March 22: Today on TownHall Jake Lancaster, Chief Medical Information Officer at Baptist Memorial Health Care interviews Nathan Yung, Clinical Informatics Fellow at UCSD. Clinical informaticists help decide what workflows are working and what workflows aren't working. What tools do we have available? How are we using them? How can we potentially use them better? When it comes to remote patient monitoring, there can be a lot of data. How do we use the data decisively? How can you visualize that data within Epic? What are the challenges with bringing in an external artificial intelligence program into a system? How do you integrate it to make it easy to onboard new algorithms as well fine tune it with your own data prior to the implementation?
Today on This Week Health.
One of the biggest challenges that clinical informaticists uniquely have to face is bridging the technology languages with clinicians and aligning some of their incentives and desires to help the patients. I think that is probably one of the hardest things because everyone is trying to help the patient. And we just need to start speaking the same language and start seeing where there are more opportunities for us to team up as opposed to saying things are imperfect and finger-pointing.
Welcome to This Week Health Community. This is TownHall a show hosted by leaders on the front lines with 📍 interviews of people making things happen in healthcare with technology. My name is Bill Russell, the creator of This Week Health, a set of channels designed to amplify great thinking to propel healthcare forward. We want to thank our show sponsors Olive, Rubrik, Trellix, Hillrom, Medigate and F5 in partnership with Sirius Healthcare for investing in our mission to develop the next generation of health leaders. Now onto our show.
Hello, I'm Jake Lancaster, an internal Medicine Physician and the Chief Medical Information Officer for Baptist Memorial Healthcare. And today I have on the program, Nathan Young, a Clinical Informatics fellow and an internal medicine physician. Nathan, welcome to the program.
Thanks for having me.
For the audience, can you tell us just a little bit about how you got into clinical informatics and what you were hoping to get out of the informatics fellowship?
Sure. I got into clinical informatics a little bit from the quality improvement stem point of things. I was very clinical in my training. I, I stayed at my medical school because I thought I would get really great clinical exposures. And what I realized while I was there was there's a lot of broken processes, a lot of things that were just classically Swiss cheese model errors, where we saw opportunities for the EHR to do a little bit better and to do a little bit more for us taking care of complex academic patients who belong at academic health centers.
And I got into informatics because I really wanted to understand the power of the tools that we had available, how we were currently using them and how we could potentially use them better. And so I went into informatics fellowship with the goal of learning some of these approaches, learning how we use some of these tools, both with the EHR and non EHR systems and what those tools.
My goal has been to understand what those tools can do for us. Going forward as, as we do more and more in the ambulatory quote unquote environment outside of the traditional walls of the hospital.
Well, that's great. So every informatics fellowship program is slightly different. Can you tell us a little bit about the structure of your program and what sort of things are you being exposed to?
So I'm a second year fellow at UCSD and our program is honestly really, really flexible. I went into fellowship looking for a place that also had an NLM program training PhDs. And I wanted some of that research exposure to understand how some of the biomedical informatics people were trained and approached some of these problems and worked together with those trainees on some of those joint projects. That was one of the electives that we had took. And then some of the more structured curriculum has been around the different areas and aspects of our main EMR, which is Epic and looking at the structure of how some of these things have been coordinated and governed across population health, across the ambulatory teams and across the inpatient services to give kind of a launch dual exposure of what are the problems that the health system is facing.
And I was lucky enough to join when we had a lot of stuff being implemented around the CA notify app, which was California's notifications for COVID exposures which was a really interesting implementation. And right now we're actually planning for a big PACS vendors switch. And so seeing tha kindt of change management and how some of these things have evolved. It has also been a big part of some of our structured curriculum.
You'll learn a lot. It's certainly from that, that PACS switch. That's not going to be a very small project I imagine.
Definitely not. Definitely not. It's very large teams and trying to make sure that all of these niche cases are established with a good workflow.
And we were chatting a little bit before we started recording. And you were telling me about a project you're doing with connected glucometers. Can you comment a little bit more on that?
Sure. So one of the things that we had noticed was there were certain patients who despite our best efforts of our primary care faculty group still needed a little bit of extra monitoring, a little bit of extra attention. And one of the ways that we wanted to monitor these high need individuals maybe their disease has been stable and then suddenly started progressing, or they've always had kind of a progressive poorly controlled state.
We wanted to stand up a connected glucometer to track these patients and have a higher touch platform. And so the pilot phase has kind of been implemented where it's a platform for endocrinologists to enroll some of their more complex patients and report out to other teams with the population health space on where they having challenges.
And give us insight into when we need to reach out to them to make sure that they're starting to move in the right direction and starting to get their diabetes under control.
Talk to me just a little bit more about how the data is used. I know when we talk about remote patient monitoring and we talk about all the data that could potentially be fed into the EHR to the primary physician, sometimes there's a hesitancy on the primary physician's side, because that's a lot of information and they're going to be nervous that they'll not be able to rack to every data point or maybe miss something. So are the endocrinologist's going to be the ones responsible for reviewing all this information coming in or is there a dedicated team that can only refer the high-risk patients back to them?
So right now, as it exists, only the endocrinologists are referring patients into this platform. We kind of carefully designed the alerting criteria and there's a dedicated team run out of the population health group that tracks and monitors, this kind of subset of the endocrinology diabetes patients.
And when we noticed that one of our triggers, our flags is being set off about one of these patients they receive another patch, another communication to see if there's something else that we can do. If they're running into a certain type of problem to kind of escalate the number of touch points and the number of times we're re-evaluating them to see if we can intervene earlier.
Okay. So you have a little bit of a certain criteria that have to be met before an alert gets sent to the team. How did y'all decide on this threshold?
A lot of it was a consensus amongst the specialists around what are the typical times that we'll be receiving data from the patients, how often we expected them to be uploading information to us. And then when we thought it would be a reasonable time to flag someone as normal or abnormal. And it was based off of tighter trends versus looser trends, depending on the patient. But one of the things that the endocrinologist also wanted was a little bit of individualized control of them as a specialist saying, okay, I want to slightly adjust this flag for this patient and see if there's a little bit of a compliance issue or if there's extra information just from those daily fluctuations.
How does this all flow back into the EHR? Is it two separate modules? And you just have to go into both or duel integrated in any way?
So there was an integration. And that part is, where it gets a little bit more challenging to discuss the details of the way that the workflow happens is the patient is given some sort of connected glucometer or their existing glucometer is given an adapter to allow it to upload more frequently just when they're at home. That information gets stored through the vendor's platform. And the vendor's platform is able to print off some of the reports that actually we set the alerting criteria on the vendor's platform simultaneously that information also flows into Epic so we can pull it into our notes and pull it into our ability to review for some of the patients.
And it also gives the population health team doing the outreach a way to demonstrate in the EHR, how some of this information is displayed and what they had seen that triggered the outreaching in the first place.
So talk to me a little bit about how you visualize that data within Epic and how it gets pulled into the nodes. I assume you're not just having a long list of glucose measurements and date and timestamps. Cause that would, that would be a long list if you're recording this for, for weeks or months. So how do you, I guess, simplify that or summarize it so that you know, the physician taking care of the patient can see at a glance quickly what the status of that patient was over the period of time?
So some of it was built into the actual flow sheets, but we had aggregated and kind of meaned some of this information at different time periods. And so when the population health team goes in and does an extra outreach, they're going to pull in a little bit of this data that they had seen, that the flag went off for around either the standard criteria or the customized criteria by the endocrinologist and said, this is when the patient broke the flag. We gathered extra data on what was going on. And then those things are also tracked and reviewed that way. Some of it is the population health team aggregating and putting context around it and then documenting it for the physicians to review later. Because it's a pretty small population, unfortunately, that, that has this extra attention.
Okay. So yeah, I guess patient enrollment may be a challenge getting patients to agree to do this or why the small population?
Primarily, because we wanted to make sure that we understood all the workflows. And we wanted to understand the different ways that the patients were going to engage with us and the different ways that we wanted to perform outreach. So doing the smaller group that we were enrolling helped keep this flexible and adaptable to see what those endocrinology patients needed. There is ongoing plans to bring it out to the primary care groups. But that's still something that we haven't quite determined.
Okay. And then, so I understand this is probably in the early days of this project, but what are the things you're hoping to track? What sort of outcomes are you looking for to, I guess, prove the worth of the program going forward?
When this was originally proposed was actually a little bit before my time joining UCSD and some of the worth of the program is to really identify the people who would need a little bit of more direct to fine tuning of the troubles that they're having, either utilizing different types of medications. Just getting the data in in the first place, sometimes addressing low health literacy through the extra touch points.
And obviously we're hoping that the patients who are enrolled in this program who received these extra touch points are educated in a way where they can get their diabetes under control, get their medication regimen a little bit more aligned with what works with their lifestyle. And then we can disenroll them in that high touch platform and they can maintain that.
Well, definitely sounds like an exciting program. I expect more places are going to be doing some of this work with remote patient monitoring. But what else can you tell us? What else are you working on as part of the fellowship program?
So one of the other things that I've been working on during the fellowship is the recently announced Center for Health Innovation.
And it's a relatively new center here at UCSD where we are looking at how AI models and different types of machine learning algorithms can be operationalized and observed at scale in the prospective sense. And so the center is really looking for how we can make more of the larger UCSD population available for these models to learn and to predict and how the performance of this model, these models can be studied for future use.
So are you looking to do this for, I guess, external vendor models that you bring into the system? Or are these internally developed models or are you looking to do it for.
Ideally, we'd like to be able to do it for both. Right now we're focused on internal models and there are some vendor developed models that are in flight. And that's a little bit separate from our platform. Our platform really wants to look at some of the really interesting research that's going on on our campus and how we can. We can get larger populations involved in that and work through the data pipelines that need to exist in order for these models to have a good prospective use.
Oh, that makes sense. I mean, there's always a lot of challenges with bringing in an external artificial intelligence program into any system. There's challenges of integration. There's also challenges in that the data that the model was trained on might not be the same as what you have within your walls. So it looks like we're hoping to address kind of making it easy to onboard new algorithms as well as to, to make sure that you can fine tune it with your own data prior to implementation.
Yeah. It's a really trying to do things thoughtfully and carefully, but apply these models at a larger scale and get them potentially up and running for study.
And so that's the ultimate goal is that we want to develop these, these pipelines for models to be pretty much at scale and see how they perform to really grow and learn as model developers, but also as clinical informaticists deciding on what workflows are working, what workflows aren't working. When do we want or need some sort of AI model, because it's easy, I guess it's relatively easy to say that you can develop a model that'll look at an image, but we have to know what to do with it after and how it's going to change things after a kind of play in the implementation of these models.
That's a great point. And thank you so much for your time and for agreeing to come on today. Going forward what do you see as the biggest challenge that you would like to address through clinical informatics?
I think one of the biggest challenges that clinical informaticists uniquely have to face is bridging the technology languages with clinicians and aligning some of their incentives and desires to to help the patients. I think that is probably one of the hardest things because everyone is trying to help the patient. And we just need to start speaking on the same language and start seeing where there are more opportunities for us to team up as opposed to saying things are imperfect and finger-pointing.
Well, thanks again. I definitely learned a lot, certainly doing some interesting work out at UCSD. And wish you luck going forward in the future.
Thank you so much for having me.
I love this show. I love hearing from people on the front lines. I love hearing from these leaders and we want to thank our hosts who continue to support the community by developing this great content. We also want to thank our show sponsors Olive, Rubrik, Trellix, Hillrom, Medigate and F5 in partnership with Sirius Healthcare for investing in our mission to develop the next generation of health leaders. If you want to support the show, let someone know about our shows. They all start with This Week Health and you can find them wherever you listen to podcasts. Keynote, TownHall, 📍 Newsroom and Academy. Check them out today. And thanks for listening. That's all for now. 📍