Objectives: To compare 3 measures of adherence to antiretroviral therapy (ART) in HIV-positive adults receiving free treatment from a public hospital in Malawi.
Methods: Adherence was measured over 1 month by pill count (PC), self-report, and a medication event monitoring system
Results: Data from 80 patients were available for analysis. The mean patient age was 38.6 years, and 57.5% were female. The mean adherence using the MEMS cap (MC) was 88.1%. Forty-six (57.5%) patients had MC adherence > or =95%, and 13 (16.2%) had <80% adherence. There was no association between MC adherence and time on ART. Mean PC adherence was 98.6%, significantly higher than MC adherence (P < 0.001). There was no clear relation between PC and MC adherence: 4 patients had MC adherence <20% but PC adherence of 100%. Self-reports of missing a tablet did not correlate with poor MC adherence.
Conclusions: The study shows the complexities of measuring adherence and probable overestimation of adherence by PC and self-report. Because these are the main methods used in developing countries, this raises concerns about the development of drug resistance. Improved methods are needed to detect nonadherence in developing countries, and validation of MC data with drug levels and virologic outcome in this setting is important.