Predicting treatment success at multiple timepoints in diverse patient populations of cocaine-dependent individuals.

The quest for predictive utility of baseline demographic and drug use characteristics has been difficult. The present article draws data from two studies of cocaine-dependent individuals (N = 297) in treatment at clinics in Los Angeles, and examines the utility of socio-demographic characteristics of patients and in-treatment performance variables as predictors of success at treatment end, 6 and 12 month follow-up assessments. Socio-demographic variables examined are age, gender, ethnicity, and educational attainment; drug use variables include years of cocaine use, self-reported days of cocaine use, the Addiction Severity Index drug composite score, and two composite measures cited in the literature. The in-treatment variables examined include cocaine urine toxicology results, number of weeks retained, and measures of compliance. The self-reported number of days of cocaine use in the past 30 days provides the most predictive utility of all baseline variables evaluated, and is the most parsimonious of the significant variables associated with substance use at all subsequent timepoints. Matching cocaine-dependent patients with treatment types or intensities based on the self-reported number of cocaine use days at intake may increase patient success rates.