A general framework for studying drug abuse treatment process factors is presented, and components are then used to predict relapse to opioid use during treatment in methadone maintenance. Major domains of the treatment process research framework include client variables at entry, program characteristics, treatment events, and client outcomes. The analyses rely on the use of proportional hazards models to identify significant outcome predictors in a sample of 590 methadone maintenance clients from 21 clinics in the Research Triangle Institute/Treatment Outcome Prospective Study (RTI/TOPS) data system who remained in treatment at least 3 months. The analyses were performed on the total sample and separately on clients from three groups of clinics classified on the basis of the distribution of client relapse rates and tenure in treatment. Relapse rates were related to dosage level, client monitoring with urinalyses, and methadone take-home privileges in some clinics, and hence, these time-varying treatment events were important factors in treatment outcomes. Even at entry to treatment, some measures were found to be related to how the client later performed during treatment. Finally, it was also found that the particular area of professional speciality of the staff making client diagnosis at intake and preparing treatment plans was associated with client outcomes.