The $19,000 Bill That Exposed Healthcare's Biggest Problem

Speaker A: We took a very explainable AI approach to it. We wanted to ensure that, hey, whatever we do should be fully explainable. Something that we internal developers use, like the ones the way we debug our actual, like, hey, is it working? Where is it not working? We realized that we want to ensure that it should be similarly explainable to our customers as well. So, in fact, the internal tool that we had to understand everything, we actually exposed to our customer. And that has been a key driver for building the trust within our system.

Speaker B: Welcome to Bombastic Health, where we challenge the business of healthcare and explore bold ideas that drive meaningful change. Well, hello everybody, and welcome back to another episode of Boom Bostic Health. We're here in the wild at Vive, which is a lot of fun, a lot of great energy here. It's an amazing ecosystem going on. And I'm joined by Nitesh Shroff, who is bringing a new approach to revenue cycle and in healthcare. Thank you so much for joining us on Boomastic Health.

Speaker A: Thank you. Uh, this is very exciting and very, very excited to be at Vie and talking to you. Thanks for having me here.

Speaker B: Absolutely. So we were just comparing notes a little bit. We were closer as neighbors out of Chicago, although this is the first time we've met. And now you're back in the Bay Area, but still have some connectivity back to Chicago. Um, so revenue cycle is this big, huge bucket of a bunch of different things. And it ranges from the part of revenue cycle that change healthcare touches that. As we all might have heard, there was a bit of a meltdown there that brought the US Healthcare system to its knees. You can kind of chuckle about it now that it's over with. That was painful. Um, but then there's a lot more that goes into actually ensuring that you're managing the way that healthcare is paid for. And historically, it's taken a lot of people and there have been a lot of pretty imperfect data systems that deal with this. I'd love to hear what the unique idea was that you had, because you started this about five and a half years ago. You're CEO now. Things are going really well. I always like to hear the origin story.

Speaker A: Yeah, uh, no, wonderful. Um, at Aryntra, our mission is to, uh, improve the financial health of hospitals such that healthcare becomes more affordable and accessible across the country. Um, we started, uh, this company when my co founder, Preeti, she ended up going to, uh, an emergency room of one of the large health systems in Bay Area, and she spent, like, four uneventful hours. Nothing much happened. And Then after a month, she gets like a bill of $19,000. Right? Like it was way more than what our savings at that point was. And like we couldn't really understand what leads to that $19,000 bill. And it turned out to be some error, uh, on the coding side of it that actually led to such large bill. And then my co founder Preeti, she just uh, continued to talk to this health system and fixed the coding error and then brought the bill down to $11,000. And then insurance paid 7,000 and Preeti paid $4,000. Right. That sort of put us on this path. It's like, why is this happening? This doesn't make sense. There has to be a better way. And uh, both Preity and I are PhDs in AI, so we just brought in the engineering lens, the PhD lens to it. Um, uh, and really, really the more we dug into it, the more we realized how beautiful the problem is. And both the impact on the patients and very, very importantly on the provider organizations is just so huge. Uh, we started solving it. Right? We started finding a better way to do this. Um, today, uh, we work with large health systems and physician groups across the country to automate their revenues, to automate their medical coding, some amount of prior authorization, uh, providing a lot of documentation, insights from the CDI angle and then also prevent their denials. Right. So all, all in all, uh, bulk of mid revenue cycle that we automate that too within the ehr. Like within epic, within, like inside epic, inside athenahealth and some other EHR as well.

Speaker B: And were you in health care before?

Speaker A: I was not. Uh, but my team has always been in healthcare. We have, uh, very seasoned team of, um, our Susan, who is the VP of revenue cycle, uh, she has been like doing this for 25, 27 years. Um, and the entire team is actually very, very healthcare focused. Nouri, uh, my whole, uh, CS and the sales team.

Speaker B: Well, I actually think a lot of times the biggest innovations come from outside of healthcare. People that aren't so immersed in the way it's always been, and they can come into healthcare and say, hey, we should do things differently, we can do it better.

Speaker A: Exactly.

Speaker B: And so is it really an army of agentic AI agents that are dealing with these components? And is there a reinforcement learning through human feedback that's mixed into this? Or like, how does it actually work?

Speaker A: Very much, very much. And exactly what you said. Like, uh, we came from outside, but we immersed ourselves deep into like we sat within a, uh, large urgent care center in Alabama and truly understood the actual pain points before we started solving this. So we really, really brought in the insider and sort of an external views. A very unique mix of the two. Um, I'm a PhD in AI. My co founder is a PhD in AI. We took a very explainable AI approach to it. We wanted to ensure that, hey, whatever we do should be fully explainable. Something that we internal developers use, like the ones the way we debug, uh, our actual, like, hey, is it working? Where is it not working? We realized that we want to ensure that it should be similarly explainable to our customers as well. So in fact the internal tool that we had to understand everything we actually exposed to our customer and that has been a key driver for building the trust within our system. Our AI is built, uh, basically, uh, as a network of clinical agents and financial agents that truly understands how the care was provided and what was the exact care that was provided. Right. Uh, for instance, is the diabetes stable? Is the hypertension chronic? Was the surgery done on the left knee or the right knee? These are the clinical things each of our agents understand, uh, and collaborate with each other. And then on the financial side, was it the resident that provided the service first and then a supervising provider came in, or was it one doctor? So we have to understand and who was, who did the lab draw? The lab draw was done by who actually and everyone in the healthcare needs to get paid. The hospital needs to get paid, the doctor needs to get paid. So you have to understand all aspect of how the service was provided and how the care was actually provided, in what setting, who all were involved, how the money should be split between the two, because everyone needs to get paid. And exactly what was the complexity of the patient and the complexity of care? Have a network of agents that, that uh, identifies individual elements and once we do that right, and uh, 86% of the time we fully automate that, 14% of the time we put it in our customer's queue bucket, the whole bucket or the Q bucket. Right. Where it's our customers, coders that are human in the loop that they actually look at it and the changes they make, we actually pull in back and continue to improve the system based on, for that specific organization. So there's a significant reinforcement learning that happens and the system keeps getting better and better at all the tasks that it performs.

Speaker B: That's very cool. Very cool. It sounds like you were a bit ahead of this huge wave that was forming.

Speaker A: Absolutely.

Speaker B: When did you finish your PhD in AI?

Speaker A: I finished in 2012 and since then I have been building. I was Building self driving cars before this, um, now, uh, it was a company called Zoox, now owned by Amazon. You probably see them driving in the Bay Area all the time. Uh, and before that another company called Light, where again I was building sensors for self driving cars. M. The key learning was how do you deploy AI in the, in the real world where safety really, really matters. Right?

Speaker B: Yeah.

Speaker A: Uh, and that's the lens we brought into health care. And it has been, it has been amazing how that sort of the safety approach that we have brought in how. And that's why explainability was the first thing that we brought in. Right. Uh, to what we call as our, uh, as the revenue assurance. Right. Like we want to ensure that health systems get paid for all the services they have provided. Uh, appropriately. That's appropriately. Appropriately and on time. Right. Because the gross, the margins at which this health system run is so thin, probably negative 1%, 2%. Right. So we want to ensure that um, hospitals get paid for all the services they're provided. Accurately, timely, and we ensure that they do it in three different steps. First of all, unfortunately, just because you have provided the service in this country doesn't mean you'll get paid for it. First you have to document what the service was provided and then you have to ensure that you could code for what services you have actually documented. Right. Not code for, uh, what was actually provided, but code for what was documented and then ensure that you get paid for exactly what was coded for. So in these three different steps, we ensure the revenue assurance for our health system partners.

Speaker B: So you have so much knowledge, uh, of how to apply these data science concepts. How do you avoid the temptation to have this scope of uh, what you're attacking grow exponentially? Because I'm sure as you look at the overall healthcare system and all the things that are touched and affected and things you could help make better, there's an unlimited amount of opportunity. How do you decide what to do and what not to do?

Speaker A: I think our lens to this is customer driven. Right. Uh, like one of the beauty of healthcare that we truly believe in is once you have won the trust of a health system partner, of a, of a client partner, they, every one of them become a design partner for you. They want you to develop, they want, they tell you their core pain points. Right. And then once you understand the pain points across different health systems, it's, it becomes very clear exactly what you should focus on. Right, Right.

Speaker B: Yeah. And so you're following that. Yeah.

Speaker A: So it's all customer driven, customer led. Um, our design Partner Mercy Health, our design partners, like all the other health systems that we work with, we understand from them and their pain point is sort of reflective of others as well. And we take the lead from that and truly decide based on that. So you're of course have certain DNAs as well. Like we have certain strengths. Of course we have to play to our strength as well.

Speaker B: Yeah, yeah. And you're at vive to get the word out and uh, form partnerships and connections. Are you using the market connect to meet health systems? Is that part of what your team is doing here?

Speaker A: Absolutely, absolutely. Um, we are uh, we are talking to health systems, large physician groups across the country. Right. Uh, showing them exactly the core. We deliver around 11x ROI. Right. So every dollar that is spent on Aryntra we deliver around 11x back like $11 back.

Speaker B: Over what period of time?

Speaker A: Because it's a usage pricing. Because in the first like our time to value is maybe two to three months. Right. Uh, so within two to three months we are already delivering 11x ROI. Okay. Uh, and ours is a usage based pricing as well. So it's rather greatly the ROI that they get back on it.

Speaker B: So you share in the value that gets created as it goes.

Speaker A: That's right.

Speaker B: Okay.

Speaker A: Um, uh, and basically because we know uh, that and the core value we have been providing has been that includes the revenue uplift, we help them uh, free up their staffing so that they can focus on some of the more complex stuff. Right. Like we'll focus on more complex but coding more, uh, like work at the top of their certificate. Right. Um, and collaborate better with the providers. Right. And then we also reduce their AR days, so on and so forth. And this is the sort of the holistic ROI approach that we take uh, for our health system partners here.

Speaker B: What is the competitive dynamic? Like what are the reasons that people say no, we aren't interested? Because it sounds like everybody should do this.

Speaker A: That's exactly right. That's exactly right. Right. And um, that's why we are seeing such large uh, growth today. Um, because it has become such a big pain point. Um, the country is not producing enough coders. The complexity of revenue cycle keeps growing and growing. Payers continue to make it more complex for the health systems to get paid. Denials are always increasing. Just because it's a denial Tuesday or a denial Wednesday, it continues to increase and the only real solution then is to actually bring in automation. And one of the approach, key design choices we have made is to fit within the workflow of the health systems within The EPIC within the athenahealth within. And when we fit in the workflow, when we provide a very, very high value, they continue to expand. Right. Uh, and that's sort of the lens we have taken towards working with our health system partners to deliver everything within the workflow of our health system partners. Uh, in terms of the competitive lens, of course, there are legacy solutions, the ones that have been more helping out on the assisted part of it. And then there are other solutions that have focused a little bit more on, uh, different care settings, like some on radiology, some on, um, more ed. We today have the broader spectrum, uh, of specialties that we actually automate for our health system partners, and that's what is driving a lot of our growth for us.

Speaker B: Okay, are you doing anything to help the actual patient be better informed about what is going on? This is a huge part of the problem is it's this big black box. It's totally mysterious. And sometimes you get the $19,000 bill, and it's not because it was coded incorrectly. It's just because health care pricing is a little bit, um, messed, uh, up at times, to use a very kind phrase. What are you doing to try to help solve that part of the equation, which, obviously there is a huge tragedy with this UnitedHealthcare Insurance CEO, and part of the basis for this insanity that led to that tragedy was this whole, oh, well, the way that, you know, you pay for health care is just so opaque and nobody understands it. Help me understand how you guys see the consumer side of this.

Speaker A: Pretty much. Uh, yeah, no, thank you. Um, look, I think healthcare is extremely complex in this country. Right? There are thousands of moving parts in the country. Right. You cannot be solving all of it. However, the one. One of the key pieces on the admin side that impacts the. The, uh, patient's bill the most is inaccuracy in the coding and the billing. Right. That's what we are solving. And it has a very direct impact on the accuracy of the. Of the bills that the patients get. Right. As part of some of the workflows that we automate, patients are informed upfront that, hey, um, if they will actually be doing, let's say, a sick visit within wellness visit, they'll actually be possibly getting a double charge for it. Right. So they are getting to know upfront, which earlier was happening and was surprising them. They're actually getting to know upfront. Right. They're getting more accurate bills. But, um, there is a significant consolidation in our. All our health system partners to do the whole SBA Effort to, to make it a single billing, uh, office, like a single payment, a single bill that goes out to the patients. Right. So a lot of things are happening. Our starting point on this has been automate coding that ensures that the billing that the patient gets is accurate. And the insurance companies, it's much more about being, being accurate.

Speaker B: Right. So if it's accurate, it's going to be more clear and the process will be run in a way that's more appropriate. So if there is a prior auth required, for example, or some other stipulation that's not a, uh, you figure it out later downstream, it's the patient knows right away that is right.

Speaker A: And in fact we have started to like. It's about being accurate, it's about being compliant. It's about coding what was actually done right. Um, and health systems are already taking significant steps in all other directions to bring all of them right. It has to be handled from multiple directions. We are the best in what we do today to ensure that it's accurate coding. And now we have started taking steps in actually ensuring that the prior authorization is done right as well, so that again, patients don't get the wrong bill, they don't get denied, uh, or uh, have to pay for the bills that they were not supposed to pay because it was prior authorized wrong. So we are starting to highlight that, make sure their prior authorization was done right.

Speaker B: Uh, and I think there's probably a whole world of opportunity in that area as you continue to grow a lot,

Speaker A: A lot actually, and revenue cycle just continues to like, the more you dig into it, the more you realize how complex revenue cycle is. Um, uh, and that's why we love.

Speaker B: That's the lifeblood of healthcare, you know. Exactly. Uh, well, thank you so much for the good work that you're doing to make healthcare better. It's a big reason our podcast exists is to promote that and thanks for sharing it with us, you know, in our audience.

Speaker A: Thank you, thank you. Very excited to be, uh, to be here and building in this space. Thank you.

Speaker B: Awesome. Well, we're glad to have you as part of the Boom Bostic health community. Thank you everybody for joining us for another episode here at Vive in la. We appreciate you being here. I'm your host, Bradley Bostick. We'll see you soon.