What is the added value of the Hypotension Prediction Index algorithm?
The Hypotension Prediction Index (HPI) is an established machine learning algorithm designed for predicting intraoperative hypotension. Recent literature highlights selection bias in the development process and therefore a significant reliance on mean arterial pressure (MAP) within the HPI algorithm. In this retrospective observational study involving 100 patients undergoing moderate- to high-risk elective non-cardiac surgery, a strong correlation between HPI and MAP signals (-0.90) with no time delay was found. Both HPI and concurrent MAP had identical predictive performance (AUC of 0.90) in predicting hypotension. This suggests that a MAP based alarm could serve as a simpler alternative to the HPI algorithm for predicting intraoperative hypotension.
Personal information
Organisation | University of Twente |
Position | PhD Candidate |
BIOGRAPHY
I am an enthusiastic and driven technical physician; currently working as lecturer and researcher at the University of Twente in the Cardiovascular and Respiratory Physiology (CRPH) group. With my PhD research I try to support clinical decision making on the ICU and OR for patients with hemodynamic instability, using advanced signal analysis and computational physiological models.
Programme
Time | Session | Location | |
---|---|---|---|
14.45 - 16.00 | TechMed Talent Talks: frontrunners of tomorrow | Full Focus | Floor 1 |