The validity of the HCR-20 and PCL:SV Spanish version in predicting inpatient violence in a civil psychiatric sample

ARBACH, K.; ANDRÉS-PUEYO, A.

Amsterdam

Congreso; 6th Annual International Association of Forensic Mental Health Services Conference; 2006

International Association of Forensic Mental Health Services

Background The problem of clinical and actuarial violent behavior prediction has revived the controversy on the relation between violence and mental illness. For a long time it was believed a mistake to consider mental patients to be violent subjects, but recently there has been verified that this population presents an increased probability to behave violently under certain circumstances. In this line of the research, in countries like USA, Canada or England the predictive validity of some actuarial methods and structured clinical judgment instruments as the HCR-20 (Historical, Clinical, and Risk Management) and the PCL:SV (Psychopathy Checklist: Screening Version) has been assessed, and today they allow to achieve very accurate evaluations of violence risk in psychiatric and forensic populations. Aim The first aim of the study is to explore the predictive potential of the HCR-20 and PCL:SV Spanish version in a group of 114 psychiatric civil subacutes and chronic inpatients of a mental health hospital. METHOD This longitudinal study was designed to carry out a 12 month follow-up; here results at 6-months follow-up will be presented. The subjects were assessed using the HCR-20, PCL:SV and an instrument designed ad hoc to collect some other demographic and clinical variables. The information was collected by means of the clinical files and interviews with the staff members. Throughout follow up violent behaviours of the subjects were coded with the Spanish version of the Modified Overt Aggression Scale (MOAS) that consider four categories of behaviour: verbal aggression, self harm, and physical aggression, against objects and other people. RESULTS The 6 months follow-up was completed in 100 subjects, in this period 75 % of them had committed one or more aggressive behaviour of any type (verbal, to objects, to other people or self-harm). HCR-20 total score and subscales scores were significantly associated to any type of violence (rs=0.32-0.47; p <0.01) with the exception of self-harm. In the regression analysis HCR-20 emerged as significant 6-months predictor for verbal violence (p<0.001), and violence towards objects (p<0.05) and persons (p<0.01), but not for self-harm. All of this behaviours occurring inside in wards. Limiting the criterion to physical violence towards persons or objects we found that 36 % had committed one or more violent behaviour. With this criterion we determined odds ratio for violence, using as predictor variable the HCR-20 global risk valuation. Those showing one or more violent behaviour towards persons or objects were 9.72 more likely of having obtained a level of moderated or high risk. Receiver operating characteristic analyses with the HCR-20 yielded strong association with violence (area under curve [AUCs] =.78). Limitations The total sample was composed by psychiatric chronic and subacute patients, and the records of violent behavior were limited to the intrahospitable context. Conclusions This study suggests that HCR-20 and PCL:SV appear to be valid perdictors of in-patient violence within 6 months follow-up to a Spanish psychiatric hospital and predictive validity is similar to that reported in studies using the English version.