In this episode

Keith Fraidenburg spent much of 2024 traveling across the United States to understand how health systems are approaching AI. He dives deep into what they’re most excited and concerned about, and where we’re likely to see healthcare AI going in practice in 2025. He called out the following themes in how health systems are approaching AI:

• Health systems are often focusing on areas where ROI can be demonstrated, like simplifying administrative tasks.
• Starting small with quick wins allows health systems to try out new AI technology without too much change management.
• Each health system is approaching AI differently, depending on their challenges and their openness to being an early adopter.

Fraidenburg discusses these trends and what to watch as AI implementations become ever more commonplace in this episode.

“This new world of AI, it’s moving so fast, and it means something different for so many different organizations…I expect the industry to change rapidly over the next one to two years.”

-Keith Fraidenburg

 

Key takeaways

Fraidenburg surveyed health systems on their attitudes toward AI in a 2024 “roadshow” over 9 months, which culminated in published AI Principles from the College of Healthcare Information Management Executives (CHIME). He discussed their feedback and his observations – on what’s being done today, barriers to using AI for health systems, and his advice based on this feedback to health systems interested in AI. Here’s what he said:

Don’t let AI hype give you “fear of missing out” on new solutions. Do what’s right for your organization.

When asked about what health systems are doing with AI, Fraidenburg said: “If you’ve seen one hospital or health system, you’ve seen exactly one. And that that’s because every organization is so different. You have to look at the the culture of the organization, their appetite for taking on some some risk, capital and resources.”

He pointed out that some organizations have the culture and resources to be early adopters of new technology, and others have far different concerns. “Speak to any CIO at a rural or critical access hospital about GenAI, and they’re like, ‘We’re just trying to keep the doors open here.'”

“Every organization is probably exactly where it needs to be based on their own organization, their culture, politics, funding, et cetera, in terms of what they’re able to do today with AI,” Fraidenburg said.

He encouraged healthcare organizations that, “If you’re blazing a path, great. We need trailblazers. If you haven’t even started yet, that’s okay too.”

If you’re not sure where to start with AI, there are resources you can leverage.

Fraidenburg noted that not every organization has the bandwidth or resources to explore AI, but for those who would like to get started, there are resources you can leverage to define your plan.

• CHIME’s AI Principles: 
”If you’re just getting started, it’s a high-level sort of playbook,” said Fraidenburg. The principles were created from the insights Fraidenburg and the CHIME team gleaned from discussions with health systems, and highlight a variety of topics including governance and cybersecurity.

• Use cases from other organizations:
”There are a lot of published use cases out there now for healthcare organizations, very specific to GenAI tools, which tend to be free, if not low cost. These use cases will help you demonstrate ROI very quickly through pilot programs.”

“If you’ve got questions about where to find some of these resources, feel free to reach out to me, or the CHIME organization,” Fraidenburg said.

Consider the potential for operational and administrative use cases, not just clinical ones. 

Fraidenburg expressed excitement about the future of “adoption and application of AI beyond administrative efficiencies…improving health outcomes directly, identifying patients at risk before they get sick.” He continued, “Improving our ability to diagnose patients and get them into more customized programs…that’s promising potential, right on the other side of this.” But he also acknowledged enormous opportunities for operational and administrative improvements using AI.

“Most of the applications [of AI today] are focused on GenAI. And those specific applications tend to focus on achieving administrative or supply chain efficiency, driving revenue cycle efficiencies. It’s maybe not rocket ship building…and a lot of people think innovation is building a rocket ship. But listen, we know that there’s a lot of administrative and operational and supply chain waste in healthcare. There’s tremendous opportunity to wring out waste and save money…organizations are driving administrative and operational efficiencies, improving revenue cycle.”

Fraidenburg specifically calls out that critical access hospitals might not have the resources or organizational push to consider AI, but could benefit significantly from this technology to help extend their resources.

“If there’s a part of our ecosystem that could really benefit from AI tools, it’s rural healthcare, and they simply don’t have the capital and resources to take advantage of it,” Fraidenburg said. “So if there’s an opportunity for us as an industry to come together and do some good, it’s supporting rural healthcare with GenAI, because I think it can help them.”