In Toulouse, artificial intelligence to anticipate blood pressure drops in anesthetized patients
Régis Fuzier, anesthetist at the Oncopole, is developing with researchers from the IRT Saint-Exupéry an assistance model to prevent arterial hypotension, a common perioperative risk.
You work on the prevention of low blood pressure, what is it?
This drop in blood pressure is a common symptom of anesthesia because the injection of an anesthetic product makes the patient observe the blood pressure, it is normal and expected. On the other hand, if the average arterial pressure drops below 65 mm of mercury for more than 5 minutes, this can have effects on different organs. This risk is increased in the context of major surgery, in patients with high blood pressure or subject to cardiovascular problems. For prevention, anesthesiologists monitor and monitor the pressure in a more or less invasive way.
How could artificial intelligence (AI) prevent this risk?
Currently we see low blood pressure when it occurs, without knowing if this drop will last over time or not. Obviously, we do not take the risk of waiting, and take close measures, in order to analyze all the patient data and provide solutions. All the data (treatments administered, blood pressure, temperature, respiratory parameters, etc.) are permanently recorded, but so far we have not interpreted them. Thanks to AI, by analyzing all these indicators in real time, we want to set up a predictive algorithm for hypotension. It could make it possible to define predictive scores of arterial hypotension for each patient; without the need for specific and required monitoring equipment.
What does your collaboration with the IRT Saint-Exupéry consist of?
Using computerized and anonymized data from several hundred patients, computer scientists at the IRT (Institute for Technological Research) will work on the development of an algorithm that defines a profile of patients at risk of preventing arterial hypotension. For us anesthesiologists, permanently on the alert, this signal sent in real time will save precious minutes of reactivity. The care will be improved, the risk reduced.