Abstract
Pharmacy refill records (PRR), are an accessible strategy for estimating adherence in low- and middle-income countries (LMICs). However, the low-cost urine-tenofovir point-of-care test opens up the possibility of an objective metric of adherence that is scalable to LMICs. This study compared adherence to tenofovir-based regimens using urine-tenofovir point-of-care (POC) test with pharmacy refill records in a Nigerian population of HIV-positive persons. This was a cross-sectional study among 94 HIV-positive adults, which was conducted from June to August 2021, in a large outpatient clinic in Lagos, Nigeria. Adherence to pharmacy appointments was automatically calculated using a computerized pharmacy appointment system (FileMaker Pro™). Urine drops on the urine-tenofovir POC test strip developed 2 lines for a negative test (tenofovir absent) and one line for a positive test. Fisher's exact test was used to examine the association between pharmacy refill record and urine-tenofovir point-of-care test. Logistic regression was performed to predict viral suppression (<1000 copies/mL, based on WHO recommendations) using both methods of adherence determination. A Receiver Operating Characteristic (ROC) curve of the association between specificity and sensitivity was generated to evaluate the predictive value of adherence determined using pharmacy-refill record and urine-tenofovir point-of-care test in forecasting viral suppression. The statistical significance level was set at 0.05. Fisher's exact test showed no statistically significant difference in adherence using urine-tenofovir point-of-care test or pharmacy refill record. The logistic regression model showed that an increase in pharmacy-refill record of ≥ 95% was associated with viral suppression (P = .019). From the ROC curve, the sensitivity was same at 95.5% for both methods, but the specificity of the urine-tenofovir point-of-care test was greater (96.6% vs 95.5%) than pharmacy refill record (P = .837). Urine-tenofovir point-of-care test provided equivalent adherence data to pharmacy refill data.