HCA Healthcare, Inc. (HCA) Earnings Call Transcript & Summary

So that's all good numbers. We've got Mike and Tim from HCA. We have an operator and we have the finance guy. And this is pure fireside. And we have lots of questions. So let's get started.

I think the -- in talking to people, the top of mind is -- and I saw your comments yesterday, but just to recap, where we are with the AI, and we'll have some follow-up questions, but just kind of the near-term, intermediate term opportunities by bucket and like the cadence of pulling that down into margin.

Good morning, everybody. So when I think about AI, I'll put it in context of our long-term strategic plan. AI is one component. And if you just kind of take AI and even broaden and think about digital transformation, of creating a step change in performance for the company over time. Given HCA size and scale, when you think about 47 million patient encounters last year, you think about the size of our revenue cycle, supply chain and related operations. We are ripe for being able to use things like artificial intelligence and automation to drive effectiveness and efficiency. And so we've been about building our internal capabilities to both identify, prioritize and then build, pilot and then scale out key use cases in 3 domains related to AI. The first one is clinical and clinical is really our Holy Grail as a company because it supports our mission. And our mission is the care and improvement of human life. So about 1/3 of our efforts related to AI are focused on improving clinical and this is the idea of supporting our physicians, our nurses and our care teams, with decision support and with use cases that reduce administrative burden. So those use cases in the clinical domain are going to take longer, they're riskier. And so we put a lot of structure with human in the loop and supports to make sure that those use cases are ready when they go to scale. And we can give some examples in clinical. So those are longer term. And you just got to think about those use cases taking multiple years to build and scale out as we roll through the enterprise. The second area is operational. Think about just the running of the hospitals. And so the throughput of the patients and the operations around labor management and the like are a big part of our operations. And so we have several use cases already in flight. And operations as a domain will be more intermediate term. So I think 3 to 5 years roughly in terms of really starting to see material impacts.

So let's just kind of break that down. An example of an operational initiative would be what?

Yes. Labor would be the first one that we're further so along the line. And we have a digital product that helps us predict demand and then schedule against that demand for all of our labor. So I think we have 100,000 nurses. We have 50,000 doctors. And so the ability to use AI to predict what the census level will be in a nursing unit, 28 days out is a huge help for us in terms of knowing how many staff members to bring in. And then to use AI to really help balance that schedule and get the right care team members scheduled to take care of those patients over time. And so the first big digital product, and we're live in about 80 hospitals right now, is a product that we call Timpani, which is all focused on using AI to predict demand and then do a better job of scheduling our care teams. And we're really excited about that digital product, that's one.

And again, just -- I remember when the staffing companies were red hot a few years ago after COVID, you guys talked about maybe offering either an employee model or a pay-per-click model. Do you -- did that really happen? Do you have internal nurses like an internal staffing company where you can flex up and down on a variable basis? Or is most of your clinical labor still kind of salary and bonus?

It's a blend. So we own our own staffing company, under HealthTrust, which is our GPO structure. It's called HealthTrust Workforce Solutions. And in that organization, we have both what you would think of as travelers or contract labor that operate as travelers and then we also have market pools that operate on per diem. That's our variable workforce.

And that's where AI could really help out with that.

Yes, 100%. Okay. And so -- and that's part of Timpani. That's part of our labor management tool, is really thinking about the Flex path and how to flex up and down to volume over time. The other area operational that I would call that is throughput. And the importance of managing length of stay in the inpatient settings, the emergency department throughput, the operating room throughput is really important. And so we are developing several digital tools that really help us predict how long a patient will need to be in the hospital, identify barriers to discharge and then help our teams orchestrate care and streamline patients getting through their care journey and then out of the hospital with their next point of care. And so throughput, given the importance of length of state management is another operational use case that we're in early innings on, but pretty excited about using AI...

How much you've been able to flex down length of stay so far using this tool?

Yes. Well, the tools are still pretty early. And our efforts around length of stay are broader than just technology. I mean we've invested heavily in people, heavily in our processes. And now our tech tools are just now starting to come out in terms of using AI. But we've had good success. I mean, last year -- I'm going by memory here, but it was about a 2% reduction in length of stay. And this is a multiyear effort. I mean I think between now and the end of the decade, we still have a lot of opportunity to reduce length of stay in our hospitals, and improve turnaround times in our emergency departments and operating rooms as well. So asset optimization is a really key part of our work. The third area for AI is administrative. And so think supply chain, rev cycle, human resources and the like. For HCA, we tend to operate big shared service platforms. So we have centralized management. We have more standardized data. And I think you're going to start seeing, and we are seeing AI use cases even in the short to intermediate term with administrative. So I think '26, '27, '28, we are seeing digital products come into our hospitals from our efforts because we have a little more ability to move quicker in our administrative platform. And so I think about short term, intermediate term and long term, broadly speaking for AI, I think we're still pretty much in the early innings.

So guys said this in say me, [indiscernible]. Let's go from the general to the specifics. So rev cycle specifically, one thing we've heard is, it's now easier to submit a rejected claim using AI, especially the small. It wasn't worth it before. You don't do a lot of small claims. But what specifically in rev cycle as you fight through increased denials, what tools does it give you that you might not have had 2 years ago? And what -- any quantification around bringing down denial rates or just at least holding your own there?

Yes, it's a great question. Large language models. And so one of the big areas that we're working on, and these are products that are in beta and starting to scale out for us now. is using large language models to summarize and synthesize the medical record and the administrative record. That allows us to do really do 3 specific use cases. The first one is to help our physician advisers prepare for peer-to-peer consults with their payer partners. And it reduces the time they need to prepare, and it creates a more comprehensive review of the record. The second is appeals of denials. And so you can use large language models, and we are to summarize the record, summarize the administrative record and prepare, a denial appeal letter. Now those denial appeal letters, then still go through reviews by our doctors and nurses before they're submitted. But think about the efficiency gains and the accuracy gains have been able to appeal thousands and thousands and thousands of denials versus human beings having to type de novo every time, and that's going to be powerful. The third area that the large language models, I think, will have a big impact on is helping us prepare for dispute resolution. So helping our legal department to summarize thousands of claims and get prepared for things like arbitration. So large language models in the revenue cycle, I think, are going to be a big help for us. The second area in rev cycle that -- and it's also related to the...

And this is being implemented, is being is fully implemented or it's being implemented?

Its not fully implemented . We're through alpha and beta, and we're starting to scale those out in '26. So that's my point about administrative a little quicker here in terms of seeing our efforts. The other one administratively in rev cycle would be what we think of as kind of intelligence. And so think about studying...

I'm not familiar with that concept.

Think about deciding of studying all of your claims. All the claims that got paid on time the amount you expected versus claims that had a denial or had friction. And then using machine learning to understand why a claim got denied from the clinical record and from the administrative record learning from those root causes and then trying to take actions earlier in a patient's stay to address any friction that you may see from a payer. So I think over time, AI is going to help us advance our work in terms of denial mitigation in a material way. It's a little early to see results yet because these products are still pretty new, but we're excited about it.

These are all -- you're not using third-party point solutions, these are all internally developed.

So for our -- we organize our efforts under an organization called digital transformation innovation. So that's our department's name. When we build digital products, it is a blend of build, buy and partner. So for the specific ones that I just mentioned in terms of large language model and intelligence, we partnered with Palantir. And so that's one of our tech partners that we partner with. We -- I think everyone has heard us mention before, we use Google pretty heavily as well for our cloud and for our data science tools. But that one was a partner use case. We have others where we build it entirely, and we have others that we buy just off the shelf.

Okay. Yes, there's a lot of ways we can go. I don't want to belabor this point, but you've spoken in the past about despite the ACA pressure, you're still looking at 2% to 3% volume growth this year. So in the simplest SEC math way possible, describe the decremental margin effect from the ACA and kind of the -- I know you quantified the balance, but maybe simplified that calculation a little bit for people.

Sure. I'll do my best. Is that work?

I got lost at the fourth step.

Understood. So when I think about the Affordable Care Act reforms. It's really a combination of 3 factors. It's the expiration of the EPCs. It's administrative rules that were passed as part of the one big beautiful bill at. And it's the administrative rules that came through rulemaking from the Trump administration last year. And so what we've ever to do is try to model what we think the potential impact of the comprehensive effect of those 3 changes are to HCA. Just in a very simplified way. First, we believe that our exchange volumes will be declining about 15% to 20% and '26 versus '25 full year. So that's the first impact that we think we'll see. Out of that decline, we think about 15% to 20% of the people who lose coverage on the exchanges will get coverage through employee-sponsored insurance. And so that's a benefit. The balance, so that 80% to 85%, we think go become uninsured. And for the uninsured population, we believe that there will be about 30% reduction in utilization of health care services on that population that becomes uninsured. The comprehensive impact or calculated impact of those changes, including our assumptions around utilization trends and around collectibility of patient amount dues leads to that $600 million to $900 million of negative impact to adjusted EBITDA in '26.

So this one...

Was that close?

That's perfect. I'd like to get a joint answer here. So one thing that intrigued me was, 13 outpatient things around 1 hospital going to 20. So from your perspective, Tim, what are the most needle-moving investments you can make? And then what assets are getting to the front of the line be at surgery centers or imaging centers? And what are kind of more incidental?

Right. It's an interesting question. I think a couple of big -- bigger players for us, and we spent a lot of time and energy focusing on. The first would be freestanding emergency rooms. This puts care closer to people's homes. It takes, frankly, decompresses many of our hospitals that are incredibly full and trying to make a throughput for the patient experience be better. . And it also, in some ways, funnels off some of those lower acuity patients into what I think of as a softer setting from being in an intercity emergency room. The second big one, and we -- today, I think we have 150 freestanding emergency rooms, I think in my -- I oversee about 1/3 of the company, have about 38 of those, and we just keep adding to those every year because it's a really efficient use of capital. But more importantly, it's a great way to get patients to come in -- and maybe first touch points within our system. The second one is really the urgent care platform. This has sort of been going on for a while, but we've really doubled down, as you see the generational differences of people accessing health care. People don't go to their 1 family practice or internal medicine doctor for 30 years, they want to go when I want to go, where I want to go and I either I'm close to my work, close to my home and access that. So we've really significantly beefed up the number of urgent care centers, some de novo, some through acquisitions across our markets. And then when ambulatory surgery centers has been a terrific way for us to partner with our doctors, also decompress some of our ORs in our hospitals as care has shifted to the outpatient setting. And so that's been very material. And then probably the fourth is expansion of clinics to find opportunities where there's unmet need. We've spent a lot of time trying to acuitize our hospitals, that's been a little bit of -- I think of it as a secret sauce. And to do that, you have to have more cardiologists, more surgeons and so forth out in the communities and not be expecting all those people will be coming back to just the home base to get that care where it's more complicated to park and the like. And so we think that those are differentiating models of how our expansion has been.

So remind me of your ASC footprint, how much is on-campus HOPD versus now freestanding and physician fee schedule?

We have ASCs. So these are not HOPDs for the most parts. I mean, there's a small handful that are HOPDs. These are ambulatory surgery centers. There are a number of them that are on or proximate with campus, but we also have a lot of ASCs in the communities broadly. So one of the things I've mentioned and from what Tim said, if you think about the net effect of this, what we intend to do over time is build out a comprehensive network of different continuums of care. So as patients need to access physician clinics, urgent care clinics, emergency rooms, surgery centers, we're able to give to our patients, to our doctors and to our payers. Different access is to price points to affordability points to provide the care that's needed at that time for those patients. So I think the idea of a network is this comprehensive nature of it over time. I think it creates competitive advantage, as Tim was mentioning.

So we were surprised your adjusted organic growth guidance this year is 8%. That's the departure from your long-standing 4% to 6%. Supplemented by resiliency, $400 million from resiliency. But would -- I think this is inevitable with just -- when somebody asked you, hey, that old 4% to 6% EBITDA guide and 19% to 20% margin with AI and all your efforts? When do you think you might be ready to revise those long-term targets?

Well, we clearly are going to have to get through this reform error. If you think about our guidance at midpoint for '26, it's just short of 3%. And that's our guidance for adjusted EBITDA growth. And that's reflective of the full nature of our guidance. And obviously, the audience knows this. We give full year guidance, not quarterly guidance. But in that full year guidance, what we highlighted on our fourth quarter call, was this reality is that we are facing $600 million to $900 million of headwinds from the Affordable Care Act reform items. We're also dealing with something like $250 million to $450 million of state-directed payment headwinds as well. Now that assumes none of the grandfathering applications get approved, a different matter. I know we'll talk about that in a minute. But -- and so resiliency, we've been working hard to put a resiliency plan in place that would help offset as much of the adverse impact from the Affordable Care Act reforms as we could. The net of all that, though, still yields is kind of just sub-3% at midpoint. And then as I think about the long term, what we've said is that we believe that between reilliency between the AI strategies and our network development and optimization strategies that we still believe that we'll be able to maintain our long-term plan. And so I think about this is advancing the company through this point of reform, getting to our long-term plan. And then over time, as our AI strategies and network strategies continue to mature, we are hoping that we bust through those margins and continuing to improve. But we're still talking about our long-term plan for now.

So a joint question. you've been a long-standing customer of MEDITECH. You're expanding from their legacy model to something more modern, mobile cloud based. How does that -- as an operator, is that how heavy a lift is that? How much help do you get from corporate? And then just practically speaking, what does it do for you? How does it play into your AI initiatives? Is it integrated or separate? And how do you partner with them on your technology?

When we think about our primary goal with MEDITECH expand is to reduce variation and optimize the care model so that we can take more accurate, more timely care and more coordinated care of our patients through our continuum of our systems. So I'll give you an example. Through COVID, you had care plans, policies, procedures change rapidly. . And when -- if you go back to a system like we had MEDITECH Magic, now we've already implemented 40 of our hospitals into our new expense plan. We expect to have all of our hospitals converted for the most part over by...

40 for the company or 40 for your...

40 for the company . A lot of those were actually happened to have been in my group. And so through that transition, what we're finding is that in a COVID example, we can change the policy and procedure literally in minutes across those enterprises that are currently on the MEDITECH expense model that would have been more clunky in the past where we'd have to go one by one and do that. A different example would be that for pharmaceuticals, there's an announced drug shortage of something. And in the past, the clumsiness of trying to go into each individual system and communicate, hey, that drug is not available, here's an alternative wasn't as readily available to do that. And so today, we can change that care paradigm immediately and then give -- and speed up the care. So we're not going back to a doctor, "Hey, we don't have this, what would you like to use as a substitute and just improve the efficiency of the care.

And from your -- the ones that's gone live, how the transition is going?

Yes. Transition is difficult, really difficult with the complex thing like the EHR. But what's happened with the companies, the resources that the company has really put behind it and MEDITECH have enabled the go-live to go, frankly, smoother than what we even envisioned they could go. And so we're doing big complex hospitals in Nashville, Centennial Medical Center. It's a 800-pound -- 800-bed hospital. And that transition we just did in the recent months went really smoothly. And that's a differentiator for physicians who are understandably anxious about this transition or nurses who are using these systems. And so now it's a much more intuitive system versus for folks and particularly when you have -- maybe you may have contract labor and others who are coming in and out of your system, to be able to utilize the system and take great care of our patients.

So last year, and I'm glad it's over way too much, Tom spend on the enhanced subsidy debate, hospital loss that round. This year, we've been looking at site neutral, and there's a little bit of a debate in the team of what could CMS do on its own versus what Congress would have to do. So talk about that -- I mean we think about it as drug administration, HOPD go into outpatient. We think about maybe tacking on a facility fee for a doctor visit maybe radiology being paid on HOPD going to half the rate. So just talk about -- and you're talking to the average Congress creditor, how do you frame that debate?

We don't call them Congress critters.

I don't know. That's not -- majority. That's majority.

When I think about site neutroal at least to this point, it's been pretty limited. It's been limited to physician clinics, as you mentioned, that where some systems set up their physician clinics as HOPDs. HCA generally does not. I mean -- so that's one big distinction. And other is drug administration. And so for those 2 kind of current areas in focus, the impact HCA is pretty immaterial, pretty limited. It would just be a bit more notable if that starts getting expanded to things like surgery centers and radiology and the like. And so we're watching that carefully. When I think about kind of what's in flight right now, there's not -- typically neutrality gets mentioned when their people are looking for offsets for other funding needs. And so right now, we don't see a lot in the near term that would drive that. But we're watching it carefully. And so for site neutrality for us, I would just say that it's not as material as a topic for HCA because of the way that we structure our operations as potentially the not-for-profits and the academics, which it would be more notable for. But it is an area that we're watching...

So you're running a 1% operating margin. Some of these are negative operating margins.

And I think a lot of the not-for-profits, for example, they've set up their cancer systems, their outpatient infusion centers and their hospital-based clinics, they've set them up as hospital based. And so given the way site neutrality works, it would be a more notable reform for those systems than it would for HCA.

Great. Speaking of doctors, Valesco was mentioned 1 million times a couple of years ago, you guys had to do kind of an emergency bailout of your staffing provider. You haven't really cited pro fees, professional fees as prominently as maybe some of your peers. But maybe just kind of reset where we are with Valesco, how much is inside, how much is outside? And what are the pressure points in the physician -- hospital-based for this, is it radiologists? Is it pathologist? Is it ER docs? Where are you seeing the pressure? Anesthesiologist, I'll throw them all in there, for sure.

Yes. I mean clearly, you go back a couple of years ago, we had to complete an acquisition of Valesco. And Valesco just for everyone's memory, was mostly emergency room physicians. It also had some hospital medicine positions. And it was important at the time that we do that. We've largely integrated the Valesco team into HCA now, I'm really proud of that team. Our hospital operators like Tim. Our physician services organization has done a wonderful job of bringing those teams on board and now starting to manage our hospital-based physician platform in the emergency room area, in the hospital medicine area in the HCA way. So people, processes, technology really making good investments. And we're seeing a good stability now in our cost factors related to the emergency room and related to hospital medicine and that -- and we're pleased with that. 2025 just to kind of size it in total, we were up about 10% on same facility pro fees. And if you go back to '24 and before, it was twice that level of growth because of the pressures we were seeing. I still think -- and we mentioned this on the call, but I still think we have some pressures coming into '26. And it's now more related to radiology and anesthesiology. And so we're guiding something like high single digits for our professional fee growth. We're working really hard as a company, both in anesthesiology and radiology, and we're making investments. But we are not as far along in terms of getting through that change curve as we are in the emergency room and in hospital medicine.

So a little bit of an ad lab. We follow a company that has developed and there are others that radiology tools that use AI improve throughput? Or have you been studying those as a potential productivity and cost lever for your radiologists?

Absolutely. We have a radiology committee that has been hard at work looking at not only technology but teleradiology and all of the various components to improve our day-to-day radiology services. We work mostly with partners in radiology. We don't really have a big employee or in-source capability. So our partnerships with the big groups and a lot of smaller local groups are really important. And we are not only studying but started to execute on things like both enhanced radiology and identifying the right AI and automation plays as well, but mostly with partners, right?

We're kind of -- we got about 3 minutes. I did want to hit, we talked last night, I wasn't aware that you had a GME program, and Galen, I'd love an update on that. But just obviously, one of the things I've always thought about you is your tech platform, let's say, you hire a nurse, who's used to work on the Epic or [indiscernible] platform. So you've got to train this person on your MEDITECH, which I've heard that sometimes can be an issue. And then there's also a bottleneck in residency spots for a graduating physician. So you guys have kind of in your kind of classic Toyota long-term thinking you've made some investments there. So maybe just kind of level set where we are on that.

Just to touch on GME first. So today, we have 5,800 residents across 81 of our hospitals. And that covers sort of 270 different residency programs. And what that does for us, first of all, is -- it helps us plant doctors in our communities. If they do to the residency there, they're more prone to stay there and live there and have their families in those markets. So the alignment -- there's obviously, a natural alignment to that component. It also -- if you go back 15, 20 years ago, most of those residencies were in academic medical centers. And this just gives a whole slew of more choices to people in some cases, may be living in their hometown that they grew up in. And so when we think about our 43 major markets across the country, as soon as we open these programs, the demand is there. As you and I talked last night, there's about 1,000 students that come out of Medical School and have no residency because of the limitation of available slots. On the Galen part, which has been really interesting, we acquired Galen, I think it was 2019. It was sort 5 bricks-and-mor campuses. Today, it's 25 bricks-and-mortar campuses, soon to be 30 over the next couple of years. And this is now -- we now have about 20,000 students in our programs. If you think about the long-standing notion that demand has way outstrip supply of nurses. And so this program, one of the interesting things that it's done, it's to really allow for 1 second careers. People who are average age and that school is about 33 years old. And frankly -- and those are -- 50% of those people have dependents relying on them. And so they may be doing switching into a different course and frankly upskilling them to meet that demand, and it's been really a powerful connection not only for HCA, but the communities we serve. And so we're really excited about that launch. And then the other real component for us is this thinking about clinical education. We spent a lot of time building sim centers and taking nurses or newer people in their field or new technologies come out. We can bring them in the sim letter, sim center. We can bring them for preceptors and get -- and continually push their skill sets in a sort of a safe teched environment that isn't directly with the patient.

And a bump of a guess sound, but what -- approximately what percent of the graduates end up at HCA hospitals versus other hospitals?

I think it's about 35% or 40%. Mike, is that about what you think it is?

Yes. I mean, obviously, some of them go out in the community, some stay with us. So over time, what we're trying to do is create a system where when some graduates from Galen, we're going to try to make it easy for them to come in to HCA. But over time, I think that will improve.

Okay. Thank you. That's good to break out.