Background: Despite agreement that restrictive intervention (RI) should only be used as a last resort and for as long as required, little is known about long-term RI among individuals with a disability. This study examines long-term RI use. Method: From the Restrictive Intervention Data System dataset from Victoria, Australia, a cohort of 1,414 people reported to be restrained or secluded between July 2008 and June 2010 were identified. The primary outcome was restraint during the follow-up period (July 2013-June 2015). Measures of the secondary outcome, reasons for restraint cessation, were assessed via a self-report survey completed by 54 service providers. Results: At follow up, 74% of the cohort was still subject to RI. Antipsychotic medication use, a diagnosis of autism, and communication difficulties were associated with the use of restrictive interventions at follow up. Conclusions Long-term RI is prevalent, but can be minimised by positive behaviour support.