Longitudinal Validity of the DUNDRUM-1 Triage Security Scale in a Canadian Not Criminally Responsible Cohort: A Retrospective Study of 85 Patients.

ObjectiveTo evaluate the longitudinal stability of DUNDRUM-1 Triage Security Scale recommendations among individuals found not criminally responsible on account of mental disorder (NCRMD) receiving care within a Canadian forensic psychiatric service.MethodsWe conducted a retrospective cohort study of adults found NCRMD who received a DUNDRUM-1 assessment at a tertiary academic forensic psychiatric centre in Quebec between 1 January 2019 and 31 December 2024. Eligible participants were aged ≥18 years and had a minimum of 12 months of follow-up. Demographic, forensic, and clinical variables were extracted from electronic medical records. Primary outcomes included changes in level of therapeutic security within 12 months. Secondary outcomes included psychiatric rehospitalization and transitions across DUNDRUM security categories. Descriptive statistics, Fisher's exact tests, Mann-Whitney U tests, and exploratory logistic regression analyses were performed.ResultsEighty-five individuals met inclusion criteria. The cohort was predominantly male (91.8%) with a mean age of 39.0 years (SD = 11.9). Schizophrenia spectrum disorders were the most common primary diagnosis (83.5%), while substance-related disorders (74.1%) and personality disorders (49.4%) were frequent comorbidities. Initial DUNDRUM recommendations included low security (56.5%), medium security (35.3%), open settings (4.7%), and maximum security (3.5%). Overall, 72 participants (84.7%) remained within their initially recommended level of care over 12 months, while 13 (15.3%) required a change in therapeutic security. Psychiatric rehospitalization occurred in 12 participants (14.1%). Exploratory analyses identified the mean violence-related DUNDRUM score as the only significant predictor of changes in level of care (OR = 2.88, 95% CI: 1.03-8.00,