A comparative study on estimation methods of multi-pollutant health effects based on simulation data
Objective To explore the assessment method of multi-pollutants effectsat the same time.Methods The data of daily cardiovascular deaths and concentration of 6 main pollutants during 2015 to 2019 in Beijing were collected.Based on the simulated death data under 10 combinations of 2 pollutant weights and 5 health effects,4 approaches were compared in evaluating health risks of multi-pollutants:Single pollutant addition method,cumulative risk index(CRI),supervised principal component analysis(SPCA)and Bayes multi-pollutant weighted model(BMP).Results The standard methods always overestimated the health effects of pollutants,while SPCA methods always underestimated the effects.When the simulated health effects are low,the BMP was the most accurate,but the CRI estimate is closer to the real value when the setting effects are high.In all scenarios,the standard deviations(SD)of the health effects estimated by BMP were consistently smaller than CRI,the SD of the total effect estimated by BMP is 69％-96％of the SD of CRI.Conclusion The BMP method can estimate the health effects of pollutants stably,and the estimation results can reasonably explain the impact of pollutants on population health.
Multi-pollutantsSimulationCumulative risk indexSupervised principal component analysisBayes multi-pollutant weighted model
NETL
NSTL
万方数据
