PEPFAR interrupted: real-world consequences of U.S foreign aid instability for HIV service delivery in South Africa

In January 2025, funding through the U.S. President's Emergency Plan for AIDS Relief was disrupted. While modeling studies have projected the consequences of these financing interruptions, models are not based on population-representative, real-world impacts on service delivery and clinic operations. We aimed to quantify disruptions to facility-level services, operations, and staffing resulting from foreign aid instability using province-representative facility audit data from South Africa. We used a probability-proportional-to-facility-size sampling approach to select public facilities providing HIV care in KwaZulu-Natal, South Africa for inclusion in Uhambo Lwami, a multisite, observational cohort study. Structured, cross-sectional surveys were conducted with clinic Operational Managers from January to July 2025 to document interruptions to clinic services, operations, and staffing. Perceived impacts of these interruptions were also captured. Province-representative estimates of the prevalence of facilities impacted and patients served by impacted facilities were quantified using inverse probability weighting. Thirty-six of 519 clinics (6.9%) were sampled. A total of 179,586 unique people with HIV were seen across sampled clinics (11.1%), representing 1,622,247 people with HIV after weighting. Fifteen sampled clinics (41.7%) serving 88,620 people with HIV, representing 200 (38.5%) clinics serving 828,413 (51.1%) people with HIV in the province, reported service, operations or staffing interruptions due to funding cuts. The most common operational interruption was in data entry/filing, affecting 13.4% of clinics (N = 70). Patient tracing, HIV testing, and treatment services were interrupted in 9.9% of clinics (N = 52) serving 429,934 people with HIV (26.5%) in the province. Nearly 30% of clinics (N = 152) in the province experienced staffing disruptions; layoffs of data capturers (N = 142; 27.3%) and community health workers (N = 121; 23.4%) were most common. Increases in clinic wait times were reported in 11.0% (N = 57) of facilities.