Future HIV epidemic trajectories in South Africa and long-term consequences of reductions in general HIV testing: a mathematical modelling study

Background Following successful intensive interventions to rapidly increase HIV awareness, antiretroviral therapy (ART) coverage, and viral suppression, HIV programmes in eastern and southern Africa must now consider whether to scale-back certain programmes, such as widespread general population HIV testing services (general HTS), without risking a resurging epidemic or substantially increasing long-term ART need through slowed incidence declines. Methods We used a mathematical model (Thembisa) to project the South African HIV epidemic to 2100 under current epidemiologic and programmatic conditions. We assessed the epidemiological impact and cost of implementing general HTS reductions at different times between 2025 and 2050, while maintaining antenatal, symptom-based, and risk-based testing modalities and other HIV prevention. We considered how future uncertainty interacted with testing reductions by assuming positive or negative changes in ART interruption rates and condom usage over 2025–2035. Findings Under the status quo scenario, HIV incidence (15-49 years) steadily declined from 4.95/1000 (95% CI: 4.40–5.34) in 2025 to 0.14/1000 (0.05–0.31) in 2100, attaining <1/1000 in 2055 (2051–2060). When general HTS was scaled-back in 2025, incidence continued declining, but time to <1/1000 was delayed by 5, 13, and 35 years for a 25%, 50%, or 75% reduction in general HTS, and not attained by 2100 with full cessation. Reducing general HTS by 25% to 100% from 2025 resulted in 10% (8–12%) to 65% (53–77%) more new HIV infections and 7% (5–8%) to 46% (38–53%) more AIDS-related deaths over 50 years. Delaying general HTS reductions for 5 to 25 years mitigated some impacts. HIV testing accounted for only 5% of total programmatic costs at baseline. Reducing testing modestly reduced short-term total costs, but increased long-term costs. Changes in ART interruption rates and condom usage levels affected incidence decline rates and general HTS levels required to control transmission but did not cause rapid resurgent incidence. Interpretation Scaling-back general HTS did not result in resurging HIV infections, but it delayed attainment of incidence reduction targets and increased long-term expected infections, deaths, ART provision, and costs. HIV programmes face decisions balancing near-term health system resource savings by reducing intensive HIV programmes with epidemic control objectives over several decades. Funding BMGF, Wellcome, UKRI