Identifying risk factors and predicting stroke using Bayesian networks: Evidence from NHANES 2011-2020.

Area of Science:

• Cerebrovascular disorders
• Biostatistics
• Machine learning for healthcare

Background:

• Stroke is a major global health concern, necessitating early risk factor identification for effective prevention.
• Understanding complex interdependencies among stroke risk factors is crucial for developing targeted interventions.
• Current predictive models may not fully capture the intricate relationships influencing stroke occurrence.

Purpose of the Study:

• To develop an interpretable Bayesian network (BN) model for predicting stroke.
• To identify key stroke risk factors and their complex interdependencies.
• To evaluate the predictive performance of the BN model against other machine learning algorithms.

Main Methods:

• Utilized cross-sectional data from the National Health and Nutrition Examination Survey (NHANES) (2011-2020).
• Employed univariate and multivariate logistic regression for feature selection.
• Constructed the Bayesian network using the hybrid HPC algorithm (H2PC) and estimated probabilities via maximum likelihood estimation.
• Assessed model performance using the area under the receiver operating characteristic curve (AUROC).

Main Results:

• The final sample included 20,535 individuals.
• Identified five direct stroke risk factors: age, sleep disorders, alcohol consumption, coronary heart disease, and diabetes.
• The BN model achieved an AUROC of 0.803, outperforming other machine learning methods.

Conclusions:

• The developed BN model offers an interpretable visualization of stroke risk factor interdependencies.
• The model demonstrates competitive predictive accuracy, highlighting its potential for stroke risk assessment.
• Further longitudinal validation is suggested to enhance its utility in future stroke prevention strategies.