An interrupted time series analysis of the cardiovascular health benefits of a coal coking operation closure

Abstract

With the widespread implementation of air pollution mitigation strategies for health and climate policy, there is an emerging interest in accountability studies to validate whether a reduction of air pollution exposure, in fact, produces the human health benefits estimated from past air pollution epidemiology. The closure of a coal coking plant provides an ideal ‘natural’ experiment opportunity to rigorously evaluate the health benefits of air pollution emissions reductions. In this study, we applied an interrupted time series model to test the hypothesis that the substantial reduction in air pollution induced by the closure of the Shenango, Inc. coke plant in Pittsburgh, PA during January, 2016 was followed by immediate and/or longer-term cumulative local cardiovascular health benefits. We observed a 90% decrease in nearby SO2 levels, as well as significant reductions in coal-related fine particulate matter constituents (sulfate and arsenic), after the closure. Statistically significant cardiovascular health benefits were documented in the local population, including a 42% immediate drop (95% CI: 33%, 51%) in cardiovascular emergency department (ED) visits from the pre-closure mean. A longer-term downward trend was also observed for overall emergency visits at −0.14 (95% CI: −0.17, −0.11) visits per week rate of decrease after the closure, vs. a rise of 0.17 (95% CI: 0.14, 0.20) visits per week before. Similarly, inpatient cardiovascular hospitalizations per year showed a decrease after closure (−27.97 [95% CI: −46.90, −9.04], as compared with a 5.09 [95% CI: −13.84, 24.02] average increase in cases/year over the prior three years). Our study provides clear evidence that this intervention lowering fossil fuel-associated air pollution benefited public health in both the short and longer term, while also providing validation of the past use of observational air pollution epidemiology effect estimates in policy analyses.