Aims: To evaluate the effects of a jail-based continuation of methadone maintenance therapy (MMT) on subsequent inmate recidivism risks.

Design: Prospective, longitudinal, observational study.

Setting: A large, Southwestern United States jail that continues MMT for heroin-addicted inmates on MMT at the time of booking.

Participants: A total of 589 inmates booked between 22 November 2005 (the start date for the MMT program) and 31 October 2006.

Measurements: The outcome measure was time from release to subsequent re-booking in the jail. Predictors included binary dosing with methadone in the jail, final dose received (mg), age, gender, race/ethnicity, previous bookings and days in jail.

Findings and conclusions: Random effects Weibull proportional hazards models were fit to the recidivism times to estimate the impact of treatment with MMT in the jail on re-booking risks. There was no statistically significant effect of receiving methadone in the jail or dosage on subsequent recidivism risks (hazard ratio = 1.16; 95% confidence interval = 0.8-1.68). Offering jail-based MMT does not increase recidivism risks by eliminating the deterrent effect of imposed withdrawal, nor does it reduce recidivism in this high-risk population.