Utility of Different Adherence Measures for PrEP: Patterns and Incremental Value

Measuring PrEP adherence remains challenging. In 2009–2010, the International AIDS Vaccine Initiative randomized phase II trial participants to daily tenofovir disoproxil fumarate/emtricitabine or placebo in Uganda and Kenya. Adherence was measured by electronic monitoring (EM), self-report (SR), and drug concentrations in plasma and hair. Each adherence measure was categorised as low, moderate, or high and also considered continuously; the incremental value of combining measures was determined. Forty-five participants were followed over 4 months. Discrimination for EM adherence by area under receiver operating curves (AROC) was poor for SR (0.53) and best for hair (AROC 0.85). When combining hair with plasma or hair with self-report, discrimination was improved (AROC > 0.9). Self-reported adherence was of low utility by itself. Hair level was the single best PK measure to predict EM-assessed adherence; the other measurements had lower discrimination values. Combining short-term (plasma) and long-term (hair) metrics could be useful to assess patterns of drug-taking in the context of PrEP.