Development and validation of an interpretable machine learning model for predicting atrial fibrillation risk in middle-aged and older patients with coronary heart disease - Takeaways - MDSpire

Development and validation of an interpretable machine learning model for predicting atrial fibrillation risk in middle-aged and older patients with coronary heart disease

  • By

  • Feng Chen

  • Qin Fu

  • Ling Li

  • Xiao Zhang

  • Yongqiong Ge

  • Jianfei Chen

  • July 8, 2026

  • 0 min

Share

  • 1

    A machine learning framework was developed to predict atrial fibrillation risk in middle-aged and elderly patients with coronary heart disease.

  • 2

    The study analyzed a cohort of 47,617 hospitalized CHD patients using electronic medical records from January 2020 to December 2025.

  • 3

    XGBoost outperformed other algorithms, achieving an AUC of 0.867 in the training set and 0.813 in the validation set for AF risk prediction.

  • 4

    Sixteen predictors were identified, with pulse rate, total cholesterol, and systolic blood pressure being the top contributors to AF risk.

  • 5

    The model utilized SHAP and LIME for interpretability, allowing for individualized assessment of AF risk in the target population.

Original Source(s)

Development and validation of an interpretable machine learning model for predicting atrial fibrillation risk in middle-aged and older patients with coronary heart disease

  • by Feng Chen, Qin Fu, Ling Li, Xiao Zhang, Yongqiong Ge, Jianfei Chen

  • July 8, 2026

Related Content

Patient-reported and clinical outcomes after first-time atrial fibrillation ablation in older patients: a real-world retrospective single-center study

Contemporary extracorporeal cardiopulmonary resuscitation programs for cardiac arrest

Patterns of temporary mechanical circulatory support, escalation, de-escalation and outcomes in cardiogenic shock