Timeliness and retention in a 6-month follow-up study were analyzed by subjects' baseline characteristics in a seroincidence study of 263 injection drug users at high risk of human immunodeficiency virus infection. Subjects were recruited from September 1997 to June 1998 in community settings in Philadelphia, Pennsylvania. Of these subjects, 93% were completers: 11% before the targeted date, 38% at the targeted date, 32% within 1 month of delay, and 12% beyond 1 month. Late completers were more likely than other completers to be younger and to live farther away from the study center, less likely to have stayed in a shelter or a welfare residence during the past year, more likely to have a lower income, and more likely to have shared rinse water, cotton, or cooker. By contrast, loss to follow-up was not associated with these variables. Subjects lost to follow-up were more likely than those retained to have a high school diploma and to have moved during the past year; their source of needles was less likely to be a needle exchange program and more likely to be a shooting gallery. None of the drug-related behaviors that increase the risk of human immunodeficiency virus infection was associated with timeliness or retention, suggesting that the study might be minimally biased.

Background HIV prevention interventions in the United States have failed to eliminate racial inequities. Here, we evaluate factors associated with racial inequities in HIV prevalence among people who inject drugs using HIV Prevention Trial Network 037 data. Methods We measured racial homophily (ie, all members share the same race), being in an HIV+ network (network with ≥1 HIV+ member), and drug and sex risk behaviors. A 2-level logistic regression with a random intercept evaluated the association between being in an HIV+ network and race adjusting for individual-level and network-level factors. Results Data from 232 index participants and 464 network members were included in the analysis. Racial homophily was high among blacks (79%) and whites (70%); 27% of all-black, 14% of all-white, and 23% of racially mixed networks included HIV+ members. Sex risk was similar across networks, but needle sharing was significantly lower in all-black (23%) compared with all-white (48%) and racially mixed (46%) networks. All-black [adjusted odds ratio (AOR), 3.6; 95% confidence interval (CI), 1.4 to 9.5] and racially mixed (AOR, 2.0; 95% CI: 1.1 to 3.7) networks were more likely to include HIV+ network members; other factors associated with being in HIV+ network included homelessness (AOR, 2.0; 95%CI, 1.2 to 3.2), recent incarceration (AOR, 0.4; 95%CI, 0.2 to 0.7), and cocaine injection (AOR, 1.7; 95% CI, 1.0 to 2.7). Risk behaviors were not associated with being in an HIV+ network. Conclusion Despite having lower drug risk behavior, all-black networks disproportionately included HIV+ members. HIV prevention interventions for people who inject drugs need to go beyond individual risk and consider the composition of risk networks.