Temporal Variation In Heat-Mortality Associations: A Multi-Country Study

Abstract

Introduction: Recent investigations have reported a decline in the heat-related mortality risk during the last decades. However, these studies are frequently based on modelling approaches that do not fully characterize the complex temperature-mortality relationship, and are limited to single cities or countries. In this contribution, we investigate the issue using a multi-country data set and flexible modelling techniques. Methods: We collected daily time series data of temperature and all-cause mortality for 272 locations in 7 countries, with a total 20,203,690 deaths occurring in summer months between 1985 and 2012. The analysis is based on two-stage time series models. The heat-mortality relationship was estimated in each location with time-varying distributed lag non-linear models, based on a bivariate spline to model the exposure-lag-response over lag 0-10. The temporal variation was expressed through an interaction between the bivariate spline variables and calendar time. The overall cumulative exposure-response curves predicted for the years 1993 and 2006 were pooled by country through multivariate meta-analysis. Results: Several countries show a significant attenuation in mortality risk due to heat. Time-varying relative risks for the 99th percentile vs the temperature of minimum mortality decrease from 1993 to 2006 in Japan (1.16 to 1.06), Spain (1.56 to 1.37) and the USA (1.12 to 1.02). Estimates from Australia and South Korea are unclear due to lack of statistical power, while no variation seems to occur in the UK (1.16 to 1.17). In the USA, there is evidence of a stronger attenuation for moderate compared to more extreme heat. Conclusions: The mortality risk associated to high ambient temperature has decreased substantially in most of the countries during the last decades. While in some populations the risk is almost completely abated for moderate heat, an excess persists for more extreme temperatures in all the countries at the end of the study period.