ABSTRACT Globally, more than half of the world's 37 million people living with HIV are on antiretroviral therapy (ART) representing immense and encouraging success with access to HIV care. ART prevents disease, death and HIV transmission and HIV-positive persons can expect to live as long as their HIV-negative peers when their viral load is undetectable. However, treatment success still lags behind goals. In South Africa alone, 8 million HIV- positive persons require ART for life and only 4.5 million are currently on ART. Patient barriers to care, such as missed wages, transport costs, and long wait times for clinic visits and ART refills, are associated with detectable viral load, the hallmark of struggling to access and take ART. HIV differentiated service delivery (DSD) has simplified ART delivery: incentives, multi-month scripts, fast-track ART, and community or home ART delivery motivate clients, reduce the frequency of clinic visits, and decongest clinics. DSD is standard for clients who achieve viral suppression and engage in care; however, DSD needs adaptation to serve clients who are not succeeding. Indeed, persons who are not engaged in care arguably need simplified, client-centered approaches even more than those who can successfully engage. A suite of adaptive DSD strategies, including community-based ART, have been tested among stable clients with viral suppression. Lottery incentives effectively change short-term behavior, increasing ART initiation. Community and home ART delivery increases ART coverage and simplify ART access overcoming clinic barriers. For stable clients, these DSD activities are as effective as clinic-based care in terms of achieving and maintaining viral suppression, although among stable clients they have not shown superiority in viral suppression or cost savings. In contrast, DSD has the potential to improve rates of viral suppression and retention in care and save costs among more hard-to-reach groups. There is great potential that DSD systems can be client- responsive and system-efficient for subgroups requiring additional services, matching services with client needs. A sequential, comprehensive package of DSD approaches, with each step increasing the intensity of service provision – adaptive DSD – has not been tested to determine the proportion and characteristics of persons who would achieve viral suppression and retention in care and to estimate the cost-effectiveness and budget impact. To increase population level viral suppression, persons with detectable viral load need responsive DSD interventions. A Sequential Multiple Assignment Randomized Trial (SMART) design facilitates evaluation of a stepped, adaptive approach to achieving viral suppression with `right-sized' interventions. We are an experienced team and propose to build on our strong partnerships to sequentially test adaptive DSD strategies for persons with detectable viral load: lottery incentives, community-based ART, and home ART delivery. Our aim is to identify the most effective and efficient HIV care delivery strategies for South Africa.

PROJECT NARRATIVE In South Africa, where 8 million persons are living with HIV, 56% are on ART, and only 45% are virally suppressed (the key indicator of treatment success) due to barriers to care including offering streamlined care only those persons demonstrating treatment success. To address this implementation gap, we propose to test the effect of adaptive differentiated service delivery by offering, in sequence, more tailored services to clients with detectable viral load and those not engaged in care. Working closely with the local HIV clinic, we will leverage ongoing changes in clinic practices to determine the impact of adapting the best clinic practices to include lottery incentives, community ART, and home ART delivery on viral suppression and cost-effectiveness at population level.