Title

Presentation

Operator: Hello, everyone, and welcome to Astrana Health’s First Quarter 2026 Earnings Call. [Operator Instructions] Today’s speakers will be Brandon Sim, President and Chief Executive Officer of Astrana Health; and Chan Basho, Chief Operating and Financial Officer. This press release announcing Astrana Health’s results for the first quarter ended March 31, 2026, is available in the Investor Relations section of the company’s website at www.astranahealth.com. The company will discuss certain non-GAAP measures during this call. Reconciliations to the most comparable GAAP measures are included in the press release. To provide some additional background on the results, the company has made a supplemental deck available on its website.

A replay of this broadcast will be available at Astrana Health’s website after the conclusion of this call. Before we get started, I would like to remind everyone that this conference call and any accompanying information discussed herein contains certain forward-looking statements within the meanings of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. These forward-looking statements can be identified by terms such as anticipate, believe, expect, future, plan, outlook, and will, and conclude, among other things. Statements regarding the company’s guidance, continued growth, acquisition strategy, ability to deliver sustainable long-term value, ability to respond to the changing environment, liquidity, operational focus, strategic growth plans, and acquisition integration efforts.

Although the company believes that the expectations reflected in these forward-looking statements are reasonable as of today, those statements are subject to risks and uncertainties that could cause the actual results to differ materially from those projected. There could be no assurance that those expectations will prove to be correct. Information about the risk associations with the investing in Astrana Health is included in the filings with the Securities and Exchange Commission, which we encourage you to review before making any investment decisions. The company does not assume any obligation to update any forward-looking statements as a result of new information, future events, change in market conditions, or otherwise, except as required by law.

Regarding the disclaimer language, if you would like to refer to Slide 2 of the conference call presentation for further information. With that, I will turn the call over to Astrana Health’s President and Chief Executive Officer, Brandon Sim. Please go ahead, Brandon.

Brandon Sim: Good afternoon, and thank you for joining us on Astrana Health’s first quarter 2026 earnings call. Today, I’ll begin with our first quarter results. Then discuss how we have built and positioned Astrana anchored in our AI-enabled platform and longitudinal payer-agnostic care model and why that model is increasingly advantaged. I’ll then provide updates on our 4 strategic pillars and our progress against each. And finally, I’ll provide some color on the Prospect integration, expansion market performance, and recent regulatory updates before turning the call over to Chan. Astrana delivered a strong start to 2026. We saw continued disciplined growth, well-controlled medical cost trend, meaningful operating leverage, and early performance from new full-risk contracts that continue to track in line with our underwriting expectations.

More importantly, this quarter reinforces our broader thesis. As the health care environment becomes more complex, advantage will accrue to organizations that can integrate care delivery, data, and financial accountability into a single operating system. Astrana has built that operating system. And we believe that advantage is widening. In the first quarter, Astrana delivered revenue of $965.1 million, up 56% year-over-year and adjusted EBITDA of $66.3 million, up 82% year-over-year. Non-GAAP adjusted EPS was $0.74, up 76% year-over-year and free cash flow was just over $64 million in the quarter. Deleveraging also continued to progress ahead of schedule with net leverage declining to approximately 2.3x on a pro forma trailing 12-month basis and to 2.2x based on the midpoint of our full year guidance.

As a reminder, when we announced the Prospect transaction, we communicated a path to deleveraging below 2.5 turns of net leverage within 24 months. We have now achieved that milestone in just 3 quarters. And we anticipate ending the year at or below 2 turns of net leverage. We are pleased with the consistency of our performance and execution against our priorities in the first quarter. And our results increasingly reflect the advantages of the platform we have built and the way we are embedding AI across our platform. Our view is straightforward. AI can improve individual tasks. But the greatest value accrues to the orchestration layer where data, workflows, clinical decisions, and financial accountability are integrated across the system. In health care, that means connecting how care is financed, coordinated, and delivered and, ultimately, improving outcomes for patients.

We believe that requires deep architectural alignment. Unlike fragmented health care technology stacks assembled across multiple third-party vendors, our platform was designed internally as an integrated operating system because an embedded orchestration across workflows, care delivery, and financial operations requires that. As a delegated payer-agnostic platform, we sit at the center of the health care ecosystem with a continuous longitudinal view of each patient across plans, settings, and time. We are not tied to a single payer or a single line of business. We follow the patient throughout their health care journey. That creates 2 structural advantages. First, it creates long-term value. The continuity we build with our patients allows us to engage and manage care over extended periods of time, driving better clinical outcomes, more efficient resource allocation, and more predictable financial performance.

Second, it creates a compounding data advantage. Our longitudinal view allows us to build a more complete and persistent understanding of each of our patients, which improves our ability to predict risk, intervene earlier, and coordinate care across settings. And on top of that foundation, we have built a proprietary data ontology and AI models that translate intelligence into action, embedding real-time insights, next best actions, and workflow orchestration directly into provider workflows and care management operations. Across our platform, our AI agents are increasingly embedded into operational and clinical workflows, helping manage authorizations, claims processing, care management, quality outreach, and next best actions in real time.

Because these agents operate within our broader platform and data infrastructure, they act with longitudinal context across the patient journey rather than within isolated workflows. And these capabilities are embedded directly into the day-to-day workflows of our providers and care teams, driving measurable improvements at the point of care. Providers actively using our platform achieve a 24% higher gap closure rate and a 30% higher annual wellness visit completion rate. And those outcomes are increasingly powered by AI-enabled patient engagement at scale, including around 500,000 automated member interactions across voice and text each month, the equivalent of several hundred personnel worth of outreach capacity. We are seeing similar leverage operationally.

For example, our AI claims agents have reduced provider payment cycle times to less than half that of manually processed claims. Taken together, these capabilities translate directly into improved clinical outcomes, more efficient operations, and ultimately, more predictable financial performance. Importantly, because we operate the system our AI is improving and because we maintain longitudinal relationships with patients across payers, the benefits compound over time within our platform. As more patients flow through our system, our models improve, our predictions sharpen, and our ability to allocate resources becomes more precise across the patient journey. That combination of longitudinal relationships, data continuity, and integrated workflows is what really enables us to translate AI into durable clinical and economic value.

We continue to see those platform advantages translates into consistent clinical performance across the enterprise. In the quarter, medical cost trends slightly outperformed our full year trend assumption of approximately 5.2%, with strong performance across both our core and legacy Prospect populations as we continue integrating Prospect onto the Astrana operating system. Our original Medicare populations in both ACO REACH and MSSP also performed well, reinforcing the scalability of our platform and the ability of our technology and clinical infrastructure to drive consistent outcomes across lines of business. We are also seeing that leverage reflected in our operating structure. In the first quarter, G&A as a percentage of revenue was 6.4%, a 70 basis point improvement year-over-year.

As we continue embedding agentic workflows and intelligence across the platform, we expect additional operating leverage over time and believe that we will exit the year at levels below where we are today. Turning to membership. We ended the quarter serving approximately 1.55 million members in value-based care arrangements. On Medicaid and Exchange, trends in the quarter remained generally in line with expectations with puts and takes across the portfolio, largely offsetting one another. Medicaid membership attrition tracks modestly below expectation, while acuity has remained favorable, reflecting less adverse selection than modeled due in part to our longitudinal patient relationships. On the exchange, attrition tracked somewhat ahead of expectations during the quarter.

And overall, we continue to manage these dynamics with a disciplined and appropriately conservative approach. And our broader assumptions and outlook for 2026 remain unchanged. On prudent risk progression, we delivered on the commitment we made in late 2025 to convert key contracts to full risk arrangements. At quarter end, approximately 80% of care partners’ revenue and around 40% of owned membership were in full risk arrangements. Importantly, new contracts that commenced this quarter are performing in line with our underwriting, reinforcing the discipline of our approach. Collectively, our results reflect continued execution across the 4 strategic pillars we have discussed consistently over the past several years: Disciplined growth, prudent risk progression, strong clinical and medical cost performance, and expanding operating leverage through our platform.

Now, turning to Prospect. Integration remains on track and continues to validate the strategic rationale for the transaction. We have completed financial standardization, established full visibility into medical economics and aligned clinical workflows under the Astrana Care model. Gross provider retention remains above 99% for the quarter. And we continue to track towards the high end of our $12 million to $15 million annual synergy target. In our expansion markets, Southern Nevada, which reached run rate profitability in 2025 with a 20% year-over-year improvement in MLR, continues to perform well. In Texas, the launch of our full risk delegated model with a large payer partner on January 1 is progressing in line with expectations. And we expect our platform and operating model to drive a similar maturation curve over time in Texas as we’ve observed in our other markets.

Finally, some quick comments on the regulatory environment. On the 2027 Medicare Advantage final rate notice, we believe there continue to be structural tailwinds for Astrana. Our model is not dependent on diagnosis sources that are being disallowed. And our historically conservative and counter-based approach to risk adjustment positions us well under the revised framework. More broadly, as regulatory changes continue to minimize risk adjustment as a source of alpha, we expect relative performance across the industry to be increasingly driven by underlying clinical execution and cost management. That is core to how we operate. To close, our first quarter results reinforce the structural advantages of the Astrana platform. We are growing with discipline, progressing risk responsibly, managing medical costs with consistency, and continuing to widen a durable technology and AI advantage that compounds with every patient we serve.

With that, I’ll turn the call over to Chan.

Chan Basho: Thank you, Brandon, and good afternoon, everyone. Our first quarter financials reflect solid execution and a strong start to 2026, driven by the commencement of new full risk contracts, continued contribution from Prospect and disciplined platform-wide performance. Total revenue for the first quarter was $965.1 million, up 56% versus the prior year period, driven by the full quarter contribution from Prospect, commencement of full risk contracts and continued organic growth across our Care Partners segment. Adjusted EBITDA for the quarter was $66.3 million, up 82% versus the prior year period. Both revenue and adjusted EBITDA came in at the higher end of our guidance range, reflecting the durability of our model.

Net income attributable to Astrana was $14.4 million and adjusted EPS was $0.74 per share. Medical cost performance in the quarter was in line with expectations. Our 2026 plan assumes a blended cost trend of approximately 5.2%. And Q1 actuals across both legacy Astrana and legacy Prospect were consistent or better than planned across all lines of business. G&A as a percentage of revenue was 6.4% compared to 7.1% in the prior year first quarter. This 70 basis point improvement reflects continued operating leverage as we scale revenue and continue to embed AI capabilities across the enterprise. Free cash flow for the quarter was $64.1 million due to strong operating performance and conversions to full risk. We continue to expect strong full year free cash flow generation as new full risk contracts ramp, working capital normalizes, and integration-related investments decline.

We ended the quarter with $478.4 million of cash and $586.8 million of net debt. Net leverage on a pro forma basis was approximately 2.3x, down from 2.6x at year-end, reflecting strong free cash flow generation and continued EBITDA growth. We remain committed to meaningful deleveraging over the next 12 months through profitable growth, free cash flow generation, and disciplined debt reduction. We are reaffirming our full year 2026 outlook. We continue to expect total revenue in the range of $3.8 billion to $4.1 billion, adjusted EBITDA between $250 million and $280 million, and free cash flow between $105 million and $132.5 million. We’re pleased with our first quarter performance and continued execution and remain disciplined in our approach to full year guidance.

Our outlook continues to assume conservative Medicaid membership trends and 0 contribution from HQAF. We expect greater clarity on both items as the year progresses. And until then, we will continue to apply an appropriately conservative approach to full year guidance. As a reminder, the midpoint of our 2026 guidance reflects our operating plan. The low end assumes a stacked downside case rather than a shift in underlying execution. On the headwind side, we have embedded expected declines in Medicaid and exchange enrollment, adverse selection, losses associated with new cohorts and expansion markets, conservative medical cost assumptions, and 0 contribution from HQAF. On the tailwind side, we have modeled improved 2026 Medicare Advantage rates, continued realization of Prospect synergies, ongoing maturization of full risk cohorts, and operating efficiencies driven by automation and AI deployment.

For the second quarter of 2026, we expect revenue between $965 million and $1 billion and adjusted EBITDA between $65 million and $70 million. Taken together, our first quarter results give us continued confidence in our ability to deliver against our 2026 framework. With that, operator, we’re happy to take questions from the audience.

Receive real-time insider trading and news alerts

Operator: [Operator Instructions] Our first question is from Jack Slevin with Jefferies.

Jack Slevin: Candidly, crazy afternoon, so a little trouble processing information. Maybe just to hit on what I think are like the 3 biggest things for everyone here. I heard the commentary on the trend better or in line with what you’re expecting across all books. If I just think about enrollment and trend in Medicare Advantage on the HIC side and in Medicaid, can you just give me the rundown on sort of where that stuff landing versus expectation and how to think about the progression there versus what you sort of already expressed at the last quarter call? [Technical Difficulty]

Operator: Will the speakers please check and see if their line is muted.

Brandon Sim: Sorry, can you guys hear me?

Operator: Yes.

Brandon Sim: I apologize. Sorry, I know that, that was — that was a busy quarter, Jack. Thank you for joining anyway. Happy to give an update per line of business on enrollment and trend. For — starting off with Medicare, enrollment came in, as we had described before, mid-single-digit growth in eligibility. I’ll start first with enrollment and then go to trend. On Medicaid, as I mentioned in my prepared remarks, enrollment or disenrollment tracked slightly ahead of the midpoint of our range. And so we’re looking at probably on the high end of our range for disenrollment for the year. And then finally, for exchange, things came in better than expected in terms of disenrollments as has been noted industry-wide. In terms of trend, we were able to come in at or above our full year range for trend, which is a blended 5.2% cost trend year-over-year.

And so trend has performed very well across all lines of business. Notably, trend came in better in Medicaid as well relative to our expectations. So there was lower adverse selection so far throughout the year than we expected even with the slightly higher disenrollment than expected. So as I mentioned in the prepared remarks, Medicaid and exchange kind of puts and takes there ended up balancing out. And trend ended up performing better than expected really across all lines of business and for both core and legacy Prospect populations.

Jack Slevin: Just one follow-up for me. The balance sheet, obviously now getting to a better position. I know you called it out and then sort of a lot of where you had been messaging and things progressing nicely and good free cash flow generation in the quarter. I guess maybe just thinking about, you had done some M&A, nothing obviously on the scale of Prospect beforehand. But as you sort of get that leverage ticking down and think about what you can do with excess free cash, would love to get your thoughts just on what you think the best use of cash is here. If there’s ample tuck-in opportunities? If the buyback is something you should look at? Just curious to sort of hear what you’re thinking about there.

Brandon Sim: Yes, of course. Overall, our approach to capital allocation, I think, is going to remain disciplined and consistent with the priorities we’ve previously communicated. First and foremost, of course, our near-term focus is on deleveraging following the Prospect transaction. As I mentioned in the remarks, we’re very pleased with the pace of progress so far. As I mentioned, net leverage already declining to approximately 2.3 turns on a pro forma TTM basis. And that’s far ahead of the timeline we originally communicated when we announced the transaction. And so when we think about capital deployment, I think our highest priority continues to be investing organically into the platform, including our technology infrastructure, AI capabilities, clinical operations, and expansion markets.

And we see strong returns and a meaningful runway ahead in those efforts. On M&A, though, it’s really a question about capital allocation efficiency. I think we already — we believe we already have the core capabilities required to operate a fully integrated AI-enabled health care operating system internally. So the question is less about acquiring technology capabilities and more about determining the most capital-efficient way to expand membership provider relationships and market density over time. So it’s going to be a bit of a buy versus build question in terms of M&A. That being said, we continue to believe the platform is extremely well positioned to integrate and scale M&A acquisitions over time. Because we’ve built that proprietary operating platform, I think we’ve proven that we’re able to operationalize acquired assets very efficiently and very consistently across the platform.

And we’ve demonstrated that capability, as you noted, with Prospect, but also with things like collaborative health systems, CFC, and more in the past. So it’s going to be an important opportunity to continue growing the platform. But we’re going to remain disciplined and highly selective in the approach. And finally, on share repo, we did continue to do share repurchases in Q1 as we have in Q4 of last year. And we’ll continue to evaluate that capital allocation strategy dynamically based on where we believe the risk-adjusted return for repo will be and where we think we can create kind of long-term shareholder value. So given the strong cash generation so far and that integration is on track and ahead of schedule, we’re pleased with where we are.

And we think we have a lot of flexibility over time as we continue growing the platform.

Operator: Our next question is from Ryan Daniels with William Blair.

Ryan Daniels: Congrats on the strong start to the year. Brandon, I thought you gave a great overview of the Astrana operating platform and the advantages it gives you both on care and operating efficiencies. So I’m curious how much more leverage do you have there to drive maybe G&A efficiencies? And what type of new programs are you launching? And then as a follow-up, I’d love to learn more about how you plan to commercialize that in the market as other vendors kind of struggle sometimes to manage care as effectively via your care enablement partner offering.

Brandon Sim: Hello, Ryan, thanks for the question. I think there’s a lot of — I described some of the examples of how we’re using technology so far. It’s really deeply integrated into the system. And it helps that we have a fully delegated capitated model where we do act as a single payer. And we have the visibility across authorizations, claims, care management, and the entire ecosystem. So far, as I mentioned, we’ve really been using a lot of AI in terms of our risk stratification models, our next best action models, creating a suite of agents on both the payer-facing and provider-facing side. On the payer side, for example, on claims adjudication and prior authorization, on the provider and patient side in terms of engagement through voice and text as well as clinical documentation and gap closure.

I think some opportunities remain in further expanding our agentic care management workflows, something we’ve developed over the last half year or so that we’re — that is already in use, but certainly can lead to further efficiencies on both the OpEx and, hopefully, over time on the cost of care line as well. We’re also looking at, of course, continuing to finish off the integration of Prospect onto the Astrana operating system, which can drive further operating leverage as well as over the medium term, medical cost leverage, and continuing to expand our clinical decision support capabilities embedded directly into the provider workflow as part of the Astrana operating platform. So I think there are going to be continued opportunities. And like I mentioned in the prepared remarks, already reduced G&A as a percentage of revenue, 70 basis points year-over-year and expect to exit the year even lower than where we came in around — sorry, lower than where we came in, 6.4% in Q1.

On the second question in terms of commercializing this in the market, I think perhaps an underappreciated part of our story is that there is a segment that we report in which we do commercialize some of these tools to the market in our Care Enablement segment. That segment continues to grow rapidly, has a strong gross margin and EBITDA margin profile. And just in this quarter, we added a new client, which we had disclosed kind of on earnings — on a previous earnings report to that client base in the Care Enablement business. So we continue to grow that business rapidly. And we think there is potential to not only improve groups and clients in that business, but also one day potentially, as we did with the Community Family Care acquisition, to look for deeper ways to partner and get them perhaps into our Care Partners business.

Ryan Daniels: And then one quick follow-up. This is more housekeeping. But with the quality assurance fund, I know that’s not included in your guidance. Has there been any update there or any thoughts on when we might get timing on that to see if there could be potential contribution to this fiscal year for you guys?

Brandon Sim: Thanks, Ryan. Yes, I think that’s unfortunately going to have to wait until later in the year. We don’t have an exact date in mind, but probably in the third or fourth quarters. So again, out of conservatism, we’ve left that contribution out of the guidance for 2026. But we look forward to hearing more and updating the street when that happens.

Operator: Our next question is from Jailendra Singh with Truist Securities.

Jailendra Singh: Congrats on a strong quarter. Brandon, I know you have been cautiously optimistic around your 2027 EBITDA target of $350 million and you’ve said that there is still a path to get there. But in recent few months, there have been some positive developments around 2027 CMS MA rule. You just said that Medicaid and HICs have been trending better to at least in line to better than expectations and then you’re also driving AI-driven efficiencies. Are you feeling better about that target now versus 3 months back? Or at least you’re willing to say that current consensus, which is around $340 million, seems to be at least in a reasonable range. Just trying to understand like how your views about 2027 might have shifted in the last couple of months or 3 months.

Brandon Sim: Hello, Jailendra, thank you for the question. When we originally provided that 2027 adjusted EBITDA framework, this was back in 2024. Of course, we’re in a meaningfully different regulatory and industry environment than the one we’re operating in today. But with that being said, I think the more important point, the more salient point is the continued strength and adaptability of the Astrana platform over all environments. Our model was designed to operate across cycles, as I’ve mentioned many times before. And we believe the consistency of our performance over really decades of performance. But certainly even in the last 5 or 6 years, certainly reflects that. As an example, from 2019 through guidance for 2026, we’ve grown revenue at approximately a 32% CAGR and adjusted EBITDA at a 25% CAGR while continuing to generate operating leverage and free cash flow along the way as we grow very, very rapidly.

And against that context, looking forward into ’27 and beyond, we think that the business has continued to be positioned to grow organically at a mid to high teens rate while continuing to deliver on free cash flow as well. We see meaningful opportunities, of course, to accelerate that growth past the mid to high teens growth rate through disciplined and selective M&A potentially over the long-term, particularly given the scalability of what we’ve built. But even without M&A, we still think that it’s a mid to high teens organic grower. And so that being said, I think the key takeaway here is really the operating model and its durability across all regulatory and economic cycles. Our ability to continue compounding growth as we have, 25%, both organically and inorganically over the last 6, 7-year period and our continued expectation that off of the 2026 number, that mid to high teens CAGR on the EBITDA line is firmly within reach over the short to medium-term future.

Jailendra Singh: And then my follow-up on the AI investments. You talked — I think in the presentation, you said that your G&A has been benefiting from AI-enabled tools. And is the message that all of the 70 basis point year-over-year improvement was driven by these AI tools, which would imply like $7 million benefit in the quarter alone. I just want to confirm that. And then as we think about broadly your AI investment strategy. How are these investments split between focus on administrative aspect of the business where savings might directly fall to bottom line right now versus investing in clinical workflow, so that these will drive more savings down the road? Just help us understand how do you allocate your AI investment strategy and the dollars there?

Brandon Sim: Yes, of course. I think it’s a little hard to say exactly how much of the 70 bps is driven directly by AI. Certainly, AI is being infused across the board. So I would say a meaningful part of that without quantifying is driven by AI and its ability to help us scale the business without increasing G&A costs associated with that rapid revenue growth. In terms of the split between more administrative functions and maybe clinical or coordination and navigation-related functions that could potentially have an impact on medical costs in the short and medium-term future. I think it certainly started off on the payer side and on the G&A side. We built agents around claims, around authorizations, around eligibility. And I think over the last probably year or 2, we’ve been building a proprietary suite of more clinical-facing tools such as risk stratification, care management, workflow orchestration, and identification that I think will lead to MLR improvements over time.

And you can see that a little bit as we — maybe getting a little off topic here. But you can see that a little bit with how Prospect has performed as we continue to onboard them onto the Astrana operating system. Prospect, prior to the acquisition had medical cost trend running 6%, 6.5% or so. We modeled around 50 basis points of improvement in 2026 versus that number. And we’re outperforming that by a bit here even in Q1, even though we’ve already improved by that 50 basis point margin. So I think you’ll really start to see even more MLR improvement in the medium term. But I would say the improvement is largely skewed towards G&A at this point in time.

Operator: Our next question is from Craig Jones with Bank of America.

Craig Jones: So Brandon, I want to follow-up on your comments around your encounter-based risk adjustment model MA. So it sounds like CMS keeps mentioning like leveling the playing field in MA and really wants to rewrite the current MA risk adjustment model. So if you were in the room with them redoing the risk adjustment model, what would you recommend changing? And then how do you think the potential changes end up making potentially going to this encounter-based model would help Astrana? And then do you think you could see something along these lines as soon as the 2028 technical notices fall?

Brandon Sim: Thanks for the question. Yes. I think the future of risk adjustment is really interesting. As you can see in the ACO lead preliminary model details. There is the phasing in of an AI inferred risk score, which would depend not necessarily on an organization’s ability to document and submit codes, but rather trying to use AI to infer the true acuity of the member and reimbursing appropriately based on that kind of “gold standard” kind of determination of a member’s risk. Again, I think, ultimately, because we’ve been conservative on risk adjustment, because we see members over a longitudinal period of time and we try to be very appropriate in terms of capturing the clinical complexity of the population. We think that either way, we’re well structured, we’re well positioned for that future.

We think that because we haven’t relied on documentation or coding optimization to generate savings and value for the health care system in the past, it may even be beneficial for us, for example, to have a true determination of what a patient’s risk is via AI that the government or CMS is going to determine rather than everyone playing a game to try to improve their risk scores over time on a relative basis. So I think really, regardless of how all that shakes out, we think we’re structurally well positioned for the long-term. That being said, if I had my way, I do think that the — that risk adjustment as a source of alpha is not really, I think, in the benefit of the health care ecosystem in the long-term and for the Medicare Trust fund in the long-term.

So I would recommend without knowing more that some of these approaches that are being suggested like AI inferred risk models seem very appropriate and seem like a much more efficient way to standardize what risk determination looks like across the American population.

Operator: Our next question is from Michael Ha with Baird.

Michael Ha: So when it comes to AI, clearly, everyone is talking about it this earnings season, all the large national payers, providers. But the thing is most of them have pretty legacy old infrastructure, fragmented data, as you said yourself. So when I think about Astrana versus, I guess, almost all of your peers. It’s the fact that you built an AI-native tech platform many years ago. And the fact that you yourself are spearheading foreseeing AI adoption across basically every facet of your company. I think that’s still widely underappreciated. So I was wondering if you could talk more about this specifically, the structural differences between you Astrana versus your peers when it comes to unlocking the power of AI? In other words, like what still has to happen — what still has to be done by your peers to get there versus what can already start to happen at Astrana?

Brandon Sim: Yes. Thanks so much for the comments, Michael. I think broadly, that’s right. I think our thesis has always been building internally. And I think that thesis is being rewarded in an era where it is easier than ever and faster than ever to build internally because of the advent of generative AI and its use in coding. And I think as long as you have the integrated data infrastructure to support that, the ontology is on top of that, the definitions, the concepts and the relational — and the relationships between those concepts so that the AI understands how to operate on each of these concepts and how they relate to each other and how they ultimately translate into actionable insights. I think that’s hard to replicate, right?

I think if you’re operating a system where you’ve acquired a bunch of stuff. And you haven’t integrated them into a unified data layer with a unified set of concepts and vocabulary on top of that, on top of which the AI and the agents can operate. You’re going to find it very difficult to kind of build the fifth floor of the building without having the structural supports in the ground floor and the lobby built out. And I think that’s a lot of what our peers are doing perhaps without getting too much into what our peers are doing. I think there’s a rush to chase the kind of the sexiest parts of AI to build the top floor, the penthouse unit without having the foundational approach, without having the pick axis, the knowledge about how to dig the hole and the foundation into the ground to build that in an effective manner.

And I think we’ve spent a lot of time, myself personally, given my engineering background to build out that foundation. And now we think that’s going to unlock our business in terms of rapidly adopting AI across the enterprise and embedding it deeply into each and every workflow, both operationally, clinically and on the quality of care side. So we’re really excited about where we can take this platform. We’re already seeing the G&A improvements. We’re starting to see some of the trend improvements as we continue to integrate new businesses onto the platform. And we’re seeing great success as well in terms of our Care Enablement business, selling the tools, and the integrated workflow that we’ve built to other provider groups and helping them succeed also in an accountable care relationship.

Michael Ha: So next question on the final MA rate notice. So I’m getting roughly like 4% net rate increase for Astrana if I exclude — on the chart reviews. And when I think about Astrana’s margin expansion, just how sensitive it is to the rate environment? I know your cost trends are running, I think, 4% to 5% roughly for MA. So at face value, right, that would imply rates are basically they match up. But if I start at 4%, add maybe 1% to 2% coding, maybe another 1% to 2% help from plans, benefit design pricing. Then we’re getting into a different sort of ballpark of 6% to 9% rate versus trend of 4% to 5%, up to 400 basis points of margin expansion. So it feels quite considerable. And that’s not even including right, your regular cohort maturation dynamics, any other trend vendors or G&A. So at a high level, am I missing any major components? Is this even the right way to start thinking about 2027?

Brandon Sim: Yes. Mike, as always, your math is great. So I would broadly agree with your comments. I think the final rate notice was constructive overall. And the overall top line kind of effective growth rate of 5.33% does more appropriately reflect underlying medical cost trend. As you mentioned, the disallowed diagnosis — or the diagnoses, sorry, are expected to be immaterial for Astrana, given our historically conservative and encounter-based approach to risk adjustment. So as you had noted correctly, the average change for us might be the 2.48% plus the 1.53% or approximately 4% in aggregate. And as we think about ’27 more broadly in our models, at our current RAF levels, we probably expect to maintain MA margins consistent with 2026 with that 4% kind of average rate book increase.

Beyond that, we continue to see tailwinds and opportunity for more accurately capturing the complexity of our populations and risk adjustments. And there’s potential tailwinds above and beyond the 4% from those sources.

Operator: Our next question is from David Larsen with BTIG.

David Larsen: Congratulations on the great quarter. Can you talk a bit about your margins for like, I guess, full cap books of business that would include inpatient? And can you remind me what regions or how many members are full cap, including pharmacy, doc, inpatient?

Brandon Sim: Thanks for the question, Dave. Thanks for tuning in. Our fully capitated arrangements start off in lower kind of EBITDA margin arrangements as we transition them from full risk because as we talked about before, you get the kind of increase in percentage of premium without yet necessarily flowing through the decrease in inpatient utilization as we take on additional portions of the risk dollar. Over time, the maturation of the full risk cohorts, as we’ve seen over the past years as we’ve moved members cohort at a time into full risk arrangements as we continue to do that as we did in Q1 of this year. You see that margin profile mature and ultimately get to hopefully a similar point as the kind of partial risk members as well.

So I think that’s what we expect as we continue to move members selectively and prudently into full risk arrangements. We underwrite kind of this margin maturation cycle. We’ve seen that happen now over several years. And each of those has matured as expected. And so we can kind of space out our membership moving into full risk as appropriate. I do want to mention that almost all of our full risk arrangements do not include Part D as in dog risk. So there are a handful that do and most of them do not. In terms of the geographies where we are full risk, it really varies. Most of our membership, 80% of the revenue approximately comes from California. So I would say still that California does have a large percentage, a majority of the full risk members.

However, we have moved over 14,000 Medicare Advantage members into a full risk delegated construct arrangement with a payer partner, for example, in Texas in the first quarter of this year. So — and we also have full risk delegated contracts in Nevada. And of course, the ACO REACH business is in some aspects, the full risk business also. So we’re really in the business of properly underwriting and then appropriately and proactively reducing the cost of care for our populations and then making sure that our financial contractual arrangements are conducive to us capturing some of the value that we’re generating for our patients and for our communities over time.

David Larsen: And then for Prospect, I think you may have mentioned this earlier. Is it still $80 million of EBITDA? Is that on track?

Brandon Sim: Yes, that’s right, Dave. Prospect was on track for around $80 million of adjusted EBITDA on an annualized basis. And at this point in time, it is currently tracking a bit ahead of those expectations.

David Larsen: And then just one quick one. It looks like your stock has been doing really well over the past couple of months. I guess what do you attribute that to just at a high level? What?

Brandon Sim: Sure. I mean we’re always happy to see that as it’s our job to continue generating shareholder value, of course. I think I hope it’s a continued recognition of our leadership and our consistency and stability of our model. The differentiation of our technology platform, the 35% revenue CAGR, the 25% adjusted EBITDA CAGR that I mentioned, which I think is fairly unheard of in health care services over a very long period of time, over 7 years. And ultimately, of course, definitely helps that there’s been — there have been regulatory tailwinds, including the adjustment, the more appropriate, in our view, 2027 Medicare Advantage final rate notice. So I think overall, a lot of positive kind of macro tailwinds lining up and hopefully, an increased recognition of the unique platform that we’ve built that has really generated free cash flow, profitability, and now rapid growth for over 3 decades.

David Larsen: The best health care is when you don’t actually have to see your doctor. And that’s the model that you guys have created. So nice quarter.

Operator: Our next question is from Ryan Langston with TD Cowen.

Christian Borgmeyer: This is Christian Borgmeyer on for Ryan. So looking at the second quarter guidance and EBITDA margin, how should we think about puts and takes within the cost of service revenue and G&A lines? For example, any seasonal considerations within medical utilization, in particular that are different this year? Or on the G&A side, any sequential savings from AI or Prospect synergies embedded in that?

Chan Basho: In terms of our 2026 guide, probably the best way to think about this is in the first half of the year, we’re probably going to see a little over 50% of profitability coming in consistent with what’s happened in historical years. As you think about puts and takes, the puts and takes, as we mentioned, it’s around HQAF. It’s around opportunities with MA and ACO as well as watching in terms of what’s going to happen around the Medicaid membership trend.

Brandon Sim: And then maybe to answer the other part of your question. We didn’t see any abnormal utilization necessarily in Q1. I know there’s been talk about weather and the flu season and whether that was heavier or lighter. I don’t think things came in pretty operationally clean is how I characterized the quarter and tracked pretty consistently both on the inpatient and outpatient side with the broader medical cost trends that we reported across the business, even drilling down into each line of business as I started off the Q&A session with. So we felt pretty comfortable this quarter. And we’re maintaining guidance primarily because we want to take a disciplined and conservative approach early in the year here.

Christian Borgmeyer: I actually had a quick balance sheet question actually. I see the accounts receivable balance and the medical liabilities balance are each up like $90 million to $100 million sequentially. Anything to call out there related to any one item or program in particular? Or is that the full risk conversions contributing to that?

Chan Basho: Yes, that’s the full risk conversion that you’re seeing in Q1.

Operator: Our next question is from Gene Mannheimer with Freedom Capital Markets.

Eugene Mannheimer: Congrats on a good start to 2026. So coming in or tracking at the high end of cost synergies with Prospect. Can you discuss potential revenue synergies there and when you may start to see that realized? And my follow-up would be on the MLR trends at or better than the 5.2% or so that you called out. Did you or can you break that out across the legacy Astrana and the Prospect book?

Brandon Sim: Hello, Eugene, thanks for calling in. Sure thing. So on the revenue synergies, we haven’t — those are not included in the $12 million to $15 million synergy range. So we haven’t quantified that yet. But we do expect over time that our partners and our providers and ultimately, our members will see the value of our denser network and our ability to drive access to care, high-quality care in a faster way because of the larger network that we now have. So over time, we do think that, that value will be realized by the platform, but we haven’t yet quantified necessarily what that looks like. On the trend item, I would say that, as I mentioned before, Prospect came in 6% to 6.5%, 6.2%, 6.3% trend prior to the acquisition.

And we are underwriting a 50 basis point improvement in that trend year-over-year. Our overall trend for the year is around 5.2% on a consolidated basis. And I would say that both core Astrana as well as legacy Prospect businesses performed better than expected here in Q1. Again, it’s still early on in the year. We don’t have perfect visibility, claims visibility yet, for example, on March. So we wanted to be conservative, but things are tracking well here to start the first quarter.

Operator: Our next question is from Matthew Gillmor with KeyBanc Capital Markets.

Matthew Gillmor: I wanted to follow-up on the full risk contract transition discussion. I think this quarter, you had 40% of members in full risk. I think last quarter, you set an expectation of 36 members in full risk as of the first quarter, so maybe a little bit ahead of schedule. I wanted to see if I had those numbers right and then just get an update in terms of how you’re thinking about the pacing of members moving to full risk over the course of the year.

Brandon Sim: Hello, thanks for the question. Yes, I think that’s approximately right. Around 40% of our members are in full risk arrangements. And that translates into around 80% of our care partners revenue being — coming from full risk arrangements. And of course, that’s because the percentage of premium that we receive in the full risk arrangements per member is obviously higher than the partial risk arrangements. So I think you’re continuing to see that the percentage of both membership and revenue continue to grow. In Q1, this took a step up because of the forwards contracts that we had started in the first quarter as we had guided to late last year. And all of those have now been completed. And so that’s what’s led to the spike here in Q1.

On a go-forward basis, I think in our supplemental presentation deck. We did note that we do expect continued growth in the percentage of full risk members. And we’ll be phasing that in over time kind of on a regular course basis.

Matthew Gillmor: And as a follow-up, we’ve been particularly interested in your ability to bring this delegated model into new markets. And so the news out of Texas that you’ve updated us on has certainly been encouraging. I did want to take your temperature in terms of expanding a delegated model either into new markets or even just new states or even just new markets within states like Texas, which many of those places traditionally haven’t had fully delegated models.

Brandon Sim: Yes, it’s a great point. And you’re absolutely right. A lot of parts of the country have not necessarily operated in a — don’t even mention delegated model. They haven’t even operated really in a value-based care setting in a broad way. And so we recognized the challenges of kind of gone 0 to 1 in a very short period of time. And I think that’s why we’ve been really working on a gradated kind of approach to helping providers and payers along as we continue to take the Astrana delegated model outside of California, outside of Nevada, outside of Texas, and through the rest of the country. And what that looks like really is, first, entering into partial risk arrangements, ensuring that the data feeds that we need are on the ground and ready to go.

Ensuring that our relationship with our downstream providers, primary care specialists and even hospitals are strong. Ensuring that our technology platform is integrated directly into the workflow of those providers. And ensuring that kind of our care management orchestration is in place and kind of allowing the economic contractual relationships to kind of follow behind the wake of the operational changes that we’re making in terms of how health care is delivered in these new states and/or geographies. So it is a created kind of stepwise approach to getting folks into the validated model. We think that it ultimately will win out because, frankly, at the end of the day, it’s just a more efficient model. It’s a more valuable model to the health care system and a more efficient one for both payers and kind of the overall system.

So we think that logic will take the day here and the economics of it will take the day. But we do recognize that change management takes time. And we’re prepared to and have engaged in Florida, I mean, sorry, in Texas and Nevada, for example, on that path forward step by step.

Operator: Our final question is from Andrew Mok with Barclays.

Thomas Walsh: This is Thomas Walsh on for Andrew. You shared that acuity in the Medicaid population remains favorable in part due to your longitudinal patient relationships. Can you help us understand how those patient relationships mitigated the acuity impact in practice? And are there any other factors on the mix of members disenrolling or otherwise that mitigated the adverse selection?

Brandon Sim: Yes, definitely. I think what I was alluding to, to put it more clearly is that the patients that tend to be Astrana members, which — and the patient attribution mechanism is the choice. The selection of a primary care provider who is an Astrana primary care provider. The members that tend to be attributed to us, tend to not be the members who have low to no utilization because they are almost making an active choice to be an Astrana member and to be engaged in our care model and to be engaged in the longitudinal nature of the care model that we have with our patients as we follow them, for example, across line of business. I mentioned before, the example during COVID, members who lost their jobs and had to switch from a commercial line of business to commercial insurer to potentially something on the exchanges or something on Medicaid, for example, could continue to be in the Astrana ecosystem, continue to have the same care management, continue to see their same PCP and same specialist network.

Those are the benefits, I think, of being in the Astrana network. And that tends to insulate us or we think partly insulate us from the level of adverse selection we expected from the disenrollment of members and kind of potentially their lower MLR. That didn’t end up playing out the way that — or to the degree that we thought it would. And that’s what’s led to some of the improved acuity in the Medicaid population.

Thomas Walsh: And following up on ACA attrition tracking better than expectations, similar to trends across the industry. Could you share the actual disenrollment you experienced there? And at what point in the year would you expect to have enough visibility to make an informed revision to the full year membership expectation?

Brandon Sim: Yes. I think at the beginning of the year, embedded in our guidance, we thought it would be a 30% to 40% disenrollment number in exchange throughout the year. And I think we had quantified that at a kind of mid-single-digit EBITDA impact headwind, of course. What we’re seeing so far this year is not quite the 30% to 40%, really closer to high single-digit attrition in the ACA population. It still is early. We’re still in May here. And there could be further disenrollments after the 90-day grace period. But across the industry, as we think about projections for actuarial firms and what others have been saying as well as our own experience. We’re now projecting a decline of, call it, 20% to 30% instead of 30% to 40% internally at least. Again, we haven’t reflected that in the guidance yet of conservatism, but that’s kind of the quantum of the numbers we’re talking about.

Operator: There are no further questions at this time. I would like to turn the conference back over to management for closing remarks.

Brandon Sim: Thank you, everyone, for joining our call today. We appreciate your time. And we hope to see you in the near future at one of our — one of several conferences we’ll be attending or we can catch up at any time if you e-mail investors@astranahealth.com. Again, thank you so much for joining and have a great evening.

Operator: Thank you. This will conclude today’s conference. You may disconnect at this time. And thank you for your participation.

Receive real-time insider trading and news alerts