Asthma is a common chronic respiratory condition that affects one in every thirteen adults in the U.S. Exacerbations, which are episodes of worsening asthma symptoms, are a source of major morbidity and healthcare costs. Many social, demographic, and environmental factors are known to predispose some asthma patients to exacerbations, though the way these factors interact to affect asthma outcomes are not fully understood. Approaches that enable the study of large and diverse cohorts of asthma patients would aid in the characterization of exacerbation risk factors. The use of electronic health record (EHR)-derived data for research offers unprecedented access to large volumes of longitudinal patient data that can be leveraged to understand various diseases, including asthma. However, while EHRs contain information about clinical encounters and patient-level demographics (i.e. gender, race, ethnicity, etc.), they do not capture detailed socioeconomic and environmental variables, which has diminished their utility for studies of social and environmental exposures. Here, we propose a novel approach to enhancing EHR-derived data that is cost-effective and broadly applicable. Using residential addresses obtained from EHRs, we will integrate patient records with rich and diverse data on socioeconomic variables, crime, tree canopy, and air pollution. This data integration will allow us to identify potential factors driving asthma exacerbations among Philadelphia residents. Moreover, our method preserves the major assets of EHR-derived data, namely its large patient numbers and diversity, which will enable us to perform stratified analyses with samples sizes sufficient to detect effect modification among exacerbation risk factors. Our central hypothesis is that social and environmental risk factors obtained from publicly available resources will augment our ability to understand geographic variations in asthma exacerbation rates observed in EHR-derived data, which we will address via the following specific aims: (1) Identify social factors associated with asthma exacerbations by geospatially linking census and municipal crime data to EHR data; (2) Determine the association of tree cover with asthma exacerbations by linking urban tree canopy data to EHR data; and (3) Determine the association between short-term changes in air pollutant levels with asthma exacerbations by linking air monitoring data to EHR data. This project will identify social and environmental risk factors of asthma exacerbations for real life patients with asthma in Philadelphia, a city with high sociodemographic diversity and high rates of asthma. The approaches we develop to cost-effectively enhance EHR-derived data will be readily applicable to study other complex diseases. The completion of the proposed research project, combined with complementary training and mentorship in bioinformatics, epidemiology, and statistical methods, will prepare the applicant to succeed as an independent investigator with expertise in big data approaches for epidemiologic research.
Asthma attacks are an urgent public health problem, and current understanding of the ways that social, demographic, and environmental factors interact to trigger asthma attacks is incomplete. Here, we propose a novel approach to link electronic health records with geospatial information on socioeconomic status, crime, tree cover, and air pollution to study whether these factors are associated with asthma attacks in a diverse cohort of real-life asthma patients in Philadelphia. Our approach is cost-effective and broadly applicable, and can be used in future studies of other complex disease outcomes.
