Welcome to the first of an ongoing series of roundtable discussions among Chartis consulting leaders around the emerging reality of artificial intelligence (AI) in healthcare.
Join Tom Kiesau, Chartis Chief Innovation Officer and Head of Chartis Digital; Jody Cervenak, Chartis Informatics and Technology Practice Leader; Cindy Lee, Chartis Chief Strategy Officer; Julie Massey MD, Chartis Clinical IT Practice Leader; Chirag Bhargava, Chartis Revenue Cycle Transformation Practice Co-Leader; and Jon Freedman, Partner in Chartis Digital, as they discuss AI, what Chartis is seeing in real time, and what they think is coming next.
Tom Kiesau: Thanks, everyone, for being here. The big news at HIMSS was Epic’s integration of GPT-4 into its EHR platform. What was your reaction to the news?
Jon Freedman:
It is exciting news about real, tangible, actionable use cases being set in motion. Significant, but it also wound up overshadowing several other notable AI-related announcements, including: Amazon’s partnership with 3M Health, HCA’s partnership with Augmedix (which has a partnership with Google Cloud), and Nuance’s announcement of integrating GPT-4. (Nuance was acquired by Microsoft in 2022.)
In terms of the Epic-GPT-4 announcement, it’s worth noting that Doximity actually came to the market first with its DocsGPT product. However, Epic, through this integration with the medical record, advances the opportunity significantly—given the implications of applying this technology directly and specifically to patient information and given the inherent advantages for inserting itself directly into provider workflows.
Jody Cervenak:
The announcement is creating dialogue about appropriate applications of generative AI with EHR solutions. Epic will approach this (as they have always approached innovation) in a thoughtful way, being considerate of the positive implications and anticipating and planning for the unintentional consequences. The focus on Epic with generative AI will likely be on suggested text generation, automation of routine tasks, auto-summarization, and translation. There are major benefits of each of these applications.
Chirag Bhargava:
It’s notable that each of the big announcements at HIMSS had one of the big tech companies directly or indirectly involved. These are the companies with the capital, expertise, and computing power necessary for just about anyone—vendor or health system—to get ahead in this space.
Tom: What are some of the specific implications of Epic integrating AI into its EHR?
Cindy Lee:
Due to Epic’s leadership position and market penetration, any collaboration they announce has significant impact on a large proportion of providers. The EHR is often cited as one of the primary reasons providers experience burnout, so making workflows easier for providers to engage while making documentation and other requirements easier will be critical to the user experience and could serve as a key point of competitive differentiation for Epic.
Julie Massey, MD:
I am both excited and concerned. The potential to relieve clinicians of the burden currently experienced is amazing. But that said, even with training, natural language processing (NLP) tools still make mistakes that can be missed by clinicians moving quickly. We need to ensure that the important, relevant details are readily available for accurate clinician-to-clinician communication and to support clinical decision-making. How to do so is still a work in progress—and likely to be so for some time.
Jon:
I agree with Julie. Incorporating AI into the EHR will require clinicians learning how to “trust” AI—when to, and when not to—and how to teach it to be better. Figuring out how to do that along the spectrum of product vetting through product use will be challenging—especially around last-mile implications: How much human interaction (e.g., accuracy verification and refinement) will need to be involved? Will this diminish over time? What metrics will be put in place to show increased or decreased AI capability and clinician confidence?
Chirag:
A real implication of these recent announcements is that while the potential is significant, organizations risk feeling overwhelmed and reluctant to turn these features on. Often the biggest struggle for organizations in leveraging AI is integrating it with the EHR platform and their existing workflows. The requirements for effective change management and training will be enormous.
Tom: You’re all touching on a key point: It’s one thing to introduce an AI-enabled EHR, but quite another thing to have it successfully deployed, adopted, and yielding tangible results. How will providers need to adapt to begin to realize the value of AI?
Jody:
An aspect that leaders of provider organizations will need to think about is the training, change management, and communication around the new technology. As Chirag noted, it will be essential to have a comprehensive change management and education approach that is fully integrated with the roll-out of an AI-enabled EHR. Especially given that this will be a first-generation technology roll-out that requires the clinician to know when to “trust” it and when to take it as a more supportive “suggestion” for consideration.
Julie:
As this type of technology continues to mature, how our providers leverage and adopt tools will be influenced by how the tool helps them get through their day—without additional burden. At the outset, there will likely be additional burden as providers spend time actively engaging with the AI-powered technology to ensure and advance its accuracy. Providers will have to believe that the ultimate benefit is worth the initial investment of their time and focus.
Tom: Terrific point. There was a study released by JAMA just last week that compared AI-chatbot responses to physician responses, and it found evaluators preferred the chatbot’s answers nearly 80% of the time. That’s a staggering result and has huge potential implications for things like provider inbox management. It’s just one study, conducted in isolation, but the potential implications are material.
Cindy:
And for AI to achieve its potential, providers must be part of the quality assurance process and solution development as well.
Jon:
All of this will put pressure on the leaders of provider organizations who will have to parse and assess the ever-evolving vendor landscape, where everyone claims to have AI as a core component of their product offerings. They will need to be able to separate the signal from the noise.
Tom: OK, final question, with all the activity and seemingly daily news related to AI, what should leaders be focusing on now?
Julie:
Leaders will need to learn to tell the difference between what is real and what isn’t. There is so much hype. They need to understand the nuances of this technology and the use cases for each that claims to have harnessed the power of AI. It will also be important to resist the urge to implement the “latest and greatest” without engaging clinical staff who will need to adopt these tools into their everyday workflow.
Chirag:
Leaders will need to invest in education for teams about AI, especially understanding generative AI, which is a whole new ballgame. They will need to create a team that can lead the way in AI—from understanding the technical implications of leveraging AI to the clinical, operational, and practical requirements for doing so.
Cindy:
Leaders should create and support forums to discuss and innovate with AI in small ways at first. The old adage of “go slow to go fast” will apply here to enable higher adoption with both providers and patients.
Jody:
It is essential for organizations to have a plan on how best to adopt it based on their strategic priorities, pain points, and immediate opportunities. It requires organization, focus, and programmatic management. The time to plan for it is now.
