While the direct health impacts of air pollution are widely discussed, its indirect effects, particularly during pandemics, are less explored. Utilizing detailed individual-level data from all designated hospitals in Wuhan during the initial COVID-19 outbreak, we examine the impact of air pollution exposure on treatment costs and health outcomes for COVID-19 patients. Our findings reveal that patients exposed more intensively to air pollution, identified by their residence in downwind areas of high-polluting enterprises, not only had worsened health outcomes but also consumed more medical resources. This increased demand is primarily due to their heightened vulnerability to cardiopulmonary conditions. Using a causal machine learning method called Causal Forests to estimate individual treatment effects, we uncover significant heterogeneity across demographic and socioeconomic characteristics, with older and economically disadvantaged patients showing particular vulnerability. These findings highlight the importance of considering environmental factors in pandemic preparedness and suggest the value of targeted interventions that account for demographic and socioeconomic variations in vulnerability.
Keywords: air pollution; causal forests; health expenditure; health outcome; pandemic.
© 2025 John Wiley & Sons Ltd.