Abstract
The reason for this study was the void of validated risk assessment screening tools for violence in adolescence psychiatry. Our aims were to test the predictive validity and feasibility of a pilot version of the Violence Risk Screening for Youth (V-RISK-Y). The V-RISK-Y was based on a violence risk screen for adults, the V-RISK-10, and adapted to adolescents, resulting in 12 risk items that are scored for (a) presence and (b) relevance for future violence. In this naturalistic, prospective observational study, the V-RISK-Y was scored at admission and compared with recorded episodes of violent acts and threats during hospital stay. The target population was all 92 patients admitted to the emergency department of adolescent psychiatry at Oslo University Hospital for 1 year, of which 67 patients were scored with the V-RISK-Y at admission and constituted the study sample. The predictive validity of the V-RISK-Y for violent behavior showed an AUC of 0.762 (p = 0.006). Staff approved the screener and found it to be equally or better usable than the V-RISK-10, which was previously used in the department. Still, a high proportion of raters failed to follow the scoring instructions of relevance scores, reducing feasibility. The results must be interpreted within the limits of a pilot study and low power. We conclude that results suggest changes of certain parts of the V-RISK-Y and provide a basis for testing a revised edition of the screener in a more comprehensive study, preferably with a multicenter design.