The January 2025 US foreign-aid freeze and PEPFAR stop-work orders affected 15,374 PEPFAR-funded HIV response staff in South Africa, nationally and in 27 priority districts, with estimated staff costs of ZAR 4.6 billion (UNAIDS, 2025). These staff supported HIV testing, linkage to care and treatment retention in the country with the world's largest HIV epidemic. Projections and early descriptive reports show service declines, but causal evidence linking these funding cuts directly to declines in HIV testing and antiretroviral therapy (ART) remains limited. This matters because falls in testing or ART refills can also reflect seasonal migration, local economic shifts or concurrent health policy changes. Descriptive data show that services declined, not that the freeze caused the decline. Causal estimation isolates the impact attributable to the funding freeze itself, which is the evidence policymakers need to target recovery. The impact is also unlikely to be uniform: districts and facilities differed in their reliance on PEPFAR support, and patients leaving affected clinics may shift pressure onto neighbouring ones. The study aims to estimate the causal effect of the disruption on HIV testing and ART outcomes, assess how this effect varies across districts and facilities, and determine whether it spilled over to neighbouring facilities. This longitudinal quasi-experimental study will analyse routine monthly facility-level programme data from January 2022 to December 2026, linked to facility locations. A facility will be classed as disrupted from the first month it lost PEPFAR-funded staff or support, using partner close-out records verified against facility service records. Because facilities lost support at different times, a staggered quasi-experimental design will compare affected facilities with similar facilities not yet or never affected, adjusting for baseline differences such as HIV burden, facility type and location. Trends before the disruption will be checked to confirm that the groups were comparable. ART continuity is the primary outcome; HIV testing and ART initiation are secondary. Effects will be tracked month by month to show whether services recovered or continued to decline. Causal machine learning will identify which facilities and communities were hardest hit, by baseline HIV burden, socioeconomic vulnerability and facility capacity. Effects will be mapped to show how they vary across districts. Spillovers capture patient movement: when a clinic loses staff, patients may seek care at nearby clinics, adding to their workload and possibly affecting their services. Neighbouring facilities will be defined by overlapping patient catchment areas, approximated by travel distance between facilities. Robustness checks will test whether the findings hold when disruption dates are shifted, when outcomes not funded by PEPFAR are examined, and when facilities with incomplete reporting are excluded. The study will produce district-level estimates of HIV tests, ART initiations and retained ART clients lost to the funding freeze, with maps showing where losses were concentrated. These estimates will show where recovery resources are most needed to protect South Africa's HIV response. The approach also offers a transferable framework for evaluating sudden donor funding withdrawals in other high-burden countries.
