Sharing common measures of the environment across continents: challenges and opportunities for global studies of aging

ObjectivesEnvironmental exposures are modifiable risk factors for accelerated aging, but research is frequently limited to individual countries due to inconsistent exposure assessment. Cross-national data provide broader perspectives but add methodological complexities. This study evaluated spatial and temporal patterns of 5 environmental measures assigned to older adult pseudopopulations in 8 countries to highlight opportunities and challenges for aging epidemiologic studies.MethodsThrough the Gateway to Global Aging Data project, we harmonized measures of air pollution (PM2.5, NO2, O3) and natural spaces (greenspace, blue space) for longitudinal aging cohorts in Brazil, Chile, England, India, Ireland, Mexico, Northern Ireland, and the United States. Global exposure data (1990–2021) derived from satellite observations, spatiotemporal models, and deterministic simulations were linked to 10,000 population-weighted points representing adults >50 years per country. We characterized urbanicity and area-level deprivation and examined spatial/temporal patterns to inform environmental aging research.ResultsExposure levels and variability differed within and between countries. Greenspace and NO2 exhibited high within-country variation (200% higher within-country vs between-country standard deviations), whereas PM2.5 and O3 had larger across-country differences (300% higher between-country than within-country standard deviation). Temporal trends were generally consistent across countries, though unique patterns emerged (e.g., increasing PM2.5 and O3 levels in India and a radical drop in greenspace in Chile). Correlations were most consistent between NO2, greenspace, and area-level deprivation, though they varied (0.1 to 0.7).DiscussionHarmonized measures facilitate cross-country comparisons of environmental exposures but require careful consideration of confounding by place and time and differential measurement to ensure robust inferences.