Performance and clinical utility of a short violence risk screening tool in U.S. adults with mental illness

Abstract

Risk assessment instruments are typically long, costly. and resource-intensive. Thus, a short, easily administered preliminary screening tool can increase the efficiency of the subsequent violence risk assessment process. A preliminary tool can identify those at low risk of violence so that they can be screened out of the process of further violence risk assessment. Recently, Singh, Grann, Lichtenstein. Lingstrom, and Fazel (2012) used data drawn from national registries to develop a short screening tool for a sample of Swedish adults diagnosed with schizophrenia. The screening tool included 5 items: male sex, previous criminal conviction, under 32 years of age, alcohol abuse, and drug abuse. The current study examines the predictive validity and clinical utility of the screening tool developed by Singh et al. (2012) in predicting community-based violence over 6-month and 12-month durations in U.S. adults with schizophrenia (n = 3,471) and the generalizability of those findings to the assessment of violence risk in adults with other primary diagnoses. Results demonstrated that the screening tool performed reasonably well at screening out individuals who did not commit violence during follow-up; however, the screening tool did not perform as well at identifying individuals who did commit violence during follow-up. Although those who screened positive were about twice as likely to engage in violence in the 6-month follow-up period, by the 12-month follow-up there was little difference in likelihood of engaging in violence between participants who were screened in and those who were screened out. Overall, findings of the present study do not provide compelling support for the clinical utility of the screening tool in its current form.