Quantification of deaths attributed to air pollution in Sweden using estimated population exposure to nitrogen dioxide as indicator

In the previous phase of this project a model was provided for quantifying the general population exposure to air pollution. From that work interpolated yearly mean concentrations of nitrogen dioxide were provided for the Swedish population. To be applied in the health impact assessment we selected an ecological study from Auckland, New Zealand, which reported a 13 % increase in non-accidental mortality (all ages) for 10 µg/m3 increase in NO2. Based on official national data we assumed a baseline rate of 1010 deaths per 100 000 persons and year at the population weighted mean level of approximately 10 µg NO2/m3. We then calculated the death rate and the yearly number of deaths expected at the population weighted mean exposure in each of four exposure classes above 10 µg/m3. Using the modelled levels of NO2 as an indicator of air pollution levels from transportation and combustion, and calculating effects on mortality only above the yearly mean 10 µg/m3, we estimated excess exposure to result in 2837 (95% CI 2400-3273) deaths per year. A recent paper presenting similar calculations estimated the local contribution to urban levels of PM in Sweden to result in around 1800 deaths per year, but the authors questioned the use of risk coefficients for regional PM to assess the effect of local traffic pollutants. The new results obtained, using locally produced nitrogen dioxide as the basis for the risk assessment, resulted in an impact estimate 55 % higher than the published estimate based on PM.In the previous phase of this project a model was provided for quantifying the general population exposure to air pollution. From that work interpolated yearly mean concentrations of nitrogen dioxide were provided for the Swedish population. To be applied in the health impact assessment we selected an ecological study from Auckland, New Zealand, which reported a 13 % increase in non-accidental mortality (all ages) for 10 µg/m3 increase in NO2. Based on official national data we assumed a baseline rate of 1010 deaths per 100 000 persons and year at the population weighted mean level of approximately 10 µg NO2/m3. We then calculated the death rate and the yearly number of deaths expected at the population weighted mean exposure in each of four exposure classes above 10 µg/m3. Using the modelled levels of NO2 as an indicator of air pollution levels from transportation and combustion, and calculating effects on mortality only above the yearly mean 10 µg/m3, we estimated excess exposure to result in 2837 (95% CI 2400-3273) deaths per year. A recent paper presenting similar calculations estimated the local contribution to urban levels of PM in Sweden to result in around 1800 deaths per year, but the authors questioned the use of risk coefficients for regional PM to assess the effect of local traffic pollutants. The new results obtained, using locally produced nitrogen dioxide as the basis for the risk assessment, resulted in an impact estimate 55 % higher than the published estimate based on PM