Environment/Climate Change
A difference-in-differences analysis evaluating the impact of a clean heating policy on inflammatory biomarkers among rural Beijing adults Wenlu Yuan* Wenlu Yuan Kaibing Xue Talia Sternbach Xiaoying Li Collin Brehmer Xiang Zhang Ellison Carter Brian Robinson Christopher Barrington-Leigh Guofeng Shen Sam Harper Jill Baumgartner Shu Tao Yuanxun Zhang
Populations with high blood inflammatory biomarkers face greater risk of chronic illness and premature death. Outdoor and indoor PM2.5 are known contributors to elevated levels of these biomarkers. Coal has been commonly used for residential space heating in northern China, substantially contributing to indoor and outdoor PM2.5. In 2016, Beijing implemented the Clean Heating Policy (CHP), to switch homes from coal to electric heating and improve air quality. We leveraged the natural experiment from the CHP’s staggered rollout across rural Beijing villages and applied a Difference-in-Differences (DiD) approach to estimate its impact on blood inflammatory biomarkers (IL-6, TNF-α, CRP).
We enrolled participants >40yrs from 50 villages that were eligible for, but not yet treated by, the CHP in Beijing during winter 2018-19. By winter 2019-20, 10 villages had joined the CHP. Sociodemographic data was collected via home visits and blood samples were taken in village clinics by trained nurses during both winters. IL-6, TNF-α, and CRP concentrations were measured using automated enzyme-linked immunosorbent assay. We used generalized linear models with Gamma distribution and a log link to estimate CHP impacts. Dummy variables for treatment exposure, post-treatment period, and their interaction were used to implement the DiD design. Standard errors were clustered at the village level.
A total of 982 participants were included. We did not find strong evidence that the CHP affected IL-6, TNF-α, or CRP among the treated (ATT). We observed stronger ATTs among males for TNF-α (2.5pg/mL, 95%CI: 0.9, 4.1) and CRP (0.9mg/L, 95%CI: 0.1, 1.7), which were different from females (TNF-α: p<0.05, CRP: p<0.01). Our study suggests the CHP increased inflammation in male residents, despite no overall effect. The findings highlight the need for future studies to understand the mechanisms through which the CHP affected inflammation and guide similar policies in the future.