The lethal grip of heat: mapping the heat wave-mortality nexus in Spain (1975-2019)

Abstract

Nowadays, the rise in the global temperatures are a source of concern, particularly in the Mediterranean region, where Spain is already witnessing notable consequences for its aging population. Predictions for the end of the XXI century reveal a persistent increase in air temperatures along with an increment of extreme episodes. Abnormal heat, once considered an 'environmental accident', is now a serious public threat. This contribution endeavors to quantify the added effects of heat wave exposure on mortality by demographic and socioeconomic strata during the period of 45 years in Spain at the provincial level. Moreover, we aim to explore the temporal evolution in these effects and variations in its spatial patterns, especially focusing on the inequality aspects that shape the health outcomes in an increasingly aging population. Here we leverage daily individual mortality data and other contextual data on population from the National Institute of Statistics of Spain and air temperature estimates from the ERA5 global reanalysis. We also use the historical settlement data as a proxy for population distribution from 1975 onward. To estimate the main and added effects of heat waves we fit a quasi-Poisson time-series regression model using a distributed lag non-linear model with 10 days of lag, controlling for trends and day of the week. We analyze approximately 15.8 million of deaths registered in Spain between 1975 and 2019. During the selected time window, we expect to see a shift in the temperature-mortality association from a V-shape in the first decades of the observation to a U-shape by the end of the period all across the provinces, thus revealing a progressive flattening of the exposure-response curve. We also expect to observe an overall reduction in the mortality burden associated with the temperatures. In particular, we anticipate more significant and rapid decline in the cold-related risks and attributable fractions in comparison with the heat-related ones, with some latitudinal variations across the country. On the other hand, we witness a steady increase in the incidence of the heat wave episodes with time all over the country. We expect to see a positive added effect of heat wave on mortality, however this effect is assumed to be smaller than the primary effect. In addition, we anticipate observing variations in the effect depending on the heat wave order, duration, intensity, geographic location and demographic strata. The largest added effects are expected for the longest and strongest heat waves in the oldest-old population in the less accustomed to extreme heat areas.