We conducted a demonstration project to design, implement, and evaluate a risk-reduction intervention delivered by medical providers to patients with HIV during routine care in 2005 and 2006. Medical providers at seven HIV clinics in the United States received training to deliver an intervention in which they screened patients for behavioral risks, gave targeted counseling, and delivered prevention messages. A longitudinal cohort (n = 767) of patients completed a baseline questionnaire and two follow-up questionnaires (6-month intervals) after the intervention was initiated. Logistic regression and generalized estimating equations (GEE) methods were used in analyses. The cohort had a median age of 41, was 58% black, 28% white, and 10% Hispanic; 32% were women and 42% self-identified as men who have sex with men. The 3-month prevalence of unprotected anal or vaginal intercourse (UAVI) with any partners declined significantly (p < 0.001) from baseline (42%) to follow-up (26% at first follow-up, 23% at second follow-up). The decline was significant with partners who were HIV-negative/unknown serostatus or HIV-positive. Cohort patients' self-reported receipt of safer-sex counseling at all, some, or no clinic visits during the interval between baseline and first follow-up showed a dose-response relationship with decline in prevalence of UAVI in that interval, with relative reductions of 45%, 35%, and 19%, respectively. All findings were confirmed in multivariate models that controlled for demographic factors and HIV clinical status of participants. This project demonstrated that with only brief training, HIV medical providers successfully conducted an HIV prevention intervention with their clinic patients. Our findings indicate that clinics that serve HIV patients can incorporate such programs as standard of care.