Global, regional, and national burden of heatwave-related mortality from 1990 to 2019: A three-stage modelling study.

Background: The regional disparity of heatwave-related mortality over a long period has not been sufficiently assessed across the globe, impeding the localisation of adaptation planning and risk management towards climate change. We quantified the global mortality burden associated with heatwaves at a spatial resolution of 0.5°×0.5° and the temporal change from 1990 to 2019. Methods and findings: We collected data on daily deaths and temperature from 750 locations of 43 countries or regions, and 5 meta-predictors in 0.5°×0.5° resolution across the world. Heatwaves were defined as location-specific daily mean temperature ≥95th percentiles of year-round temperature range with duration ≥2 days. We first estimated the location-specific heatwave-mortality association. Secondly, a multivariate meta-regression was fitted between location-specific associations and 5 meta-predictors, which was in the third stage used with grid cell-specific meta-predictors to predict grid cell-specific association. Heatwave-related excess deaths were calculated for each grid and aggregated. During 1990 to 2019, 0.94% (95% CI: 0.68–1.19) of deaths [i.e., 153,078 cases (95% eCI: 109,950–194,227)] per warm season were estimated to be from heatwaves, accounting for 236 (95% eCI: 170–300) deaths per 10 million residents. The ratio between heatwave-related excess deaths and all premature deaths per warm season remained relatively unchanged over the 30 years, while the number of heatwave-related excess deaths per 10 million residents per warm season declined by 7.2% per decade in comparison to the 30-year average. Locations with the highest heatwave-related death ratio and rate were in Southern and Eastern Europe or areas had polar and alpine climates, and/or their residents had high incomes. The temporal change of heatwave-related mortality burden showed geographic disparities, such that locations with tropical climate or low incomes were observed with the greatest decline. The main limitation of this study was the lack of data from certain regions, e.g., Arabian Peninsula and South Asia. Conclusions: Heatwaves were associated with substantial mortality burden that varied spatiotemporally over the globe in the past 30 years. The findings indicate the potential benefit of governmental actions to enhance health sector adaptation and resilience, accounting for inequalities across communities. Qi Zhao and colleagues modelled the global, regional and national burden of heatwave-related mortality from 1990 to 2019. Author summary: Why was this study done?: Although exposure to heatwave has been associated with increased risk of excess deaths and certain temporal attenuation observed, the evidence mainly comes from limited locations. The regional disparity in heatwave-related mortality burden over a long period has not been adequately quantified across the globe. What did the researchers do and find?: Heatwave-related excess deaths were predicted to each grid cell (0.5°×0.5°) across the globe between 1990 and 2019 using a three-stage modelling framework. During 1990 to 2019 warm seasons, 153,078 deaths were associated with heatwaves (nearly half in Asia), which accounted for 0.94% of all deaths and equated 236 deaths per 10 million residents. The global heatwave-related excess death rate declined by 7.2% per decade in comparison to the 30-year average. Heatwave-related mortality showed complex regional disparities, such that Southern and Eastern Europe or areas had polar and alpine climates, and/or their residents had high incomes had the greatest cumulative burden while certain locations in Oceania or those with tropical climate or low incomes were observed with the greatest decline over decades. What do these findings mean?: These regional disparities suggest the necessity of considering the cumulation and temporal change of mortality burden for evaluating the historical health impact of climate change and developing time-efficient adaptation strategies for a certain area. Localised adaptation planning and risk management should be encouraged across all levels of government, with considering subnational capacities and inequalities. Lack of time-series death data from certain regions, e.g., Arabian Peninsula and South Asia, may reduce the prediction accuracy in those regions. [ABSTRACT FROM AUTHOR]