Prediction of Blood Transfusion Need and Dose in Patients With Upper Gastrointestinal Bleeding: Retrospective Multicenter Prediction Model Study.
Machine-learning model predicted transfusion need in upper gastrointestinal bleeding with near-perfect accuracy
Prediction of Blood Transfusion Need and Dose in Patients With Upper Gastrointestinal Bleeding: Retrospective Multicenter Prediction Model Study.
Transfusion thresholds in upper gastrointestinal bleeding are debated; hemoglobin cutoffs of 70-80 g/L are widely cited yet inconsistently applied.
This study aimed to develop and validate a two-stage, clinically constrained gradient-boosting framework (Medically Constrained Gradient Boosting [MCGB]) that predicts transfusion need and estimates transfusion dose with quantified uncertainty and to implement a prototype recommendation system for clinical use.
We analyzed a retrospective multicenter cohort of 849 adults with endoscopically confirmed upper gastrointestinal bleeding admitted to 3 hospitals in Chongqing, China (January 2019 to August 2025).
model told apart transfusion need from no need almost perfectly
At a reference probability threshold of .50, sensitivity, specificity, and F1-scores were 0.99, 0.87, and 0.85, respectively, providing a representative operating point for comparison.
For dose prediction among transfused patients, MCGB achieved R² of 0.95 and mean absolute error 0.04; 95% prediction-interval coverage was 0.94, indicating accurate point estimates with reliable uncertainty quantification.
The software prototype further demonstrated feasibility of real-time decision support at the bedside.
MCGB provides calibrated, interpretable predictions of transfusion need and individualized dose in upper gastrointestinal bleeding and may support bedside decision-making and blood-bank planning, with a prototype interface demonstrating potential for clinical deployment.
External validation in additional settings is warranted to confirm generalizability.