Predictive Validity of the Brøset Violence Checklist in a Secured Institution for Offenders With Intellectual...

The integration of systematic risk assessments into care settings for offenders with intellectual disabilities (IDs) is a growing priority, yet evidence on the validity of existing tools in this population remains limited. The Brøset Violence Checklist (BVC) is widely used to assess short-term violence risk, but its predictive accuracy in individuals with IDs remains uncertain. This study investigates the predictive validity of the BVC in a specialized institutional context for offenders with IDs, focusing on its performance across different shifts and demographic subgroups.

A retrospective observational design was used to analyse 153 262 BVC assessments and 1325 documented severe violent incidents over 2.5 years in a secured Danish institution for offenders with IDs. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated across different BVC thresholds. Generalized linear mixed models (GLMMs) were applied to assess the influence of sex and time of day on predictive performance.

The BVC demonstrated high specificity and NPV across all shifts, confirming its utility in identifying low-risk scenarios. However, sensitivity and PPV were limited, particularly during night shifts and at higher score thresholds. Predictive accuracy was highest during evening shifts, aligning with periods of increased staff-resident interactions. GLMM analyses indicated that the relationship between BVC scores and violence risk was moderated by sex and time of day, with female residents and low-activity periods presenting unique challenges to prediction.

The findings underscore the need for population-specific adaptations to the BVC, particularly to address contextual and demographic factors influencing aggression in ID populations. Recommendations include supplemental assessment strategies for low-activity shifts, sex-specific behavioural indicators and the development of tailored tools for ID care settings. This study advances the understanding of violence risk dynamics in ID populations and informs interventions to enhance safety for residents and staff.