Both cocaine withdrawal symptoms, measured by an instrument called the Cocaine Selective Severity Assessment (CSSA), and urine toxicology results obtained at the start of treatment have been shown to predict treatment outcome in outpatient cocaine dependence treatment. This study further evaluates the predictive validity of the CSSA and urine toxicology results, alone and in combination. Subjects included 76 cocaine-dependent individuals who participated in 7-week, outpatient, pilot medication trials for cocaine dependence. Predictor variables included CSSA scores and results from a urine toxicology screen obtained on the first day of medication treatment. Successful outcome was defined as 3 continuous weeks of self-reported abstinence from cocaine confirmed by urine toxicology screens. Predictive validity was assessed by logistic regression analysis. Both the urine toxicology screen and the CSSA scores were significant predictors of 3 weeks of continuous abstinence from cocaine, and the inclusion of both variables significantly improved the predictive validity of either variable alone. Urine toxicology results and CSSA scores obtained at treatment entry are useful predictors of outcome in outpatient cocaine dependence treatment.