← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Cheng-Kai Kao, MD, FACP, SFHM, Chief Medical Information Officer at UChicago Medicine
00:10:52
Loading summary
Transcript13 lines
- [00:01]
A
Philips is a health tech leader focused on innovation that improves the health and well being of people. Our healthcare technology and informatics solutions help care teams diagnose, treat and manage more patients with greater precision, speed and confidence. Across the care journey with Philips, clinicians are empowered with streamlined insights in the moments that matter for every patient. Better care for more people. Philips.
- [00:29]
B
This is Grace Lynn Keller with the Beckers Healthcare Podcast and we are live at the 10th annual Health IT, Digital Health and RCM meeting. I am currently joined by Kai Kao who is the Chief Medical Information Officer at University of Chicago. So Kai, thanks for being here. Let's have you start off by sharing a little bit more about yourself and your work in healthcare.
- [00:49]
C
Thank you so much for having me. So yes, my name is Shanghai Kao, I'll just go by Kai. I'm a practicing hospitalist and also Chief Medical Information Officer at University Chicago Medicine. Have been practicing medicine for about 15 years and so have been working a lot on the inpatient care, but also developed the hospital home program at University of Chicago Medicine. And as a cmio, I am involved in lots of EHR optimization, particularly epic and most recently a lot around clinical AI and digital health solutions and how do we intake, select and actually figure out the right tool for the right problem.
- [01:22]
B
Wonderful. Well, thanks for being here and let's start our conversation talking about AI, because nearly half of medical practices reported using AI in some capacity in the last year and that obviously remains a key topic for health IT leaders. So from your perspective, what are the use cases that are making a difference right now and how are you leveraging them in your organization?
- [01:41]
C
Yeah, I think right now it's actually hard to find a company that doesn't have anything that says AI in their distribution. So it's literally everywhere. And so we. Which creates a lot of potentials but also some of the challenges. So even just properly vetting them before we adopt the solution is a key piece. But we definitely think there's a lot of use cases for AI to play a huge role both now and also even more so in the future. And we see a lot of developments of genetic AI and these additional platforms that can be even more powerful. We just need to give it more time, at least for now. We already seeing AI, you know, having some great impact right now in a couple of different places. Well, first, ambient documentation. We know that, you know, a lot of institutions are doing that. We did see a great, you know, patient satisfaction scores and also, you know, physicians really love that, you know, help them, you know, prepare Better notes to save the pajama time and being able to focus on the patient, which is what we sign up for in the medical school. But in the meantime, also see there's additional tools such as, you know, being able to use AI to improve our revenue cycle piece. Especially now financial pressure has been on every health institution. So being able to use AI, for example to draft denial letters, appeal letters, or for example, to figure out, you know, if there are certain areas where we could, you know, potentially get more revenue back is a very important piece. And we have, you know, quite a lot of success in combining some of the tools that EPIC provides, for example, for denial letter appeal drafting and also one of our Hong Kong tools called chat UCM. UCM is UChicago medicine. So it's essentially HIPAA compliant version of ChatGPT that allows people to actually just, you know, use PHI freely on the platform. They don't need to worry about it being leaked to any sort of other third parties or generate any privacy or security concerns, but you can just basically leverage that to, for example, you know, put in the patient story and just say based on insurance policy for this particular patient, you know, does this patient potentially can meet the criteria? If so, can we generate denial letter from that? So that has being quite successful and useful for our revenue cycle specialists. And all in all, just with that platform, we also see a lot of people using that to draft their email communications to, for example, translate patient information into different languages or different sort of versions for the patient or family to take home with. So we think this is very helpful in terms of improving the staff efficiency in terms of augmenting their work. Whatever they do with almost like everyone now have inexperienced little AI intern working besides you. So whatever you do, you can ask the AI intern questions regarding how do I think about this work? And it's going to give you some potentially immature answers. But in the meantime, you could have something that you either wouldn't think about or you would think about, but you're just too busy, you forgot about certain things. So now you have actually a better comprehensive plan in terms of the things that you do. So we see a lot of benefit in that too. Just in three months when we roll out Chatucm, we already see more than 150 users sign up and using that pretty regularly. So we think there's a lot of potential and we anticipate to see even more in the future.
- [04:47]
B
Absolutely. And as virtual care expands from AI enabled tools to remote monitoring to broader digital health platforms, introducing new technology does bring Challenges. So what advice do you have for leaders navigating everything from governance to patient engagement? And can you share an example of how your organization has balanced innovation with operational constraints?
- [05:07]
C
Yeah, great question and especially regarding the AI governance piece, because as we talk about that, almost every product have AI component now, so do we just trust that they will work fine or work as it's intended to? In our experience, when we actually validate some of the AI models that we have trialed before, either from EPIC or from others, there is always a degree of validation we have to do because not all of them perform well. Especially in Chicago medicine, we take care of a very vulnerable population in South South Chicago, potentially with different geography, different profiles or demographics that could potentially have different outcome based on the other AI model that may be stem from a very different population in other places. So what we have done is we establish innovation intake process. If there's a solution with AI components, they'll go through a standardized intake process. We have a few questionnaires regarding like AIs, what model did you use, when did you validate it and what sort of the publication or anything you can show about your validation results. Then we have a team actually review these in the results, you know, staffed by our AI specialists, you know, in our data analytics team to talk about, you know, is this something safe or do we think there's additional validation we need to do? And moving forward, once we implement this system, we want to make sure we'll be able to continue to monitor the performance of these AI models as well. So I would just say in general, like, you know, as we advance these AI tools, a very important piece is always validate and making sure they work as intended to. And overall for virtual care, as you mentioned, I often said the best technology or the best integrated technology is the one is invisible. You don't even see that in a workflow. It just naturally happened because it's being integrated so well, so intuitively that you don't even see that. So I think that's also the same goal that we want when we look at, for example, we have household home program at UChicago Medicine. So when we deploy additional technology tool, can we make sure it creates the least impact on patient adoption in terms of the complexity of the tool and in terms of AI or even technology literacy of our patients, just making sure that this is something that actually helping them instead of adding additional burden to them for them to use the services. Because we all want the best for our patients, regardless where they come from and what their background Is and how.
- [07:30]
B
Are you seeing recent legislation, both state and federal affect healthcare organizations and healthcare IT specifically. And have you adjusted strategies in response?
- [07:40]
C
Yeah, very interesting time. Right. Given the government shutdown and everything. So a lot of speculation about how things would go. But I think in general we feel very confident that I think largely we want to stay on course. We don't anticipate we'll be changing dramatically just with all the hopefully temporary hurdles we are facing right now, we think there's a lot of opportunity to still continue what we are doing, but we just need to be carefully watching what's going on. So for example, Hazard Home has to be put on hold because right now there's literally no waiver extension at this point to continue the service per the CMS waiver requirements. But we remain helpful that this will be resolved hopefully soon in the next few months. So I wouldn't necessarily think that change our overall roadmap in terms of virtual care, in terms of how we use AI and how we adopt AI. But it is obviously always important piece. We constantly consult our legal and compliance about what's the latest legislation both at the federal and state level and where we can help uniting with other health system to push to make that right change. Because we are seeing too often in this area where technology is in advance of the policies and legislation. So things need to catch up and that otherwise is just hindering a lot of developments of a lot of benefit we could have in healthcare.
- [09:12]
B
As we wrap up our conversation, I'd love to know your top piece of advice for healthcare leaders as they prepare for further advancements in technology and rising demands for care.
- [09:21]
C
Yeah, I would just say like I mentioned earlier, I think the best technology is something that people don't even recognize as something I need to learn or adapt. It's just naturally fit into the workflow. So whenever we think about deploying a technology, a new technology, whether it's AI or not, always making sure it's human centered, always making sure it's patient centered. And that often means that you want to engage the frontline clinicians, you want to engage the actual patient to look at how this works for them. For example, when we roll out hospital home programs, we know the patient population we serve, it's very vulnerable. They don't have the best technology literacy to start with. So we actually work with a design school in noise tech to really redevelop actually patient facing material. We asked the patient to sort of play with the Bluetooth enabled devices they will be going home with and we actually create patient facing material based on their feedback. And then we ask them to look at those materials and tell us like is this clear enough to you or do we need to make additional adjustment to make it even clearer to you? So want to make sure all of this become less of a burden for our patients. So I just think that human centeredness is always going to be key, both for clinicians and for our patients.
- [10:37]
B
Wonderful. Well Kai, thanks so much for taking the time to share these thoughts and insights with us on the Backers Healthcare podcast. Again, we are live at the 10th annual Health IT Digital Health and RCM meeting.
- [10:48]
C
Thank you so much for having me.