Socioeconomic disparity in the association between long-term exposure to PM2.5 and mortality in 2640 Chinese counties

• Published in: Environment international 2021-01, Vol.146, p.106241, Article 106241
• Main Authors: Han, Chunlei, Xu, Rongbin, Gao, Caroline X., Yu, Wenhua, Zhang, Yajuan, Han, Kun, Yu, Pei, Guo, Yuming, Li, Shanshan
• Format: Article
• Language: English
• Subjects: Difference-in-differences; PM2.5-mortality association; Socioeconomic status
• ISSN: 0160-4120; 1873-6750
• DOI: 10.1016/j.envint.2020.106241

•Little is known about the socioeconomic disparity in PM2.5-mortality association.•Counties of lower socioeconomic levels showed stronger PM2.5-mortality associations.•Ambient air pollution may exacerbate the existing health inequality in China.•Less developed areas need more attention to reduce the mortality burden of PM2.5. Although the association between long-term exposure to PM2.5 and mortality has been evaluated intensively, little is known about the socioeconomic disparity in the association. We collected data on annual all-cause mortality, PM2.5 concentration, socioeconomic and demographic characteristics of 2640 counties from the two most recent Chinese censuses in 2000 and 2010. We applied the difference-in-differences (DID) method to estimate PM2.5-mortality association for counties at different quartiles of literacy rate, college rate, urbanization rate and GDP per capita, respectively. Overall, every 10 µg/m3 increase in annual average PM2.5 was associated with 3.8% (95% confidence interval [CI]: 3.0–5.0) increase of all-cause mortality. The stratified analysis suggested higher health impact of exposure in counties with lower socioeconomic status. For counties of the lowest quartile (Q1) of literacy rate, college rate, urbanization rate and GDP per capita, the effect estimates were 6.0% (95% CI: 4.2–7.7), 4.4% (95% CI: 2.8–6.0), 3.5% (95% CI: 2.0–5.1) and 4.9% (95% CI: 2.7–7.1), respectively. There was strong evidence for elevated risk in mortality associated with PM2.5 of all socioeconomic factors in the lowest quartile (Q1) compared with the highest quartile counties (Q4) (p-value for difference