Artificial intelligence

Team seeks solution for ‘fighting the ventilator’

It’s a common situation in intensive care: patients who are “fighting the ventilator”. A team from Erasmus MC and the startup Deep Breath is working with AI-driven software to find a solution for this harmful phenomenon.

Share
0 likes
Reading time 2 min
beademing

The lung function of many patients in the intensive care unit (ICU) is supported by controlled mechanical ventilation because patients are too ill or weakened to breathe on their own. As patients gradually awaken, a situation often arises that ICU doctors and nurses refer to as “fighting the ventilator”. Patients begin to breathe alongside the ventilator, but their breathing often does not sync with the settings of the ventilator. For instance, the patient may want to inhale earlier, later, or deeper than the machine allows. This is very uncomfortable for the patient, increases the risk of lung damage, disrupts sleep, and potentially increases the risk of death.

Real-time detection

A team from the ICU of Erasmus MC and the startup Deep Breath is developing a way to improve the dyssynchrony between the patient and the ventilator. ‘Currently, as a quick fix, the patient is often put back to sleep so that the machine can fully control the breathing again. This is not a good solution, as we know that ICU patients who are kept artificially asleep longer have poorer outcomes’, explains research leader and technical physician Annemijn Jonkman from Erasmus MC.

The team from Deep Breath develops software that recognizes when the breathing of the awakening patient is not in sync with the ventilator. They use data from the ventilator for this purpose. ‘These devices generate a lot of data that are currently not being utilized. We use this data to detect in real-time a disruption in the synchrony between the patient and the ventilator using artificial intelligence’, explains Anton Balakirev from Deep Breath.

Adjust the settings

The software will be improved and tested over the next four years in a study at the ICU of Erasmus MC. ‘We will investigate whether the software helps ICU doctors and nurses recognize disruptions in synchrony and the effect of this dyssynchrony on the patient,” says Jonkman.

The team’s ultimate goal is for the software to alert ICU staff when a disruption in synchrony occurs and to immediately suggest how better to adjust ventilator settings to the individual patient’s needs. ‘We hope to make it easier to recognize dyssynchrony at the bedside and improve the quality of care in the ICU’, concludes Jonkman.

This research is supported by the Top Sector Life Sciences & Health (Health~Holland).

Also read