OBJECTIVE: Over the past decade, a body of observational research has accrued about the effects of outreach-based human immunodeficiency virus (HIV) interventions for drug users. The authors reviewed the findings related to postintervention behavior changes and integrated findings across studies to provide the best estimate of program impact. METHODS: The authors conducted a computerized literature search to locate published accounts of HIV intervention effects on drug users. Thirty-six publications covered outreach-based HIV risk reduction interventions for out-of-treatment injecting drug users (IDUs) and reported intervention effects on HIV-related behaviors or HIV seroincidence. Two-thirds of the publications reported that participation in street-based outreach interventions was followed with office-based HIV testing and counseling. The authors described the theoretical underpinnings of outreach intervention components, the content of the interventions, and the outcome measures that investigators used most frequently. The authors also described and critiqued the evaluation study designs that were in place. Because most of the evaluations were based on pretest and posttest measures of behavior rather than on controlled studies, results were examined with respect to accepted criteria for attributing intervention causality, that is, the plausibility of cause and effect, correct temporal sequence, consistency of findings across reports, strength of associations observed, specifically of associations, and dose-response relationships between interventions and observed outcomes. RESULTS: The majority of the published evaluations showed that IDUs in a variety of places and time periods changed their baseline drug-related and sex-related risk behaviors following their participation in a outreach-based HIV risk reduction intervention. More specifically, the publications indicated that IDUs regularly reported significant follow-up reductions in drug injection, multiperson reuse of syringes and needles, multiperson reuse of other injection equipment (cookers, cotton, rinse water), and crack use. The studies also showed significant intervention effects in promoting entry into drug treatment and increasing needle disinfection. Although drug users also significantly reduced sex-related risks and increased condom use, the majority still practiced unsafe sex. One quasi-experimental study found that reductions in injection risk led to significantly reduced HIV seroincidence among outreach participants. Few investigators looked at dosage effects, but two reports suggested that the longer the exposure to outreach-based interventions, the greater the reductions in drug injection frequency. CONCLUSIONS: Accumulated evidence from observational and quasi-experimental studies strongly indicate that outreach-based interventions have been effective in reaching out-of-treatment IDUs, providing the means for behavior changes and inducing behavior change in the desired direction. The findings provide sound evidence that participation in outreach-based prevention programs can lead to lower HIV incidence rates among program participants.