Predicting coronary heart disease in Chinese diabetics using machine learning

Area of Science:

• Cardiology
• Endocrinology
• Artificial Intelligence in Medicine

Background:

• Diabetes mellitus is a global chronic disease associated with significant vascular complications, including coronary heart disease (CHD).
• CHD is a leading cause of mortality worldwide, and its comorbidity with diabetes requires further investigation for effective management.
• Understanding the factors contributing to diabetes and CHD comorbidity is crucial for public health strategies.

Purpose of the Study:

• To develop and evaluate an artificial intelligence (AI) model for predicting the risk of coronary heart disease (CHD) in patients with diabetes.
• To identify key risk factors associated with the comorbidity of diabetes and CHD.
• To provide personalized early warning guidance for diabetic populations at risk of CHD.

Main Methods:

• Analysis of a large cohort (>300,000) of diabetes patients in southwest China.
• Statistical analysis of demographic, laboratory, medical examination, and questionnaire data.
• Evaluation of machine learning models including eXtreme Gradient Boosting, Random Forest, and Logistic Regression.

Main Results:

• Identification of nine significant predictive features for diabetes/CHD comorbidity: age, waist-to-height ratio (WHtR), body mass index (BMI), history of stroke, smoking, chronic lung disease, alcohol consumption, and medical service utilization (MSP).
• The developed AI model achieved an Area Under the Receiver Operating Characteristic Curve (AUC) of 0.701 on test samples.
• Traditional machine learning methods were assessed for their predictive performance.

Conclusions:

• The AI model demonstrates potential for predicting CHD risk in diabetic patients.
• Key identified features provide insights into the multifactorial nature of diabetes/CHD comorbidity.
• The findings can support personalized risk assessment and early intervention strategies for diabetic individuals.