Predictors of methadone maintenance treatment outcome have not been extensively studied as they relate to variations in program philosophy, nor have such predictors received much examination among recently treated, older cohorts of opioid addicts for whom drug use patterns have changed. Predictors of outcome were examined at 18 months post-treatment entry for 353 admissions to methadone maintenance who received random assignment to one of three counseling conditions: (1) medication only, (2) standard counseling and (3) enhanced services; and one of two contingency conditions: (1) no contingencies, and (2) contingency contracting in a six-cell 3x2 design. Subjects in contingency contracting conditions were placed on contingency contracts for positive urine toxicology results and ultimately discharged for unremitting drug use. All subjects completed the Addiction Severity Index (ASI) and provided weekly urine specimens. Predictors of urinalysis results and treatment retention were determined using bivariate and multivariate techniques. Interactions between subject characteristics by experimental condition assignment were also examined as predictors. Higher rates of total positive urine specimens were predicted by younger age, greater pre-treatment frequency of smoking cocaine, lower ASI psychiatric composite scores, and higher ASI legal composite scores. Higher rates of opiate positive specimens were predicted by younger age, lower pre-treatment frequency of alcohol intoxication, higher ASI legal and lower ASI employment and psychiatric composite scores, and assignment to medication only/no contingencies condition. Higher rates of cocaine positives were predicted by younger age, black race, lower ASI psychiatric composite score, greater pre-treatment frequency of intravenous and smoked cocaine use, less pre-treatment frequency of marijuana use, and lower methadone dose level. Assignment to enhanced/contingency contracting predicted lower rates of cocaine positives. Treatment retention was predicted by older age, non-black race, lower ASI legal composite score, higher methadone dose level and assignment to non-contingent conditions. While subject variables over which treatment providers have little control were, thus, related to outcome, type of treatment provided and methadone dose also influenced outcome.