This study examined the characteristics of 447 cocaine- dependent, first-admission outpatients in relation to their AIDS risky behavior at intake. The study sought to ascertain whether there was a reduction in risky behavior at follow-up 9 months after admission and to determine whether reductions in risky behavior were related to patient characteristics, group as compared to individual treatment, or time in treatment. At 9-month follow-up, AIDS-risk behaviors had decreased significantly. The degree of improvement was not associated with demographic variables but was predicted by higher intake problem severity and psychological symptomatology scale scores. While improvement in risky behavior was not related to type or duration of treatment, it was related to decreased substance use. Tables, references