SHORT-TERM EFFECTS OF AIR-POLLUTION ON HEALTH - A EUROPEAN APPROACH USING EPIDEMIOLOGIC TIME-SERIES DATA - THE APHEA PROJECT - BACKGROUND, OBJECTIVES, DESIGN

Recent studies investigating the adverse health effects of air pollution indicate that effects exist around and below the current national and international air quality guidelines and standards, However, the difficult methodological issues involved, and the diversity of analytical techniques so far applied, hinder direct between-study comparability and the drawing of clear conclusions. The APHEA (Air Pollution on Health: European Approach) project is an attempt to provide quantitative estimates of the short-term health effects of air pollution, using an extensive data base from 10 different European countries, which represent various social, environmental and air pollution situations. Within the framework of the project, the methodology of analysing epidemiological time series data, as well as that of performing meta-analysis, are further developed and standardized. Data have been collected from 15 European cities with a total population exceeding 25 million. The exposure data consist of daily measurements of black smoke, sulphur dioxide, suspended particles, nitrogen dioxide and ozone (each available in several, though not all, cities) from already existing monitoring networks. There is substantial variability in air pollution mixtures and air pollutant levels in participating cities. The mean (24 h) levels of SO2 range 27-327 mu g . m(3) in the winter season, and those of black smoke range 15-292 mu g . m(3). The mean (1 h) levels of ozone in the summer season range 32-166 mu g . m(3). The outcome data are daily counts of total and cause-specific deaths and hospital emergency admissions. Data on potential confounders (mainly meteorological and chronological variables) are also used. There is large diversity in the climatic conditions in the different cities. Thus, the mean winter temperature ranges -4 to 10 degrees C (minimum -1 to -37 degrees C) and the mean summer temperature 16 degrees to 26 degrees C (maximum 26 degrees to 35 degrees C). Poisson regression allowing for autocorrelation and overdispersion is used in the analysis, after careful control of seasonality and other periodic patterns, and other confounding effects. The protocol and procedures followed are described, and the advantages and expectations from such an extensive European collaborative effort are presented and discussed.