Background: Measuring adherence to PrEP (pre-exposure prophylaxis) remains challenging. Biological adherence measurements are reported to be more accurate than self-reports and pill counts but can be expensive and not suitable on a daily basis in resource-limited countries. Using data from a demonstration project on PrEP among female sex workers in Benin, we aimed to measure adherence to PrEP and compare self-report and pill count adherence to tenofovir (TFV) disoproxil fumarate (TDF) concentration in plasma to determine if these 2 measures are reliable and correlate well with biological adherence measurements.
Methods: Plasma TFV concentrations were analyzed in samples collected at day 14 follow-up visit and months 6, 12, 18, and 24 (or at last visit when follow-up was shorter). Self-reported adherence was captured at day 14 follow-up visit and then quarterly by asking participants to report the number of missed pills within the last week. For pill count, medications were refilled monthly and participants were asked to bring in their medication bottles at each follow-up visit. Using generalized estimating equations adherence measured by self-report and pill count was compared to plasma drug concentrations.
Results: Of 255 participants, 47.1% completed follow-up. Weighted optimal adherence combining data from all visits was 26.8% for TFV concentration, 56.0% by self-report and 18.9% by pill count. Adherence measured by both TFV concentrations and self-report decreased over time (P = .009 and P = .019, respectively), while the decreasing trend in adherence by pill count was not significant (P = .087). The decrease in adherence was greater using TFV concentrations than the other 2 adherence measures.
Conclusion: With high levels of misreporting of adherence using self-report and pill count, the objective biomedical assessment of adherence via laboratory testing is optimal and more accurately reflects PrEP uptake and persistence. Alternative inexpensive and accurate approaches to monitor PrEP adherence should be investigated.