Excess Winter Mortality in France: Influence of Temperature, Influenza Like Illness, and Residential Care Status

Abstract

Despite a surge of 15,000 deaths caused by the 2003 heat wave in France, the annual winter excess mortality rate remains a recurring phenomenon that is quantitatively greater than the isolated summer event. Coefficients of Seasonal Variations in Mortality (CSVMs) were calculated using monthly mortality data from 1998 to 2007 in France. CSVM was a percentage representing the excess death rate from December to March inclusively, against average, monthly mortality from the other 8 non-winter months. Sociodemographic parameters and environmental factors were also obtained. Univariate and multivariate analyses were performed to identify risk factors of increased winter mortality. The data revealed an annual winter excess death of 23,836 (± 7951) (mean ± 1 standard deviation) cases. On average, CSVM in France was+14.94% (13.54 [12.03; 19.70]) (mean, median, and interquartile intervals). Multivariate analysis results revealed that several factors contributed to the CSVM: sociodemographics, such as age (CSVM higher for the population older than 75) and death location (CSVM higher in nursing homes), environmental factors, such as the severity of the winter season (per monthly minimal temperature), and estimated number of influenza-like illnesses (ILI). Correlation between observed and predicted CSVMs was extremely consistent (R(2)= 0.91). There was a fundamental belief that residents in nursing homes were well protected from cold spells and their consequences. Our results revealed this to be a mere misperception.