Data on each participant came from the medical charts. Urinalysis at the clinic was used as an objective measure of drug use. Heroin and cocaine use was measured by the proportion of positive urine samples during each 3-month period after clinic admission. The research used cluster analysis to examine the data. Results revealed distinct groups with very different heroin and cocaine usage patterns during treatment. About 80 percent of the participants either started in or transitioned into a low-heroin-use group; 50 percent either started in or transitioned into a low-cocaine-use group during treatment. In addition, one- third of the participants used cocaine extensively during treatment. Findings indicated that certain groups at high risk for poor outcomes should receive more intensive services immediately after intake; these groups include younger persons, those with no recent employment, those involved in the criminal justice system, and those dependent on multiple drugs. Greater therapeutic attention to those groups could further increase the already substantial success rates of methadone programs. Tables, figures, author biographies and photographs, and 45 references (Author abstract modified)