Comparative diagnostic performance of machine learning models and traditional scores for HFpEF in older adults

Machine learning algorithms, particularly RF and XGBoost, demonstrated superior diagnostic accuracy compared to established HFpEF scoring systems. These findings support the potential integration of ML-based tools into clinical workflows to facilitate earlier identification of HFpEF and prompt initiation of guideline-recommended therapies.