Case Study 2 "Temperature-related human mortality in European regions": Product Development Report (D5.8)

In order to describe and model the relationship between ambient temperature and mortality, there are several different statistical and epidemiological methodologies. The objective of this deliverable is to describe the methodology that we are going to use to develop and implement a prototype of integrated heat health early warning system for Europe. 

The definition of the risk of death associated with temperature depends on the intensity of the exposure (temperature) for a range of days leading to the health outcome. This has been described as the exposure-lag-response association. In this report we briefly describe the Distributed Lag-Non Linear Model (DLNM) framework, which is here used for the analysis of the delayed short-terms association between temperatures and mortality.

In order to develop the statistical temperature-mortality association, we calibrated the DLNM model with two different datasets. On the one hand, we used our own database of mortality, which contains information on daily mean temperature and daily counts of all-cause mortality for the period 1998-2012 in 147 regions from 16 European countries, representing around 420 million people. This mortality dataset is of restricted use and is not being made open, neither to project partners nor to the general public. On the other hand, our partners from the City Council of Almada provided a dataset with information on daily mean temperature and daily counts of all-cause mortality in Almada for the period 2000-2015. By using these two sources of information, we calibrated the temperature-mortality model that is going to be used to predict the temperature-attributable mortality in Europe as a part of the heat health early warning system. This prototype of early warning system will use weather forecasts with lead times up to 15 days to predict the temperature-attributable mortality in Europe at the regional scale.

As a result of the research involved in the case study (CS2), we published several scientific articles which validate the methodology (DLNM) to be used in the temperature-mortality predictions (next step of CS2). The results we here report also highlights the importance of implementing public health preventive measures, such as heat health early warning systems, that have a positive impact on reducing heat-related mortality.