Changes in the associations between heatwaves and human mortality during two extreme hot summers in Shanghai, China

Heatwaves are projected to increase in frequency, intensity, and duration as the climate warms. However, it is unclear whether human mortality from heatwaves is changing in frequency with time. We used Quasi-Poisson regression with a distributed non-linear model (DLNM) to examined associations of heatwaves and their characteristics (intensity, duration, and timing) with human mortality due to different diseases and total non-accidental diseases (TND) for different sociodemographic subgroups between 2002 and 2004 and 2012–2014 in Shanghai, China. We found that heatwaves showed a significant association with cause-specific mortality and TND for socio-demographic subgroups during the two study periods. Relative risks (RR) of mortality decreased for most demographic subgroups from 2002 to 2004 to 2012–2014, while RR of respiratory diseases (RD) increased over time. The association between heatwave characteristics and human mortality changed over time. RRs of heatwaves on mortality were higher for females, the elderly, and low- and middle-educational level populations than for males, younger and highly educated counterparts, respectively. Overall, heatwaves had a stronger association with the mortality of RD from 2002 to 2004 to 2012–2014. Heatwave duration also had an enhanced association with all subgroups over time. Our research findings could provide insights into the design of sustainable cities and society.