Key result
Multimodal ML integrating clinical and ECG data predicts 30-day heart failure mortality with ~0.88 AUC.
Why the study?
Heart failure is a major cause of global morbidity and mortality, necessitating accurate tools for health outcome prediction and risk stratification.
Does an interpretable multimodal machine learning framework accurately predict 30-day all-cause mortality and hospital readmission in patients with heart failure?
Does an interpretable multimodal machine learning framework accurately predict 30-day all-cause mortality and hospital readmission in patients with heart failure?
Effect estimate: AUC 0.881 for mortality prediction; AUC 0.709 for readmission prediction
A multimodal machine learning model integrating clinical data, particularly laboratory tests and ECG features, can effectively predict 30-day mortality and readmission in heart failure patients.
May enhance HF risk stratification; extends unimodal models but remains hypothesis-generating pending prospective validation.
Heart failure (HF) is one of the major causes of morbidity and mortality globally, necessitating accurate tools for health outcome prediction and risk stratification. In this study, we propose an interpretable multimodal machine learning framework integrating four clinical data modalities (i.e., demographics, medications, laboratory tests, and electrocardiograms [ECGs]) to predict 30-day all-cause mortality and hospital readmission in HF patients. Using clinical data from 2868 HF patients across 43 local hospitals in Hong Kong, we trained and evaluated ten machine learning models for HF risk prediction, with the best performing model achieving an area under the receiver operating characteristic curve (AUC) of 0.881 for mortality and 0.709 for readmission. Notably, laboratory tests and ECG features dominate predictive power, and their combination alone yielded near-optimal results (AUC: 0.872), suggesting that these two modalities may be adequate for effective risk prediction in resource-constrained settings. The SHapley Additive exPlanations (SHAP) analysis identified serum albumin, high-sensitivity troponin I, lactate dehydrogenase, and QT interval dispersion as key predictors. Feature redundancy analysis further revealed strong correlations within laboratory tests and ECG features, suggesting opportunities for model simplification. To the best of our knowledge, this is the first study that comprehensively evaluates diverse configurations of four data modalities for HF risk prediction through ablation analysis, quantifying the marginal gains of each data modality and their combinations. Our findings demonstrate that an interpretable multimodal machine learning model can enhance risk prediction in HF patients, supporting personalized management and scalable deployment across diverse healthcare settings.
No takes yet. Share an insight, caveat, or question.
Chae et al. (2026) studied Heart failure patients with a median age of 77.37 years, with available 12-lead ECG on admission, from 43 public hospitals in Hong Kong (n=2,868). Interpretable multimodal machine learning model using demographics, medications, laboratory tests, and ECG features vs. Other ML models and models with fewer data modalities was evaluated on 30-day all-cause mortality and 30-day all-cause hospital readmission (AUC 0.881 for mortality prediction; AUC 0.709 for readmission prediction). An interpretable multimodal machine learning model integrating demographics, medications, laboratory tests, and ECG features achieved an AUC of 0.881 for predicting 30-day mortality and 0.709 for predicting 30-day hospital readmission in heart failure patients.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: