Generative artificial intelligence is being touted as technology that can significantly improve healthcare delivery, but it doesn't come without difficulties; clinical workflows must either integrate the new technology or organizations must change their workflows entirely to incorporate it.
UC San Diego Health is working with generative artificial intelligence, built by Epic and Microsoft, to help physicians respond to patients' questions in online portals.
The current pilot with the technology is going well, but Christopher Longhurst, MD, chief medical officer and chief digital officer at UC San Diego Health, told Becker's there are two difficulties that can arise when working with the new tech in healthcare.
"We have a lot of clinical workflows today, so you either have to change those workflows, or you have to integrate it [generative AI] into the existing workflow," Dr. Longhurst said.
Dr. Longhurst said the current generative AI technology from Epic and Microsoft is integrated into UC San Diego Health's existing workflow, which allows the technology to work.
"If you had to take a piece of a message, and then copy it into GPT to get the response, physicians wouldn't want to do that," Dr. Longhurst said. "So it's literally just the integration in the workflow that makes it work, that's one really important piece."
The second piece is making sure hospitals and health systems don't run too quickly toward a "shiny object," according to Dr. Longhurst.
"We've got to be thoughtful and careful, even though we've got this really cool sounding tool," Dr. Longhurst said. "We've got doctors here at UC San Diego wanting to quickly turn this on, but we have to ensure this technology doesn't have downsides."
Dr. Longhurst said the team at UC San Diego Health is working closely to examine the AI for potential for bias and worsening of health inequities, as well as collecting data to see if the tool actually makes clinicians' jobs more efficient and to see if patients find ChatGPT's responses helpful.