Hydrometeorological Disasters and Climate-Sensitive Disease Burden in Pacific Island Countries: Evidence on Disease-Specific and Temporal Heterogeneity from a Multi-Database Panel Study

Abstract

Background: Pacific Island countries are highly vulnerable to hydrometeorological disasters, but comparable quantitative evidence on how such disasters affect climate-sensitive disease burden across countries remains limited. This study examined whether disaster-related health associations varied across diseases, outcome dimensions, and lag structures.

Methods: We conducted an ecological country-year panel study of 11 Pacific Island countries from 2000 to 2023 using data from the Global Burden of Disease Study, EM-DAT, the World Development Indicators, the Global Health Expenditure Database, and the ND-GAIN Country Index. Two-way fixed-effects models with country-clustered standard errors were used to estimate associations of disaster occurrence and disaster count with aggregate and disease-specific disease burden outcomes, with alternative incidence and death outcomes, lag terms, supplementary sensitivity analyses, wild cluster bootstrap tests for selected coefficients, and interaction models involving health expenditure and ND-GAIN readiness.

Results: No stable short-term average association was observed between hydrometeorological disasters and the aggregate climate-sensitive age-standardized DALY rate. In fully adjusted models, the coefficients for disaster occurrence and disaster count were 17.56 (95% CI: −12.43 to 47.54) and 9.39 (95% CI: −9.46 to 28.25), respectively. Disease-specific analyses revealed marked heterogeneity. The most robust finding was a delayed positive association between disaster occurrence and protein-energy malnutrition burden, with coefficients of 10.68 (95% CI: −0.13 to 21.49) at a 1-year lag and 13.65 (95% CI: 0.73 to 26.56) at a 2-year lag; both remained significant in wild cluster bootstrap tests (p = 0.026 and p = 0.032). By contrast, diarrheal disease signals were weaker and not robust under stricter inference. Disaster-related signals were more evident for the aggregate climate-sensitive incidence rate than for the aggregate death rate, although the incidence association remained borderline under bootstrap testing.

Conclusions: Hydrometeorological disasters were not associated with a stable short-term increase in the aggregate climate-sensitive DALY burden in Pacific Island countries, but their health effects showed clear disease-specific and time-dependent heterogeneity. The most robust evidence was concentrated in delayed nutrition-related burden, suggesting that climate-resilient health strategies in the Pacific should extend beyond emergency response to include stronger nutritional resilience, surveillance capacity, and cross-sector coordination.