Predicting prolonged hospitalization in asthma patients: model development and external validation

Area of Science:

• Artificial Intelligence in Medicine
• Clinical Informatics
• Respiratory Medicine

Background:

• Prolonged hospitalization in asthma patients poses a significant clinical and economic burden.
• Accurate prediction of prolonged stays is crucial for resource allocation and patient care optimization.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for predicting prolonged hospitalization in asthma patients.
• To identify key clinical factors associated with extended hospital stays.

Main Methods:

• A retrospective cohort study involving 2820 asthma patients for internal validation and 1714 patients for external validation.
• Utilized LASSO and logistic regression for feature selection, employing nine ML algorithms.
• The Extreme Gradient Boosting (XGBoost) model was selected based on performance metrics.

Main Results:

• The XGBoost model demonstrated strong predictive performance with an AUC of 0.829 (internal) and 0.745 (external).
• Key predictors included age, oxygen saturation, red blood cell count, hemoglobin, and comorbidities like pneumonia and COPD.
• Decision curve analysis confirmed the model's good clinical utility.

Conclusions:

• The XGBoost model effectively predicts prolonged hospitalization in asthma patients.
• This tool can aid clinicians in identifying at-risk individuals for proactive management.