Last week, I attended “Tech on Tap: AI in Healthcare” - a panel session hosted by the KC Tech Council in Kansas City. It was a great opportunity to hear from several Kansas City-area leaders in the healthcare space on how their organizations, including Oracle, Children’s Mercy Kansas City, and University of Kansas Health System, are approaching and implementing AI.

I left the nearly 90-minute event with a couple of unexpected takeaways after listening to Carl Chesser, Principal Engineer at Oracle; Bill Saltmarsh, VP and Chief Data Officer at Children’s Mercy Kansas City; and Dr. Denton Shanks, DO, MPH, Medical Director of Innovation Digital Health and Virtual Care, University of Kansas Health System.

The first takeaway for me was the conversation’s quick honing in on AI’s role in improving efficiencies - things like gathering and organizing patient notes in order to free up time, specifically. I took the panel’s collective nodding heads when the topic came up — and its willingness to stay on the topic for much of the session — as a sign that this is a common pain point for healthcare professionals. 

And while no fully detailed examples were given, it was mentioned that AI implementation has been hampered a bit in healthcare organizations because of liability concerns. To that, Saltmarsh referenced a conversation he had with his organization’s general counsel in recent months.

“So right now, if you are using artificial intelligence to make diagnostic decisions, or intervention decisions as a clinician, and it's wrong, there’s a higher margin of error that you're willing to accept, right?”, Saltmarsh said. “So, you don't want to be wrong, and so you don't necessarily rely on the technology. 

“What happens with the technology is so much better than the human in that context? We're just a liability shift for the provider. It's a really interesting conversation.”

Shanks followed with his thoughts on liability.

“There's liability that I use this generative AI tool, and it produced something wrong, or there's liability that I didn’t and I missed something that it could have caught. It’s a really difficult situation,” Shanks said. “There could be a syndrome of symptoms that this patient was experiencing, and I can't quite put it together. If I would have thrown those into Chat GPT, it would have said, ‘Here's the differential. Why don’t you look for these?’ 

“If I didn't use [Chat GPT], then maybe I missed it. But if I do use it, and maybe it hallucinated something … so yeah, it gives me a kind of tricky data point.”

It all was, and is, really fascinating to consider. However, I noticed what I felt was a rather glaring omission from the implementation conversation, or at all, and that was no real mention of education. That’s incredibly important because, to me, in order to fully leverage the benefits of any tool, not just AI, those using it must be sufficiently trained to use it.

So then, my thoughts go to things like wondering what happens if healthcare simulation programs and tools continue to incorporate AI tools into their education offerings? Couldn’t or wouldn’t that be an incredible value proposition for simulation programs? To be able to assist system-wide AI implementation efforts?

Maybe it’s because I see and hear about the incredible advances many companies working with SSH through the International Meeting on Simulation in Healthcare and/or at SimOps, but I know AI is already an increasingly important piece to many of the simulation tools in the market. 

And, knowing SSH’s vision for the next five years, which includes heavy focus on growth and advocacy, it all felt as if the opportunity for the healthcare simulation industry is timed exactly right and *right there*. Pretty exciting!

I found myself thinking along those lines as Shanks offered his view on the future of the industry.

“Right now, really, the message is that AI won’t replace doctors, but doctors that use AI will replace those that don’t,” Shanks said. “Or else, AI won’t replace workplace health systems, but health systems that integrate with AI, that build centers of excellence and do AI research, that are actively engaged in AI in clinical care will replace those that don't because of all of the efficiency gains. 

“There really are, now, symptom checkers does that lead you through algorithms, and it's not really AI. But as soon as you apply AI to that, there's a lot that it can do. But it's still, still by far, is at a place of a draft and then verify. So yeah, how we interact with it absolutely will continue to evolve rapidly.”

I mentioned earlier I’ve seen and heard about some of the AI tools coming into the marketplace. I’d love to hear more. Please feel free to let me know about what you and your organization are doing! Email me: ckitchen (at) ssih (dot) org!

Until next time,

Curtis Kitchen
SSH Director of Marketing