Surmortality Related to the August 2003 Heatwave: A Contextual Risk Factors Analysis in Paris, France

Abstract

MAA1-O-06 Introduction: During the August 2003 heat wave in France, almost 15,000 excess deaths were recorded. Paris was severely affected, with 1053 excess deaths corresponding to an excess death rate of 190%. An ecological study was undertaken to describe spatial distribution of deaths within Paris and the impact of neighborhood socioeconomic characteristics on the excess death rates according to the place of residence. Methods: The study population included all death between August 1 and 20, 2003 (n = 1562) and during the same period in reference years (2000, 2001, 2002; n = 509). Spatial heterogeneity in mortality distribution was analyzed by means of standardized mortality ratio (SMR) within Paris at the quartier level (n = 80). Average household incomes and the proportion citizens 60 years of age and older people living alone were used as indicators of socioeconomic and sociodemographic status. The analysis is based on a hierarchical Bayesian model at the quartier level. Results: Some differences were observed in SMR computed by quartier[r] of residence. Maps showed a gradient of excess deaths increasing from the north–west to the south–east in 2003 and a classical zone of excess deaths in the north–east during reference years. We therefore observed a shift in excess mortality toward the south in August 2003. Contextual socioeconomic conditions had a significant impact on mortality both during heat wave and referral years. The main contextual factor in August 2003 was the average household income: between the richest and the poorest districts, the relative risk of mortality was 0.60 (95% confidence interval = 0.49–0.74). However, during the heat wave, the socioeconomic factors impact was weaker than during reference years. Analysis of residual relative risks of mortality underlined the finding that the excess deaths observed in the south part of the town were not explained by temperature and air pollution. Conclusion: The spatial distribution of the excess deaths observed in August 2003 generated a new map of the capital. The impact of socioeconomic context independently of the individual sociodemographic characteristics was proved. These results suggest that the heat wave was not simply the consequence of high temperature. They confirm a double level of risk during heat wave, individual and contextual, both of which must be taken into account to understand the effects of a heat wave on mortality and its prevention.