Hemodynamics, fluids & transfusion
Development and validation of a dynamic early warning system with time-varying machine learning models for predicting hemodynamic instability in critical care: a multicohort study
Authors' conclusion
TvHEWS addresses the challenge of early detection of hemodynamic instability by integrating multiple clinical parameters and offering continuous, dynamic risk assessments.
Study context
- Population
- Intensive care unit (ICU) patients
- Outcome
- Prediction of hemodynamic instability by Time-varying Hemodynamic Early Warning Score (TvHEWS) (hourly updating time-varying ML model)
Methodological detailsmachine learning, retrospective development and internal validation, prospective validation, external validation, multicohort
- Machine learning
- Retrospective development and internal validation
- Prospective validation
- External validation
- Multicohort
Performance
As reported in the abstract
Predictive performance of TvHEWS
| Test or model | AUROC | Precision | Recall | Brier score |
|---|---|---|---|---|
| VGHTPE 2010 | 0.93 | 0.94 | 0.77 | 0.082 |
| VGHTPE 2022 | 0.92 | 0.74 | 0.76 | 0.085 |
| MIMIC IV | 0.82 | 0.72 | 0.36 | 0.116 |
AUROC
VGHTPE 20100.93
VGHTPE 20220.92
MIMIC IV0.82
Precision
VGHTPE 20100.94
VGHTPE 20220.74
MIMIC IV0.72
Recall
VGHTPE 20100.77
VGHTPE 20220.76
MIMIC IV0.36
Brier score
VGHTPE 20100.082
VGHTPE 20220.085
MIMIC IV0.116
Further findings
Lead time of prediction before intervention · up to
Lead time of prediction before intervention
- Low false alarm rates
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Development and validation of a dynamic early warning system with time-varying machine learning models for predicting hemodynamic instability in critical care: a multicohort study. Crit Care. Published 23 July 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →