Health literacy, antiretroviral adherence, and HIV-RNA suppression: a longitudinal perspective

Abstract

Background: Low health literacy has been associated with worse adherence to antiretroviral therapy (ART) and higher HIV-RNA levels, but these relationships have not been evaluated in longitudinal analyses.

Methods: We evaluated literacy using the Rapid Estimate of Adult Literacy in Medicine (REALM) (< or = 6th grade, 7th to 8th grade, > or = 9th grade) in the HIV-Alcohol Longitudinal Cohort study of HIV-infected persons with a history of alcohol problems, conducted from 1997 to 2001. We tested HIV-RNA levels and administered a standardized questionnaire regarding demographics, substance use, receipt of ART, and adherence with ART, every 6 months for up to 7 occasions. Among the 235 subjects on ART, we investigated the relationship between literacy and 2 outcomes: 100% 3-day self-reported adherence and HIV-RNA suppression (<500 copies).

Results: Subjects' literacy levels were the following: 14% < or = 6th grade, 29% 7th to 8th grade, and 57% > or = 9th grade. In 66% of the observations (478/725), subjects reported 100% 3-day adherence with ART. Of the 685 HIV-RNA assays from these subjects, 62% had <500 copies. In unadjusted analyses, subjects with the lowest literacy level (< or = 6th grade) had a higher odds of adherence (odds ratio [OR] 2.23, 95% confidence interval 1.15 to 4.30) and HIV-RNA suppression (OR 2.01, 95% confidence interval 1.03 to 3.90) compared with those with > or = 9th grade literacy. This trend persisted but was no longer statistically significant in adjusted models of adherence (AOR 1.93, 95% confidence interval 0.86 to 4.31) and HIV-RNA suppression (AOR 1.70, 95% confidence interval 0.79 to 3.65).

Conclusion: Contrary to our hypothesis, low literacy was not associated with a lower odds of adherence or virologic suppression in this longitudinal analysis of HIV-infected patients with a history of alcohol problems. Indeed, trends in these data suggest the possibility that low literacy may be associated with a higher odds of adherence and virologic suppression. These counterintuitive findings underscore the need to pursue a fuller understanding of the mechanisms by which literacy affects health outcomes.

Figures

FIGURE 1

Sampling cascade for the analytic cohort.