Related Experiment Video
Updated: May 5, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Development and validation of a risk prediction model for 30-day readmission in elderly type 2 diabetes patients
Yuxin He1, Yuan Yuan2, Qingzhu Tan3
1Department of Medical Administration, Affiliated Banan Hospital of Chongqing Medical University, Chongqing, China.
Background:
Elderly type 2 diabetes mellitus (T2DM) patients complicated with heart failure (HF) exhibit a high rate of 30-day readmission. Predictive models have been suggested as tools for identifying high-risk patients. Thus, we aimed to develop and validate a predictive model using multicenter electronic medical records (EMRs) data to estimate the risk of 30-day readmission in elderly T2DM patients complicated with HF.
Methods:
EMRs data of elderly T2DM patients complicated with HF from five tertiary hospitals, spanning 2012 to 2023, were utilized to develop and validate the 30-day readmission model. The model were evaluated using holdout data with the area under the receiver operating characteristic curve (AUROC), calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC).
Results:
A total of 1899 patients were included, with 955, 409, and 535 in the derivation, internal validation, and external validation cohorts, respectively. Pulmonary infections (odds ratio [OR]: 3.816, 95% confidence interval [CI]: 2.377-6.128, P < 0.001), anti-hypertensive drug use (OR: 5.536, 95% CI: 1.658-18.486, P = 0.005), and neutrophil percentage-to-albumin ratio (NPAR) (OR: 1.144, 95% CI: 1.093-1.197, P < 0.001) were independent predictors of 30-day readmission risk. AUROC in the derivation, internal validation, and external validation cohorts were 0.782 (95% CI: 0.737-0.826), 0.746 (95% CI: 0.654-0.838), and 0.753 (95% CI: 0.684-0.813), respectively. The calibration curve, DCA results, and CIC results indicated that the model also possessed good predictive power. Additionally, an operation interface on a web page (https://cqykdxtjt.shinyapps.io/readmission/) was created for clinical practitioners to apply.
Conclusion:
A 30-day readmission risk prediction model was developed and externally validated. This model facilitates the targeting of interventions for elderly T2DM patients complicated with HF who are at high risk of an early readmission.
