Background Violent behaviour among persons with severe mental illness (SMI) causes public concern and is associated with illness relapse, hospital recidivism and poor outcomes in community-based treatment. Aims To test whether involuntary outpatient commitment (OPC) may help to reduce the incidence of violence among persons with SMI. Method One-year randomised trial of the effectiveness of OPC in 262 subjects with psychotic or major mood disorders and a history of hospital recidivism. Involuntarily hospitalised subjects awaiting OPC were randomly assigned to release or court-ordered treatment after discharge. Those with a recent history of serious assault remained under OPC until expiry of the court order (up to 90 days); then OPC orders were renewed at clinical/court discretion. Control subjects had no OPC. Four-monthly follow-up interviews with subject, case manager and collateral informant-took place and service records were collected. Results A significantly lower incidence of violent behaviour occurred in subjects with greater than or equal to 36 months' OPC. Lowest risk of violence was associated with extended OPC combined with regular out-patient services, adherence to prescribed medications and no substance misuse. Conclusions OPC may significantly reduce risk of violent behaviour in persons with SMI, in part by improving adherence to medications while diminishing substance misuse. Declaration of interest Funded by the National institute of Mental Health.