[00:00:02] Speaker A: Welcome to Health Data Talks where industry experts offer bite sized tips and trends for managing legacy data.
[00:00:12] Speaker B: Thanks for joining us everybody. I'm Eric Johnson, the VP of marketing here at Harmony Healthcare IT and today I am joined by Travis Gregg.
Thanks for being here today, Travis.
[00:00:24] Speaker A: Thank you Eric.
[00:00:27] Speaker B: Really interested in, in, in, in joining our podcast here, Travis, and if you don't mind, could you just provide a little bit of a brief background and introduction of yourself?
[00:00:42] Speaker A: Sure.
So you know, I've been in it my entire career.
Basically I focused on process automation.
So if you think about in healthcare and actually most enterprises, there's just a ton of just manual process people have so many just menial tasks, copying data from one system to another, juggling spreadsheets or clumsy reports.
All these things have to be done to keep the organization moving forward. But it eats up so much time and effort from people and you know, people are, people are creative and talented and they could be doing much more impactful things. And the key to that is just automation like taking a lot of that off their plate. And that's just what I've been doing my entire career, 25 plus years. Wow, that's a long time.
[00:01:40] Speaker B: That's amazing.
Maybe just touching on that a little bit too. Travis, can you talk a bit more about your journey into healthcare applications? Right, and you start with Trinis and maybe you don't mind just talking about that a little bit, how that manifests your passion for automation into melding that into a business.
[00:02:03] Speaker A: Oh wow. So yeah, really there's a big part of that was what wanting to be excellent and work with people that were very good. So over the years you work with people that you really like, you get along with and you're able to achieve great things. And so slowly over time we kind of coalesced all those people together and we all worked on this organization as one and we built a lot of great things primarily around process automation.
Specifically when you think about like large data sets, you know, document based workflows and things of that nature. Down.
[00:02:52] Speaker B: Yeah, excellent.
That's really interesting. Well, maybe we can, we can talk a little bit about a product that you've been working on here for quite some time now. That's really exciting.
And that's the Clearway product.
Could you just start a little bit with an overview on just what clearway is, kind of its background and history and some of the problems that you're trying to solve. And if you kind of tie back in that passion that you have around process automation, maybe kind of talk about some of the origin associated with Clearway?
[00:03:36] Speaker A: Sure. Yeah, definitely. So, you know, we went to a longtime partner of ours and presented, you know, some of our capabilities. And she had a good handle on what we were really, you know, what we excelled at. And we asked her, you know, what, what challenge could we undertake that would really be impactful? And she very quickly said, you should automate clinical registries.
So we went back and started researching and working on proof of concepts and testing. But very quickly we began collaborating closely with our partners and nurse abstractors. And fast forward to two plus years later. We have a really compelling solution to really kind of get into it. You kind of have to describe what is a registry, especially in this context.
So a clinical registry, like a specialty registry, it's a centralized repository of data for specific patient cohort, so patients that received a particular procedure or patients that were treated with a specific primary diagnosis. You know, these different bodies kind of are the, you know, house this data and a hospital that participates in that registry, they submit data for those eligible patients, which, you know, typically involves nurses abstracting, you know, pulling that information out of the chart, filling out all these data elements and then submitting it to the registry. And these data elements, these registries, they can require hundreds of what I would call complex fields. They're not just discrete data elements.
They require searching through all the unstructured text that make up a health record and pulling together several different bits of information to kind of synthesize an answer for a lot of these.
So, you know, why, why would I can just talk about the benefits of participating in a registry? You know, just the fact that, you know, you participate in a registry can increase overall quality.
You know, because basically you're shining a light on your process, your documentation, you're tracking, you know, best practices and standards of care. You also get a lot of really good reporting on important metrics.
You're able to benchmark your organization against other organizations that participate. So there's a lot of, you know, things that you get from that, you know, in addition to just those quality improvements, it can either directly or indirectly help you with the different accreditation efforts. So a lot of great reasons to participate in the registry, but it's a challenge.
It comes with a cost. It's very resource intensive.
These registries have hundreds of fields. The answers are difficult to find, very manual effort. You typically need a seasoned nurse with experience in that relative specialty.
The abstraction is very time consuming.
Since it's not easy to do. It makes it very difficult to cross train your staff to deal with that work and then so dealing with unexpected departures or just the general surges of a workload, it can just be a real challenge.
Additionally, most places we talk to, there's a demand for participating in additional registries and organizations can't handle that with their current staffing. So that's, you know, part of the reason why, you know, Clearway was, was we've been working on this is to kind of help with those challenges.
And really, when it comes down to it, our solution, we automate the entire process of registry participation, from the patient identification and management to the submission to the registry body. And in between, there's a really important part and that is automating the extraction of all the relevant data elements required. And we auto, auto complete all the fields for that registry and, you know, the auto completion, that's easier said than done.
Typically there are some discrete fields that map one to one from an EHR to a registry, but the majority require hunting through and searching, some kind of, you know, translation or several aspects that, that go into a lot of reasoning for some of the answers. So it's not a straightforward exercise.
So one of the things that we. Oh, yeah, go ahead.
[00:08:18] Speaker B: No, sorry, I was just going to ask you, how long does it take, like on average, how long does it take a clinician to abstract in your experience? So when we talk about that level of manual effort, right, to finding those fields, can you characterize just that in a length of time from what you've seen?
[00:08:42] Speaker A: Sure. And there's a lot of variability and it depends on the registry, it depends on the location and how complex the care is.
It's, you know, oftentimes we, we get numbers like this where they may say, well, it takes me about 30 minutes to do one case, but sometimes there are easy cases that are more like 20 or sometimes it's like, yeah, it's usually 60 minutes to do these cases for this registry. Sometimes it could be a little bit less. There's some registers that take even longer to do the abstraction and a lot of times those numbers you get are from seasoned abstractors that have been doing it for a long time and they get really good at it because it's very complex. So if you have someone new, it takes even longer. And there's a lot of hand holding and you know that, that gets into like, it's very, you can get into situations where you have a handful of nurses that can do this one registry really very well. But then if you're growing, it's very hard to add staff. Sometimes you Just have one nurse that is the key person for this. And if, you know, they leave or if they're out sick or something like that, you can really be in a bind because it's very difficult and time consuming to kind of train other people to be able to handle that. And that's one thing that's really great about Clearway is since we put everything on rails, it makes it much easier, much more streamlined, much more consistent, so that, you know, you can have a, you know, your, your normal team, they can cross train and they can understand how to do these other registries. And then it takes a lot of that burden off as far as, from a time perspective, you know, that same team can handle additional registries. And it just gives you a lot of flexibility in kind of how you handle surges of work. And, you know, as you, you grow and add additional registries or, you know, you, you have additional locations that become accredited for certain procedures and things like that.
[00:10:43] Speaker B: That's, that's excellent. Maybe we can talk a little bit technical here. Right. And you know, to your point, an area that was ripe for process automation and an opportunity to take all the experience that, that you've had and lend itself to this particular problem.
Can you talk a bit about the kind of application related to artificial intelligence and natural language processing that, you know, that the team has incorporated to help with this process? So if you, if you don't mind talking about that in general and then talk about how, how you navigated this notion of building that clinician trust in the use of these tools and avoiding kind of the black box that sort of gets associated with AI in general. Right. Which is send it over to the black box and send me a result that I'm supposed to take. I'm supposed to trust that how did the team address and take, you know, use the tools and then address that in clear way?
[00:11:59] Speaker A: No, sure. So I actually have a background in AI that goes back into the 90s, doing a lot of neural network. So done, you know, got some real world experience there. So once we started doing Clearway, we knew that we would need to leverage some natural language processing and different AI techniques. But I have to admit I was very skeptical about it.
Just with my experience, those kinds of AI is fantastic when you're thinking about huge data sets and finding the needles in a haystack. And if you find 80% of those needles, that's a huge win. When you're thinking about an individual use case like this, 80% isn't good enough. You need a much Higher level of, of accuracy and confidence.
So it was actually, you know, our team of architects and engineers, they were much more optimistic and it took some convincing. You know, we used several different techniques and it would get us so far. You know, of course we're, you know, being very skeptical. We're benchmarking things. We tried additional techniques and that would get us further.
So it took us a while to kind of hone in on that. And I think that that coming at it from a very skeptical standpoint really helped us evaluate and really kind of set us up for success down the road with that. As we really honed that, another key thing was instead of focusing on that technology, you know, just almost, you know, sort of a dual path, we were focusing on the experience of the users, the, the nurses, you know, what was going to be important to them as they used the system. You know, how do we instill confidence, put everything at their fingertips, make it. It's a complex problem, but how do you make it as simple as possible?
And then a real key thing is to make it very transparent so the users can see everything that we're doing and kind of, you know, they can review anything that they need to. And that really instills that confidence. As they're kind of using the system, they can get faster at those review tasks that are within Clearway.
[00:14:16] Speaker B: Awesome. Excellent.
Sounds like it's been a fascinating application of some of your, your practical experience. Right. And using science and, and, and focusing in on, you know, all aspects of proving our own selves wrong in terms of can we meet our own thresholds. Right. And while balancing the experience of the user. Right. So they've got something that is, is a practical, practical application to that point, maybe. Let's just skip kind of to maybe some of the results. And can, you know, when you talked about the complexity of. And the use of time. Right. And how long it takes even seasoned users to complete an abstraction, kind of more so on the manual side, can you talk a little bit about what we've seen so far through our studies and early adopters of the products that we've seen and how clearway has been able to affect that in a positive way, both from a time savings, but also accuracy and then maybe even just other benefits that we've seen so far?
[00:15:31] Speaker A: Sure. So I won't bury the lead here.
We did a couple of studies and we're seeing anywhere between 60 and 80% reduction in time per chart for the patient abstraction. So when you're looking at those kinds of savings, that's Huge. That gives you a lot of flexibility to be able to a, you know, leverage your nursing staff in more interesting ways. You know, a lot of times these teams, they're focused on quality, but they're spending the majority of their time on this sort of menial task of doing the abstraction. And it's time consuming, but it's also kind of mentally draining. So if you take, if you give back some of the time and give back some of that mental capacity instead of this menial task, they can think about the trends and where things are going and what can we do to impact this number and can we get our outcomes to be more positive, things like that. That's something that a lot of the astronauts we've talked to, they're excited about that, that ability to be able to do that.
It also just gives you so much more flexibility to bring in new users because it's easier, it's kind of on rails, and we have a lot of capabilities that makes it easy to kind of train those users to use the system and, you know, have oversight and things like that as well. So it really, the, the key is enabling these users that, that, that typically have a passion for quality improvement, being able to give them more time to do that, but then the organization as a whole, give them more flexibility to participate in additional registries, spearhead additional, you know, quality initiatives.
So it's, it's really exciting, you know, the, the feedback we've gotten as we've worked with these nurses to kind of see, you know, you know, these impacts which, you know, are really, really, you know, wonderful and, and we're very excited about it.
[00:17:50] Speaker B: Fantastic. Well, Travis, this has been a great conversation. Appreciate you sharing all this information, fascinating about your, your experience and how you've been able to apply it to this particular problem and even broader process automation. So thanks for joining today to our audience. Thanks for listening. Join us next time for continued conversations and meaningful discussions around the management of health data. So thank you.
[00:18:21] Speaker A: Thanks, Eric.
That's it for this session of Health Data Talks. Check out helpful
[email protected] and follow us in your favorite podcast app to catch future episodes. We'll see you next time.
