Background and Aim: The objective of this study was to determine factors associated with time to relapse and, thus, retention time, of a cohort of opioid-dependents under methadone maintenance treatment, using survival models, in the Iranian National Center for Addiction Studies (INCAS).

  Materials and Methods: A total of 198 opioid-dependent clients participating in the Methadone Maintenance Treatment Program, implemented by INCAS between April 2007 and March 2011, were included in the study. A Cox proportional hazard (PH) model was applied to determine predictors of relapse time among the patients.

  Results: The data showed that 86 clients relapsed into drug use during the treatment program. The proportional hazard assumption was satisfied according to the goodness of fit test showing that Cox proportional hazard model was appropriate. Estimates of the PH model indicated that an increase of 1mg in the methadone dosage could lead to a decrease of 0.15 in hazard ratio and an increase in the length of treatment (p<0.001). Predictors for raising the probability of drug relapse included suffering from mental disorders (compared to mentally healthy: hazard ratio = 2.29, p<0.001), being a poly-substance user (compared to mono-substance users: hazard ratio = 4.80, p<0.001), and having retention experience in the past (compared to those with no previous therapy: hazard ratio = 1.90, p<0.001). Other variables entered in the model, including social and demographic variables, had no statistically significant effect on hazard ratio.

  Conclusions: Although higher methadone dosages are associated with a longer time to relapse, we recommend highly to pay special attention to providing more therapeutic and consultive services to mental health patients, poly-substance users, and individuals with a past therapy record.