Development and Validation of a Nomogram for Predicting Risk of Emergency Department Revisits in Chinese Older Patients

Abstract

Objective: The identification of older patients at risk of repeated emergency department (ED) visits is crucial for managing preventable adverse outcomes. This study aims to identify risk factors associated with ED revisits and to develop and validate a nomogram for predicting risk of geriatric ED revisits. Methods: This was a cohort study comprising 553 older patients, who attended the two tertiary hospitals EDs in China from August 2018 to February 2019 and were prospectively followed for any unplanned revisit within 1 year after discharge. Patients were randomly assigned to a training or validation set at a ratio of 2:1. Stepwise selection procedure was applied to select factors associated with ED revisits for inclusion in a multivariable logistic model from which a nomogram was elaborated. Discrimination, calibration and clinical utility of the nomogram were assessed using C-statistic, calibration plot, Hosmer-Lemeshow test, and decision curve analysis (DCA). Results: The final nomogram included four predictors for ED revisits: age, BMI, frailty and polypharmacy. Older patients having revisits were more likely to be frail (OR = 1.17, p = 0.031), have polypharmacy (OR = 1.69, p = 0.049) or BMI <18.5 kg/m2 (OR = 2.45, p = 0.025), and were less likely to be older than 90 years (OR = 0.21, p = 0.002). The nomogram demonstrated acceptable discrimination ability in the training (C-index = 0.661) and validation sets (C-index = 0.651), satisfactory calibration (p > 0.05), and good clinical applicability. Conclusion: A nomogram incorporating four obtainable variables was constructed to individualize ED readmission risk in older patients. These patients may benefit from early triage and better-targeted care if considering the nomogram as a clinical decision aid.