Clients in Uganda accessing preferred differentiated antiretroviral therapy models achieve higher viral suppression and are less likely to miss appointments: a cross‐sectional analysis

INTRODUCTION: The Uganda Ministry of Health recommends facility‐ and community‐based differentiated antiretroviral therapy (DART) models to support person‐centred care for eligible clients receiving antiretroviral therapy (ART). Healthcare workers assess client eligibility for one of six DART models upon initial enrolment; however, client circumstances evolve, and their preferences are not routinely adjusted. We developed a tool to understand the proportion of clients accessing preferred DART models and compared the outcomes of clients accessing preferred DART models to the outcomes of clients not receiving preferred DART models. METHODS: We conducted a cross‐sectional study. A sample of 6376 clients was selected from 113 referrals, general hospitals and health centres purposely selected from 74 districts. Clients receiving ART accessing care from the sampled sites were eligible for inclusion. Healthcare workers interviewed clients (caretakers of clients under 18), over a 2‐week period between January and February 2022 using a client preference tool to elicit whether clients were receiving DART services through their preferred model. Treatment outcomes of viral load test, viral load suppression and missed appointment date were extracted from clients’ medical files before or immediately after the interview and de‐identified. The descriptive analysis determined the interaction between client preferences and predefined treatment outcomes by comparing outcomes of clients whose care aligned with their preferences to outcomes of clients whose care misaligned with their preferences. RESULTS: Of 25% (1573/6376) of clients not accessing their preferred DART model, 56% were on facility‐based individual management and 35% preferred fast‐track drug refills model. Viral load coverage was 87% for clients accessing preferred DART models compared to 68% among clients not accessing their preferred model. Viral load suppression was higher among clients who accessed the preferred DART model (85%) compared to (68%) clients who did not access their preferred DART model. Missed appointments were lower at 29% for clients who accessed preferred DART models compared to 40% among clients not enrolled in the DART model of their choice. CONCLUSIONS: Clients who accessed their preferred DART model have better clinical outcomes. Preferences should be integrated throughout health systems, improvement interventions, policies and research efforts to ensure client‐centred care and client autonomy.