The tempo-spatial heterogeneity in mortality burden attributable to non-optimal temperatures in China and its provinces from 1990 to 2021

Existing studies have not characterized the temporal patterns and underlying drivers of temperature-related mortality across Chinese provinces. We utilized data from the Global Burden of Diseases Study (GBD) 2021 to analyze the tempo-spatial changes in mortality burden attributed to low and high temperatures from 1990 to 2021 and employed a random forest regression model to investigate the impact of socio-economic and demographic factors on these changes. We found that total deaths attributable to non-optimal temperatures rose from 0.376 million in 1990 to 0.589 million in 2021, of which 0.542 million were cold-related and 0.049 million heat-related, while males and those aged 70+ years bore disproportionately higher burdens. Nevertheless, age-standardized mortality due to non-optimal temperatures declined by 50.8% (95% uncertainty interval (UI): 40.8, 58.5), with cold- and heat-related mortality decreasing by 51.8% (95% UI: 42.5, 59.8) and 35.5% (95% UI: 0.7, 56.0), respectively. Eastern China, notably Shanghai, exhibited the most pronounced decline (−61.0%, 95% UI: −70.4, −49.5). In addition, we identified GDP, air-conditioner ownership per 100 households, sex ratio, and health facilities as primary drivers of the cold-related mortality, whereas air-conditioner ownership per 100 households and medical-bed availability per 10,000 persons chiefly influenced heat-related trends. Our findings underscore the importance to promote health equity and reduce regional disparities in temperature-related mortality burdens. Future strategies should focus on improving heat-health surveillance, expanding early warning and response systems, and implementing targeted adaptation measures for heat-vulnerable communities.