A Spatiotemporal Causal Machine Learning Framework for Quantifying the Impact of PM2.5 on Asthma at the Community Level
- 1Department of Geography and the Environment University of Denver Denver CO USA
- 2Department of Geosciences Texas Tech University Lubbock TX USA
- 3National Jewish Health Denver CO USA
Abstract
Air pollution, particularly fine particulate matter with an aerodynamic diameter of 2.5 μm or smaller (PM2.5), is a leading environmental risk factor for respiratory diseases. While previous studies have investigated the biological mechanisms linking PM2.5 to asthma exacerbation, these studies are unable to scale for population‐level assessments. In contrast, most regional analyses rely on traditional statistical models that primarily quantify associations, limiting estimation of health changes under hypothetical pollution interventions. This study introduces a spatiotemporal causal machine learning framework to estimate PM2.5‐related changes in asthma outcomes at the county level. Using county‐level monthly hospitalization and emergency department records from 2011 to 2019 in Colorado, the framework integrates spatial and temporal representation learning to model geographic interdependence and temporal dynamics in exposure–response relationships. Further, through a causal inference approach, the model estimates model‐based counterfactual health outcomes under hypothetical PM2.5 interventions, moving beyond regression‐based correlation analysis. Results show that the model‐based counterfactual estimates exhibit spatial and population heterogeneity. Larger estimated effects were observed in several communities experiencing compounded environmental and social stressors, while uncertainty was greater in data‐sparse counties.
Related
- Machine learning-based fall risk prediction in community-dwelling older adults using dual-task spatiotemporal gait parameters
- Quantifying Reproducibility of Common Laboratory Tests in a Tertiary Care Hospital for More Realistic Synthetic Data in Machine Learning
- A spatiotemporal deep learning framework for detecting unusual human activity in surveillance videos
- Primary Care Quality and Inappropriate Community Antibiotic Use: A Double Machine Learning Instrumental Variable Approach
- An integrated evolutionary optimization and machine learning framework for predicting pharmacokinetics of nanocarrier-encapsulated drugs
BibTeX
@article{li2026spatiotemporal,
title = {{A Spatiotemporal Causal Machine Learning Framework for Quantifying the Impact of PM2.5 on Asthma at the Community Level}},
author = {Jing Li and Xuantong Wang and Hong Wei Chu},
journal = {GeoHealth},
year = {2026},
month = oct,
doi = {10.1029/2025gh001735},
url = {https://europepmc.org/article/PMC/PMC13723011},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - A Spatiotemporal Causal Machine Learning Framework for Quantifying the Impact of PM2.5 on Asthma at the Community Level
AU - Jing Li
AU - Xuantong Wang
AU - Hong Wei Chu
PY - 2026
DA - 2026/10/31
JO - GeoHealth
DO - 10.1029/2025gh001735
UR - https://europepmc.org/article/PMC/PMC13723011
AB - Air pollution, particularly fine particulate matter with an aerodynamic diameter of 2.5 μm or smaller (PM2.5), is a leading environmental risk factor for respiratory diseases. While previous studies have investigated the biological mechanisms linking PM2.5 to asthma exacerbation, these studies are unable to scale for population‐level assessments. In contrast, most regional analyses rely on traditional statistical models that primarily quantify associations, limiting estimation of health changes under hypothetical pollution interventions. This study introduces a spatiotemporal causal machine learning framework to estimate PM2.5‐related changes in asthma outcomes at the county level. Using county‐level monthly hospitalization and emergency department records from 2011 to 2019 in Colorado, the framework integrates spatial and temporal representation learning to model geographic interdependence and temporal dynamics in exposure–response relationships. Further, through a causal inference approach, the model estimates model‐based counterfactual health outcomes under hypothetical PM2.5 interventions, moving beyond regression‐based correlation analysis. Results show that the model‐based counterfactual estimates exhibit spatial and population heterogeneity. Larger estimated effects were observed in several communities experiencing compounded environmental and social stressors, while uncertainty was greater in data‐sparse counties.
SN - 2471-1403
ER -