We collected time-series daily data, including mortality, weather variables, and air pollution measures, from 384 locations in 13 countries: Australia (three cities, 1988–2009), Brazil (18 cities, 1997–2011), Canada (21 cities, 1986–2009), China (15 cities, 1996–2008), Italy (11 cities, 1987–2010), Japan (47 prefectures, 1985–2012), South Korea (seven cities, 1992–2010), Spain (51 cities, 1990–2010), Sweden (one county, 1990–2002), Taiwan (three cities, 1994–2007), Thailand (62 provinces, 1999–2008), UK (ten regions, 1993–2006), and USA (135 cities, 1985–2009). Mortality was represented by daily counts of deaths for either all causes or, where not available, non-external causes only (International Classification of Diseases [ICD]-9 0-799, ICD-10 A00-R99). We chose mean daily temperature as the exposure index, calculated from central monitor stations, either as the average between maximum and minimum values or the 24 h average. We did a sensitivity analysis by modifying the modelling choices, replacing all-cause with non-external mortality, and controlling for air pollution and humidity in the subset of countries that provided such information. The appendix contains details of the exact study periods, further information on data collection, additional results, and results from the sensitivity analysis.
Free full text: Click here
