Machine Learning Pipeline Exposes 60% Global Disease Burden Receiving 2.5% of Targeted Aid

The study deployed a multistage machine-learning pipeline to map aid against disease burden at country and disease levels using granular OECD data on 320,000 projects. It identified significant overall correlation between funding and burden yet revealed stark regional imbalances in Central Africa, South Asia, and West Africa. Noncommunicable diseases such as cardiovascular conditions, diabetes, and kidney disease each drew just 0.2% of total aid despite rising incidence in low- and middle-income countries.

HIV and sexually transmitted infections captured 34% of aid while representing 6% of burden; similar over-allocation occurred for malaria and neglected tropical diseases. This pattern persists even as recent U.S. aid reductions widen resource gaps against UN Sustainable Development Goals. The analysis extends prior work by combining global scope with high-resolution disease and geography tracking.

Disparities heighten epidemic risks because underfunded health systems in high-burden regions struggle with both infectious and rising chronic conditions, amplifying inequality and cross-border transmission. Observational design limits causal claims, yet the mechanism linking aid misalignment to system fragility is plausible.

Next steps require prospective trials of reallocation frameworks and integration of IHME burden estimates to test whether targeted shifts reduce mortality differentials within five years.