Related Experiment Video
Updated: Jul 9, 2026
07:31
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 External Validation of a Machine Learning Model for 90-Day Readmission in Hospitalized Older Patients
Guibin Zhang1, Dan Wang1, Hang Chen1
1Department of Respiratory and Critical Care Medicine, Affiliated Hospital of North Sichuan Medical College, Nanchong, Sichuan, 637000, People's Republic of China.
Summary
This study developed a predictive model for 90-day readmission in older adults with acute exacerbation of chronic obstructive pulmonary disease (AECOPD). The XGBoost model demonstrated strong performance, aiding early risk stratification for AECOPD patients.
Area of Science:
- Pulmonary Medicine
- Health Informatics
- Biostatistics
Background:
- Short-term readmission is common in older adults hospitalized with acute exacerbation of chronic obstructive pulmonary disease (AECOPD).
- Early risk stratification for AECOPD readmission remains limited.
- This study aimed to develop and validate a 90-day readmission prediction model using early admission data.
Purpose of the Study:
- To develop and externally validate a predictive model for 90-day hospital readmission in older adults diagnosed with AECOPD.
- To identify key predictors of readmission using early admission data.
- To compare the performance of various machine learning models for AECOPD readmission prediction.
Main Methods:
- A retrospective two-center study involving 692 patients (513 development, 179 external validation).
- Predictors were limited to early admission variables, including neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), D-dimer, frequent exacerbations (FE), body mass index (BMI), and albumin (ALB).
- The Extreme Gradient Boosting (XGBoost) model was trained and validated, with performance assessed using Area Under the Receiver Operating Characteristic Curve (AUC) and Brier score. Shapley Additive Explanations (SHAP) were used for interpretability.
Main Results:
- The XGBoost model achieved an AUC of 0.892 in the development cohort and 0.847 in the external validation cohort.
- Key predictors identified were D-dimer, NLR, and FE, with SHAP analyses suggesting non-linear effects.
- The model demonstrated stable performance and good discrimination for predicting 90-day readmission in AECOPD patients.
Conclusions:
- An externally validated 90-day readmission prediction model for AECOPD was developed using early admission data.
- The XGBoost model exhibited stable performance and clinically interpretable risk patterns.
- This model shows potential for early risk stratification in older adults hospitalized with AECOPD.
