Analysis of factors associated with acceptability of pre-exposure prophylaxis among people at risk for HIV based on Bayesian networks

Abstract This study aimed to assess the acceptability of Pre-exposure prophylaxis (PrEP) and related influencing factors. A cross-sectional survey was conducted among female sex workers (FSW), people who inject drugs (PWID), and men who have sex with men (MSM) using respondent driven sampling. Factors influencing PrEP acceptability were estimated using multivariable logistic regression and Bayesian networks. The survey included 765 eligible participants. Overall, 33.5% and 25.3% showed moderate and high levels of PrEP acceptability, respectively. Multivariable logistic regression analysis revealed age, income, insurance, perceived utility, ease of use and risk of PrEP, perceived privacy in using PrEP, perceived comparative advantage of PrEP, and perceived image of PrEP user were associated with PrEP acceptability. The Bayesian network model showed perceived ease of use, perceived image as being health conscious, and perceived comparative advantage of PrEP over other methods concerning having sex whenever wanted were directly associated with acceptability of PrEP. If these three factors were at a high level, 74.6% of the participants would have a high level of acceptability of PrEP. By building complex models, Bayesian networks could identify the important factors associated with PrEP acceptability and provide inference based on prior probability of each factor.