Citation

Strand S, Belfrage H, Fransson G, Levander S. Leg. Crim. Psychol. 1999; 4(1): 67-76.

Copyright

(Copyright © 1999, British Psychological Society, Publisher John Wiley and Sons)

DOI

10.1348/135532599167798

PMID

unavailable

Abstract

Purpose. The predictive validity of the risk prediction instrument HCR-20 was studied. Methods. Two matched groups of discharged forensic psychiatric patients, one who had recidivated into violent criminality and the other not, were assessed with the HCR-20. This was done retrospective and blind to the outcome.Results. The results show an overall high predictive validity (AUC = .80). However, historical data had none, or a low, validity while clinical and risk management data had a very high validity.Conclusions. One of the most interesting findings in this study is that clinical and risk management factors came out as more predictive of future violence than historical factors, which is very much contrary to findings in past research. We think that one has to bear in mind that both clinical and risk management factors are heavily influenced by historical information. Thus, historical data are probably as important as is generally shown in follow-up studies of violence in various offender groups. However, using the HCR-20, which allows systematic and reliable coding of clinical and risk management factors, seems to make it possible to use these factors more successfully than has been hitherto possible.

Language: en