The development of medications for the treatment of cocaine dependence has been a high priority of the U.S. National Institute on Drug Abuse, National Institutes of Health. One of the main strategies has been to test available, marketed medications that affect CNS function and have a rationale for testing in a cocaine dependent population. The Cocaine Rapid Evaluation Screening Trials (CREST) utilized a randomized, controlled, parallel group, blinded methodology for comparing one or more medications against a placebo. Subjects were evaluated for a 2-4 week baseline and then randomized to a treatment group for 8 weeks. Standardized measures of outcome were used: urinary benzoylecgonine, retention, craving, depression, clinical global impressions, HIV risk behaviors. Counseling and procedures were also standardized across studies to facilitate data comparisons across drug classes. A total of 19 drugs were evaluated in 5 research clinics. Results from the studies suggested that cabergoline and reserpine should be further evaluated. Less robust effects were seen with sertraline and tiagabine although the sample size in each group was small (n= 15/group). Trials were analyzed separately and then a pooled analysis was performed. For example, an analysis of characteristics leading to at least 2 weeks of abstinence was performed. Being female with at most 5 years of prior use and being over 40, being a non-African-American male with at least four baseline uses and more than 4 years of prior use, and males with at most three baseline uses and less than 20 years of prior use were associated with abstinence rates of 25, 30, and 42 %, respectively. One of the lessons learned was the potential value of assessing cocaine use during the baseline period prior to randomization. Another lesson learned was the use of both standardized assessments across sites and outcome measures that were also employed within individual sites. This allowed exploratory analyses within sites to determine sensitivity of outcome measures.