Machine learning prediction of 1-year mortality in older patients with heart failure: a nationwide, multicenter, prospective cohort study - PubMed
3 months ago
- #Heart Failure
- #Machine Learning
- #Prognosis
- Study aimed to develop a machine learning model for predicting 1-year mortality in older heart failure patients using functional assessments.
- Data from the J-Proof HF Registry in Japan, involving 9700 patients aged ≥65, was analyzed.
- An XGBoost model with 77 predictors achieved an AUC of 0.76, outperforming traditional risk scores like AHEAD and BIOSTAT.
- Key predictors included functional measures at discharge such as Barthel index, gait speed, and handgrip strength.
- The model demonstrated improved risk stratification and clinical utility over established scores.
- Functional status at discharge was highlighted as a critical prognostic indicator for post-discharge care planning.