HIV risk difference between condom users and nonusers among U.S. heterosexual women

Using data from the National Survey of Family Growth, we estimate that among 3,498,060 U.S. reproductive-age women least likely to be protected from HIV, 12% rely on condoms for birth control. We have modeled the risk difference between condom users and nonusers and projected the number of preventable and nonpreventable HIV infections likely to occur among the 419,201 condom users as a function of 50 HIV-incidence/relative risk assumptions. Results of the attributable-risk model suggest that at the current low HIV-incidence level in U.S. women, condom-user failure rates will be less than 1% per year, substantially lower than the 10% condom-user failure rate for pregnancy. As few as 1% but up to 11% of all new HIV cases may be prevented by the current low level of condom use, depending on the exact degree of condom effectiveness in this population at risk. However, the model further projects that up to 45% of all new HIV cases may be prevented if condoms are maximally effective and far more widely used. Women with seropositive partners may enjoy the same protective benefits of condoms, but the costs in terms of condom-user failures will be much higher than in the remainder of the population at risk. Among serious and reliable users, condom-user failure rates for HIV may approach those for pregnancy, but only in women who have known seropositive partners.