UNC Health CIO Brent Lamm told Becker's that the health system is looking to reduce physicians' pajama time — time spent on answering patient messages within its My UNC Chart — with Epic and Microsoft's generative artificial intelligence tool that automatically drafts message responses.
On May 23, Chapel Hill, N.C.-based UNC Health joined UC San Diego Health; Madison, Wis.-based UW Health; and Palo Alto, Calif.-based Stanford Health Care in piloting generative AI technology, built by Microsoft and Epic, to help physicians respond to patients' questions in online portals.
"I think this has the power to help our physicians, and eventually other care team members, to spend more time with the patient and less time with the computer," Mr. Lamm said. "And hopefully it will give them more time at home with their families at night without having to do documentation."
The health system will initially begin piloting the technology with five to 10 physicians and will assess how the tool can draft an initial response back to a patient's inquiry or message.
"The physician in this initial set of use cases would not need to spend cognitive time having to craft an initial draft of patient messages," Mr. Lamm said. "The technology will craft the initial draft and physicians can review it and determine if they want to edit, review and then hit send."
Administration burdens such as EHR documentation cost U.S. healthcare $1 trillion annually. But with Epic and Microsoft's generative AI technology, hospitals and health systems are optimistic that this will help significantly reduce administrative responsibilities for their providers.
"We want to go fast in rolling this out," Mr. Lamm said. "So we're working very closely with Epic to identify phase two of this pilot and beyond so we can make this broadly available."
Mr. Lamm said UNC Health will be conducting a "specialty specific" approach with the technology to assess what areas it is most effective in.
"Maybe it works really well for primary care, but then as we continue to pilot we could find out that it doesn't work as well for cardiology, that's the kind of things we will look at," Mr. Lamm said. "We will provide that feedback to Epic and will hopefully be able to co-create with them."
Mr. Lamm said the integration, though in its early stages, is exciting and has the potential to be a "game changer" for care teams.
"If we can quickly take the learnings from this initial work and begin to apply it with Epic, we could eventually apply it to the other clinicians like nurses and dramatically reduce the amount of time our care teams are having to spend in front of the computer," Mr. Lamm said. "We're talking about real time back for them to spend with our patients, and ideally, I think this could result in substantially higher patient satisfaction and patient experience."