Development and validation of a nomogram model for predicting unplanned readmission in patients with acute pancreatitis.

OBJECTIVE

The objective of this study was to develop and validate a nomogram for predicting 1-year unplanned readmission in patients with acute pancreatitis (AP) to identify high-risk populations.

METHODS

We retrospectively selected 474 AP patients who were treated and discharged from the First People's Hospital of Linping District, Hangzhou City, from 1 January 2021 to 31 December 2023. These patients were randomly divided into a training cohort (n = 332) and an internal validation cohort (n = 142) in a 7:3 ratio. In addition, 218 AP patients treated during the same period at the People's Hospital of Jiande City were selected as an external validation cohort. The least absolute shrinkage and selection operator (LASSO) was used for variable selection, and multivariable logistic regression was applied for model development. A nomogram was then constructed to estimate the risk of 1-year unplanned readmission. Model performance was evaluated using the consistency index (C-index), calibration curves, receiver operating characteristic (ROC) curves, and decision curve analysis (DCA).

RESULTS

Within 1 year, the incidence of unplanned readmission was 36.1% (120/332) in the training cohort, 40.1% (57/142) in the internal validation cohort, and 42.7% (93/218) in the external validation cohort. Six independent predictors of unplanned readmission in patients with AP were identified, including biliary AP, diabetes, alcohol, infected pancreatic necrosis (IPN) at first admission, acute peripancreatic fluid collection (APFC), and readmission score. The nomogram demonstrated sufficient predictive accuracy, with area under the curve (AUC) values of 0.739 (95% confidence interval [CI]: 0.684-0.794), 0.836 (95% CI: 0.770-0.902), and 0.704 (95% CI: 0.636-0.772) in the training cohort, internal validation cohort, and external validation cohort, respectively. The calibration curve showed good agreement between the predicted risk and the actual risk observed.

CONCLUSIONS

The nomogram developed in this study demonstrates good predictive value for unplanned readmission in patients with AP and may help identify high-risk populations.