← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Eric Poon, Chief Health Information Officer, Duke University Health System
00:15:21
Loading summary
Transcript10 lines
- [00:00]
A
Welcome, everyone. My name is Giles Bruce, assistant editor with Becker's Healthcare. I'm pleased to be joined today by Dr. Eric Poon, Chief Health Information Officer of Duke University Health System. We are here live in Chicago at Becker's 10th annual Health IT, Digital Health and RCM Conference. Thanks, Eric, for joining us for the podcast today.
- [00:22]
B
Well, Giles, thanks for having me. Good to see you again.
- [00:25]
A
Yeah, good to see you too. And I know the last time we talked, you were doing a lot with AI as Duke has been doing, been doing for years. Your health system is definitely a leader in this space. So tell me some of your big successes in that area with that technology these days.
- [00:40]
B
Yes, I think since we last talked, we have been really leaning into how AI could not only be a tool to help us become more efficient, it's really a tool for us to think about how to transform healthcare. So we have really made a lot of progress in different areas. So I think last time we talked back in the spring, we were in the throes of implementing ambient technology, and that has been a great hit. I think we are getting such great feedback from our clinicians that this is a technology that will allow us to record the conversations with patients and generate a note. And I might have mentioned that since using that technology myself, we are now seeing clinicians getting done with all the paperwork on time, going home early so that they can enjoy time with the family. So I think this has really been a great success story. We've also beginning to get data showing that even without us asking, clinicians are becoming more productive. So that is certainly something that is making our leadership really happy. But we're not stopping there. With ambient technology, I think AI can do so much more to reduce the friction in the process of care. So we are leaning into AI to help us summarize charts so that as we go walk into a patient room or start seeing a patient, we have the context in mind. Sometimes patients come visit, particularly our specialists, with a stack of paper from the outside and trying to get a sense of what's happening so that clinicians can make a sound. But rapid judgment is a challenge. So we've actually done some really exciting internal innovation work to summarize charts coming in from the outside and for example, for oncologists to get a sense of what treatments have had the patient gone through in the past, maybe outside of Duke, so that we can decide what options might be the best for the patient. So that's been really exciting. And then we also trying out some of the chart summarization tools coming straight out of EPIC for our inpatient hospitalists. And that's really beginning to save them time. So I think that's in more ways than one, shaving minutes off of every encounter, maybe not in front of the patient, but in the preparation phase. So that's been definitely a winner. One thing that we are also leaning into is how we reimagine the model of care for our, our hospitalized patients. And one of the things that we've been doing recently is to, as we have renovated a few units in an older part of the hospital, we've actually decided to make them into what we call the innovation units. So these are three patient care units in the northern part of our hospital where we are putting in new technologies, try them out and prove value before we scale that beyond. So one of the things that we're doing is that we're working with the company Autosight to put in cameras and microphones so that we have more eyes and ears on what's happening with the patient. And of course, all that data is being fed into AI so we're able to detect if the patients might have the arms leaning out of the bed rails, who might be at risk of falls. We can make sure that the patients are turned adequately so that we can prevent pressure ulcers. And of course, the videos and cameras will allow us to pipe in our virtual nurses that might be deployed centrally to help reassure patients or provide education for patients. I think the goal of these innovation units really is to use technology, AI or otherwise, to help us think about what would be the new way of delivering health. Because healthcare is very labor intensive. Healthcare right now is very geographically constrained. And so if we can provide a more flexible staffing model, we can really do a lot of good for patients. So that's really exciting, those innovation units. And we're trying other things, including using computer vision, to help us, help us with supply chain management. I don't know whether you've walked into a supply area on the patient care unit. A lot of times I tell people, people, it's like a candy store that is not very well organized. It's hard to figure out where the supplies are, where things are running out. So we're hoping that by deploying computer vision, we can be more proactive in making sure that all our doctors and nurses have the right supplies to take care of the patients and we don't overstock unnecessarily. So that's something that we are also leaning into. And then on the administrative side, I know we are under a lot of pressures financially. So I think appropriate to this particular conference, we are leaning into AI to think about how can we be more effective in our revenue cycle operations. So we are doing quite a few things. We are now working with companies to help us with pre authorization of procedures so that our pre auth staff don't have to troll through the chart for a long time to justify why a patient might need a procedure. So that AI is now really helping us gain efficiencies there and paving the way for a smoother patient experience if they were to get surgeries. At Duke, we are using AI to help us with coding so that we are appropriately reimbursed for the care that we deliver and help our physicians document the right details in the chart so that we get credit not just for the care that we deliver, that we also get credit for the quality of care that we deliver. So lots of exciting things there. So those are things that we're actively doing and thinking about what we can do around the corner. We actually actively exploring using AI to help us with clinical trials management. As you may or may not know, clinical trials, there's not always a central repository. And actually we've had some team members who've had some personal experience with cancer in the family. They have discovered the hard way how hard it is to find the right trials for their loved ones. So we are actually leaning into that to explore how we can use AI to provide not just the appropriate trials that is suitable for that particular type of cancer, that it is also going to be respectful of the patient's wishes as to what objectives they are looking for in participating in trials. Is it about contributing to science or is it about providing relief or controlling the cancer? And we're also thinking about using AI to help us abstract data from the chart, both for research and for quality registries. So that's in some ways a sampling of some of the things that we are looking to do with AI.
- [08:06]
A
Excellent. Yeah, a lot of interesting projects there and yeah, interesting to learn about your innovation units. I'm actually moderating a panel discussion later today about smart hospitals and smart room. So I know a lot of those technologies will be, will be brought up there. You know, as far as some of the in room, you know, cameras and AI, etc. There's definitely a growing, growing space these days. Yeah. And in talking about, you know, how you make decisions here, you know, we both know there are a lot of AI, you know, tech solutions in health care. You know, a lot of the companies are even here at the conference in our exhibit hall. So, you know, how do you narrow it down and decide on, you know, what the best solution is going to be that's really going to, you know, drive value and outcomes at Duke?
- [08:54]
B
Yeah, I agree with you. The number of solutions that leverage AI and other technologies is really exploding. And in some ways, for folks like us, it's going to a conference like this is like a kid walking into a candy store. And I think an approach that we have found to be really helpful at Duke is to really think about what are the problems we're trying to solve and then creating the appropriate foci in terms of governance and decision making. So, for example, we are very interested in reimagining the care model in the inpatient setting. And therefore we decided to set up the innovation units to try out new things. And we have set up a group that's really overseeing what we put out. Put in those innovation units. Obviously, we cannot put every technology in. We need to sequence it in the appropriate way so that not only would the technology work well for our clinicians and patients, that we can actually measure the outcomes. In some ways, we are dividing the world into categories of problems we're trying to solve and creating the right group of folks, multidisciplinary folks, to think about what technologies to try. And how do we make the case for scaling the technology beyond the innovation units. I think revenue cycle is also an important focus. That's why we are now leaning into many of the technologies, some of which are actually at this conference, to think about how can we begin the journey of automating the process of pre authorization and coding, understanding that it's still going to be labor intensive for the time being. But this is something that takes literally an army of folks. So how can we help everybody practice at the top of the license and take some of the drudgery out of work and then improve the patient's experience? It's a win win that we are creating. And then, of course, clinician burnout. That goes without saying. This is why we're creating foci around ambient with for our providers, which certainly has been very successful. We're also looking at the same technology for nursing.
- [11:20]
A
Oh, great. Yeah, yeah. I've been hearing more and more about how ambient listening is expanding into the nursing and inpatient space these days. So that'll be exciting to see how that evolves. So let's end here with a fun question because as a health tech reporter, you know, I love to look to the future as I'm sure you do as a Chief Health Information Officer. So, you know, what will, what will healthcare look like in five to 10 years? I know that's, you know, a long ways away and with the way technology is developing, but what are your predictions there?
- [11:52]
B
Yes, well, it's going to be really interesting to see how the world evolves because the pace of change is only accelerated and continues to accelerate. I do think that ultimately, overall, the challenges we face in healthcare today, which are quite numerous, is going to drive the pace of innovation. So I think that over the next few years, we'll all be thinking about how healthcare will be different. My overall prediction is that we are going to use technology and AI to remove a lot of the friction in healthcare so that care will be available at the most convenient location for the patient. Many of our listeners, I'm sure, have had experience talking, having a clinical question or an issue where they might have to call into a nursing or physician answering service and trying and being put on hold and having the care team trying to figure out what to do offline. I do think that I can see a world where a lot of the interactions could be handled by AI so that patients can access care that they need and still have a timely response from the clinicians who will then be able to, after the patient has interacted with the chatbot, be able to have a summary in front of them and be able to very quickly help the patients figure out what to do next. I do think that the more care will be delivered in the homes and we will continue to see the trend of less care being delivered in the acute setting and then moving into the home setting. And I think AI will provide the monitoring and feedback for the care team. I do think that more and more the knowledge that's inside our heads as clinicians is going to be available to the masses. I can see a time when a lot of the questions from patients can really be answered by this technology. So that clinicians job will really be focusing on the tough situations that well defined pathways and algorithms cannot address. But I think the technology can take care of a lot of things that are well defined that have been distilled from the best of medical science so that clinicians can take care of the personal aspects of medicine and help patients navigate the unknown and the uncertainty.
- [14:46]
A
Great. Yeah. And I hope, you know, we're all hoping that all those things come true because, you know, if they do, they will definitely make, you know, the experience better for patients and the care team alike. So, yeah, definitely will be interesting to see, see how all this evolves, but interesting stuff here. So, yeah, I want to thank Eric for joining us for the podcast today and also for being a keynote speaker here at the conference. We really appreciate your participation. And yeah, thanks again. We look forward to talking to you in the future.
- [15:16]
B
Thank you, Giles.