Abstract
BACKGROUND: Female sex workers (FSWs) in Uganda are at high risk of HIV infection. Scaling up oral pre-exposure prophylaxis (PrEP) will reduce HIV incidence if high levels of adherence are maintained. This study evaluates PrEP adherence using clinic-based pill counts and self-reported measures, and factors associated with protective levels of adherence. METHODS: Participants were sex workers who had been taking PrEP for at least 5 months and were attending routine follow-up visits for PrEP care in fishing communities and along the Trans-African Highway. Participants who had a pill count showing at least 85% use since their last clinic visit and those who reported taking their PrEP every day in the last 5 months were categorised as having 'protective adherence'. Spearman's correlation and weighted kappa assessed the relationship between pill count and self-reported measures. Bivariate and multivariate logistic regression was used to determine factors associated with protective adherence as measured by pill count. RESULTS: We recruited 524 FSWs, with a median age of 29 years (IQR 23-35). Participants were recruited from fishing communities and Trans-African Highway towns (n = 297, 56.7%, and n = 227, 43.0%). Nearly three quarters (n = 372, 71.0%) of women were estimated to have protective adherence based on pill count (i.e., a pill count of >85%) and 50.4% by self-report in last 3 months. There was a strong positive association between self-reported measures and pill count measures (rest = 0.6453, 95% CI = 0.5924-0.6927) and a moderate agreement between self-reported measures and pill count measures, κ = 0.544 (95%CI = 0.4869-0.6011, p < 0.001). Factors associated with protective adherence included being aged 35 years or older (aOR = 2.40, 95% CI = 1.17-4.86), living in a fishing community (aOR = 1.45, 95% CI = 0.62-3.38), and having an STI in last 3 months (aOR = 1.64, 95% CI = 1.07-2.49). CONCLUSION: Our findings indicate that PrEP-experienced FSWs attending clinical follow-up visits reported high protective levels of oral pre-exposure prophylaxis, as measured by both pill count and self-reported measures, and a moderate agreement between pill count and self-reported measures.