Statistical Analysis

We assessed the three VL measures by the following stratification variables: sex (male, female), race and ethnicity (black/African American, Hispanic/Latino, white, and other), age (based on the person’s age at the end of 2010, with persons assigned to one of 5 groups: 13–24, 25–34, 35–44, 45–54, and ≥ 55 years), transmission category (based on a presumed hierarchical order of probability of infection by sex: for males, male-to-male sexual contact, injection drug use, male-to-male sexual contact and injection drug use, heterosexual contact; for females, heterosexual contact, injection drug use), year of diagnosis (diagnosed before 2008 or in 2008–2010), and gaps in care (had a gap [consecutive VL tests >12 months apart during the 2-year period], vs. no gap).

We determined the numbers and percentages of persons in HIV care who had durable viral suppression over two years and estimated univariate and multivariate prevalence ratios (PRs) with confidence intervals (CI), derived from binomial regression models, to identify differences between groups. For viremia copy-years and person-time measures, we used t-tests and multivariate regression models to examine group differences. Because the transmission categories were stratified by male and female, a separate sex variable (male vs. female) was not included in the multivariate models. All analyses were conducted in SAS version 9.3 (SAS Institute Inc, Cary, NC).