AMA Update covers a range of health care topics affecting the lives of physicians, residents, medical students and patients. From private practice and health system leaders to scientists and public health officials, hear from the experts in medicine on COVID-19, medical education, advocacy issues, burnout, vaccines and more.
Featured topic and speakers
How is AI currently used in health care? How will AI impact health care in the future? Can AI be used to predict cancer risk? What is ambient AI in health care?
Our guest is Jeremy Cauwels, MD, chief medical officer at Sanford Health. AMA Chief Experience Officer Todd Unger hosts.
- Watch AMA Update for health care news for physicians, residents and medical students.
- Not a physician? Join the Patients Action Network.
- Learn more about the AMA Health System Program.
- The AMA is your powerful ally, focused on addressing the issues important to you, so you can focus on what matters most—patients. We will meet this challenge together. Join us.
- Learn more about our AMA advocacy priorities, including:
- Reforming Medicare payment
- Fighting scope creep
- Fixing prior authorization
- Reducing physician burnout
- Making technology work for physicians
Speaker
- Jeremy Cauwels, MD, chief medical officer, Sanford Health
Transcript
Dr. Cauwels: The people that we were looking at were actually five or six times more likely on the high end of the spectrum than our normal risk population, really allows us to say, we believe we can drive people in a better way to making decisions about who needs colon cancer screening and who needs colon cancer screening sooner rather than later.
Unger: Hello and welcome to the AMA Update video and podcast. Today, we're talking about two new ways that Sanford Health is using AI, and the impact that they're having on patients and physicians. Our guest today is Dr. Jeremy Cauwels, chief medical officer at Sanford Health in Sioux Falls, South Dakota. Dr. Cauwels has asked me to call him Jeremy. I'm Todd Unger, AMA's chief experience officer in Chicago. It's great to have you with us today, Jeremy.
Dr. Cauwels: It's great to be here, Todd. Thanks for the opportunity.
Unger: Well, we've been talking a lot lately about colon cancer, including how it's on the rise in younger adults and what we can do about that. One of the ways that you happen to be using AI is to identify patients who are at risk for colon cancer. Let's just start there. Tell us a little bit more about how it works.
Dr. Cauwels: Sure. So as you're probably aware, the upper Midwest, especially North Dakota, has some of the highest rates of colon cancer in the country. And so as the USPFTF moved the recommendation down to age 45, we gained 100,000 new people in our footprint that needed to be appropriately screened for colon cancer. What we also did not gain was a closet full of GI doctors.
And so it was important for us to say there are probably better ways we can go after this. And one of the groups we requested was actually our data and analytics team to say, hey, is there a better way to find all of these people in the normal risk group and sort of, if you will, risk stratify them to say some people in normal risk are on the higher end and some people are on the lower end?
And they were able to identify roughly 85 individual items that you can find in any electronic medical record, but that would take doctors hours to pore through on their own and weigh them out and really get into the fact that we can probably get a score. That while not making decisions for doctors, because that's why we're all here, but it can help us understand a little bit about the risk of the patient sitting in front of you, and something that we can bring to them more real time, and when they're actually face to face with that patient.
Unger: Well, first of all, that's—what a smart use of AI. Because as you point out, you throw 100,000 new people into the pool and it's already hard really, to get something set up to get a colonoscopy, for instance. That's a lot of lead time there. So to be able to take a look at the data and create a model to be more predictive about who is at risk makes a lot of sense. Was there anything in that model you were surprised by?
Dr. Cauwels: Well, what I was surprised by was the magnitude. So if we figure going through 450,000 patients in our database would yield about a 2.5% colon cancer detection rate overall, the fact that we were able to take these risk factors and find our normal risk population and increase that by a factor of five or six, so the people that we were looking at were actually five or six times more likely on the high end of the spectrum than our normal risk population, really allows us to say we believe we can drive people in a better way to making decisions about who needs colon cancer screening and who needs colon cancer screening sooner rather than later.
Unger: Wow, that is really impressive in terms of the impact. Tell us a little bit more about the results that you're starting to see with the model.
Dr. Cauwels: So first and foremost, I should say that we, just like every other system or academic area, want to make sure we're really careful about this. So when I heard about the model the first time, the first thing I asked him to do was—our EDA team, and some of our GI doctors actually took this to Digestive Disease Week. They presented it retrospectively. We came back then six months later and showed prospective data that was able to identify these patients in both of those populations.
And now, we're in the process of saying, we're ready. We think we can pilot this in front of our own clinicians on a regular basis and put this in. So that is the phase of this that we're in right now, is to say, what does this look like in the day-to-day of a primary care doctor's workflow.
Unger: Well, I'm curious. You mentioned something at the very beginning of this conversation, that there's a higher prevalence of colon cancer in the area where you are in North Dakota. You're in a rural health system. GI specialists might be harder to come by. I guess just right there, there are a lot of reasons for why AI makes sense with a health system like yours but tell us more about that fit.
Dr. Cauwels: Certainly. So in one of our markets, in Fargo, North Dakota, for instance, we knew we had a backup of patients for colonoscopy that was approaching 18 months. And we knew that we had to find a better way to appropriately screen people, not to take people in or out of the line, but to make sure that the line was appropriately risk adjusted.
What we discovered was in advancing both AI and advancing stool-based testing, we were able to decrease the total number of people in that line, which gets those higher risk patients into and in front of our GI doctor sooner, as well as making sure that those other resources in a rural health care setting, like primary care or general surgery, who do do some scopes as well, we're also able to help out and lift that patient care load just a little bit faster.
Unger: That makes a lot of sense. So we've talked about the use of AI with colonoscopy and colon cancer testing. But that's not the only AI tool that you've had success with recently. You also completed a pilot of an ambient AI listening tool. Tell us more about what the pilot looked like.
Dr. Cauwels: So we started out with 100 doctors. We sort of carefully curated for those folks that we knew that were already busy, that already had great primary care practices, and people who already had either fairly good patient experience or wanted to improve patient experience. And we put a bunch of those folks throughout our footprint into the first 100 people in the pilot.
And what we were able to show was, if you take those folks and you give them the opportunity and you give the patients the opportunity to not have to worry so much about turning to sideways face the patient while you're typing on a screen and bring them into that node a little bit better, not only did we see a tremendous improvement in our physicians understanding and our physicians joy in work, but we also had a tremendous improvement in patient experience during that same time. So it really did aim towards those things that we really always wanted to do as doctors and patients in health care.
Unger: And it's funny you should mention that, because I just had a recent visit with a specialist, and I had my appointment. And afterwards, I was like, what was different here? And number one, no computer or laptop at all. We just had a discussion; we had the examination and that was it. And I will tell you, it was a great experience. I'm curious to find out more. When you take a look, for instance, at physicians and you ask them about their experience with the tool, what are you finding out?
Dr. Cauwels: What we're finding out is what I would call resoundingly positive numbers. 100% of those docs in that first pilot said they would be disappointed if they couldn't use the tool. 88% of them would recommend it to all of their partners. And we saw 95% of our docs said there is a reduction in the amount of cognitive burden, and 80% of them said they're more likely to continue practicing either longer or busier, or just generally enjoy the practice more than there ever was before the ambient listening.
I had a couple of specific anecdotes of one of our docs who point blank said, "I haven't been home in time to do something between my office and dinner and doing charting at night for years, and I got home the other day at 5 o'clock. My daughter had recently turned 16, and we went car shopping before dinner." So for him, it was a real game changer in just life and that work life balance.
Unger: That's really a moving story and I just love it. And it's really along or on the same theme that I've seen in these discussions, which is finally, we're seeing technology that's really put to use to reduce these administrative burdens and free up physicians to what they love to do, which is take care of people. Obviously, AI is already having a big impact on your organization, and that's probably only going to grow. What are some of the other ways you see Sanford Health using AI this year and in the future?
Dr. Cauwels: One we've already incorporated that, I would say, differs just a little bit from our GI scoring because it was already evidence-based, there was no debate about it, was chronic kidney disease. We were able to put a calculator in that uses AI to look up things that all doctors could look up. It just looks them up faster. And it can give you a score to say you know what? Jeremy's chronic kidney disease might progress faster than the person next to him in the waiting room, so we need to do a better job of referring that person to nephrology.
We've also had additional things like better triage and looking to figure out how we can do triage and results in a more smooth way. As every doctor knows, the bane of their existence right now is their in-basket. So how can we help them by getting the information into that in-basket that needs to be there because they have to talk to their patients, and getting that information out of their in-basket that doesn't need to be there because some other member of the team can handle it? And so all of those are potential options where we continue to look forward and say, how does the technology that we currently have do a better job of moving us forward?
Unger: That's pretty exciting. And so clearly, a lot of opportunities ahead. Jeremy, thank you so much for joining us today, and we can't wait to see what you do next. Making technology work for physicians is a key focus of the AMA, and you can support our efforts by becoming an AMA member at ama-assn.org/joinnow. That wraps up today's episode and we'll be back soon with another AMA Update. Be sure to subscribe for new episodes and find all our videos and podcasts at ama-assn.org/podcasts. Thanks for joining us today. Please take care.
Disclaimer: The viewpoints expressed in this video are those of the participants and/or do not necessarily reflect the views and policies of the AMA.
Subscribe to AMA Update
Get videos with expert opinions from the AMA on the most important health care topics affecting physicians, residents, medical students and patients—delivered to your inbox.
