Applications of AI in Healthcare: Augmented Intelligence and Artificial Intelligence in Healthcare

AMA Update covers a wide range of medical topics that impact the lives of physicians, residents, medical students, and patients. Hear from medical experts, from practitioners and health system leaders to scientists and public health officials, on COVID-19, medical education, advocacy issues, burnout, vaccines and more.

speaker

Vincent Liu, MD, MS, Chief Data Officer, The Permanente Medical Group

Unger: Hello. Welcome to AMA Update videos and podcasts. These days, it feels like everyone is thinking about how to leverage AI in their practice. Today we're talking about what it takes to make your AI initiative a reality. Today's guest is Dr. Vincent Liu, chief data officer of Kaiser Permanente's Permanente Medical Group, from Santa Clara, California. I'm Todd Unger, his AMA Chief Experience Officer, and I live in Chicago. Dr. Liu, I'm so glad you're back.

Dr. Liu: thank you. Always fun, Todd.

Unger: Now, last time we spoke, you talked about some of the AI programs that you have at Kaiser Permanente. First, for some background, let's quickly summarize how we're using AI today (at least the most important ways).

Dr. Liu: Yes, we are doing a lot of research in AI and machine learning-based clinical decision support. That means stitching together complex data from EHRs and other places to make information more accessible to clinicians, improve the way they think about risk stratification, and help them reach patients and the public.

We are also considering large-scale language models. A lot of innovation and excitement around using things like ambient AI. These are devices that listen to clinician-patient conversations and summarize their notes, reducing the clinical burden of documentation that many physicians find overwhelming.

Unger: About a year ago, I had the opportunity to visit The Permanente Medical Group. And one of the things that really struck me was how involved and deeply involved physicians were in technology issues. This of course makes a lot of sense. How does it feel to be able to contribute as a leader in that field?

Dr. Liu: amazing. It can be said that the greatest asset of this medical group is its doctors. They are on the front lines and are also leaders. And once we build that structure, we have a huge opportunity to understand how technology can enhance our daily work to achieve four goals: patient outcomes, population management, and health care. Opportunities are provided. Cost and provider health are all more important than ever. And I think it's really important that physicians be deeply involved in decision-making about this technology and how it should be implemented.

Unger: It's so great to hear that it's starting from there because, you know, not all technology initiatives in healthcare started that way. Let's talk about AI. There's so much potential. Obviously still new. And many practices are trying to figure out where to start. I'm curious, when you think about the AI efforts that you've outlined and other efforts that are currently underway, how do you prioritize what you want to pursue?

Dr. Liu: I think that's a great question. Again, a lot of that has to do with her quadruple goal. It's just that we have the opportunity to leverage best-in-class technology to improve one of our goals. I'll use ambient AI as an example since it's fresh in my mind. We recently published some of our experiences from The Permanente Medical Group. In doing so, we're rolling this out to all of our physicians in Northern California to help them have conversations with their patients. And the overwhelmingly positive feedback we've heard in terms of its ability to reduce the administrative burden of documentation and improve engagement and conversation between patients and providers.

So this is just a step towards overcoming some of the administrative documentation and administrative burden, allowing doctors to do what they want to do, such as communicate and interact with patients. You will be able to do this.

One more thing, Todd, I think you talked a little bit about our Aim High program the last time you joined us, where we're funding five external health systems to use AI in randomized trials. was introduced and conducted a rigorous real-world evaluation. . This is another way to consider what's ready in terms of technology adoption and use cases. When there is a compelling need, when technology improves the way patients are cared for or improves patient outcomes, evidence is needed. So that we can really support that kind of work.

Unger: So I'm going to add “shovel ready” to my dictionary. I love the way you say that.

One, and I think this is consistent with the leaders I've talked to at Permanente Medical Group, is that it's not just about technology. It's about people and it's about operations. For example, when you prioritize and think about the programs you're trying to execute, what are the teams you're assembling to make those things happen?

Dr. Liu: Yes, I like to use the term “augmented intelligence.” Because augmented intelligence puts people, not algorithms, at the center: patients, clinicians, and communities. So for us, the core competencies for this team are in three areas: We call this a tripod stool. The first is clinical integration. How this will actually be deployed into your workflow, understanding all aspects of it and ensuring this is seamless, preferably efficient, and ultimately effective and secure, or how best to do so. It's all about dedication.

The second is technology realization. That has become a recent issue. It's about getting the data in the right place, running the algorithms, feeding the data back, and presenting the data in a way that clinicians can access and understand. That's really important. It's our information technology partners, our data and analytics partners, our enterprise architecture and infrastructure, all of which enable valuable data streams to flow and power our operations.

And ultimately, and I think this is the new part, is the data scientist. That's all, what's the method? How can we rigorously test and evaluate the performance metrics of these methods? And ultimately, how can this technology be integrated with clinical staff and data scientists to ultimately improve patient health and healthcare professionals? How can we evaluate this to know that we are producing something of value for the health of our customers?

Unger: That's very interesting. Physicians aside, that sounds a lot like people on my team, who are combining data science into their work to ensure that what they learn actually feeds into their future strategy, and to test that. It is the ability to make the environment a part of the work. Of course, new technology and such new processes require a lot of training. How do you train your team to use these new tools? How do you deploy them?

Dr. Liu: Yeah, I mean, I think it's important to note that we've been working on improving quality and performance for a long time. So the fact that we're introducing other types of population health, like cancer screening, hypertension management, blood, we've been rolling those out to clinicians and teams for a long time. Therefore, it is built on the foundation of how to effectively educate, maintain, and achieve performance goals in terms of mindset for large numbers of patients who want to improve their health. I am.

Second, there has been a long history of technology adoption, even before AI. So technology took over before AI in terms of EHRs and other communication channels. These channels are critical to improving coordination between these complex teams. So it's actually built on that foundation.

I think the difference between AI and previous AI is the uncertainty of how the method will work and, more importantly, where the shortcomings of AI are. Neural networks and deep he What is learning or what is this kind of regression, I think this is a lot of part of education. And where should you apply it with confidence? So where should you pay attention to what this output shows?

Currently, clinicians and humans are still in every loop and making decisions about what these data are telling us. We think this will give us a little more insight than we had before. But again, we want them to be able to care for the patient in front of them, and to provide specialized care to this patient, and to draw from different streams of data and experience and knowledge. It teaches you to integrate other functions.

Unger: absolutely. In some ways, the result will be that they won't even know it exists. But as you said earlier, it increases the ability of doctors to get back to doing what they want to do. I'm curious, you've already done so much, what's next? What will be the next big move in AI?

Dr. Liu: I mean, I'm really excited about large-scale language models, so I think that's going to be a big focus. How do we identify the best use cases, the ones that will best help us achieve our 4x goal? I think it's a top priority for the system, for all medical groups. There are a lot of other traditional things that we've been working on for a while, like clinical decision support and computer vision. I think we're going to see these benefits really start to integrate in the kind of seamless way that you described.

Therefore, it's important to strike a balance between making sure your pipeline is robust to enable clinical decision support, computer vision applications, and also thinking seriously about how best to use things like large-scale language models. And that's probably where most of my time is spent in terms of how to operationalize these technologies.

Unger: Well, Dr. Liu, I can't wait to hear more about what you're up to. We will contact you later. In the meantime, thank you for joining us today. We will contact you immediately.

It's all in support of the AMA's efforts to make technology an asset, not a burden, for physicians. He can become an AMA member at ama-assn.org/join.

That concludes today's episode. And we'll be back soon with another her AMA update. Be sure to subscribe for new episodes and find all videos and podcasts at ama-assn.org/podcasts. Thank you for joining us today. Please be careful.

Disclaimer: The views expressed in this video are those of the participants and do not necessarily reflect the views or policies of the AMA.

Source link