As the use of artificial intelligence becomes more widespread, the UNC Health system has increasingly developed and utilized AI technology that researchers, professors and medical professionals alike say could change the future of healthcare.
According to UNC Health Chief Information Officer Brent Lamm, UNC Health is an early adopter and pilot site for a variety of AI technologies.
Lamm said many of the tools UNC Hospitals uses primarily focus on increasing efficiency by recording and analyzing patient interaction, better scheduling patient visits, drafting end-of-shift documentation for nurses and handling other administrative tasks.
“We're being aggressive in our exploration of these capabilities,” Lamm said. “We're trying things and piloting and we want to do that, but we're doing it in a very safe and thoughtful way.”
Jessica Zègre-Hemsey, a nurse scientist and UNC School of Nursing professor with current research in emergency cardiac care, has been working with AI to develop strategies to supplement clinical decision-making in emergency cardiac care.
Zègre-Hemsey said working with cardiac arrest cases requires quick decision making about diagnoses and interventions. She said AI technologies help medical staff by compiling multiple data points to create predictions of outcomes for patients when they reach the hospital and even after they are discharged.
“In the large landscape of healthcare, I do think the emergency setting is one of the specific areas that AI machine learning could most benefit, because it is a unique space where we're making really fast decisions with not always a lot of information or time,” Zègre-Hemsey said.
However, Research Director for Digital Health at the Duke-Margolis Institute for Health Policy Christina Silcox said privacy concerns, among other risks and hesitations, remain unclear because AI technology is still so new.
“It's [AI] tested with data, and a lot of our healthcare data is unrepresentative and potentially also includes systemic bias and other types of bias within the data itself,” Silcox said.