Introduction: This study examined the association between program duration and rate of criminal conviction and hospitalisation for substance use up to 15 years later among young people admitted to a short-term residential program for drug and alcohol use. Methods: Data were derived from linked administrative records of all clients referred to a modified therapeutic community for young people from January 2001 to December 2016 in New South Wales, Australia (n = 3059). Cox proportional hazards regression analyses examined the rate of conviction (separately for any offence, violent offence, non-violent offence and administrative offence) and hospitalisation for substance use, up to 15 years post-program among young people who attended treatment for 1–29 days, 30–59 days, 60–89 days and 90–120 days. Results: Thirty days or more in treatment was independently associated with a lower rate of conviction for any offence and a non-violent offence, as well as hospitalisation for substance use, while 60 days or more was associated with a lower rate of conviction for a violent and administrative offence, relative to those who spent 1–29 days in the program. Additional months in the program were also associated with reduced rates of conviction and hospitalisation, although 90–120 days appeared to confer no additional benefits than 60–89 days. Discussion and Conclusions: At least 60 days may be the minimum duration needed for short-term, therapeutic community programs to reduce the risk of conviction across all crime types and hospitalisation for substance use.