Modification of the association between high ambient temperature and health by urban microclimate indicators: A systematic review and meta-analysis

Abstract

Background Landscape characteristics, including vegetation and impervious surfaces, influence urban microclimates and may lead to within-city differences in the adverse health effects of high ambient temperatures. Objective Our objective was to quantitatively summarize the epidemiologic literature that assessed microclimate indicators as effect measure modifiers (EMM) of the association between ambient temperature and mortality or morbidity. Methods We systematically identified papers and abstracted relative risk estimates for hot and cool microclimate indicator strata. We calculated the ratio of the relative risks (RRR) and 95% confidence intervals (95% CI) to assess differences in health effects across strata, and pooled the RRR estimates using random effects meta-analyses. Results Eleven papers were retained. In the pooled analyses, people living in hotter areas within cities (based on land surface temperature or modeled estimates of air temperature) had 6% higher risk of mortality/morbidity compared to those in cooler areas (95% CI: 1.03–1.09). Those living in less vegetated areas had 5% higher risk compared to those living in more vegetated areas (95% CI: 1.00–1.11). Discussion There is epidemiologic evidence that those living in hotter, and less vegetated areas of cities have higher risk of morbidity or mortality from higher ambient temperature. Further research with improved assessment of landscape characteristics and investigation of the joint effects of physiologic adaptation and landscape will advance the current understanding. Conclusion This review provides quantitative evidence that intra-urban differences in landscape characteristics and micro-urban heat islands contribute to within-city variability in the health effects of high ambient temperatures.