Summer Heat, Hospital Admissions and Mortality among the Elderly in Michigan and the United States.

Abstract

Background: With climate change and population aging, the health burden of extreme heat (EH) is a present and future concern. However, characteristics of vulnerability to EH and associations between EH and morbidity vs. mortality are not well understood, and previous studies have focused on small numbers of cities and/or used sociodemographic and land cover information at a course geographic resolution (e.g., city-level). Objectives: For ages 65 and over, we examined 1) vulnerability to EH-associated mortality in 8 Michigan cities by individual and neighborhood socio-demographic characteristics and green space, 2) associations between heat and HAs in over 100 U.S. cities and 3) vulnerability to EH-associated HAs by individual and neighborhood characteristics and citywide air-conditioning prevalence in over 100 U.S. cities. Methods: In time-stratified case-crossover designs, we regressed natural-cause mortality or HAs against moderate heat and EH, or an indicator for maximum or apparent temperature above a city-specific threshold, over multiple lag days. We examined vulnerability by personal and ZIP-code (neighborhood) characteristics using multiple interaction terms between each of the characteristics and EH terms and by citywide air-conditioning prevalence in meta-analyses. Results: In Michigan, we observed increased vulnerability to heat-associated mortality among men, blacks and the very old (78 years or older). Nationally, EH was associated with higher HAs for cardiovascular, renal and respiratory diseases, with rates increasing by 43% (95% CI: 37%, 50%) for renal diseases, though over longer lags, substantial displacement of the initial effect was observed for cardiovascular and respiratory HAs. We observed increased vulnerability to EH-associated HAs among men and the very old, in neighborhoods with more residents of nonwhite race and lower education and in cities with lower air-conditioning prevalence. Significant heterogeneity existed between cities for many of the vulnerability characteristics. Results were suggestive of increased vulnerability to heat-associated mortality and HAs in neighborhoods with less green space. Conclusions: Individuals with respiratory and renal conditions may benefit from heat adaptation interventions, and heat health warning systems and vulnerability maps may benefit from incorporating information on heat-associated HAs. Socio-demographic characteristics and possibly green space modify the association between EH and HAs and mortality, though effects vary by city.