Related Experiment Video
Updated: Jul 16, 2025
05:37
An R-Based Landscape Validation of a Competing Risk Model
Published on: September 16, 2022
2.1K
National Multi-Institutional Validation of a Surgical Transfusion Risk Prediction Model
Sunny S Lou1, Yaoming Liu2, Mark E Cohen2
1From the Department of Anesthesiology, Washington University School of Medicine, St Louis, MO (Lou, Kannampallil).
Journal of the American College of Surgeons
|September 22, 2023
Summary
The S-PATH tool accurately predicts surgical transfusion risk, demonstrating excellent performance across diverse hospitals. While performance varied, S-PATH shows promise as a generalizable prediction tool for perioperative planning.
Area of Science:
- Medical informatics
- Surgical outcomes research
- Health services research
Background:
- Accurate estimation of surgical transfusion risk is crucial for effective surgical planning and resource stewardship.
- Existing methods for transfusion risk stratification are limited, highlighting a need for validated tools.
- Perioperative planning and resource management benefit from reliable transfusion risk assessment.
Purpose of the Study:
- To evaluate the performance of the Surgical PAtient Transfusion risk (S-PATH) model in predicting surgical transfusion risk.
- To assess the generalizability of the S-PATH model across different hospitals.
- To determine if hospital characteristics influence S-PATH model performance.
Main Methods:
- Utilized the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) dataset from 2019.
- Evaluated S-PATH performance with and without hospital-specific model tuning.
- Employed linear regression to analyze the relationship between hospital characteristics and the area under the receiver operating characteristic (AUROC) curve.
Main Results:
- Analyzed 1,000,927 surgical cases from 414 hospitals.
- Achieved an aggregate AUROC of 0.910 without tuning and 0.925 with tuning.
- Observed variation in AUROC across hospitals (median 0.900), but no significant correlation with studied hospital characteristics.
Conclusions:
- S-PATH demonstrated excellent discriminative performance for surgical transfusion risk.
- Performance variation across hospitals was not explained by the studied hospital-level characteristics.
- S-PATH shows viability as a generalizable tool for surgical transfusion risk prediction.
