The Role of Environmental Legislation on Asthma Prevalence: Insights from Natural Language Processing

• In Event: The Role of Big Data and Machine Learning in Health Care: Challenges and Opportunities

Abstract

Asthma remains a pressing public health concern in the United States, affecting approximately 25 million individuals. A growing body of research has established strong links between air pollution and asthma prevalence, particularly emphasizing the role of pollutants such as PM2.5 and NO2. Studies by Paciencia et al. (2022) and Garcia et al. (2019) highlight biological mechanisms—including oxidative stress and epigenetic alterations—and the positive health outcomes resulting from air quality improvements. However, less is known about the broader legislative landscape’s influence on asthma, particularly at the state level. This study addresses that gap by analyzing how state-level environmental legislation affects asthma prevalence over time, focusing on themes such as air quality, waste management, and urban planning.Using the Environment and Natural Resources State Bill Tracking Database Archive (2009–2022) from the National Conference of State Legislatures (NCSL), we apply Natural Language Processing (Latent Dirichlet Allocation) to extract legislative themes from environmental bills across U.S. states. These thematic proportions are combined with air quality data from the EPA, demographic data from the BEA, and annual asthma prevalence rates to develop predictive models using machine learning techniques including XGBoost, Random Forest, Logistic Regression, and SVM. XGBoost achieved the highest predictive performance, identifying asthma prevalence, Regional Price Parity, number of unhealthy days, and the proportion of bills focused on environmental and waste management as the strongest predictors.To move beyond correlation, we employed a causal analysis using a Two-Stage Least Squares (2SLS) approach with Democratic seat share in state legislatures as an instrumental variable. Results show that a 1% increase in the legislative focus on environmental and waste management correlates with a 0.028% reduction in asthma prevalence in the following year.Our findings reveal regional differences in both asthma trends and legislative priorities: the Northeast, with its high population density, emphasized waste management, while the Midwest prioritized urban planning. These insights underscore the potential for targeted environmental legislation to improve respiratory health outcomes and advocate for integrating public health considerations into environmental policy. Future research should examine the pathways through which legislation influences health outcomes and explore additional environmental determinants of asthma.

Author

• Tianxiang Shen, Northwestern University
  Presenting Author