Patterns and predictors of high-risk sexual behavior in female partners of HIV-infected men with hemophilia
Dublin, S.; Rosenberg, P.S.; Goedert, J.J.
Aids 6(5): 475-482
1992
ISSN/ISBN: 0269-9370 PMID: 1616653 DOI: 10.1097/00002030-199205000-00006Document Number: 318859
This study sought to characterize and quantify the high-risk heterosexual activity in HIV-discordant couples. An analysis of cross-sectional and longitudinal questionnaire data from 217 HIV-negative female sexual partners of HIV-infected hemophiliac men were included in this study. There was a comparison of prevalence rates of anal sex, oral sex, vaginal intercourse with or without condoms, and use of other contraceptives between 1985-91. Logistic regression analysis of demographic, sexual, and clinical variables was used to predict unprotected vaginal sex. Actuarial estimates of semiannual relapse rates to unsafe sex were used. The proportion of women at low risk increases from 7 to 69% between 1985-91, mainly because more women were using condoms during all acts of vaginal intercourse. Other contraceptive practices did not change during this time. The proportion who engaged in oral or anal sex decreased from 26 to 13% and from 13% to 4%, respectively. Unprotected vaginal sex was more common among women who enrolled earlier, had less education, engaged in oral or anal sex, and among those whose partners had not had AIDS. Unprotected vaginal sex before enrollment was the strongest predictor of this high-risk activity during followup. 2-year rates of relapse to high-risk behavior were significantly higher among women who enrolled at high risk compared with those who enrolled at low risk (39 vs 8%, p=0.005). Although high risk sexual behavior become much less prevalent in this population between 1985-91, many continued to have unprotected vaginal sex occasionally. Counseling efforts should target couples who have been the most sexually active or have less education, and should emphasize not only initial risk reduction but also maintenance of low-risk behavior.