This report assesses the effects of a community outreach program in reducing HIV risk behaviors among injection drug users (IDUs) in San Juan, Puerto Rico.

Participants were 1,113 IDUs, 88.3% of whom were also assessed at post-intervention.

The analytic strategy consisted in modeling pre-and post-intervention trends in risk behavior levels from successive cohorts of IDUs.

The trend analyses showed that secular trends, unrelated to the direct effects of the outreach intervention, accounted for significant portions of the reported risk reductions.

Nevertbeless, the post-intervention trends in the shared use of cookers and in needle bleaching showed shifts that could not be accounted for by the modeled secular trends.

The outreach intervention appears to have bad significant but partial effects on the behavioral risks associated to drug injection and no effect on sexual behaviors.

The strengths and weaknesses of time-trend analyses for evaluating community HIV-prevention programs lacking experimental designs are discussed.