A comparison of different community models of Antiretroviral Therapy delivery among stable HIV+ patients in an urban setting, Zambia. A cluster-randomized non-inferiority trial.

Background: Community models of antiretroviral therapy (ART) delivery also known as differentiated service delivery (DSD) models are a novel innovative strategy to increase sustainable ART coverage for people living with HIV (PLHIV) in resource-limited settings. We compared two different models of ART delivery with the health care facility to gather evidence on the impact of these models’ patients’ clinical and virological outcomes, operational feasibility, and acceptability to guide policy makers on which models to roll out in the context of universal treatment. Methods: A three-arm cluster randomized non-inferiority trial was conducted in two urban HPTN 071 trial communities in Zambia comparing three different models of ART delivery: Standard of Care (SoC), Home-Based delivery (HBD) and Adherence Clubs (AC). Adult HIV+ patients defined as “stable” on ART, were eligible for inclusion. The primary endpoint was the proportion of PLHIV with virological suppression (≤1000 copies HIV RNA/ml) at 12 months (+/- 3 months) after study entry across all three arms. Analysis of our outcomes used statistical methods for CRT. Results: A total of 2,489 participants were enrolled in the study (781 SoC, 852 HBD, and 856 AC). There was a strong evidence (p<0.001) that both community models of ART delivery were non-inferior to SoC. The proportion of virological suppression in our three study arms > 95% compares favourably or superiorly with results published from literature. This trial also identified gaps in the evidence base and programmatic priorities for DSD implementation in SSA in the coming years with respect to viral load testing and monitoring and evaluation of DSD models embedded in routine HIV service delivery. Discussion: Community models of ART delivery were as effective as facility-based care in terms of viral suppression. However, availability of viral load test results remains a challenge to HIV programmes and could undermine gains from universal treatment. Offering PLHIV choices of these different models of ART is feasible and acceptable.