County Level Mediators and Moderators of Latino HIV Prevalence Disparities in the United States

Introduction: Latinos/Hispanics in the U.S. are disproportionately affected by HIV. The prevalence rate of diagnosed HIV infection for Latinos in 2015 was nearly 2.5 times that of non-Latino whites. Latinos account for 25% of new HIV infection diagnoses yet only comprise 18% of the population. Latinos also have poorer outcomes along the HIV care continuum and fall behind non-Latino whites in the percentage of persons with diagnosed HIV infection, the percentage who have received any medical care, and the percentage living with HIV infection who are virally suppressed. It is important to understand the community-level factors that contribute to these disparities.

Methods: To evaluate HIV disparities between Latinos and non-Latino (NL) whites, we used publicly available data to model county-level, HIV disparities in prevalence rates. Data included race/ethnicity specific estimates of prevalence of diagnosed HIV infection from CDC’s National HIV Surveillance System, and 38 other county-level variables potentially associated with HIV prevalence that could serve as ecological mediators and modifiers of disparity. Prevalence rates of diagnosed HIV infection were fit using Poisson modeling of observed counts for both ethnic groups at the county level. Counties were treated as independent, and we include random effects for each racial/ethnic group to allow for extra-Poisson variation.

Results: A total of 781 counties were included in the analysis. Latinos had higher prevalence rates of diagnosed HIV infection than did NL whites with nearly 93% of counties having a Latino-NL white prevalence rate ratio (PRR) >1 and 6% were>10. In single covariate models associated with Latino-NL white HIV disparities, the % of people who inject drugs (IDU) and % Latinos who are Puerto Rican had the largest positive association with disparities, whereas % of men who have sex with men (MSM) and % uninsured had the largest negative associations. In the multicovariate model, 9 predictors were associated with Latino-NL white disparity (P < 0.05). An increase in Latino-NL white disparities was associated with an increase in: % IDU, % Latinos who are Puerto Rican, % rural, non-white/white segregation index, and % living below poverty. Disparity was greater in counties with lower: % uninsured, violent crime rate, % severe housing problems, and gonorrhea diagnosis rate.

Conclusions: This study is among the first to comprehensively examine a broad array of factors associated with Latino HIV prevalence disparities at the county-level. The findings highlight key factors associated with Latino disparities that can inform development and implementation of prevention and care interventions that can help achieve the ambitious goal of reducing health disparities in the US.