Abstract
INTRODUCTION: Little is known about how people living with HIV would choose if offered different tuberculosis preventive treatment (TPT) regimens, and under which conditions they would accept treatment. Actionable evidence regarding preference for TPT is needed to inform policy and the development of novel TPT regimens. METHODS: Adults engaged in care at an HIV clinic in Kampala, Uganda, completed a discrete choice experiment survey with nine random choice tasks. In each task, participants first chose between two hypothetical TPT regimens with differing treatment features (number of pills, frequency, duration, adjusted antiretroviral dosage, and side effects). Second, they answered if they would accept the selected treatment, versus taking no treatment. We simulated predicted TPT regimen choice based on hierarchical Bayesian estimation of individual preference weights. RESULTS: Among 400 participants, 394 gave high-quality answers and were included (median age 44, 71.8% female, 91.4% previously received TPT). Across nine tasks, 60.2% (237/394) accepted all selected TPT regimens, 39.3% (155/394) accepted some regimens, and 0.5% (2/394) accepted none. Regimens requiring antiretroviral dosage adjustment were more likely to be unacceptable (adjusted odds ratio, aOR 27.4, 95% confidence interval [CI] 18.5 - 40.7), as were regimens requiring more pills per dose (aOR 24.5 [95% CI 16.6 - 36.3] for 10 pills compared to 1 or 5 pills per dose). Choice simulations showed that if only 6 months of daily isoniazid (6H) was available, 11.9% would prefer no TPT. However, offering a 4-pill, fixed-dose combination 3HP regimen in addition to 6H increased the acceptability from 88.1% to 98.8% (predicted choice of 3HP 94.5%, 6H 4.4%, no TPT 1.2%). CONCLUSIONS: While adults living with HIV in Uganda demonstrate a high willingness to accept different TPT regimens, offering regimens with preferred features, such as 3HP as a fixed-dose combination, could drive TPT acceptance and uptake from high to nearly universal.