Improving Healthcare Outcomes With a Hybrid Approach w/ Carta Healthcare Co-Founder, Matt Hollingsworth | HealthBiz Brief

Matt Hollingsworth (Co-Founder, Carta Healthcare) discusses how his company is dedicated to harnessing the combined strength of people, technology, and data to improve healthcare outcomes

🎙️⚕️ABOUT CARETALK
CareTalk is a weekly podcast that provides an incisive, no B.S. view of the US healthcare industry. Join co-hosts John Driscoll (President U.S. Healthcare and EVP, Walgreens Boots Alliance) and David Williams (President, Health Business Group) as they debate the latest in US healthcare news, business and policy.

🎙️⚕️ABOUT MATT HOLLINGSWORTH
Matt, the current CEO of Carta Healthcare, leads initiatives aimed at enhancing patient care through data insights derived from provider IT systems. Holding an MBA from Stanford, Matt's academic background also encompasses high-energy physics, with research conducted at CERN as part of the team credited with discovering the Higgs boson.

 He co-founded Global Dressage Analytics, secured and managed projects for the Department of Defense, provided technical guidance for a telecom analytics startup (Deepfield), and undertook product management roles for diverse Internet of Things applications at Samsung.

GET IN TOUCH
Become a CareTalk sponsor
Guest appearance requests
Visit us on the web
Subscribe to the CareTalk Newsletter
Shop official CareTalk merch

FOLLOW CARETALK
Spotify
Apple Podcasts
Google Podcasts
Follow us on LinkedIn

#healthcare #pathology #artificialintelligence #ai #healthcaretechnology #healthcarebusiness #healthcarepodcast 

Transcript

0:00

Welcome to Health Biz Briefs. I'm David Williams, President of Health Business Group here with CARTA Healthcare co -founder and CEO, Matt Hollingsworth. CARTA combines people, technology, and data to improve healthcare outcomes. So just how much of a difference can AI and data really make? Well, welcome, Matt. Thanks for having me. You know, I want to ask you, first of all, what makes your abstraction services better than the competition? Great question. So first for the folks not familiar with data abstraction, which is perfectly fair, also, it's a strange, strange terms, very not abstract. It is a very hard working job. Basically, the general theme behind that is that it is skilled. data entry, filling out the forms that are required to do by and large quality submissions and sort of the way to think about it. So how do we monitor quality? We collect data. How do we collect data? Data abstractors is sort of the way to think about it. The reason that it's not just called form entry is because it takes a lot of energy and skill, particularly clinical skill. So these are extremely long forms. They often take more than an hour to complete. So it's not your average sort of DMV form. And it's asking really detailed clinical questions that the general population couldn't have a prayer to answer. And because it takes a long time and it requires very high skill, it can be very expensive. And so there's a whole market out there of providers that can fill these forms out on behalf of the hospital so that they don't have to tax their own nurses to do it. And that's sort of what data abstraction services are all about. And that's what we do as well. The difference is we figured out a way to do a hybrid human AI team to deliver this better, faster, cheaper, and sort of more liquid, which I'll define that in just a second, compared to the other options. So there are many groups out there that have manual labor to do this. Because we have a different ability to do this, we can compete and be differentiated amongst that crowd of folks that are using just people to do it. And in particular, the things that we focus on are reducing costs. So, The way to think about it is how much labor does it take to collect one form? We're typically between 30 and 60 % cheaper than the other alternatives. So it can be a very substantial cost savings on that front. Time, so turnaround time, meaning the time between when, like say a surgery was performed and the form was completed. We can do that in days and then typically it's more like months or quarters behind, which can make the data a lot more useful, which I'll get into in a little bit. And also quality, which is a... The term of art for that is something called inter -rater reliability, which is if you get two nurses filling out the same form, how often, like how much does it agree? And a really great number is 95%. We target greater than 98 % in it. And the last one that I mentioned is data liquidity. So what that means to me is, okay, we filled out this form. Now we have a lot of great information about a patient. How... quickly can we take that information and get it into the hands of somebody in a form that is useful to them. So form that is useful for a doctor might be a report, for a data scientist might be a raw data dump, these sort of things. Because we're using software in the loop for this, we have a lot of great solutions for getting the data where it needs to go really quickly in ways that sort of manual labor would not be able to deliver on. So those are the sort of four bullet points, cost, turnaround time, quality, and data liquidity. That sounds good. You know, in consulting, there's always three things. So you added a fourth one. So, you know, and when we started off talking about how abstraction is not very abstract, I thought you can say it's more concrete. You started saying it's more concrete, but you didn't use that word. And then I said, it's actually liquid. So it's good to go from the abstract to the concrete to the liquid all in one answer to the question. So I respect that now. All right. So. AI, so we've been hearing about AI and people have had an opportunity to use it with chat GPT understand what it is. And I think a lot of people can say, well, gee, that could do my job a lot faster. And in fact, some of the layoffs we've seen in the tech industry have been specifically citing, you know, technology improvements, productivity improvements from AI. But when I hear you talk, you're talking about bringing people in to be involved with AI. So is this like a counter, you know, intuitive approach or what's going on? Yeah, so there are many reasons for this. First off, I would say just as a truism that AI does not remove subject matter experts from the loop. It will make them more productive, but not remove them. And trying to remove them is almost always a mistake because the key here, like for us, just very practically, if you just throw AI at filling these forms out, you're not going to hit a 98 % in or out. There's no way. There's no model out there that will do that. And I don't think there will be within the next decade or so. But the problem is you can, like for some, So the way to think about these forms, there are hundreds of fields and each field is sort of its own ballgame. Like one of them will ask, was the patient hypertensive upon admission? And AI will be a 98 % or greater iterator liability or not for that field. And then the next one will be, is patient diabetic. Same story. It'll be 98 % or greater or not. The challenge is that it's different at every site because every site has different documentation practices and you can't really know which ones perform until you do it. So. No matter how powerful the AI is, until we have generalized AI that can score greater than 98 % on everything, which we're nowhere close to, even academically nowhere close to it, we're going to have to validate it somehow. And so you need people that are involved in that. So that's sort of the goal here is that we inject the subject matter expertise to be able to validate when things are working and when they're not. When they are working, then we don't need the subject matter expert anymore. So maybe it works for figuring out hypertension and doesn't for diabetes. Great. They don't have to deal with. hypertension anymore now they just need to deal with diabetes. So just sort of the way that we do it is our subject matter experts will fill out the things that the AI can't do and validate the things that it can. And then that's where the subject matter expert involvement comes into play. Got it. Well, let's talk about the patient and how that relates to the use of AI. Patients are some skeptical, some scared about the use of AI, some, I suppose, excited about it. What sort of a say should patients have in the use of AI in health care? Great question. So I sort of bucket this into two buckets, things that touch the patient directly and things that are about hospital operations and how they do their work. I think it makes no sense. Like you can't, a patient coming in and being like, thou shalt not use a whiteboard to like, no, that doesn't make a lot of sense. Like they shouldn't be able to talk about how the hospital does their own stuff. Hospitals should be held and already are. are held by the Joint Commission and all sorts of entities, the very high standard in terms of what they do behind the scenes. And we ought to trust that framework. And I don't think patients should have any say in that because everyone's going to have a different opinion that's what can work in practice. In terms of what goes into their body or like is directed, like there are things that are sort of options in terms of devices that touch them or things like that. It makes perfect sense for patients to have a say in that. And AI is no exception to it. Just like, People can opt out of having blood products if they want to opt out of AI, then they should. Sure. But why not? If they want to not take drugs, same story. But I would encourage my own personal opinion on this is that we should rely on the clinical trial system for this. And if the clinical trial system said everything's good, like if you really try to cut AI out of your life, then you're not going to be able to have CT scans or MRIs or ultrasound. Like just the AI is everywhere. Even if you don't understand that it's there, false oximeters have it like... If you truly want to opt out of that, you're going to have a real problem with the modern health system. But like, I think people have the right to do that. So, you know, I guess point is, I think you should have the right, but it should be very judiciously exercised that I don't think it's practical to have a like, no AI touching me form just because I mean, first off, docs don't even know where the AI is living in a lot of cases. And just again, we ought to trust the clinical trial system. Great. Well, Matt Hollingsworth. CEO and co -founder of CARTA Healthcare has been my guest today on Health Biz Briefs. If you like what you hear, check out my full podcast on Health Biz with Matt about his life and career. Matt, thanks so much for joining me today. Thanks for having me.