AI tool predicts when to remove ventilator from patient

OTTAWA – The Ottawa Hospital is the first hospital in the world to evaluate an innovative medical device that uses artificial intelligence (AI) to predict when critically ill patients are ready to breathe on their own, called the Extubation Advisor. The device was developed by Dr. Andrew Seely (pictured), a critical care physician, thoracic surgeon and scientist at The Ottawa Hospital and founder and CEO of Therapeutic Monitoring Systems (TMS).

Over the last two years of the pandemic, more Canadians than ever have experienced extreme difficulty breathing, requiring mechanical ventilation (intubation) in critical care units across the country. According to the Ontario Science Table, at the peak of wave three, the number of patients on ventilators was more than 180 percent of pre-pandemic historical averages, and the recovery is not easy.

A mechanical ventilator is a machine that helps patients breathe when they cannot breathe on their own due to critical illness, such as COVID-19, or surgery. The machine is connected to a breathing tube that is inserted into the patient’s trachea. The process of intubating (inserting the tube) and extubating (removing the tube) is very complex, and anyone requiring ventilation will require months of recovery and rehabilitation to learn how to swallow, eat, talk and breathe again. Currently, there is no patient monitoring equipment to help physicians decide the best time to remove a patient from a ventilator to improve their outcomes, but Dr. Seely and his team are well on their way to changing that.

Over the last 13 years, Dr. Seely and his colleagues have made major progress in using complex mathematics, artificial intelligence and routinely collected vital sign data to predict when patients are ready to be extubated. Dr. Seely founded TMS in 2007 to further develop the research and bring it closer to helping patients. In 2019, TMS partnered with manufacturer OBS Medical to bring their design and concept to life, and this year, the system was used for three months at the bedside of ventilated patients in The Ottawa Hospital’s ICU, with permission from their families.

“Currently, one in every seven ICU patients experiences extubation failure. Prolonged ventilation harms patients, and early extubation requiring reintubation can be a devastating blow to their recovery,” says Dr. Seely, who is also a professor at the University of Ottawa. “We’ve developed the first medical device to offer extubation decision support, which we believe will help standardize and improve care.”

The device constantly monitors and analyzes a patient’s vital signs, including blood pressure, oxygen levels, breathing rhythms and heart rate during their ventilation. Then, it uses AI to provide doctors with a specific read of when the patient can be safely removed from the ventilator. This is the first time that real-time predictive analytics based on this type of high-frequency data is being used and evaluated at the bedside.

Dr. Seely and his team hope this tool will help improve patient safety and outcomes in the near future. After the successful three-month initial evaluation at The Ottawa Hospital, the metrics are looking promising, and the feedback received from physicians was very positive.

“What Dr. Seely and TMS are doing is revolutionary. It’s the future of medicine,” says Dr. Kwadwo Kyeremanteng, department head of Critical Care at The Ottawa Hospital. “This kind of technology will allow us to make more informed decisions that are precise and patient centered. This will allow for better outcomes and improved resource use. Honestly, this is groundbreaking research.”

Dr. Seely and his team’s next steps include a randomized controlled trial. With each milestone, they are one step closer to transforming care for some of the sickest patients treated at hospitals.

“Although there is still work to be done to evaluate clinical impact, this first-time bedside implementation is a step in the right direction of transforming routine monitoring into improved care,” says Dr. Seely. “We are grateful to our partners, collaborators, consenting patients and research team who are enabling this innovation.”