Evidence for the effects of HIV antibody counseling and testing on risk behaviors

Objective. —To review published abstracts, journal articles, and presentations for evidence of the effects of human immunodeficiency virus (HIV) antibody counseling and testing on risk behaviors. Studies reviewed focused on homosexual men, intravenous drug users in treatment programs, pregnant women, and other heterosexuals. Data Sources. —Peer-reviewed journals (January 1986 through July 1990) and published abstracts and oral presentations from the second (1986) through the sixth (1990) International Conferences on AIDS. Study Selection. —We identified 66 studies that included data on the behavioral effects of HIV antibody counseling and testing. By consensus of the authors, 16 of these were excluded because of small sample size or inadequate study design. Data Extraction. —Studies were assessed by the authors according to methodological strength (sample selection, inclusion of appropriate comparison groups, and inclusion of statistical tests of significance). Data Synthesis. —All longitudinal studies of homosexual men reported reductions in risky behavior among both tested and untested men, and a few reported greater decreases among seropositive men than among seronegative men and those untested or unaware of their serostatus. For intravenous drug users in treatment, we found reductions in intravenous drug use and sexual risk behaviors regardless of counseling and testing experience. We found little evidence for the impact of counseling and testing on pregnancy and/or pregnancy termination rates for either seropositive or seronegative high-risk women. We noted substantial risk reduction among heterosexual couples with one infected partner. Findings among other heterosexuals at increased risk were scanty and mixed. Conclusions. —Further studies should specifically address the behavioral consequences of counseling and testing in various settings. (JAMA. 1991;266:2419-2429)