Magnitude and determinants of behavioral risk reduction: Longitudinal analysis of a cohort at risk for AIDS

The magnitude and predictors of longitudinal behavioral change are reported in a cohort of homosexual men at risk for AIDS. Self-reports of sexual behavior were obtained at two points in time separated by an interval of approximately six months. These self-reports were used to construct both dichotomous and continuous measures of changes in behavior consistent with reduction in the transmission of the AIDS virus (HIV). Although there was considerable variability in behavior. mean changes were consistently in the desired direction. Avoidance of anonymous sexual partners, monogamy, and modification of receptive anal sex to reduce exposure to semen by condom use or withdrawal prior to ejaculation appeared to be especially important in this cohort. Both multiple linear regression and multiple logistic regression were used to examine the relationship between a model of health behavior and these outcomes. Variables examined included knowledge of AIDS. perceived risk of AIDS. the perceived efficacy of behavior in reducing AIDS risk, difficulties with sexual impulse control, belief in biomedical technology to provide a prevention or cure, social norms supportive of behavioral change, and gay network affiliation. Of all these factors, only the availability of supportive peer norms was consistently, significantly and positively related to multiple measures of outcome. Differences between these analyses and longitudinal analyses reported elsewhere are discussed. These results suggest the policies regarding HIV antibody testing should be developed cautiously. taking account of the failure of a sense of risk to predict subsequent behavioral change. They also emphasize the important role of gay organizations in developing social norms supportive of behavioral risk reduction.