Hospital Re-Admission Prediction Using Named Entity Recognition and Explainable Machine Learning.

Area of Science:

• Clinical informatics
• Artificial Intelligence in Healthcare
• Natural Language Processing

Background:

• Early hospital readmission poses significant financial and clinical challenges.
• Predictive models for readmission risk are crucial for patient management and resource allocation.
• Discharge notes contain valuable information for readmission prediction.

Purpose of the Study:

• To develop and evaluate a novel approach for predicting unplanned hospital readmissions.
• To leverage clinical discharge notes for enhanced readmission risk prediction.
• To improve the accuracy and interpretability of readmission prediction models.

Main Methods:

• Utilized the MIMIC-III database for clinical discharge notes.
• Employed a pretrained Named Entity Recognition (BioMedical-NER) model based on Bidirectional Encoder Representations from Transformers (BERT).
• Trained machine learning models, including Gradient Boosting, on extracted features for readmission prediction.

Main Results:

• Achieved an average precision of 88.4% using the Gradient Boosting algorithm.
• Demonstrated superior performance compared to existing state-of-the-art methods on clinical reports.
• Applied explainable Artificial Intelligence (AI) techniques for enhanced model interpretability.

Conclusions:

• The proposed approach effectively predicts unplanned hospital readmissions using discharge notes.
• The integration of BioMedical-NER and machine learning offers a powerful tool for clinical decision support.
• Explainable AI enhances the understanding and trustworthiness of predictive models in healthcare.