AI is one of the ways Connecticut hospitals hope to ease ER crowding and improve care

The Hour – Sunday, June 22, 2025
By Staff Writer

What kinds of weather conditions send waves of kids to the emergency room with colds and flu? Which approaching holidays tend to lead to higher rates of mental illness in teens? How many nurses should be on duty when school lets out and parents of sick children may rush to seek emergency care?

Predicting trends and offering suggestions to help systems work more efficiently are both features of current artificial intelligence platforms, and Connecticut health experts are seeking new ways to make the technology work to improve care.

“AI offers you the ability to see patterns that may not be recognizable by humans,” said Dr. Christine Finck, chief of pediatric general and thoracic surgery at Connecticut Children’s. “I think AI is definitely a tool that can be leveraged to help us better predict the needs that we’re going to have.”

A team of doctors, nurses and tech staff at Connecticut Children’s developed an AI model that uses weather forecasts and other factors to generate staffing levels for the Hartford hospital’s emergency department. AI was deployed with a computer simulation and mathematical optimization to create a cloud-based operations management platform to ensure that enough doctors, nurses and other staff are on hand to prevent overcrowding and long wait times, Finck said.

“There’s a whole host of things that AI can be used for to help better resource allocations and optimize the care we’re providing to our patients,” Finck said.

The Connecticut Children’s staffing model won third prize at the first Health AI Championship, an event hosted last month by Yale New Haven Health’s Center for Health Care Innovation.

Hospitals and researchers statewide were invited to present ideas for using AI to improve care and help alleviate issues like emergency department boarding, in which admitted patients must wait for beds.

“Part of our focus is on really identifying and prioritizing the unmet needs,” said Walter Lindop, lead consultant for the YNHH innovation center and an organizer of the contest. “Not only did it bring together the different institutions in Connecticut, but it provided a very focused and competitive way for an idea (addressing) ED boarding to surface to the fore and then be funded.”

First prize in the inaugural Health AI Championship went to a group of researchers from Yale University who used a deep machine learning model to determine the best time to procure organs to improve the results of transplants. Out of 54 initial entries from across the state, 12 finalists competed for a $100,000 first prize and “the opportunity to validate their solution in YNHHS data ecosystem,” according to a statement on the results.

Considering the huge amounts of data generated in modern health care through electronic medical records and other sources, AI can be a valuable partner with humans to improve medical treatment, said Dr. Lee Schwamm, Yale New Haven Health chief digital health officer.

“It is not AI replacing humans, this is AI augmenting humans,” Schwamm said. “(It’s) letting AI do what it’s good at, and letting humans then do what they’re good at, which is make decisions, make connections, develop rapport with patients.”

In the future, AI may interact with patients in the waiting room through tablets or other means, prioritizing treatment for the patients most at risk for deterioration in the emergency department by scanning medical records for warning signs amid thousands of data points, once regulatory and privacy concerns are resolved.

Finck at Connectictut Children’s sees a future where you may be scanned by AI-linked technology as you walk into an emergency room, monitored remotely for temperature and heart rate and approached by staff sooner if you show signs of serious illness.

In the nearer future, AI in the form of a “virtual intake coordinator” interface may work through your smartphone or a website to assess your symptoms in context of your past care and health record and recommend next steps in consultation with a doctor.

Some of these innovations are likely to be pioneered in Connecticut, Schwamm said.

“We have a real advantage in Connecticut, in that we’re small, which means we can work together and do some experimenting,” Schwamm said. “I think the Health AI Championship was a great example of health systems working together ‒ not in competition with one another ‒ to try to think about solutions that really improve the health of our patients.”

Access this article at its original source.