Client preferences for service delivery during the early treatment period in South Africa and Zambia: Mixed-method findings from a discrete choice experiment and concurrent focus group discussions

BackgroundDisengagement from antiretroviral therapy (ART) is common in the first 6 months of HIV treatment in sub-Saharan Africa. Using mixed-methods we aimed to understand preferences during this early treatment period. MethodsBetween 8/2023-11/2023, adults who had initiated/re-initiated ART a median of 8 months prior were enrolled at 18 healthcare facilities across South Africa (SA) and Zambia to participate in a discrete choice experiment (DCE) and focus group discussion (FGD). In the DCE, participants made 9 choices between unique service delivery scenarios (each comprised of 8 attributes). Analyzed using a conditional logit model, we report findings using odds ratios (95% confidence intervals). Following the DCE, FGDs explored barriers to care seeking and care preferences. Thematic analysis was used to interpret FGDs. DCE and FGD findings were triangulated to understand preferences. ResultsWe enrolled 250 respondents: 128 in Zambia (55% female, median age 35); 122 in SA (83% female, median age 33). Community-based services were less favorable to respondents than clinic-based care (SA: 0.62 (95% CI 0.52, 0.75); Zambia: 0.44 (0.36, 0.53)). Respondents preferred 6-month dispensing (SA: 1.3 (1.1, 1.6); Zambia: 2.1 (1.8, 2.6)) to 1-month intervals. Respondents also preferred accessing services from friendly providers. Qualitative insights corroborated DCE findings. They also revealed frustrations with long wait times at clinics. ConclusionUtilizing a decision experiment with qualitative methods allowed us to uniquely capture drivers of client decision-making and the nuanced factors that shape experiences. Results suggest that enrollment in lower-intensity models may improve client experiences during the early treatment period.