Vérène Wagner

Has the impact of heat waves on mortality changed in France since the European heat wave of summer 2003? A study of the 2006 heat wave

BACKGROUND In July 2006, a lasting and severe heat wave occurred in Western Europe. Since the 2003 heat wave, several preventive measures and an alert system aiming at reducing the risks related to high temperatures have been set up in France by the health authorities and institutions. In order to evaluate the effectiveness of those measures, the observed excess mortality during the 2006 heat wave was compared to the expected excess mortality. METHODS A Poisson regression model relating the daily fluctuations in summer temperature and mortality in France from 1975 to 2003 was used to estimate the daily expected number of deaths over the period 2004-2006 as a function of the observed temperatures. RESULTS During the 2006 heat wave (from 11 to 28 July), about 2065 excess deaths occurred in France. Considering the observed temperatures and with the hypothesis that heat-related mortality had not changed since 2003, 6452 excess deaths were predicted for the period. The observed mortality during the 2006 heat wave was thus markedly less than the expected mortality (approximately 4400 less deaths). CONCLUSIONS The excess mortality during the 2006 heat wave, which was markedly lower than that predicted by the model, may be interpreted as a decrease in the populations vulnerability to heat, together with, since 2003, increased awareness of the risk related to extreme temperatures, preventive measures and the set-up of the warning system.

Short-term associations between fine and coarse particles and hospital admissions for cardiorespiratory diseases in six French cities

Objectives: Little is known about the potential health effects of the coarse fraction of ambient particles. The aim of this study is to estimate the links between fine (PM2.5) and coarse particle (PM2.5−10) levels and cardiorespiratory hospitalisations in six French cities during 2000–2003. Methods: Data on the daily numbers of hospitalisations for respiratory, cardiovascular, cardiac and ischaemic heart diseases were collected. Associations between exposure indicators and hospitalisations were estimated in each city using a Poisson regression model, controlling for confounding factors (seasons, days of the week, holidays, influenza epidemics, pollen counts, temperature) and temporal trends. City-specific findings were combined to obtain excess relative risks (ERRs) associated with a 10 μg/m3 increase in PM2.5 and PM2.5−10 levels. Results: We found positive associations between indicators of particulate pollution and hospitalisations for respiratory infection, with an ERR of 4.4% (95% CI 0.9 to 8.0) for PM2.5−10 and 2.5% (95% CI 0.1 to 4.8) for PM2.5. Concerning respiratory diseases, no association was observed with PM2.5, whereas positive trends were found with PM2.5−10, with a significant association for the 0–14-year-old age group (ERR 6.2%, 95% CI 0.4 to 12.3). Concerning cardiovascular diseases, positive associations were observed between PM2.5 levels and each indicator, although some did not reach significance; trends with PM2.5−10 were weaker and non-significant except for ischaemic heart disease in the elderly (ERR 6.4%, 95% CI 1.6 to 11.4). Conclusions: In accordance with other studies, our results indicate that the coarse fraction may have a stronger effect than the fine fraction on some morbidity endpoints, especially respiratory diseases.

Bayesian nonlinear model selection and neural networks: a conjugate prior approach

In order to select the best predictive neural-network architecture in a set of several candidate networks, we propose a general Bayesian nonlinear regression model comparison procedure, based on the maximization of an expected utility criterion. This criterion selects the model under which the training set achieves the highest level of internal consistency, through the predictive probability distribution of each model. The density of this distribution is computed as the model posterior predictive density and is asymptotically approximated from the assumed Gaussian likelihood of the data set and the related conjugate prior density of the parameters. The use of such a conjugate prior allows the analytic calculation of the parameter posterior and predictive posterior densities, in an empirical-Bayes-like approach. This Bayesian selection procedure allows us to compare general nonlinear regression models and in particular feedforward neural networks, in addition to embedded models as usual with asymptotic comparison tests.

Association between long-term exposure to air pollution and mortality in France: A 25-year follow-up study

INTRODUCTION Long-term exposure to air pollution (AP) has been shown to have an impact on mortality in numerous countries, but since 2005 no data exists for France. OBJECTIVES We analyzed the association between long-term exposure to air pollution and mortality at the individual level in a large French cohort followed from 1989 to 2013. METHODS The study sample consisted of 20,327 adults working at the French national electricity and gas company EDF-GDF. Annual exposure to PM10, PM10–2.5, PM2.5, NO2, O3, SO2, and benzene was assessed for the place of residence of participants using a chemistry-transport model and taking residential history into account. Hazard ratios were estimated using a Cox proportional-hazards regression model, adjusted for selected individual and contextual risk factors. Hazard ratios were computed for an interquartile range (IQR) increase in air pollutant concentrations. RESULTS The cohort recorded 1967 non-accidental deaths. Long-term exposures to b aseline PM2.5, PM10-25, NO2 and benzene were associated with an increase in non-accidental mortality (Hazard Ratio, HR = 1.09; 95% CI: 0.99, 1.20 per 5.9 μg/m3, PM10-25; HR=1.09; 95% CI: 1.04, 1.15 per 2.2 μg/m3, NO2: HR=1.14; 95% CI: 0.99, 1.31 per 19.3 μg/m3 and benzene: HR=1.10; 95% CI: 1.00, 1.22 per 1.7 μg/m3).The strongest association was found for PM10: HR = 1.14; 95% CI: 1.05, 1.25 per 7.8 μg/m3. PM10, PM10-25 and SO2 were associated with non-accidental mortality when using time varying exposure. No significant associations were observed between air pollution and cardiovascular and respiratory mortality. CONCLUSION Long-term exposure to fine particles, nitrogen dioxide, sulfur dioxide and benzene is associated with an increased risk of non-accidental mortality in France. Our results strengthen existing evidence that outdoor air pollution is a significant environmental risk factor for mortality. Due to the limited sample size and the nature of our study (occupational), further investigations are needed in France with a larger representative population sample.

Definition of temperature thresholds: the example of the French heat wave warning system

Heat-related deaths should be somewhat preventable. In France, some prevention measures are activated when minimum and maximum temperatures averaged over three days reach city-specific thresholds. The current thresholds were computed based on a descriptive analysis of past heat waves and on local expert judgement. We tested whether a different method would confirm these thresholds. The study was set in the six cities of Paris, Lyon, Marseille, Nantes, Strasbourg and Limoges between 1973 and 2003. For each city, we estimated the excess in mortality associated with different temperature thresholds, using a generalised additive model, controlling for long-time trends, seasons and days of the week. These models were used to compute the mortality predicted by different percentiles of temperatures. The thresholds were chosen as the percentiles associated with a significant excess mortality. In all cities, there was a good correlation between current thresholds and the thresholds derived from the models, with 0°C to 3°C differences for averaged maximum temperatures. Both set of thresholds were able to anticipate the main periods of excess mortality during the summers of 1973 to 2003. A simple method relying on descriptive analysis and expert judgement is sufficient to define protective temperature thresholds and to prevent heat wave mortality. As temperatures are increasing along with the climate change and adaptation is ongoing, more research is required to understand if and when thresholds should be modified.

A review of the epidemiological methods used to investigate the health impacts of air pollution around major industrial areas.

We performed a literature review to investigate how epidemiological studies have been used to assess the health consequences of living in the vicinity of industries. 77 papers on the chronic effects of air pollution around major industrial areas were reviewed. Major health themes were cancers (27 studies), morbidity (25 studies), mortality (7 studies), and birth outcome (7 studies). Only 3 studies investigated mental health. While studies were available from many different countries, a majority of papers came from the United Kingdom, Italy, and Spain. Several studies were motivated by concerns from the population or by previous observations of an overincidence of cases. Geographical ecological designs were largely used for studying cancer and mortality, including statistical designs to quantify a relationship between health indicators and exposure. Morbidity was frequently investigated through cross-sectional surveys on the respiratory health of children. Few multicenter studies were performed. In a majority of papers, exposed areas were defined based on the distance to the industry and were located from <2 km to >20 km from the plants. Improving the exposure assessment would be an asset to future studies. Criteria to include industries in multicenter studies should be defined.

A simple indicator to rapidly assess the short-term impact of heat waves on mortality within the French heat warning system

We propose a simple method to provide a rapid and robust estimate of the short-term impacts of heat waves on mortality, to be used for communication within a heat warning system. The excess mortality during a heat wave is defined as the difference between the observed mortality over the period and the observed mortality over the same period during the N preceding years. This method was tested on 19 French cities between 1973 and 2007. In six cities, we compared the excess mortality to that obtained using a modelling of the temperature-mortality relationship. There was a good agreement between the excess mortalities estimated by the simple indicator and by the models. Major differences were observed during the most extreme heat waves, in 1983 and 2003, and after the implementation of the heat prevention plan in 2006. Excluding these events, the mean difference between the estimates obtained by the two methods was of 13 deaths [1:45]. A comparison of mortality with the previous years provides a simple estimate of the mortality impact of heat waves. It can be used to provide early and reliable information to stakeholders of the heat prevention plan, and to select heat waves that should be further investigated.

The mortality impacts of fine particles in France

INTRODUCTION Worldwide, air pollution has become a main environmental cause of premature mortality. This burden is largely due to fine particles. Recent cohort studies have confirmed the health risks associated with chronic exposure to PM2.5 for European and French populations. We assessed the mortality impact of PM2.5 in continental France using these new results. METHODS Based on a meta-analysis of French and European cohorts, we computed a shrunken estimate of PM2.5-mortality relationship for the French population (RR 1.15 [1.05:1.25] for a 10μg/m(3) increase in PM2.5). This RR was applied to PM2.5 annual concentrations estimated at a fine spatial scale, using a classical health impacts assessment method. The health benefits associated with alternative scenarios of improving air quality were computed for 36,219 French municipalities for 2007-2008. RESULTS 9% of the total mortality in continental France is attributable to anthropogenic PM2.5. This represents >48,000 deaths per year, and 950,000years of life lost per year, more than half occurring in urban areas larger than 100,000 inhabitants. If none of the municipalities exceeded the World Health Organization guideline value for PM2.5 (10μg/m(3)), the total mortality could be decreased by 3%, corresponding to 400,000years of life saved per year. CONCLUSION Results were consistent with previous estimates of the long-term mortality impacts of fine particles in France. These findings show that further actions to improve air quality in France would substantially improve health.

How to use near real-time health indicators to support decision-making during a heat wave: the example of the French heat wave warning system

Introduction The French warning system for heat waves is based on meteorological forecasts. Near real-time health indicators are used to support decision-making, e.g. to extend the warning period, or to choose the most appropriate preventive measures. They must be analysed rapidly to provide decision-makers useful and in-time information. The objective of the study was to evaluate such health indicators. Methods A literature review identified a range of possible mortality and morbidity indicators. A reduced number were selected, based on several criteria including sensitivity to heat, reactivity, representativity and data quality. Two methods were proposed to identify indicator-based statistical alarms: historical limits or control charts, depending on data availability. The use of the indicators was examined using the 2006 and 2009 heat waves. Results Out of 25 possible indicators, 5 were selected: total mortality, total emergency calls, total emergency visits, emergency visits for people aged 75 and over and emergency visits for causes linked to heat. In 2006 and 2009, no clear increases were observed during the heat waves. The analyses of real-time health indicators showed there was no need to modify warning proposals based on meteorological parameters. Discussion These findings suggest that forecasted temperatures can be used to anticipate heat waves and promote preventive actions. Health indicators may not be needed to issue a heat wave alert, but daily surveillance of health indicators may be useful for decision-makers to adapt prevention measures.

French Heat Health Watch Warning System: Validation of Temperature Thresholds

Contrary to the empirical method used in the French HHWWS, our method takes into account confounding factors and temporal events. Thus, it is not surprising that the thresholds resulting from a time series analysis are somewhat different to those from the HHWWS. Moreover, these predictive models made it possible to estimate daily excess mortality that could be used as an additional criterion in the proposal of an alert in the framework of the French warning system for heat waves. French Heat Health Watch Warning System: Validation of Temperature Thresholds