Andrew Lockhart, Co-Founder and CEO of Fathom

OR – Andrew Lockhart: Video automatically transcribed by Sonix

OR – Andrew Lockhart: this mp4 video file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.

Saul Marquez:
Hey, everybody! Saul Marquez with the Outcomes Rocket. Welcome back! It’s such a pleasure to have you tune in to our podcast once again. Today, I’ve got the privilege of hosting an awesome leader in healthcare. His name is Andrew Lockhart. He is the co-founder and CEO of Fathom. He’s a leader in autonomous medical coding. Fathom applies cutting-edge deep learning and natural language processing to produce complete coding results for patient encounters with zero human intervention. This is an area of major need, so I’m excited to dive into the work that Andrew and his team at Fathom are up to to make lives easier for all of you. Andrew, thanks for joining us today.

Andrew Lockhart:
Thanks for having me, Saul.

Saul Marquez:
It’s a pleasure. It’s a pleasure. So, look, before we dive into the specifics of what you guys do, talk to us a little bit more about your story. What is it that inspires your work in healthcare?

Andrew Lockhart:
Yeah. So, I actually, interestingly enough, I’m Canadian, and born and raised in the Toronto area, uh, for folks on video; hence the hat. And I got started working with the US healthcare system with a design firm before I went to grad school and began to get exposed to just like the scale and complexity of the system, and it was really attractive to me just how big it was and how many different challenges it’s facing. And then just the nuances between the states, between the payer-provider, the mixed public-private system. It was just exciting and kind of like, you know, a never-ending journey and learning how this all works, that originally got me excited about it.

Saul Marquez:
Yeah, it’s complex. It’s got a lot of problems, a lot of opportunities. So it’s a great place for entrepreneurship for sure. And so, you know, one of the different areas you were doing the design work, you’re now in the coding and reimbursement space. So talk to us about that and then how your company, Fathom, is adding value.

Andrew Lockhart:
I don’t love telling people this, but basically, Fathom was started in about the worst way possible in the sense that my co-founder and I, we recognized that, he had the intuition that the intersection of deep learning, natural language processing was about to be very exciting in terms of capabilities. He and I both wanted to do something in healthcare, and so we then, you know, wanted to look for the place where we could have enormous impact. And interestingly, the clinical side is really difficult to build scalable solutions because care environments are so idiosyncratic, but the administrative side is very normalized, and we start looking at that. And as we dug deeper, you know, you start looking at data that, you know, 15% of the US healthcare spend, which is, you know, enormous, is administrative costs. We start breaking that apart. Coding was the thing that we arrived at that felt very ripe for a really new solution and also a place to have enormous impact because it basically touches everyone, right? You know, it’s work that’s often performed by providers. So if you can take that away from the clinicians, you’re freeing up physician or nurse time, you know, the impact of it goes downstream, the payers end up paying for what happens if things go wrong, then the patient’s impacted. And it just seemed like this thing that was, one, by and large standardized, you know, you’ve got international standards to the world, international standards driving the ICDs and then CMS largely normalizing how these codes are occurring in those practices. And so it just seemed like an enormous opportunity to, you know, have impact at, you know, a really large scale.

Saul Marquez:
Yeah, and it’s certainly, you guys certainly have. So, you know, help us understand and differentiate the specific things that you guys do with regard to coding and, you know, how are you putting deep learning to practice.

Andrew Lockhart:
Yeah, so historically, you know, people have been trying to automate medical coding for decades. And historically, what that’s resulted in is somewhat disappointing results in what largely has looked like workflow assistance. So, tooling from the Optum’s and three M’s of the world, which very helpful, and that it, you know, makes it easy to work through coding cases and helpful suggestion and lookup tools but doesn’t actually do the work. What Fathom is something that, you know, does the work end-to-end. So for our clients, they see somewhere between, you know, 80% to 99% of encounters fully coded, and we seamlessly slide into whatever their existing workflows are. So, we’re not another tool to pick up some degree patterns defined by the absence of tools. It’s just like overnight, you hire someone who’s insanely productive and doing an enormous amount of your work very quickly.

Saul Marquez:
Well, it sounds like you guys slide right into whatever system is being used. Can you call out any particular examples, maybe, that’ll help put the pieces together here on the very productive, what seems to be very productive, platform and service you provide? How have you been able to help people?

Andrew Lockhart:
Yeah, I mean, so the big thing is, what’s maybe not appreciated is so, pre-pandemic, so, by and large, medical coding in the United States is either performed by clinicians so your doctors, or nurses, or trained medical coders. Pre-pandemic, the average medical coder in the United States was 54 years old and typically a woman. What happened during the pandemic is that demographic, so 55 to 64-year-old women had the highest rate of workforce abandonment. Economists are still a little confused by that, but the going theory is a lot of folks left to help with childcare, grandparents specifically. And so what you had is a labor force that was already very constrained that then got cut again. And then, on top of it, they’re seeing just widespread clinician burnout, and, you know, health systems are coming to the realization that having your most expensive people, doctors, do something that, you know, they don’t like and they’re not good at isn’t really that effective either. And so, you don’t have people, you no longer have a labor force you can turn to to do this that’s trained to do it, and this is where the opportunity for automation is stepped in. And we’re seeing a lot of largest, you know, primary care operations in the country turn to Fathom so that they can finally say, hey, you know what? Our doctors don’t have to code anymore. And then further, for the health systems that are just like we’re backed up 35, 50 days on payment because no one’s coding these encounters because we don’t have people to do it. We can turn this on and get it done, and not just more cost-effectively than we had in the past, but also with a higher level of accuracy, which is very helpful in an environment where payers are also, you know, tightening their approaches and being probably more assertive and with denials related to coding errors than they had historically.

Saul Marquez:
That’s fascinating. And so, how does it work? Like, how are you able to automate it?

Andrew Lockhart:
Yeah, I mean, the premise behind deep learning is, you can think of like historically approaches to coding automation were to like manually have people write rules, and that worked pretty well, or it gave you some leverage in ICD-9 world where you had, I think it was 13,000 codes. You moved to ICD-10, you now have 69,000 codes, so you’ve got a massive increase, and those are just the ICD codes, let alone the procedure codes. And so you’ve got to cover an enormous space of potential codes against very nuanced language, and, you know, humans writing rules is no longer sufficient. The simplistic way to think about deep learning, what it is, if you can get enough data and feed it against cloud compute, then basically the machine looks at the input, so clinical documentation and the output final coding, and is able to do its own pattern recognition, write its own rules. And that’s ultimately what has gotten us to where we are with Fathom today. So, you know, we code about 15% of emergency department encounters in the country. And it’s really on the back of, you know, a lot of hard work on the engineering side, but also the fact that we’ve been able to work with a lot of largest revenue cycle operations in the country. So Fathom today, now, is trained across over 450 million coded encounters. So it’s like effectively, you know, the world’s most experienced medical coder by orders of magnitude.

Saul Marquez:
I love that. Thank you for explaining that to us. You know, oftentimes it’s, and you did it in a simple way, the best of what they do, figure out the simplest way to explain it, right? So you definitely simplified it for us. Who can benefit most from your solution? Is it large health systems? Is it small practices? Is it all of them or?

Andrew Lockhart:
Yeah, I mean, it’s, you know, we do have an implementation cost in the near term. The people who benefit from us are large revenue cycle operations, so your big health systems, your large ambulatory practices, your large physician groups. Increasingly, we’re doing work with payers to perform risk adjustment coding so people who would otherwise need to employ a lot of coders or spend a lot with, you know, coding vendors. I think over time, what we’re beginning to see, we’re working in partnerships with various EHRs that do service a lot of smaller providers. And my hope is that, you know, over the next couple years, Fathom becomes a switch that, you know, you’re a two-physician practice on Athena, you can just turn on Fathom and benefit from the technology because the deployment has already happened at the EHR level.

Saul Marquez:
That’s awesome. Thank you for that. So it sounds like some partnerships there with EHR companies that potentially becomes plug-and-play for anybody that’s on the system.

Andrew Lockhart:
Yeah, exactly.

Saul Marquez:
And then is it like a software thing that they have to pay, like just a, like a like a plug-in type of software fee?

Andrew Lockhart:
Yes, so we model ourselves a lot like coding services companies. And so we pay, we get paid per encounter we code, unlike traditional coding software, which, you know, is going to charge clients for every encounter they touch, whether or not they do anything useful. We actually only charge in scenarios where we deliver final coding and that aligns incentives nicely, and that, you know, we’re very incentivized to get to these super high automation rates, and the client only pays when we deliver.

Saul Marquez:
That’s awesome, man. Well, well, look, all businesses that add major value run into things. From an entrepreneurial standpoint, I think we learn more from our mistakes than our wins. What’s a setback you’ve run into that has become a big learning and maybe even made you guys better?

Andrew Lockhart:
It was less a mistake, but by far and away, the biggest setback for Fathom was the pandemic. So, at that point in time, we had gone live with our first couple large clients, you know, late 2018, and we had this very active pipeline of folks that we felt like, oh, there’s this enormous opportunity. And then, all of a sudden, the pandemic hit, and, you know, everyone in healthcare, you know, had other priorities. The billing companies we were talking to had volumes way low, they were, you know, shutting doors, getting acquired. You know, the health systems themselves obviously had much bigger fires. And, you know, probably like seven months, people weren’t returning my calls. A lot of, like a lot of the orgs I was dealing with, whether the physician groups or billing companies went under or got acquired. And it did create a real crisis in confidence, and that, it felt like that was the moment where we had hit. It seemed like we had hit product market fit, and we were about to really, you know, go nuts. And there’s a lot of sleepless nights, and there was a question like, hey, is this related to the pandemic? Maybe the product market fit isn’t there, and it was really tough. But I think the takeaway from that is to just like re-anchor to what your hypothesis was. And I think the thing that my co-founder and I always believed is that humans aren’t going to be doing medical coding in the long run. Like this is just something that, you know, is a necessity, but there’s, if you can hold it to be true that technology will exist to do this, then this will be inevitable. And then we knew that the technology worked, and so it just forced us to basically like double down and hold fast. But, you know, think there’s a lot of worlds where, now, as a founder, you believe that you can do this big thing. And inherently, if you believe you can do this big thing, then you think the opportunity cost for your time is very high. And so it would have been very easy at that point to be like, hey, this doesn’t seem to be working, like I’m going to move on to the next thing. But we really, you know, sort of re-examined our conviction around the inevitability of this and the ultimate impact of this, and that, you know, led us to kind of stay the course. And ultimately, what ended up happening is, you know, the, you know, I recognize COVID is currently rampaging, but the other side of this, it’s created tailwinds. And that, as I was alluding to before, the staffing shortages, things like that have made, you know, Fathom much more necessary than it was pre-COVID.

Saul Marquez:
Yeah. Thanks for sharing that, Andrew. And it’s tough. And at what point did you start seeing things kind of get back on track? Like how long? You know, because COVID, I mean, lasted way longer than we wanted it.

Andrew Lockhart:
Yeah, I mean, it was really, you know, things sort of like froze, whatever, February, March, and then, and I think a little before Thanksgiving started seeing signs of life, and then by, you know, early January, things were firing. And then I think really, it’s just probably over the last 6 to 12 months where I feel like health systems are like, hey, okay, we now have accepted whatever this is, this is the new normal, and now we need to adapt to that.

Saul Marquez:
Yeah, no, that’s awesome, man, and kudos to you. You know, it’s a good lesson for all of us, folks listening, is the importance of that vision. You know, set the vision, repeat it often, even to yourself, especially in those moments of big challenge. Like, is what you’re doing something you deeply believe in? And Andrew, like no doubt, you and your co-founder, and your team you guys believe in this and, you know, definitely have made me a believer, and I’m sure a lot of folks listening, no human needs to code ever again. And you guys are making that possible with the work that you do. I can’t thank you enough for what you’ve shared with us today, and the opportunity is big. So if people want to learn more about you, your company, and ways to engage, how do they do it?

Andrew Lockhart:
Yeah. I mean one, FathomHealth.com. You can go visit, learn about the company. You want to ping me, andrew@fathomhealth.com. Also, you can look me up on LinkedIn, Andrew Lockhart. But any of those channels feel free to ping me. Always excited to talk about how we can move the future of healthcare forward.

Saul Marquez:
Love it. Andrew, look, I really want to thank you. And folks, you’re on the Outcomes Rocket. Don’t stop at listening. If you want outcomes, take action. And Andrew, I want to thank you for taking action with us today by being on this podcast. And I encourage everybody listening to take action, look Andrew up, and see what he could do for you. Andrew, thanks again.

Andrew Lockhart:
All right. Thanks so much, Saul.

Sonix is the world’s most advanced automated transcription, translation, and subtitling platform. Fast, accurate, and affordable.

Automatically convert your mp4 files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.

Sonix has many features that you’d love including transcribe multiple languages, secure transcription and file storage, generate automated summaries powered by AI, automated subtitles, and easily transcribe your Zoom meetings. Try Sonix for free today.