Development and validation of a nomogram for predicting hospitalization-associated disability in older patients with acute heart failure

TL;DR

This study created a 12-variable tool to predict disability risk in older heart failure patients during hospitalization, helping improve personalized care.

Contribution

A validated nomogram was developed to predict hospitalization-associated disability in older acute heart failure patients.

Findings

The incidence of hospitalization-associated disability was 41.88%.

The nomogram showed strong discrimination with an AUC of 0.841 in the training set.

Calibration and clinical utility were confirmed, supporting the nomogram's practical use.

Abstract

To develop and validate a nomogram for predicting hospitalization-associated disability in older patients with acute heart failure.

A single-center cohort study.

This study was carried out in the Cardiovascular Disease Center of a large tertiary-care hospital in China. Between December 2024 and February 2025, 480 older patients with acute heart failure were enrolled.

Hospitalization-associated disability was defined as a decline of 5 points or more in the Barthel Index from admission to discharge. Predictor screening involved univariable logistic regression, Spearman's correlation, and Least Absolute Shrinkage and Selection Operator regression. Variables retained were entered into a multivariable logistic regression model, and significant predictors were used to construct a nomogram for predicting hospitalization-associated disability. Model performance was assessed in terms of discrimination, calibration, and clinical utility.

The incidence of hospitalization-associated disability was 41.88%. A 12-variable nomogram was developed, incorporating age, ejection fraction, emergency admission, comorbidity burden, cognitive function, nutritional risk, pre-admission activities of daily living, pre-admission instrumental activities of daily living, physical mobility, sleep disturbance, depressive symptoms, and perceived social support. The nomogram demonstrated robust discrimination, with the area under the receiver operating characteristic curve of 0.841 and 0.786 in the training and testing sets, respectively. Calibration was accurate in both sets. The training set achieved a mean absolute error of 0.037 and a Brier score of 0.154, while the testing set showed 0.026 and 0.188, accompanied by non-significant Hosmer-Lemeshow test results. Decision curve and clinical impact analyses further supported favorable clinical utility.

A 12-variable nomogram was developed and validated in older individuals with acute heart failure, enabling early identification of hospitalization-associated disability risk and supporting personalized care strategies.

Figures