Individual and neighborhood predictors of mortality among HIV-positive Latinos with history of injection drug use, Florida, 2000-2011

Background: The objectives are to examine disparities in all-cause mortality risk among HIV-positive Latinos with injection drug use (IDU) history, and to identify individual- and neighborhood-level predictors. Methods: Florida surveillance data for persons diagnosed with HIV 2000-2008 were merged with 2007-2011 administrative data from the American Community Survey. Hazard ratios (HR) were calculated using multi-level weighted Cox regression adjusting for individual and neighborhood (ZCTA-level) factors. Results: Of 10,989 HIV-positive Latinos, 10.3% had IDU history. Latinos with IDU history were at increased mortality risk compared with Latinos without IDU history after controlling for individual and neighborhood factors (adjusted HR [aHR] 1.61, 95% confidence interval [CI] 1.43-1.80). Factors associated with mortality for those with IDU history included: being 40-59 (aHR 6.48, 95% CI 1.41-121.05) and ≥ 60 years (aHR 18.75, 95% CI 3.83-356.45) compared with 13-19 years of age; being diagnosed with AIDS within 3 months of HIV (aHR 2.31, 95% CI 1.87-2.86); residing in an area with ≥50% Latinos compared with <25% Latinos (aHR 1.56, 95% CI 1.19-2.04); and residing in a rural compared with an urban area at the time of diagnosis (aHR 1.73, 95% CI 1.06-2.70). Race and neighborhood poverty were not predictors among those with IDU, but were among those without. Conclusion: HIV-positive Latinos with IDU history are at increased mortality risk and have unique contributing factors. Tertiary prevention strategies should target those who are older, diagnosed at later stages, and those who live in predominantly Latino and rural areas.