Predictors of Retention in HIV Care Among a National Cohort of US Veterans

Background: Poor retention in HIV care leads to poor survival. The predictors of poor retention in HIV care are not well understood, especially from US nationwide data-sets. We determined the predictors of poor retention in HIV care among a group of US veterans and examined whether poor retention was confounded by other predictors of survival. Methods: We conducted a retrospective cohort study of 2,619 male US veterans who started antiretroviral therapy after January 1, 1998. Poor retention in HIV care was defined as having had at least 1 quarter-year without any primary care visit in the year after starting antiretroviral therapy. Survival was assessed through 2002. Logistic regression and Cox models were constructed. Results: Thirty-six percent of patients had poor retention in care. In multivariable analysis, younger age, Black race/ethnicity, CD4 cell count >350 x 10(6)/L, hepatitis C infection, and illicit drug use were predictive of poor retention in care. Having a chronic medical comorbidity and being identified as a man having sex with men (MSM) were associated with improved retention in care. In multivariable survival analyses, poor retention in care was not a confounder or moderator for other variables that predicted survival. Conclusions: Retention in HIV care is an independent predictor of survival. As routine HIV screening increases, more people with the characteristics predictive of poor retention in care will be identified. Interventions to improve retention in care are needed.