© 2024 WVIK
Listen at 90.3 FM and 98.3 FM in the Quad Cities, 95.9 FM in Dubuque, or on the WVIK app!
Play Live Radio
Next Up:
0:00 0:00
Available On Air Stations

Research in Peoria explores how artificial intelligence could reduce emergency department wait times

WCBU News/Jeff Smudde

Could artificial intelligence be part of the equation to solving the problem of emergency department overcrowding? That's what OSF HealthCare and the University of Illinois are working together to find out.

Dr. William Bond is an emergency department physician at OSF Saint Francis Medical Center in Peoria. He's also directs simulation research at Jump Simulation. He's working with University of Illinois Urbana-Champaign visiting assistant professor Dr. Hyojung Kang and a team of researchers to develop models aimed at reducing ED wait times.

Overcrowding is a long-time problem for emergency departments. The post-pandemic national nursing shortage has only exacerbated the issues.

"As you're boarding patients in the ER, you physically have less beds in which to see patients because you have admitted patients sitting in the emergency department waiting to go upstairs," Bond said, noting the problem is particularly acute in tertiary care centers like OSF Saint Francis that are also receiving patients transferred in from other hospitals for advanced care.

Bond said his team's work with machine learning is better thought of as a model than an artificial intelligence like ChatGPT.

"In your mind, that model might be as simple as a Lego diagram, where you sort of laid out Legos and the orientation, and you had little patients and little Lego beds, and you could move them around," he said.

The computer-built model is called a discrete event simulation. It allows Bond's team to change variables and test different scenarios, like adding or subtracting nurses and bed space.

The technology also allows the computer to predict which types of patients might be arriving in the emergency department at any given time, and what resources are needed to treat them.

"That involves looking at past historical data, applying these machine learning prediction techniques, and suggesting that well, instead of just saying our average number of patients on a Tuesday, saying, 'Well, this is a Tuesday in August,' and adding as many different variables in that we think are relevant, to predict how many patients we'll have at the door," he said.

In a way, Bond said it's similar to weather forecasting. He said the emergency department can apply lessons from the simulation to better prepare when those real-world situations emerge.

The research is funded through a $100,000 JUMP Arches grant. Bond said there are some small-scale experiments happening now, but he's looking towards bigger results as work proceeds over the next year.

"If you have a terribly painful condition, you want to get pain relief quickly. You'd like to be diagnosed quickly. That's your goal as a patient or a family member of a patient," Bond said. "And we share that goal, right? We acknowledge that suffering in the waiting room is every bit as important as suffering once you've come back from the waiting room. And we'd like to address people's needs as soon as we can."

Tim is the News Director at WCBU Peoria Public Radio.