Hemodynamics, fluids & transfusion
Mean arterial pressure is all you need in a machine learning model for mean arterial pressure prediction
Main finding · ROC AUC, complete training data, internal datasetResult — go to the resultMaximum difference of ROC AUC 0.006Authors' conclusion
Predicting MAP below 65 mmHg within 5, 10, 15 and 20 min for patients with and without a MAP above 65 mmHg is possible and requires only MAP as input for machine learning models.
Study context
- Population
- Adults with invasively measured BP, Kepler University Hospital, Linz, Austria, plus MIMIC-III-WDB and VitalDBanalysed
- Participants
- 2,346 / 4,741 / 3,357
- Exposure
- Machine learning models using previous IBP/mean arterial pressure (MAP) values vs last measured MAP as trivial estimator (baseline)
- Internal datasetn = 2,346
- MIMIC-III-WDBn = 4,741
- VitalDBn = 3,357
- Outcome
- Model performance in predicting MAP <65 mmHg within a given time frame (prediction horizon 5 to 20 min)
- Follow-up
- 5, 10, 15 and 20 min
Methodological detailsmonocentre, exploratory, machine learning
- Monocentre
- Exploratory
- Machine learning
Main finding
As reported in the abstract
ROC AUC, complete training data, internal dataset
Maximum difference of ROC AUC
0.006
By group
Internal dataset
0.963, 0.946, 0.934 and 0.923
Observational data: associations, not effects.
Further findings
| Outcome | Internal dataset | Maximum difference of ROC AUC |
|---|---|---|
| ROC AUC, stable patients with current MAP >65 mmHg (secondary analysis) · 5, 10, 15 and 20 minMachine learning model vs Last measured MAP (baseline) · stable patients | 0.856, 0.837, 0.821 and 0.804 | 0.051 |
| ROC AUC 5 min | 0.856 | |
| ROC AUC 10 min | 0.837 | |
| ROC AUC 15 min | 0.821 | |
| ROC AUC 20 min | 0.804 |
Original publication
Mean arterial pressure is all you need in a machine learning model for mean arterial pressure prediction. Eur J Anaesthesiol. Published 7 July 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →