Abstract
Air pollution causes premature death and disease and disproportionately harms non-white and lower-income groups in the United States. Government policies are responsible for the racial disparity in air pollution exposure and related health outcomes. Investigating complex relationships between policies, air pollution, and health requires (i) harmonized data connecting policies, environmental exposures, socioeconomic characteristics, and health at the individual and area level; (ii) interpretable estimands accounting for the complex interplay between policies and disparities in exposures and health outcomes; and (iii) data science approaches that can elucidate direct and indirect policy effects on disparities to identify effective interventions. We review statistical considerations and new data science approaches needed to scrutinize the policy impacts on disparities in air pollution exposure and health outcomes.