The delivery format of an HIV prevention program that had been efficacious with youths at high risk was modified from seven relatively short sessions over a one- to two-month period to three longer sessions over a one- to two-week period. In contrast to youths in the seven-session program, no change in sexual behavior was observed among those in the redesigned program at three-month follow-up. However, change was apparent in the mediators of high-risk behavior over the same period.