Key result
EHR-based XGBoost model identifies cardiac arrhythmias with ~0.89 AUROC in high-risk adults.
Why the study?
Cardiac arrhythmias may go unrecognized in adults with chronic disease, while established risk scores primarily address atrial fibrillation, motivating development of an electronic health record-based machine-learning model to identify patients more likely to receive a documented cardiac arrhythmia diagnosis.
Does an EHR-based XGBoost machine-learning model accurately identify patients likely to receive a documented cardiac arrhythmia diagnosis compared with other machine-learning models and established clinical risk scores (CHARGE-AF and HARMS2-AF) in adults with high-risk chronic disease (T2DM, COPD, CKD, or CAD)?
Cohort (n=30,356)
No
Does an EHR-based XGBoost machine-learning model accurately identify patients likely to receive a documented cardiac arrhythmia diagnosis compared with other machine-learning models and established clinical risk scores (CHARGE-AF and HARMS2-AF) in adults with high-risk chronic disease (T2DM, COPD, CKD, or CAD)?
An EHR-based 30-feature XGBoost model demonstrated strong internal discrimination (AUROC 0.894) for identifying undiagnosed cardiac arrhythmias among adults with high-risk chronic conditions and outperformed established clinical risk scores (CHARGE-AF and HARMS2-AF) for atrial fibrillation/flutter.
Answers are grounded in this paper and cite their sources.
Supports EHR-based ML arrhythmia screening in high-risk adults; leaves open clinical utility pending prospective validation.
Background: Cardiac arrhythmias may remain unrecognized in adults with chronic disease. Established risk scores primarily address atrial fibrillation. Objective: To develop and internally evaluate an electronic health record (EHR)-based machine-learning model to identify patients more likely to receive a documented cardiac arrhythmia diagnosis. Methods: We retrospectively evaluated EHR data (2018-2025) from one healthcare organization among adults with type 2 diabetes, chronic obstructive pulmonary disease, chronic kidney disease, or coronary artery disease. Logistic regression, random forest, and XGBoost were compared. XGBoost was selected using internal validation, reduced to 30 features using Shapley additive explanations, and evaluated in a holdout set with data ending one day before the index date. Sensitivity analyses excluded information recorded within 30, 90, or 180 days before the index date. An atrial fibrillation/flutter-specific analysis compared XGBoost with CHARGE-AF and HARMS2-AF. Results: Among 30,356 patients, 6,835 (22.5%) had a qualifying arrhythmia diagnosis. In 3,051 holdout patients, XGBoost achieved an area under the receiver operating characteristic curve (AUROC) of 0.894 (95% CI 0.879-0.908) and an area under the precision-recall curve of 0.792 (95% CI 0.763-0.822). At a 0.5 threshold, sensitivity was 0.564, specificity 0.971, and precision 0.857. AUROC remained 0.896-0.897 across earlier cutoffs. In the common score-calculable atrial fibrillation/flutter subset (n = 2,436), AUROCs were 0.883 for XGBoost, 0.712 for CHARGE-AF, and 0.554 for HARMS2-AF. Conclusions: XGBoost demonstrated strong internal discrimination for documented arrhythmias. External validation and prospective evaluation are needed to assess generalizability and clinical utility.
No takes yet. Share an insight, caveat, or question.
Goel et al. (2026) conducted a cohort in High-risk chronic disease (type 2 diabetes mellitus, chronic obstructive pulmonary disease, chronic kidney disease, or coronary artery disease) (n=30,356). EHR-based XGBoost machine-learning model (30 SHAP-selected features) vs. Logistic regression, random forest, CHARGE-AF score, and HARMS2-AF score was evaluated on First recorded qualifying cardiac arrhythmia diagnosis in the EHR (ICD-10-CM categories I44 atrioventricular and left bundle-branch block, I47 paroxysmal tachycardia, I48 atrial fibrillation and flutter, or I49 other cardiac arrhythmias) (AUROC 0.8935, 95% CI 0.8786-0.9077). The 30-feature EHR-based XGBoost model achieved an AUROC of 0.894 and AUPRC of 0.792 in holdout validation for identifying documented cardiac arrhythmias in high-risk adults.
Synapse has enriched 10 related papers. Consider them for comparative context: