Validation of an Electronic Health Record-Based Machine Learning Model Compared to Clinical Risk Scores for Gastrointestinal Bleeding (Gastro Journal)

The EHR-based machine learning model not only outperforms traditional risk scores but also identifies a significantly higher proportion of very-low-risk GIB patients eligible for discharge from emergency departments. This advancement could streamline patient management and improve resource utilization in healthcare settings.

Guidelines recommend using risk stratification scores to identify very-low-risk patients with GIB for discharge from emergency departments. Traditional scores like the Glasgow-Blatchford Score (GBS) and Oakland Score require manual data entry and might not utilize all available patient information. This study presents a machine learning model that integrates with the EHR for real-time risk assessment, potentially outperforming existing scores.