Universal thermal climate index (UTCI) applied to determine thresholds for temperature-related mortality

Our research goal was to examine a response in mortality among population in Arkhangelsk caused by exposure to high and low temperatures. We determined the best available mortality predictor out of air temperature and Universal Thermal Climate Index that characterizes how people feel temperature and detected threshold temperatures depending on sex, age, and cause of death; under exposure to such temperatures there was a statistically authentic increase in mortality. We analyzed data on daily mortality among population and meteorological data collected in 1999-2016. Relative preciseness in calculating attributive fractions of additional mortality during all hot and cold days was taken as a numeric criterion for selecting the best predictor. All the calculations were accomplished basing on Poisson's regression model taking into account a non-linear dependence between mortality and weather with a distributed lag up to 21 days long. Although people in Arkhangelsk live in a climate with cold summer and moderately cold winter, we determined attributive fractions of mortality both for cold and heat...