First steps for new extreme temperatures Health Early Warning System in Portugal

Abstract

Both extremely hot and cold temperatures are considered to have significant effects on population’s health. More likely heat waves and cold spells due to climate change could result in excess deaths. Public health institutions play a crucial role in assessing the impacts of such events and, subsequently, in providing adequate early warnings and suitable mitigation recommendations. In this work we propose an update of the heat and cold health early warning systems currently in use in Portugal. The aim was to develop a risk indicator, active throughout the whole year, and easily understood by the entire population, with the highest possible spatial resolution. Daily data of all-cause mortality and maximum, minimum and mean temperatures was gathered from public data sources for the 1995-2019 time period. District-specific temperature-mortality associations were estimated using quasi-Poisson with linear threshold distributed lag models for cold and hot semesters (minimum temperatures were considered in autumn/winter and maximum temperatures in spring/summer, to identify worst case exposure scenarios). Regressions included seasonality and long-term trends and year population estimates as an offset. Influenza incidence was also included in the model to improve predictive performance. Cold and hot thresholds were defined for each semester and for each district independently based on best data fit criteria. Results show good predictive ability of the district-specific models and allowed the identification of different temperature-mortality associations between regions. The overall cumulative RRs estimated for low (4ºC) and high (34ºC) temperatures were 1.62 (95% CI: 1.65 - 1.69) and 2.35 (95% CI: 2.19 - 2.51) for the most populated district and 1.07 (95% CI: 1.04 - 1.11) and 1.03 (95% CI: 1.02 - 1.03) for the least populated district. Variation of optimum cold and hot thresholds between districts provided interesting insights about varied regional population vulnerability. Keywords: Heat waves, Cold waves, Mortality, Prevention plans, DLM