The structural geography of epidemic financing since 2000: harmonised country-level dataset and analysis code

This deposit accompanies the manuscript Arnal & Roche, "The structural geography of epidemic financing since 2000: divergent regional regimes and implications for pandemic preparedness", submitted to The Lancet Global Health (2026).

It contains: (i) the harmonised country-level dataset for 76 low- and middle-income countries, integrating the IHME Development Assistance for Health database, UNAIDS, WHO Global Tuberculosis and World Malaria reports, the Global Fund disbursement data, the WHO Global Health Expenditure Database, Policy Cures Research G-FINDER, and the Global Burden of Disease 2021 study; (ii) the full reproducible analysis pipeline (Python 3, pandas / numpy / matplotlib) for the regional summary, Spearman correlations, four-quadrant taxonomy, pairwise Mann-Whitney tests, sensitivity analyses, and Figures 1 and 2 of the manuscript plus the appendix supplementary figures; (iii) publication-ready PNG and PDF outputs at 300 dpi.

The headline empirical finding reproduced by this code is the strong inverse correlation between domestic financing share and infectious-disease vulnerability (Spearman ρ = −0·95, p < 0·001), which holds within every region individually and which the manuscript uses to map the structural exposure of LMIC health systems to the post-2021 contraction of development assistance for health.