Thank you Hadi Javeed, Mike Logan, MD, Rachel, and many others for tuning into my live video with Sarah Habeeb! Join me for my next live video in the app.
🗓️ TEAM Connect Virtual Summit — April 30, 2026 A full-day virtual event for hospital leaders navigating TEAM implementation. Speakers from Mass General Brigham, CommonSpirit, AdventHealth, Duke-Margolis, and more. I’ll be presenting on technology-driven care coordination. Early bird registration closes March 20th. → Reserve your spot at apmconnect.com/virtual-summit
Sarah Habeeb is System Director for Medicare Value-Based Care Products at Baylor Scott & White Health and co-founder of APM Connect. Connect with her on LinkedIn or at
Dr. Christian Pean is an orthopedic trauma surgeon and faculty member at Duke University School of Medicine, core faculty at the Duke-Margolis Institute for Health Policy, and CEO and Co-Founder of RevelAi Health. He writes Techy Surgeon at the intersection of clinical AI, health policy, and care coordination. You can find him at the TEAM Connect Virtual Summit on April 30, 2026.
Sarah Habeeb started pre-med at Texas A&M. Then she drove an ambulance, decided she didn’t want to touch patients, and pivoted into health administration. She was sitting in a grad school classroom around 2014 when someone started explaining the Medicare Shared Savings Program—a program that was, at the time, so new that almost no one understood it. She listened, thought it made intuitive sense, and made a career bet on it.
That bet paid off. Today, she oversees a program that generates roughly $110 million in annual savings against a CMS benchmark, retaining about 75 cents of every dollar through MSSP’s Enhanced Track. That’s approximately $77 million flowing back out to physicians annually, for a program covering 120,000–125,000 Medicare beneficiaries across an entire major health system.
She also co-founded APM Connect—a free community for hospitals navigating mandatory payment models—and will be speaking at the TEAM Connect Virtual Summit on April 30, 2026. I’ll be there too. If you’re a hospital leader trying to make sense of what TEAM actually requires in practice, this event is where you want to be.
But before the summit, here’s what an hour with Sarah Habeeb taught me about what value-based care actually looks like from the inside—and what that means for every clinician and operator who is about to be pulled into it, whether they’re ready or not.
The Mechanics Most Clinicians Never Learn
Risk, in the value-based care sense, is a word that gets thrown around like clinicians should already know what it means. Most don’t. Here’s the actual structure:
When you enter a total cost of care contract, a payer assigns you a benchmark—an expected spend per member, per year, risk-adjusted based on patient complexity. If you keep costs under that benchmark, you share in the savings. If you don’t, you may owe back a portion of the overage. The HCC risk adjustment model is how CMS calibrates those benchmarks—it’s essentially a formula that assigns a risk score to each patient based on documented diagnoses. A patient with diabetes, COPD, and heart failure carries a higher score than one with no documented conditions, so their expected cost is set higher.
This is where documentation integrity enters the picture. Sarah is direct about it: “If you’re a cardiologist in the heart failure cohort, you need to be sure that you’re getting credit for the risk of your patients because that directly affects your benchmark, which then directly affects your performance in the contract, which affects your Part B adjustments.”
Physicians often experience risk coding conversations as administrative irritation—another box to check, another form to sign. The translation layer is missing. If your patients are genuinely sicker than their documented diagnoses suggest, you’re being benchmarked against a population that looks healthier than yours. The comparison doesn’t hold. You look like you’re mismanaging costs when you’re actually managing a high-acuity panel with inadequate documentation.
The fix isn’t gaming the system. It’s accuracy. And it starts with clinicians understanding why it matters.
What “Operator” Actually Means on a Monday Morning
Baylor Scott & White’s ACO is structured around what Sarah calls product owners—people responsible for specific contracts (MSSP, Medicare Advantage risk agreements, direct-to-employer deals) who identify their contract’s cost drivers and design remediation plans. Care management, quality teams, and marketing function as internal vendors to those product owners, not parallel departments chasing their own initiative lists.
This sounds obvious until you’ve seen the alternative, which is how most health systems actually operate: multiple teams, multiple initiatives, no clear accountability, and no way to know what’s actually moving the needle.
“Because if everybody’s working on a bunch of things and we’re not talking to each other,” Sarah explains, “you can’t figure out what actually made the difference.”
The answer for Baylor was forced prioritization. Pick three initiatives. Measure them monthly. Make every stakeholder meeting about those three things. It’s not sophisticated—but discipline consistently outperforms sophistication in operations.
The Inpatient Rehab Problem (Which Is Probably Your Problem Too)
If you work in orthopedics, you already know what Sarah is about to say. If you don’t, here’s the version that will make you understand it.
At Baylor Scott & White, inpatient rehab facility utilization is—in Sarah’s words—”completely unmanaged.” Against Milliman benchmarks, they’re over-utilizing inpatient rehab while simultaneously running near-zero skilled nursing facility use. They’re tracking 55% of hip fracture patients going to inpatient rehab, with a 30-something percent readmission rate that isn’t meaningfully better than SNF. “So was that the right decision? Were they even ready to discharge from the hospital?” she asks. And the honest answer is: often, no.
This is the central tension in any ACO that’s embedded in a health system with joint ventures. The health system may have financial interests in keeping patients flowing through high-cost post-acute settings. The ACO’s job is to reduce unnecessary utilization of those same settings. Getting alignment between those two forces is genuinely hard. Sarah doesn’t pretend otherwise. “You have to find a way.”
What Baylor has built: a six-to-eight-person team of post-acute care nurses who follow ACO patients through preferred SNFs, with direct EMR access (a requirement of network participation), weekly interdisciplinary calls with each facility, expected discharge dates set within seven days, and a target average stay under 28 days. They track return-to-acute rates and share scorecards with SNF partners quarterly. It’s brute-force infrastructure, but it’s working better than the alternative—which is patients in a black box beyond hospital discharge.
From the orthopedic side, I’ve been doing telephone visits at two weeks for all my hip fracture patients, closing them with absorbable sutures so the visit doesn’t require a trip in. Half the time, I’m chasing down whether they’re still in the SNF, whether they’ve bounced back to a different ED, whether anyone has even talked to them since discharge. The clinical relationship doesn’t just end at the door of the facility. But the information infrastructure does.
TEAM, ASM, and the Art of Not Panicking
Here’s Sarah’s read on the two mandatory models that are consuming health system bandwidth right now:
On TEAM: Most hospitals are in year one, upside only, and either don’t know they’re in it or aren’t taking it seriously. This is a mistake. The regional benchmarking structure means your performance is being measured against peers in your geography. “If your regional peers are paying attention and you’re not,” Sarah says, “that affects your benchmark.” By the time year two arrives with downside risk up to 20%, you’ll be starting from behind, not from neutral. Only four hospitals voluntarily opted into TEAM early from prior CJR participation—which tells you something about the expected economics. But ignoring it isn’t a viable option.
On ASM: CMS’s Ambulatory Specialty Model is designed as physician-level accountability—NPI-specific participation, measuring cardiologists on heart failure costs and a range of specialists on low back pain. The design is conceptually provocative: specialists competing against each other within the same market for Part B adjustments. The implementation, however, came in lighter than expected. At Baylor, the initial projection was 200–400 physicians selected. The actual list: 51. Nationally, organizations are reporting five or six physicians per system. The cost of the required quality reporting infrastructure may exceed the penalty exposure for smaller lists.
Baylor’s response: build a shadow bundle internally. Treat ASM as an MSSP workstream. Develop heart failure and low back pain strategies that produce dividends in 2027 regardless of what the formal adjustment looks like. It’s the right call. These conditions were selected because they represent genuine opportunities to improve care and reduce low-value utilization—guideline-directed medical therapy for heart failure, fewer unnecessary MRIs and high-risk opioid prescriptions for low back pain. The model may be imperfect, but the clinical direction is sound.
Where AI Actually Fits—and Where It Doesn’t
Sarah told me upfront: she doesn’t spend much of her day thinking about AI. Her ACO generates north of $100 million in savings the old-fashioned way, through data discipline, care management operations, and physician engagement. “I try not to [use AI],” she said. “I’m like, don’t be the reason we need to build data centers.”
That grounded skepticism makes her view of where AI could help more credible, not less.
Her highest-value use case: discharge disposition prediction. A tool that could tell a case manager at the bedside—in real time—that this specific hip fracture patient, given their comorbidities and functional baseline, has a 38% readmission rate if sent to inpatient rehab versus a projected outcome curve if they went home with home health instead. “That sounds like magic to me. If that can be populated, that would be really helpful.” The connection between value-based care economics and bedside decision-making is currently made by humans who don’t have time to make it well.
Her second use case: home health appropriateness review. Some patients are on home health for two years. The authorization review documents run ten pages of wall-to-wall all-caps text. Physicians batch-sign them without time to read them. AI that could synthesize that documentation into a structured summary—flagging patients who are likely no longer appropriate for the benefit—and send it to the physician for a judgment call, not a rubber stamp, would be genuinely valuable. “There needs to be some type of tool rather than a pre-review.”
Both use cases share a structure: AI doing the parsing, pattern recognition, and synthesis; humans making the clinical judgment call. Augmented intelligence, as she put it. Not a replacement for the nurse who calls the SNF. A better brief before the nurse calls.
The Summit You Should Know About
APM Connect—which Sarah co-founded with two former Common Spirit colleagues—has built the closest thing to a peer community for TEAM participants. It’s free for hospitals and clinicians. About 200 of the 729 mandatory TEAM hospitals have already joined. The resources are real: best-practice webinars, forums, a participant community, and expertise Sarah and her co-founders have spent years developing.
On April 30, 2026, APM Connect is hosting the TEAM Connect Virtual Summit—a full-day virtual event for hospital leaders working through TEAM implementation. The speaker roster includes colleagues from Mass General Brigham, CommonSpirit, AdventHealth, Duke-Margolis, and Bass, Berry & Sims. I’ll be presenting as well, focused on technology-driven approaches to TEAM care coordination.
Early bird registration closes March 20th. TEAM Connect members register at $250; non-members at $375. Group rates available for teams of three or more.
If your hospital is in TEAM and you haven’t found your footing yet, this is a low-cost way to learn from people who are further along than you. The fact that only four hospitals voluntarily opted in early tells you something about the economics. But it also tells you that most of the 725 hospitals that didn’t raise their hand are now navigating this without a roadmap. APM Connect is trying to build that roadmap in public. Register here.
What $110M Means in Context
Before getting into ACCESS strategy, it’s worth grounding the Baylor Scott & White number in the broader MSSP landscape—because “top-performing ACO” is a phrase that gets used loosely, and the data tells a more interesting story than the label suggests.
See Data Visualization of ACO savings here
In 2024, the MSSP delivered $6.6 billion in gross savings—a record, and more than 20 times what the program generated in its first full year. Of 476 participating ACOs, 75% earned performance payments totaling $4.1 billion, while Medicare netted $2.4 billion in savings relative to benchmarks—the highest share of ACOs generating savings since the program’s inception.
Here’s what makes BSW unusual: the median ACO generated roughly $8M in earned savings in 2024; top-quartile performers averaged closer to $20–25M. BSW’s ~$110M in gross savings against benchmark places it in a category occupied by fewer than a handful of systems nationally. That’s not a function of luck or favorable geography. It’s 10-plus years of compound operational discipline—the product owner model, the disciplined SNF network, the monthly initiative tracking—applied consistently at scale.
ACOs that entered the program in 2012 or 2013 and sustained participation earned an average of $7.6 million net to the Medicare Trust Fund—versus $4.6 million for ACOs that started in 2019 or 2020. Time in program matters. There is no shortcut. Which is precisely why hospitals in TEAM that are treating year one as a grace period are setting themselves up poorly for year two.
Two-sided risk is the engine: ACOs in Level E and Enhanced tracks generated more than two-thirds of all MSSP savings in 2024 ($5.4 billion of the $6.6 billion total) despite representing a minority of total ACO participation. The lesson is structural: downside risk changes behavior in ways upside-only contracts don’t. TEAM’s mandatory downside exposure beginning in year two isn’t punitive policy—it’s the mechanism that produces results.
The Data Pipeline Problem
One of the most practically useful things Sarah said, almost as an aside, is that the data infrastructure underneath a high-performing ACO is invisible to most clinicians and is harder to build than the clinical programs sitting on top of it.
At BSW, claims data comes in through CCLF files from CMS and through bespoke payer feeds for Medicare Advantage contracts. A dedicated analytics team ingests all of it into a centralized repository, then layers in a vendor like Milliman for external benchmarking—so Sarah can see not just how BSW is performing against its own benchmark, but how it compares to peers nationally. The result is the “contracting grid” she described: a single view of every risk contract, every provision, every performance lever, with monthly measurement against the initiatives that matter.
What this requires: a data team that speaks both ACO contract language and engineering. A vendor relationship for external benchmarking. A governance model that connects product owners to that data without creating a meeting-intensive bureaucracy that slows down action. And institutional patience—because the ROI on this infrastructure takes two to three years to materialize.
For smaller hospitals entering TEAM for the first time, almost none of this exists. They’re getting quarterly CMS files, trying to reconcile them with what’s in their EHR, and hoping someone has time to run a query. The gap between those organizations and BSW isn’t primarily clinical. It’s informational.
The ACO-ACCESS Playbook: What Operators Need to Know
The CMS ACCESS Model generated an underappreciated moment in health system boardrooms. As Sarah put it, her health system went through “the five stages of grief once they figured out that they could not apply as an ACCESS participant themselves.”
That reaction reveals something structurally important. Health systems are excluded from ACCESS participation because their financial incentives—keeping patients within their network, returning them for follow-up care—run counter to what the model is designed to do. An ACO, by contrast, is measured on keeping patients out of high-cost settings. These are genuinely different businesses, even when they share a tax ID.
For ACOs considering whether to engage with ACCESS participants proactively, three operating principles apply:
Move before the OAPs hit your total cost of care. Outcome Aligned Payments(OAP) from the ACCESS Service will count toward total cost of care calculations for attributed ACO patients after the grace period. The model is designed so that ACCESS participants substitute for low-value care such as imaging, injections, and certain surgeries that conservative management makes unnecessary. If it works as intended, OAP reimbursement costs should be more than offset by reductions in downstream utilization. But an ACO that hasn’t established a measurement relationship with an ACCESS participant before those services start appearing in claims data is flying blind. The time to start measuring is now, before the formal performance period creates noise in the baseline.
Come with validated outcomes, not a vendor ROI deck. ACOs have actuaries. They have their own analytical infrastructure. The last thing a Sarah Habeeb wants to receive is a projection her team has to spend two weeks independently validating. ACCESS participants approaching ACOs should arrive with outcomes already assessed by a credible third party—something like a Validation Institute report—rather than a proprietary model the ACO has no way to audit. Shift the burden of proof to the participant, where it belongs.
Understand what “substitute spend adjustment” means for your ACO partner. The ACCESS model financially penalizes participants who generate OAP utilization without a corresponding reduction in what CMS calls “substitute spend”—the downstream care that better conservative management was supposed to prevent. ACOs have a direct stake in this math. An ACCESS participant whose patients continue to receive the imaging, injections, or surgeries they would have gotten anyway is a cost increase to the ACO, not a savings lever. The right question for any ACO evaluating an ACCESS partnership: can this organization demonstrate that their model bends the utilization curve downstream? And can they do it with real claims data, not projections?
The structural alignment between ACOs and ACCESS participants is genuine. But it only materializes if the ACCESS participant can prove it.
The Physician Champion Arbitrage
One more thing Sarah said that deserves its own heading: if you’re a clinician who wants to have influence in value-based care, the path is easier than most people think, because there’s a scarcity of engaged physicians on the ACO side.
“If you are going to be the physician in value-based care that’s engaged, you’re going to have a best friend in the ACO who are always looking for this—the physician that gets it, they’re engaged. And all of our ideas are going to be filtered through that person before we go take it to this broad group.”
ACOs run on data and initiatives, but initiatives require physician behavior change, and behavior change requires trust. The ACO administrator who has a cardiologist champion who actually reads the quality reports is in a fundamentally different position than the one who doesn’t. If you’re an orthopedic surgeon, a cardiologist, or a low back pain specialist in a TEAM or ASM hospital, you have more leverage than you probably realize, and the ACO on the other end of that relationship is actively looking for you.
Dr. Christian Pean is an orthopedic trauma surgeon and faculty member at Duke University School of Medicine, core faculty at the Duke-Margolis Institute for Health Policy, and CEO and Co-Founder of RevelAi Health. He writes Techy Surgeon at the intersection of clinical AI, health policy, and care coordination. Find him at the TEAM Connect Virtual Summit on April 30, 2026.
Sarah Habeeb is System Director for Medicare Value-Based Care Products at Baylor Scott & White Health and co-founder of APM Connect. Connect on LinkedIn or at apmconnect.com.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit techysurgeon.substack.com/subscribe
