accuracy
Author(s) -
O'Shea L. E.,
Picchioni M. M.,
McCarthy J.,
Mason F. L.,
Dickens G. L.
Publication year - 2015
Publication title -
journal of intellectual disability research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.941
H-Index - 104
eISSN - 1365-2788
pISSN - 0964-2633
DOI - 10.1111/jir.12184
Subject(s) - aggression , intellectual disability , predictive validity , psychology , cohort , clinical psychology , psychiatry , challenging behaviour , cohort study , poison control , injury prevention , covariate , occupational safety and health , medicine , medical emergency , statistics , mathematics , pathology
Background People with intellectual disability ( ID ) account for a large proportion of aggressive incidents in secure and forensic psychiatric services. Although the Historical, Clinical, Risk Management 20 ( HCR ‐20) has good predictive validity in inpatient settings, it does not perform equally in all groups and there is little evidence for its efficacy in those with ID . Method A pseudo‐prospective cohort study of the predictive efficacy of the HCR ‐20 for those with ID ( n = 109) was conducted in a UK secure mental health setting using routinely collected risk data. Performance of the HCR ‐20 in the ID group was compared with a comparison group of adult inpatients without an ID ( n = 504). Analysis controlled for potential covariates including security level, length of stay, gender and diagnosis. Results The HCR ‐20 total score was a significant predictor of any aggression and of physical aggression for both groups, although the area under the curve values did not reach the threshold for a large effect size. The clinical subscale performed significantly better in those without an ID compared with those with. The ID group had a greater number of relevant historical and risk management items. The clinicians' summary judgment significantly predicted both types of aggressive outcomes in the ID group, but did not predict either in those without an ID . Conclusions This study demonstrates that, after controlling for a range of potential covariates, the HCR ‐20 is a significant predictor of inpatient aggression in people with an ID and performs as well as for a comparison group of mentally disordered individuals without ID . The potency of HCR ‐20 subscales and items varied between the ID and comparison groups suggesting important target areas for improved prediction and risk management interventions in those with ID .