Hemodynamics, fluids & transfusion
Development of a machine learning model to predict intraoperative transfusion and guide type and screen ordering
Authors' conclusion
Machine learning approaches are feasible in predicting intraoperative transfusion from preoperative variables and may improve preoperative type and screen ordering practices when incorporated into the electronic health record.
Study context
- Population
- Adults ≥18 years undergoing non-emergency surgery, 2019–2022, single centrecases 2019 to 2021
- Participants
- 100,813 adults
- Comparison
- Seven machine learning algorithms predicting intraoperative Red Blood Cell (RBC) transfusion
- Outcome
- Prediction of intraoperative transfusion, external temporal validation
Methodological detailsmachine learning, retrospective, single-centre, external temporal validation
- Machine learning
- Retrospective
- Single-centre
- External temporal validation
Performance
As reported in the abstract
| Test or model | Overall accuracy | Sensitivity |
|---|---|---|
| LightGBM | 76.1% (95% CI, 75.6-76.5) | 91.2% (95% CI, 89.8-92.5) |
LightGBM
Overall accuracy76.1% (95% CI, 75.6-76.5)
Sensitivity91.2% (95% CI, 89.8-92.5)
Further findings
Intraoperative transfusion incidence, 2019-2021
- Intraoperative transfusion
- 5488 (5.4%)
Predicted type and screen to transfusion ratio with LightGBM-based ordering
- Ratio would improve from 8.4 to 5.1
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Development of a machine learning model to predict intraoperative transfusion and guide type and screen ordering. J Clin Anesth. Published 27 September 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →