This study examines factors affecting the validity of a violence risk screen tool.
This study assesses the Fordham Risk Screening Tool (FRST), the first instrument specifically designed for violence risk screening in low-resource settings, in terms of how patient demographics, clinical factors, or contextual factors might affect the predictive validity of the FRST. This study examines the predictive validity of the FRST in a sample of 154 adult psychiatric inpatients in the Bronx, NY, with special attention to the influence of demographic, clinical, and contextual factors. Each participant was screened with the FRST upon admission, and was rated on the HCR-20V3 in the week after screening. Based on the results of classification analysis, logistic regression, and multiple regression, the FRST appears to be a significant predictor of violence risk as measured by the HCR-20V3. The FRST has high sensitivity, and a negative predictive value of 90%, and is effective with various populations and in different contexts. More specifically, gender, prior hospitalizations, psychosis, symptom severity, threat control override symptoms, homelessness, referral source, and legal status did not influence the effectiveness of the FRST. High school education was the only variable that impacted FRST efficacy, with better predictive validity for participants who had not graduated high school. Thus, the FRST is a promising violence screening tool for low-resource settings, and it can be used with a wide range of psychiatric inpatients. Classification accuracy results of this study have implications for implementation, depending on clinical resources and priorities. Future research should focus on the effectiveness of the FRST with different populations, in different settings, and when followed-up with different comprehensive violence risk assessment tools.