Making mechanical ventilators smarter with data-driven algorithms
The reality is more poignant
“IC staff work incredibly hard. The corona pandemic has certainly given everyone insight into intensive care and artificial respiration. For a long time, a shortage of ventilators for corona patients posed an imminent threat worldwide. The reality is more heartbreaking than the images broadcast on the news suggest. in intensive care, patients are on the verge of death. And even if as an engineer you are not at the patient’s bedside, you certainly see the social relevance of technological developments. My research is often related to the corona pandemic, but apart from that, I hope to be able to contribute to reducing the work pressure in the intensive care unit and improving patient care.
Basically, a breathing machine is nothing more than a small fan that blows pressurized air into the lungs, Reinders explains. “The doctor establishes a particular pressure profile. This means setting a higher pressure for an inspiration, then reducing the pressure for an expiration. In order to optimize the treatment of the patient, it is important that this pressure profile is respected as precisely as possible, minimizing any tilting.
To achieve this, Reinders first improved the control technology used with the breathing apparatus, with the ultimate goal of ensuring that the patient receives the pressure originally set by the physician. Certainly a challenge given the diversity of patients – from premature babies to adults – each with their own respiratory needs. Then, he started working on self-learning algorithms, to create room for adapting the device to the patient.
The technique used by Reinders to do this is called repetitive control. This learns from equipment errors recorded in previous breaths, then corrects them over a certain number of breaths. In simulations with artificial lungs in the laboratory, it was revealed that using this technique the pressure profile can be tracked more accurately than with current mechanical ventilation. Reinders has also developed a few self-learning algorithms that can help choose the optimal treatment for intensive care patients who are able to breathe with partial support. “The breathing of patients under sedation is very regular. But if the patients breathe themselves, the interaction with the device is all the more important. Thus, the new algorithms can estimate the patient’s own breathing and determine whether the patient and the ventilator are breathing in synchrony. »
Reinders and his colleagues hope these self-learning algorithms will bring patients one step closer to autonomous breathing. “In short, this means intubating the patient, turning on the mechanical ventilation, which itself determines the best treatment, and waiting several days until it signals that the patient can be taken off the breathing machine.”
Critical care doctors and nurses are initially not very fond of self-guided systems, saying they would rather hold the reins, so Reinders observed during his days observing intensive care run by Gommers.
The corona pandemic has changed this view and specialists have come to understand that equipment with more autonomy can really reduce their care workload and can allow the patient to receive even better care. In a follow-up study, researchers from TU/e and Demcon are now working with Erasmus MC to develop Reinders’ self-learning algorithms. “We are making a concerted effort to help patients and nurses breathe easier.”