The University of Montreal Hospital (CHUM) Emergency Department is transforming as researchers roll out a new triage system powered by artificial intelligence.
Index Sante listed CHUM at 131 percent capacity as of Wednesday afternoon. Doctors hope that AI assistance will help get people in need of care through the waiting room faster.
For example, “When you walk into the ER, you’re 88 years old, you have poor vital signs, you’re short of breath, and the algorithm predicts a 95 percent chance you’ll need hospitalization,” said Dr. Elyse Berger, doctor in the emergency room, “We want to use this knowledge to complete all steps faster so that the patient can see the patient earlier. So we have less wasted time and probably better care for the patient.”
dr Berger is working as a consultant on the project, which the team hopes will be rolled out at the CHUM ER sometime in 2023. After that, they want to expand it to other departments and then to other hospitals.
The AI system uses large amounts of ER data to predict patient needs. If all goes well, they can allocate resources to departments to receive patients before they arrive.
“What we want to do is predict the admissions every day, so the doctors are informed, so the nurses are aware, and we can plan resources differently,” Berger said.
“This is as easy as can be imagined,” said tech expert Carmi Levy. “That’s what AI was born for, and it’ll be interesting to see how that plays out, but it’s really smart given where we are in healthcare today.”
AVOIDING PREREQUISITES
Experts say that as the system rolls out, the AI will need to be regularly monitored and reviewed across the network to iron out any bias in the data, such as when certain groups weren’t treated equally.
“Artificial intelligence is only as smart as the data that feeds it,” Levy said. “So if you already had an infrastructure that was designed, for example, not to deliver services properly to one group over another, then that’s just going to be amplified.”
“It’s really a problem in the whole world of AI, especially in healthcare,” said Berger, who said a variety of healthcare professionals would be involved in reviewing the data as the system is tested.
“When I see as an emergency doctor [a trend] and it doesn’t make sense, like ‘that kind of population is usually allowed, how come the computer didn’t figure that out?’ ‘Oh, that’s because we have a bias in the data’.”
If implemented correctly, AI could help further avoid bias during triage, according to Levy.
“In other words, introducing artificial intelligence into the equation gives us an opportunity to address some of the bias issues that we may have had in healthcare in the past,” he said.
