--- - |2+ Abstract Background There is limited evidence for treatment programmes for offenders with a diagnosis of personality disorder deemed to be of high risk of violence towards others. Aim This study sought to evaluate change in clinician‐rated risk of violence in males among offenders accepted to a specialist medium‐secure hospital providing treatment based on the Violence Reduction Programme (VRP). Method A retrospective, records‐based cohort study design was used to explore change in risk of violence in a sample including five VRP cohorts composed of 33 offenders. Violence risk scale (VRS) score was the primary outcome measure. Correlates of treatment dropout were also explored. Results Twenty‐seven men completed treatment. Paired pretreatment and posttreatment VRS scores were available for 19 men. Six men disengaged with treatment. There was a significant mean reduction in risk of violence score (d = 1.08). At the individual level, eleven (57%) participants in the VRP showed reliable reduction in risk of violence. No treatment completers with a complete rating pair demonstrated an increased risk of violence. Younger age and referral from prison were significantly associated with treatment dropout. Conclusion VRP‐based treatment appears to be a promising intervention for specialist services for offenders with a personality disorder diagnosis in the United Kingdom. Longer term follow‐up to ascertain whether these apparent treatment gains are accompanied by actual reduction in violent recidivism is now required, and a full scale trial warranted. - 'Criminal Behaviour and Mental Health, EarlyView. '