Area Socioeconomic Factors and Relative Survival after a Diagnosis of HIV, United States, 1996–2003

Purpose HIV/AIDS survival differs by race/ethnicity and has been linked to low socioeconomic status (SES) in limited U.S. studies. We estimated the survival of HIV-diagnosed persons by race/ethnicity and SES factors using population-based data. Methods We estimated 5-year relative survival (RS) by 10-year age groups, race/ethnicity, transmission category, sex, diagnosis year, CD4 count within 6 months of AIDS diagnosis, and 4 county-level SES variables from the U.S. Census for all HIV cases >12 years diagnosed 1996-2003 and followed-up through 2005 in the national HIV/AIDS Reporting System. We calculated RS proportions using a maximum likelihood algorithm (expected survival from life tables), and modeled the relative excess hazard of death (REH) using generalized linear models, with poverty as a random effect. Results Black males had worse 5-year RS [86.5% (95% confidence interval (CI) 86.2, 86.6)] than whites and Hispanics (90% for both). RS was worse for males in counties with larger proportions of people living below the 2000 U.S. poverty level [87.7% (95% CI 87.2, 88.3) for poverty ≥ 20% vs. 90.1% (95% CI 88.9, 91.2) for poverty <5%], and where unemployment was greater [87.8% (95% CI 87.5, 88.1) where unemployment was >7.1% vs. 90.5% (95% CI 90.0, 91.0) where unemployment was < 4%]. Similar effects of SES on RS were noted for females. REH for males and females within 5 years of an HIV diagnosis was significantly worse in counties that were 10-19.9% below the poverty level (REH=1.3, 95% CI 1.2,1.5 and REH=1.8, 95% CI 1.4, 2.2, respectively) compared with < 5%, controlling for race/ethnicity, transmission category, age, CD4 count, and AIDS at HIV diagnosis. Conclusions HIV survival was worse in lower SES areas, but SES as used here did not explain race/ethnic differences in HIV survival. To help address the impact of underlying economic issues, HIV/AIDS resources for testing, care, and proven economic interventions should be targeted to areas with concentrations of economically disadvantaged people, especially females.