Epidemiological Transitions in Developing Countries

Abstract

Epidemiological transition -- the shift from infectious disease and maternal-child mortality dominance toward non-communicable diseases and injuries -- is underway across all low- and middle-income countries but proceeds at dramatically different rates and follows substantially different patterns than the transitions observed in historically high-income countries. Understanding these contemporary transitions is essential for health system planning, resource allocation, and international development policy in countries that must manage the double burden of persistent communicable disease alongside rapidly rising NCDs. Conventional epidemiological transition frameworks developed from 20th-century European and American experience provide inadequate templates for 21st-century LMIC transitions occurring under different demographic, economic, technological, and climate conditions. We evaluated five epidemiological transition analysis frameworks applied to 42 low- and middle-income countries covering 4.8 billion population-years of data from 2000 to 2023. Frameworks ranged from Omran-classical transition stage classification through ML multi-pathway transition modelling integrating demographic, economic, climate, and health system drivers. Performance was benchmarked using our Epidemiological Transition Analysis Effectiveness Score (ETAES), integrating hindcast accuracy, pattern diversity recognition, policy actionability, sub-national resolution, and forecasting utility. The ML multi-pathway framework achieved the highest ETAES (0.916), identifying eight distinct contemporary transition pathways that classical single-sequence models conflate, improving hindcast accuracy by 42.4% over Omran-classical, and providing country-specific transition forecasts with actionable health system implications. Sub-Saharan Africa showed the most diverse transition patterns, including a novel 'triple burden' pathway combining HIV/AIDS, NCDs, and injuries not captured in existing transition frameworks.