The year 2019 saw exceptionally high temperatures, which led to the issuance of "red heat wave" alerts for the first time in several departments. The decision to place a department under a red alert is based on an in-depth contextual analysis drawing on the combined expertise of Météo-France and Santé publique France. Criteria were identified to determine which days and departments should be subject to such an analysis. This research relied, on the one hand, on modeling the excess risk associated with high mortality (the 98th percentile of mortality, i.e., a level approximately 50% higher than normal) potentially already linked to a heat effect, in order to identify temperature thresholds and possible threshold values. Additionally, a simulation of the number of situations meeting different threshold values was conducted. The modeling covered all metropolitan departments for the period 2001–2019. The 98th percentile of mortality was modeled for different temperature terms, with relevant variables selected using the gradient boosting method. The model notably identifies a significant role for the average daily temperature (expressed as the deviation from the mean to normalize the results). A gradual increase in excess mortality risk was observed for deviations from the mean exceeding +5°C, without any marked break in the trend. Based on the simulations conducted, with a threshold of 7°C or 8°C, a red alert would have had to be issued for a few situations each year since 2015.