A Machine Learning Approach to Predict Same-Day Discharge after Angiography Procedures.

Area of Science:

• Medical informatics
• Health services research
• Machine learning in healthcare

Background:

• Same-day discharge (SDD) after hospitalization is increasingly recognized as safe and efficient.
• Selecting appropriate patients for SDD, especially after procedures like angiography, presents a significant clinical challenge due to numerous influencing factors.

Purpose of the Study:

• To develop and evaluate machine learning (ML) models for predicting same-day discharge (SDD) eligibility in patients undergoing angiography.
• To identify the most effective ML algorithms for accurate SDD prediction in this patient cohort.

Main Methods:

• Utilized data from 3227 patients undergoing angiography, including demographics, comorbidities, and procedure details.
• Trained and tested several ML algorithms: logistic regression (LR), K-Nearest Neighbor, Naive Bayes, Multilayer Perceptron, and Support Vector Machine (SVM).
• Evaluated algorithm performance using metrics like accuracy, precision, recall, F1-score, learning time, and inference time.

Main Results:

• Logistic Regression (LR) and Support Vector Machine (SVM) demonstrated the highest performance, achieving F1-scores of 0.800 and 0.806, respectively.
• LR exhibited superior performance in terms of both learning and inference time compared to SVM.
• The study successfully identified predictive models for same-day discharge after angiography.

Conclusions:

• Machine learning models, specifically LR and SVM, are effective tools for predicting same-day discharge eligibility after angiography.
• LR offers a favorable balance of predictive accuracy and computational efficiency for SDD prediction.
• These findings support the potential for ML-driven decision support to optimize patient selection for SDD, enhancing healthcare resource utilization.