Hospital admissions risks from fine particulate matter by source sector, fuel type, and components: A national analysis - ABSTRACT There is an urgent need to understand the health impacts of alternative air quality policies. Such policies can impact various source sectors (e.g., transportation), fuel type and chemical composition of fine particulate matter (PM2.5). Detailed characterization of the health consequences of exposure to PM2.5 sources and chemical composition will provide timely guidance to develop the most cost effective interventions for air pollution mitigation. The National Academy of Sciences, Health Effects Institute, US EPA, and Global Burden of Disease have all identified understanding health outcomes of air pollution sources as a critical research need. Prior studies suffer from the inability to identify relative contributions of sources and fuel types to pollutant mixtures, quantify uncertainties, or apply rigorous cross-validation, often due to lack of appropriate data and approaches. Most past analyses have inherently assumed that all PM2.5 mass is equally toxic, regardless of source or composition; in fact, PM2.5 is the only air pollutant regulated without regard to chemical form. To address this knowledge gap, we will conduct a nationwide analysis to determine which types of PM2.5 (e.g., source sectors, fuel types) are most associated with cardiovascular and respiratory hospital admissions. We will employ 2 large, well-characterized datasets (Medicare: >60 million participants ≥65 years, and Medicare Current Beneficiary Survey: >200,000 participants with detailed data on >100 individual-level variables). Using multiple approaches, including causal inference, we will test the hypothesis that health impacts of PM2.5 vary by pollution source and/or combustion type for cardiovascular and respiratory hospital admissions. We will estimate and validate PM2.5 exposure by chemical components (e.g., sulfate, organics, black carbon), source sectors, and fuel types for the continental US for 2000-2023 at monthly, ZIP code-level resolution by harmonizing data from satellites, air quality models, emissions inventories, and monitors (Aim 1). We will apply well-established approaches, modified for our analysis, to estimate exposure-response curves (ERCs), with causal interpretation, for PM2.5 and components (Aim 2). We will ensure computational scalability and account for uncertainty in estimated exposures, unmeasured confounding bias, and co-pollutants. We will estimate ERCs for cause-specific (cardiovascular, neurological, asthma, other respiratory) hospital admissions by PM2.5 type (Aim 3). This work will help air quality and health policymakers prioritize mitigation efforts to maximize health benefits. Findings will inform assessments of health effects of PM2.5 exposure. We will disseminate new exposure data, causal inference applications, and statistical code.
