Beyond the Average: A Geospatial Analysis Revealing Bimodal and Trimodal Patterns in the Community-Level Drivers of CABG Mortality
Presented During:
Monday, May 4, 2026: 9:00AM - 4:00PM
McCormick Place Lakeside Center
Posted Room Name:
Exhibit Hall, Poster Area
Abstract No:
P0119
Submission Type:
Abstract Submission
Authors:
Alyster Alcudia (1), Glade Adams (1), Aryan Gupta (1), Priya Joshi (1), Kristina Lowndes (1), Subodha Kumar (2), William Moser (3), Suyog Mokashi (3)
Institutions:
(1) Lewis Katz School of Medicine at Temple University, Philadelphia, PA, (2) Fox School of Business at Temple University, Philadelphia, PA, (3) Temple University Hospital, Philadelphia, PA
Submitting Author:
Alyster Alcudia
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Co-Author(s):
Glade Adams
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Aryan Gupta
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Priya Joshi
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Kristina Lowndes
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Subodha Kumar
-
Contact Me
Fox School of Business at Temple University
Fox School of Business at Temple University
William Moser
-
Contact Me
Temple University Hospital
Temple University Hospital
Suyog Mokashi
-
Contact Me
Temple University Hospital
Temple University Hospital
Presenting Author:
Alyster Alcudia
-
Contact Me
Lewis Katz School of Medicine at Temple University
Lewis Katz School of Medicine at Temple University
Abstract:
Objective: Current risk-adjustment models in cardiac surgery are geographically blind, treating community-level risks as a uniform national average. We hypothesized that the impact of social deprivation and air pollution on CABG mortality is not constant, but clusters into distinct, geographically-specific regimes.
Methods: Using public Centers for Medicare & Medicaid Services (CMS), Census, and Environmental Protection Agency (EPA) data, we identified a 2023 cohort of 850 US hospitals performing CABG. We defined granular, patient-flow-derived Hospital Service Areas (HSAs) for each. We calculated a population-weighted Area Deprivation Index (ADI) and Particulate Matter (PM2.5) pollution exposure for each HSA. Geographically Weighted Regression (GWR) was used to model the local association between risk-adjusted 30-day CABG mortality and these community factors, controlling for hospital ownership.
Results: A standard ordinary least squares (OLS) model confirmed that higher ADI was a significant predictor of increased mortality nationally (coefficient: +0.0089, p<0.001). However, GWR revealed that the magnitude of this harm follows a bimodal distribution. A "Compounded Risk" regime, with a strong link between deprivation and mortality, was identified in hospitals clustered in the Great Plains. In contrast, a "Mitigated Risk" regime with a much weaker association was found in other regions, including the Pacific Northwest. The relationship with PM2.5 was trimodal, unmasking a "Delta Paradox" of confounding in the Deep South and, most critically, a distinct environmental "hot spot" where higher PM2.5 was associated with a significant increase in mortality, clustered along the Southeastern Atlantic Seaboard.
Conclusions: The impact of community determinants on CABG outcomes is not a simple continuum but a set of distinct, geographically coherent regimes. Geographically blind risk models are insufficient, as they fail to capture critical, place-based drivers of surgical outcomes. Health policy aimed at improving surgical quality must be tailored to the unique social and environmental realities of the communities that hospitals serve.
Methods: Using public Centers for Medicare & Medicaid Services (CMS), Census, and Environmental Protection Agency (EPA) data, we identified a 2023 cohort of 850 US hospitals performing CABG. We defined granular, patient-flow-derived Hospital Service Areas (HSAs) for each. We calculated a population-weighted Area Deprivation Index (ADI) and Particulate Matter (PM2.5) pollution exposure for each HSA. Geographically Weighted Regression (GWR) was used to model the local association between risk-adjusted 30-day CABG mortality and these community factors, controlling for hospital ownership.
Results: A standard ordinary least squares (OLS) model confirmed that higher ADI was a significant predictor of increased mortality nationally (coefficient: +0.0089, p<0.001). However, GWR revealed that the magnitude of this harm follows a bimodal distribution. A "Compounded Risk" regime, with a strong link between deprivation and mortality, was identified in hospitals clustered in the Great Plains. In contrast, a "Mitigated Risk" regime with a much weaker association was found in other regions, including the Pacific Northwest. The relationship with PM2.5 was trimodal, unmasking a "Delta Paradox" of confounding in the Deep South and, most critically, a distinct environmental "hot spot" where higher PM2.5 was associated with a significant increase in mortality, clustered along the Southeastern Atlantic Seaboard.
Conclusions: The impact of community determinants on CABG outcomes is not a simple continuum but a set of distinct, geographically coherent regimes. Geographically blind risk models are insufficient, as they fail to capture critical, place-based drivers of surgical outcomes. Health policy aimed at improving surgical quality must be tailored to the unique social and environmental realities of the communities that hospitals serve.
ADULT CARDIAC:
Coronary Artery Disease
Image or Table
