Development and validation of a predicting model for risk of rehospitalization within 30 days in heart failure patients with preserved ejection fraction

Abstract Objective This study aimed to develop and validate a predicting model for risk of rehospitalization within 30 days post discharge in patients diagnosed with Heart Failure with Preserved Ejection Fraction (HFpEF). Methods Clinical data of 9555 HFpEF patients registered in the heart failure registry from 2014 to 2022 at our hospital were analyzed. A total of 8160 patients (from 2014-2021) were designated as the training set, and 1395 patients (from 2022) as the validation set. A predictive model was established based on risk factors associated with 30-day rehospitalization due to heart failure identified by the Least Absolute Shrinkage and Selection Operator (LASSO) regression and logistic regression analysis. The model's stability was ensured through Bootstrap internal validation and Temporal validation external validation. Results LASSO and logistic regression analyses identified eight risk factors associated with increased risk of 30-day rehospitalization, including New York Heart Association (NYHA) classification, albumin, c-reactive protein, N-terminal pro b-type natriuretic peptide (NT-proBNP) divided by 100, anemia, pleural effusion, atrial fibrillation, and pulmonary hypertension. The established predictive model demonstrated good discriminative ability in the internal validation with an Area Under the Curve (AUC) value of 0.714, and a 95% confidence interval (CI) of 0.693-0.736 after 500 Bootstrap replicates. In the Temporal validation external validation, the model exhibited robust stability with an AUC value of 0.740 and a 95% CI of 0.693-0.781. When compared to analogous models from prior studies, present model showcased superior predictive accuracy in that both Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI) values were positive and within their respective 95% confidence intervals (P-values <0.05). Conclusion This study successfully developed and validated a predictive model for 30-day rehospitalization due to heart failure in patients diagnosed with HFpEF. This model might allow the timely identification of high-risk HFpEF patients for 30-day rehospitalization due to heart failure and optimizes the patient management accordingly.