Abstract
This study: 1) examined the rate and correlates of human immunodeficiency virus (HIV) seropositivity; and 2) assessed whether self-selection in HIV testing influenced the rate and correlates of HIV seropositivity in a group of out-of-treatment drug users. Data were collected from 856 out-of-treatment drug users in Philadelphia between January 1993 and August 1994. Seventy-four percent of the sample elected to take an HIV test that was included in the project in which the drug users were enrolled, and of these, 11% were HIV positive. Multivariate analyses indicated that those who were younger, those who had an injection-drug-using sex partner, and those who reported no recent sexual activity were more likely to be HIV positive. An examination of the multivariate correlates of HIV testing indicated that subjects who took an HIV test had higher rates of participation in some risk behaviors than did subjects who did not take an HIV test, but lower rates for other risk behaviors. None of the correlates of HIV seropositivity were correlates of HIV testing. Although the generalizability of the HIV seroprevalence rate is unclear, it is probable that the correlates of HIV seropositivity are generalizable to the total sample. The results of this study indicate the importance of interventions that target sexual risk behavior among out-of-treatment drug users, and of assessing the impact of self-selection bias whenever the rate and correlates of HIV seropositivity are examined.