Impact of an international HIV funding crisis on HIV infections and mortality in low-and middle-income countries: a modelling study

  1. Debra ten Brink
  2. Rowan Martin-Hughes
  3. Anna L Bowring
  4. Nisaa Wulan
  5. Kelvin Burke
  6. Tom Tidhar
  7. Shona Dalal
  8. Nick Scott

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Background

International funding for HIV has been critical in reducing new HIV transmissions and deaths. Five countries providing over 90% of international HIV funding announced reductions in international aid of 10%–70% between 2025–2026, with the US government ceasing aid on 20-January-2025. We investigated the potential impact of these funding reductions on HIV incidence and mortality through mathematical modelling.

Methods

We used 26 country-validated Optima HIV models (Albania, Armenia, Azerbaijan, Belarus, Bhutan, Cambodia, Colombia, Costa Rica, Côte d’Ivoire, Dominican Republic, Eswatini, Georgia, Kazakhstan, Kenya, Kyrgyzstan, Malawi, Malaysia, Moldova, Mongolia, Mozambique, South Africa, Sri Lanka, Tajikistan, Uganda, Uzbekistan, Zimbabwe). HIV incidence and mortality were projected across 2025–2030 for a status-quo scenario (most recent HIV spending continued) and scenarios capturing the impact of anticipated international aid reductions for HIV prevention and testing; plus additional impact on treatment and facility-based testing resulting from immediate discontinuation of President’s Emergency Fund for AIDS Relief (PEPFAR) support. Country-specific impacts were estimated using sources of country-reported HIV funding. We extrapolated scenario outcomes to all low-and middle-income countries based on the modelled proportion of globally reported international aid by source (49% overall, 54% PEPFAR). Upper and lower bounds reflected different mitigation and absorption assumptions.

Findings

Across all low-and middle-income countries, anticipated 24% (weighted average) international aid reductions plus discontinued PEPFAR support could cause an additional 4·43–10·75 million new HIV infections and 0·77–2·93 million HIV-related deaths between 2025–2030 compared with the status-quo. If PEPFAR support could be reinstated or equivalently recovered, this reduced to 0·07–1·73 million additional new HIV infections and 0·005–0·06 million HIV-related deaths. Impacts were greatest in countries with a higher percentage of international funding and those with increasing incidence among key populations.

Interpretation

Unmitigated funding reductions could significantly reverse progress in the HIV response by 2030, disproportionately affecting sub-Saharan African countries and key and vulnerable populations. Sustainable financing mechanisms are critical to ensure people have continued access to HIV prevention, testing, and treatment programs, thereby reducing new infections and deaths.

Funding

None.

Article activity feed

  1. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/15282576.

    Impact of an international HIV funding crisis on HIV infections and mortality in low- and middle-income countries: a modelling study

    https://www.medrxiv.org/content/10.1101/2025.02.27.25323033v3

    This preprint describes the potential impact of international funding cuts on new HIV infections and mortality in low- and middle-income countries (LMICs), cutting across Eastern Europe, Africa, Asia, and South America.

    Twenty-six (26) country-validated Optima HIV models were used and extrapolated to other LMICs around the world. The Optima HIV models have been used to inform national HIV/AIDS strategies, impact evaluations, and allocative efficiencies. 

    Five funding scenarios were modeled in this study: Scenario 1: status quo; Scenario 2: proportional cuts by 4.4% in 2025 and a further 19.6% by 2026 and maintained at this level for all years after 2026; Scenario 3: prevention budget reallocated to treatment; Scenario 4: discontinued PEPFAR support with mitigation; Scenario 5: discontinued PEPFAR support without mitigation. The paper focused on children, all adult populations and key adult populations.

    This study projects that about 4% of international aid reductions plus discontinued support from PEPFAR could lead to an additional 4.43 - 10.75 million new HIV infections and 0.77-2.93 million HIV-related deaths between 2025-2030 compared with the status quo (secnario 1), which potentially will reverse years of progress in HIV response and control globally.

    The models used in this study demonstrate diverse country representation and a comprehensive scenario design, while focusing on key HIV populations. However, the assumptions of similar effects and extrapolation to all LMICs from 26 countries are overly simplistic and may not account for regional and contextual variations. This study, however, comes at a time when many HIV-burdened countries are battling major funding cuts on health projects, and will be essential in informing HIV control programs on the best strategies to stay on track.

    Major Considerations

    1. Assumption of equal funding cuts: The assumption that funding cuts (4.4% in 2025 and 19.6% in 2026) would affect each country's international funding equally is simplistic. International aid varies by country. The authors can address this by refining their methodology to replace uniform funding cuts with weighted reductions based on each country's proportion of international HIV funding. The countries can also be grouped by dependency level and the funding cuts modeled proportionally.

    2. Model calibration transparency: The authors should provide detailed calibration data sources, procedures and metrics (could be an appendix or table) listing the exact sources, calibration targets or parameter values for each country model. The paper will benefit from quantitative measures such as goodness of fit on how well each model reproduced historical HIV trends mentioned in the paper, plus details on what specific "country-validated" modifications were made.

    3. Focus on PEPFAR: While PEPFAR is a major funder (92% of the US's 75% HIV Aid), the scenarios focus primarily on PEPFAR funds, with the potential of overemphasizing its role while underrepresenting the potential impact of other international funding sources.

    4. Predictive ability of the model: While the authors cite other uses of the Optima HIV model in allocative efficiency, they should provide more evidence on the model's predictive ability for HIV funding cuts by citing literature (if any exist) where the Optima HIV model succesfully predicted HIV cuts or a succesful prediction in countries that have expereinced changes in HIV funding. 

    Minor Issues:

    1. Aggregating results across all 26 modeled countries obscures country-specific burdens. Presenting disaggregated data in the main text or a supporting document would help in understanding which countries are most vulnerable to the cuts.

    2. Figure 3, which shows the incidence of new HIV infections, combines all regions, making it difficult to discern specific impacts or burden. The authors acknowledged the disregard of the smaller values, but may want to consider separating the regions into different figures.

    Opportunities for Improvement:

    1. Lack of other statistical methods beyond absolute numbers: The authors acknowledge the focus on absolute numbers rather than relative density, which may mask impacts on key populations and should explore other statistical approaches such as hazard ratios.

    2. Pregnant women living with HIV (PWLHIV) are not explicitly categorized as a key population but are considered vulnerable and should be included.

    3. Sensitivity analysis: The authors should conduct and present sensitivity analyses to address uncertainties in model parameters and assumptions, such as the equal funding cuts assumption.

    Competing interests

    The authors declare that they have no competing interests.

    Use of Artificial Intelligence (AI)

    The authors declare that they used generative AI to come up with new ideas for their review.

  2. Version published to 10.1101/2025.02.27.25323033v4 on medRxiv
  3. Version published to 10.1101/2025.02.27.25323033v3 on medRxiv
  4. Version published to 10.1101/2025.02.27.25323033v2 on medRxiv
  5. Version published to 10.1101/2025.02.27.25323033v1 on medRxiv