Characteristics and Comorbidities Influencing Mortality Risk Among Patients with Hereditary Angioedema

Subhan Khalid, M.S. Applied Statistics
Ph.D. Candidate, Data Science, Harrisburg University of Science & Technology
Senior Associate Manager Data Science, IQVIA Inc.

Abstract

Background

Patients suffering from hereditary angioedema (HA) face a heightened mortality risk due to multiple factors.

Objective

The purpose of this study was to identify patient demographics or comorbidities associated with higher mortality risk using Bayesian network analysis.

Methods

Data from the 2021 Nationwide Inpatient Sample were used to identify hospitalized patients with hereditary angioedema. Patient demographics, comorbidities, and severity measures were analyzed, and a Bayesian network model was developed to assess factors contributing to mortality risk. Structure learning was performed using a directed acyclic graph and probability estimating using Bayesian Inference. Model performance was validated using a 70/30 training-testing split and assessed via area under the curve.

Results

Older HA patients and those with autoimmune conditions, hypertension, or low income were at higher risk of mortality. Elevated risk was also observed across certain racial groups, insurance types, and income levels. Notably, older Black patients from the Midwest exhibited the highest estimated mortality risk. The Bayesian Network demonstrated strong predictive performance, highlighting its potential for identifying high-risk subgroups and supporting targeted clinical interventions.

Conclusion

The findings of this study provide valuable insights into the factors influencing mortality risk for HA patients, with BN analysis offering a detailed understanding of complex dependencies among patient demographics and comorbidities. These results have ramifications for both patients and physicians to improve HA symptom management and preventing onset of life-threatening situations.

About the Presenter

Subhan Khalid is a Ph.D. candidate in Data Science at Harrisburg University of Science & Technology, where his research focuses on Bayesian modeling, health outcomes and causal inference. He holds a Master of Science in Applied Statistics from Villanova University and currently serves as a Senior Associate Manager of Data Science at IQVIA Inc., a leading global provider of advanced analytics and clinical research services. With a strong background in real-world evidence and health outcomes research, Subhan brings a data-driven perspective to improving patient outcomes and informing evidence-based decision-making in healthcare.