The EHR vendor is building AI that can identify care gaps and proactively reach out to patients to schedule appointments, and foresee hospital capacity bottlenecks.
“The next evolution of how we can leverage generative AI is to help streamline workflows, make the software easier, simpler, faster and more helpful when it comes to actually going through patient care and even helping patients,” Epic Director of AI Sean McGunigal told Becker’s during the HIMSS conference in Las Vegas. “You can think about it like a generative AI workflow, where instead of us codifying the steps in that workflow, we instead give an agent a purpose.”
He gave an example of an AI agent that can close care gaps: “That agent finds that a patient is behind on their vaccines or immunizations, and that agent can send a notification to the patient via MyChart that they might want to get a vaccination appointment scheduled, because it’s gone through their chart and identified that care gap.”
AI agents might also help with patient flow, proactively identifying an emergency department bottleneck, for instance, he said.
Epic released its first AI agent — the Slicer Dicer Sidekick — as part of its November 2024 EHR update.
Jackie Gerhart, MD, vice president of clinical informatics at Epic, explained how it works: “When you’re doing analytics, one of the self-service tools you can use is called Slicer Dicer. It lets you take a population, either your hospital’s population or only your patients, and then you search a research question. Before I would have to choose the population, then I would have to choose the measure, and I would have to choose how to slice it. I now can use a conversational agent to do the steps for me.”
Clinicians and researchers can go back and forth with the agent until they drill down to the precise data they need, she said.
Epic is developing its own AI agents rather than partnering with other companies, like it has with AI scribes from Abridge and Microsoft. “To build a really strong, powerful agent, you have to have a good understanding of the data model, a good understanding of the workflow and a good understanding of the actions of the system you want that agent to take,” Mr. McGunigal said. “And with our integrated system, we’re pretty positioned to do that.”
Epic also partners with “early adopter” health systems that help codevelop the AI models. “These organizations are excited to turn on the AI functionality and are willing to volunteer their time to help validate and ultimately improve these tools before they become generally available,” Mr. McGunigal said.