What actually works, what’s getting commoditized, and the governance gap most organizations bought technology without closing
Two years ago AI scribes were a breakthrough; today new physicians ask in job interviews whether a health system has them. That commoditization is both proof that AI can penetrate clinical workflows and a warning that the easy wins are already priced in. The clinicians and operators who’ve navigated the gap between a promising model and a functioning clinical tool were specific about what separates the two — and almost none of it is about the technology.
Bob Wachter, Chair of Medicine at UCSF and author of A Giant Leap, described the adoption curve precisely: two years ago essentially no large health system had AI scribes; today they’re a near-universal price of admission, with new physicians asking “where are your scribes?” as a condition of employment. That curve proves AI can penetrate clinical workflows at scale.
But the ROI math resists easy answers. Wachter is candid that scribes “don’t save us that much time” — because the physician is now actually looking the patient in the eye and having a human conversation, which takes slightly longer than ignoring them while typing. The measurable value lands in physician satisfaction, recruitment, and retention — real but hard to put in a spreadsheet. Health systems that bought scribes as a cost play are likely disappointed; those that bought them as a physician-experience play are not. And the vendors who built the category have watched prices plummet to commoditization. The lesson: the next AI wins won’t come from the obvious, easily-copied tools.
Pelu Tran, CEO of Ferrum Health, surfaced the single most actionable finding in the AI conversation: 81% of FDA-cleared AI solutions perform more poorly in real-world hospital settings than their FDA submission data indicated. FDA clearance tells you a model worked on a specific study population, with specific scanners and protocols — not how it will perform on your patients, your equipment, your workflows. As Tran put it, even something as mundane as a hospital’s hanging protocols can change a model’s performance.
The motivation behind Ferrum is not abstract: Tran’s uncle died from lung cancer that was missed over several years by physicians who didn’t have, or didn’t use, AI tools already available on the market that would have cost pennies. The clinical cost of slow and ungoverned AI adoption is measured in missed diagnoses. The organizations doing this right validate models on their own populations before deployment, monitor performance after go-live, and govern the full lifecycle — rather than sending patient data to each of 1,000+ point vendors.
Ali Parsa — founder of Babylon Health and now Quadrivia AI — framed the structural problem more precisely than most: healthcare uniquely combines elastic demand (parents will spend anything for a sick child) with constrained supply (finite clinicians, more retiring than entering, universities can’t produce them fast enough). “If you have that in any sector, you will have inaccessibility and unaffordability.” This is a supply problem at root, not a funding or policy problem.
Parsa put a number on the waste that follows: the US spends roughly $2.5 trillion a year on clinicians, and 20-30% of that goes to highly repetitive tasks AI can already do at far lower cost. That’s on the order of $600 billion a year spent on work that doesn’t require a human clinician — while patients wait for the clinician who’s busy doing it. Crucially, his vision isn’t a chatbot running a fixed script: he distinguishes real generative AI (which can hold an open clinical conversation and decide whether a patient with a headache needs to be seen today) from “if-this-then-that with a bit of AI on top.” The former adds clinical judgment; the latter just completes a task.
Parsa is also disciplined about the limits: AI cannot represent a nurse or doctor today. His calibration on accuracy is worth holding onto — the best AI scribes are ~94% accurate against a ~96% human baseline; the best AI summaries are 80-85% accurate; AI coding produces “50% and then the rest is garbage that humans need to fix.” Co-pilot, not autopilot.
Two guests independently diagnosed the same failure. Wachter called EMRs “foundational but insufficient” — they turned physicians into “expensive data entry clerks” and added 1.5-2 hours of daily inbox work through patient portals alone. Glen Tullman, founder of Livongo and former CEO of Allscripts, was blunter: Epic’s EMR was sold to hospital boards and CEOs on billing and integration, not to the users — so bad that health systems hired human scribes just to enter data into it.
The through-line: EMRs were digitization sold as transformation, optimized for the buyer rather than the user. You need data in digital form to do anything that resembles transformation — but the EMR was the prerequisite, not the destination. The AI era risks the identical mistake: tools bought by boards and CIOs on productivity promises, without clinicians leading deployment, without validation on local populations, without lifecycle governance. The organizations that will get durable value learned the EMR lesson — start with the clinical problem, involve the clinicians, validate on your population, govern before you scale.
Some of the most concrete AI value described across these conversations was unglamorous. Joe Corkery’s Jaide Health is building “Google Translate for healthcare” — real-time AI medical interpretation for the 25-26 million Americans (about 8% of the population) with limited English proficiency, a group with measurably higher readmission rates, longer length of stay, and poorer adherence directly attributable to communication gaps.
Marc Paradis at Northwell described Optain — a retina-scanning camera the size of Star Wars binoculars that any untrained person can operate, with on-device AI (no cloud round-trip) assessing diabetic retinopathy, cardiovascular, and neurological risk, deployable in primary care offices. And Tom Andriola, formerly CDO at UC Irvine, built UC-wide data platforms aggregating EMRs, medical images, and genomics across UCSF, UCLA and others — the unsexy infrastructure that makes everything downstream possible. The pattern: the highest-value AI is often the least visible, solving a specific operational or access problem rather than chasing a flashy clinical headline.
Andrew Lacy, CEO of the whole-body preventive imaging company Prenuvo, offered a useful corrective to the hype. Asked whether AI is what made whole-body MRI possible, he called that “a massive oversimplification” — the real unlock over the past decade has been hardware. MRI is a 30-year-old technology, and the difference between today’s research-grade machines and older ones is like the difference between a 1995 Nokia camera and an iPhone 16. AI helps speed image acquisition, but it forces a choice: acquire fuzzier images and use AI to fill in quality, or take the same time and use AI to increase accuracy.
Lacy’s warning is that “there’s a lot of AI washing going on” — companies inflating AI’s role in their business — and that in the wrong hands AI can produce a screening that looks impressive but isn’t diagnostic. Prenuvo’s positioning is that it’s a real clinical practice with skilled radiologists first, and an AI company second; the AI’s biggest impact on radiology is real but long-run. The reframe that matters for executives: catching cancer at stage 1-2 versus the typical stage 3-4 detection is a clinical-practice achievement that AI supports, not one AI delivers on its own. The discipline of asking “what is the clinical practice here, and where does AI actually help?” separates durable value from marketing.
Geoff Cook took over Noom — one of the most recognized consumer health brands in the country — at the moment GLP-1 medications were disrupting its core business. Noom was built on behavior change (‘it’s not a diet, it’s a mindset’); Ozempic, Wegovy, and Mounjaro made medically assisted weight loss the market default, and Noom had no answer. The rebuild is a case study in strategic response to platform disruption: rather than fighting the medications, Noom built the best GLP-1 companion — a habits program addressing protein intake and resistance training to counter the muscle wastage and sarcopenic obesity risks the drugs carry (GLP-1 labels literally state they’re to be used as an adjunct to diet and exercise) — and then added affordable access to the medications themselves as one integrated program.
The result: more than 60% of Noom’s revenue now comes from medicated programs, and the company has repositioned from dieting app to longevity platform — ‘vitality today’ rather than abstract healthspan promises. The transferable lesson for every digital health company watching AI or pharma disrupt their category: the winning move is rarely to defend the old model. It’s to become the essential companion to the disruption — the layer that makes the new technology work better and safer — while it’s still early.
Bob Wachter — Chair of Medicine, UCSF; author of A Giant Leap
Pelu Tran — CEO, Ferrum Health
Ali Parsa — founder of Quadrivia AI; founder of Babylon Health
Andrew Lacy — CEO, Prenuvo
What actually works, what’s getting commoditized, and the governance gap most organizations bought technology without closing
Two years ago AI scribes were a breakthrough; today new physicians ask in job interviews whether a health system has them. That commoditization is both proof that AI can penetrate clinical workflows and a warning that the easy wins are already priced in. The clinicians and operators who’ve navigated the gap between a promising model and a functioning clinical tool were specific about what separates the two — and almost none of it is about the technology.
Doug Jacobs, former Chief Transformation Officer at CMS and a practicing primary care physician, pointed to the clearest performance signal in the data: the Medicare Shared Savings Program — the largest accountable care program in the country with more than 11 million attributed beneficiaries — shows that ACOs where 75% or more of participating providers are in primary care consistently outperform specialist-heavy ones.
This isn’t rhetoric about primary care’s importance; it’s a numerical pattern in the nation’s largest value-based program. Primary care physicians manage the relationships, close the care gaps, manage the medications, and prevent the high-cost acute episodes that determine whether an ACO saves or loses. The policy direction reinforces this: CMS is proposing to shorten the upside-only ACO participation window from seven years to five, pushing organizations faster toward genuine downside risk — and most ACOs are already in two-sided models. An ACO built around specialists, or one treating primary care as subsidiary, is structurally disadvantaged from day one.
The math is striking: ACA premium subsidies were projected at $138 billion in 2025, a near-tenfold increase that bears no relationship to a tenfold increase in care delivered. As healthcare analyst Tony Lozazo explained, insurers operating in ACA markets faced limited price competition because the government underwrote the premiums — “happy times to be an insurer,” with little reason to compete on price.
The result was a decade of premium inflation partly hidden from consumers by expanding subsidies. As those subsidies are reduced or capped, the true cost of individual-market coverage is becoming visible to a large population for the first time — and the predictable effect is migration toward higher-deductible coverage or no coverage. Aaron Carroll of AcademyHealth put the underlying reality plainly: US healthcare is expensive because of prices, and there are no solutions to high prices that involve no trade-offs. For health systems and physician groups, the practical implication is a payer-mix shift toward self-pay and Medicaid in the markets where subsidy reduction bites hardest.
Per-capita healthcare spend varies dramatically by state in ways that track policy, not just demographics. Doug Jacobs cited the figures: Pennsylvania spends $8,700 per capita (near the national average), Oregon $12,000, California $10,500. Pennsylvania’s relatively low spend despite high inpatient utilization (30%+ more inpatient days per 1,000 than the national average) comes largely from how aggressively the state managed its Medicaid MCOs — negotiating pharmaceutical prices, standardizing prior authorization, holding plans accountable. That’s a replicable model that requires administrative will to execute.
On measuring state performance, Liz Fowler, former Director of CMMI, was emphatic that ChatGPT-generated state comparisons are unreliable — they took days to produce and the numbers couldn’t be trusted. The Commonwealth Fund state scorecard, not an AI tool, is the credible source. Massachusetts has insurance coverage north of 95% versus a national average around 90% — a policy choice, not an inevitability.
While the industry debates whether value-based care can work in complex specialties, Daniel Virnich’s Oncology Institute has managed full capitated risk on oncology spend since 2007 — contracts still intact 18 years later — across more than 2 million lives in Medicare, commercial, and managed Medicaid.
The model rests on a population-math insight: in a senior population only 9-11% of members need oncology in a given year (commercial populations, half to a quarter of that). That makes episodic contracting on cancer inherently volatile — one high-cost case can break a small panel — while population-level risk becomes predictable above a scale threshold. Scale is not just strategy in specialty capitation; it’s a financial prerequisite. Just as telling is how TOI grows: not through cold outreach, but through referrals from existing payer partners to their sister groups and plans in new markets. That is what 18 years of demonstrated value produces — a business-development engine powered by proof rather than pitch.
To understand value-based care, start with the scale of the program driving it. Mina Seshamani, who ran the Center for Medicare as CMS Deputy Administrator, put the dimensions plainly: Medicare covers 68 million people and processes about $1 trillion in claims every year, making it the nation’s single largest payer. It partners with more than a million clinicians and over 6,000 hospitals, and 98% of clinicians participate. When Medicare changes a payment rule, the entire delivery system moves.
Roughly half of Medicare beneficiaries now choose Medicare Advantage — where the government pays private plans to administer benefits — and half remain in traditional fee-for-service. Seshamani’s team focused on three things: transforming care so the system cares for people rather than just treating a disease (Medicare now pays for community health services, care navigation, and behavioral health integration into primary care for the first time ever); making Medicare Advantage achieve its potential, with particular scrutiny on supplemental benefits and how they’re actually used; and improving transparency across the program. For any organization operating in value-based care, the strategic reality is that CMS sets the direction and everyone else adapts to it.
Sachin Jain, CEO of SCAN Health Plan, cut through the polarized headlines about Medicare Advantage with the clearest frame in the series. The program, 25 years into its history, has split into two fundamentally different businesses. Good Medicare Advantage genuinely coordinates care, invests in chronic disease management, keeps patients out of the hospital, and in doing so lowers total cost of care while giving seniors more generous benefits. Bad Medicare Advantage is focused on risk adjustment — gaming the arbitrage in the government’s payment formulas, which pay plans more to cover sicker patients, and pocketing the margin without doing the intensive care coordination.
As Jain described it, bad MA plans are essentially “aggregating clinicians and beneficiaries” and recalculating risk to win an arbitrage between what the government pays and what care actually costs. This distinction matters enormously for investors and operators: the regulatory scrutiny now falling on Medicare Advantage is aimed squarely at the risk-adjustment arbitrage, and plans built on that model face mounting policy risk. Plans built on genuine care coordination — like SCAN’s not-for-profit model, which reinvests surplus into member benefits and addresses social determinants — are aligned with where the program is heading. Jain’s larger candor is worth holding onto: “This industry is broken beyond repair — there’s plenty to learn still.”
Senator Richard Burr — 30 years in Congress and former chair of the Senate HELP Committee — offered a window into how healthcare policy is actually made, and what’s lost when experience leaves the room. When he arrived in the House, healthcare was the largest employer in his North Carolina district; he had no expertise, so he built it the hard way — scheduling himself into operating rooms on weekends, standing in on heart bypasses to see firsthand how many people it takes before voting on reimbursement cuts for those procedures. He drove pandemic preparedness legislation in the early 2000s, decades before COVID proved its necessity.
His warning is about institutional memory: when policies with a history of working come under attack, it’s often the people who created them who remember why the structures work — and the people who weren’t there who don’t understand what they’re dismantling. For healthcare executives navigating today’s policy volatility (subsidy expirations, Medicaid changes, site-neutral payment debates), the practical implication is that the durability of any policy-dependent strategy depends on understanding why the policy exists, not just what it currently pays. Strategies built on reimbursement arbitrage that policymakers don’t understand the origins of are the most fragile.
Liz Fowler — former Director, CMS Innovation Center (CMMI)
Doug Jacobs — former Chief Transformation Officer, CMS; practicing primary care physician
Daniel Virnich — CEO, The Oncology Institute (2M+ lives under risk since 2007)
Sachin Jain — CEO, SCAN Health Plan
Mina Seshamani — former CMS Deputy Administrator & Director, Center for Medicare
What actually works, what’s getting commoditized, and the governance gap most organizations bought technology without closing
Two years ago AI scribes were a breakthrough; today new physicians ask in job interviews whether a health system has them. That commoditization is both proof that AI can penetrate clinical workflows and a warning that the easy wins are already priced in. The clinicians and operators who’ve navigated the gap between a promising model and a functioning clinical tool were specific about what separates the two — and almost none of it is about the technology.
Across specialties, the operators were consistent that physician ownership and governance are operational prerequisites, not soft preferences. Rothman Orthopedics — the largest doctor-owned orthopedic practice in the US, with ~200 providers across Pennsylvania, New Jersey, and Florida and a full continuity of care from rehab to ASCs — is built entirely on physician ownership. Gary Kirsh built Solaris Health in urology by acquiring large, sophisticated local-market players (not mom-and-pops), preserving local management depth.
The capitation math reinforces the need for scale. The Oncology Institute’s model works because, in a senior population, only 9-11% need oncology in a given year — too small a panel makes the economics volatile; above threshold it becomes predictable. The same logic appears in mental health: Mark Frank’s SonderMind found direct-to-consumer acquisition is the most expensive channel, while payer-direct partnerships — delivering the app to an entire member population preventatively — are the most efficient. Reaching a population before crisis is cheaper and clinically better than acquiring patients in acute need.
The cardiology and dermatology playbook is now visible across every specialty, at different maturity levels. In cardiology, Robbie Allen’s US Heart and Vascular spans 400+ providers across six states along the I-10 corridor (largest in Texas), serving 1.3-1.4 million patients across cardiology, vascular, and CT surgery. Tim Atterbury’s Cardiovascular Associates of America runs 160+ locations and 550+ providers across eight states, founded as recently as December 2021 — what he calls “structural innovation.”
In orthopedics, David Jacofsky’s HOPCo has evolved into the only vertically integrated MSK outcomes platform in the world, with four verticals — and a telling shift: physician practice management went from 95% of earnings a decade ago to roughly 40% today, as value-based care and digital became the growth engines. In dental, Daniel Burke’s PDS Health (formerly Pacific Dental Services, founded 30+ years ago) is pioneering dental-medical integration on the thesis that oral health predicts systemic disease. The maturity varies; the direction is uniform.
The employer market keeps surfacing as the next disruption, and the lessons are concrete. Glen Tullman’s Transcarent operates as an open network bundling surgery centers, specialty care, and navigation for self-insured employers — and partners directly with specialty MSOs, which he calls great partners. But the cautionary tale is Haven, the JP Morgan / Berkshire Hathaway / Amazon joint venture. Tullman, who was recruited to lead it and declined, assessed it as “right idea, wrong partners” — three great companies with three different objectives, which made execution impossible.
Rivka Friedman, who runs JPMorgan’s Morgan Health (launched in 2021 after Haven wound down), describes employer-sponsored insurance as “the Wild West” — thousands of employers each running a small insurance business with misaligned incentives, versus Medicare’s uniformity. For specialty MSOs, the implication is that the employer channel rewards quality outcomes data and patient experience in ways traditional fee-for-service contracting does not — Centers of Excellence designations and direct carve-outs for high-cost specialties favor MSOs that can prove their results.
Paul Teirstein, Chief of Cardiology at Scripps Clinic, was precise: the physician shortage is the direct result of identifiable policy decisions. Residency funding was effectively capped by Congress in the early 1990s and hasn’t meaningfully grown since. Board certification requirements function as a supply-restricting monopoly — which is why Teirstein founded an alternative certification board, NBPAS. Telemedicine remains technically illegal under a 1998 statute that expired and hasn’t been renewed, even though it worked demonstrably during COVID.
The supply math is already painful — an 80,000+ physician shortage, four-month waits for consults — and residents earn under $100,000 a year for grueling work. One proposal Teirstein endorses, regardless of its political source: adjust the Medicare fee schedule to pay primary care physicians more and certain specialists less. The directional logic is sound — the current schedule has undervalued primary care coordination and overvalued procedure volume for decades, steering new physicians away from the primary care that value-based models depend on. As Teirstein put it, anytime you see singular supply in healthcare — one board, one payer setting rates — tread carefully; monopolies harm patients.
Sean Griffin, CEO of URAC, made a point that’s easy to overlook in the consolidation rush: as MSOs scale and as new technologies enter care delivery, the mechanism that lets payers, employers, and patients trust an unfamiliar organization is accreditation. URAC has run for 35 years with more than 40 accreditation programs, governed by a board spanning the AMA, the American Nurses Association, AHIP, and Blue Cross Blue Shield — “all the voices around the table, but nobody gets a veto.”
Two developments make this newly relevant. First, accreditation is increasingly written into law and contracts — the No Surprises Act recommended URAC accreditation for independent review organizations, and marketplace health plans must be accredited. Second, URAC is now building an AI accreditation program through an open call for contributing organizations — an early signal that the governance gap around clinical AI (the same gap Ferrum Health flagged) is starting to be addressed through formal standards. Specialty pharmacy is among the fastest-growing accreditation areas, and URAC’s telehealth accreditation already extends into Africa, South America, and the Middle East. For consolidating specialty groups, the takeaway is that quality standards are becoming a competitive differentiator and a contracting prerequisite, not a back-office formality.
The series surfaced two fundamentally different ways to build a specialty platform, and the contrast is instructive. US Heart and Vascular grew through aggressive acquisition — roughly 10x in scale in under two years, going from fewer than 40 physicians, one state, and one ASC at formation to 400+ providers across six states with about 15 ASCs (over half of them built de novo). Robbie Allen described still acquiring five to seven practices a year, though the large practices are mostly gone — “there really are no large practices left except one in Virginia” — so the targets are now smaller and more accretive.
PDS Health took the opposite road. Over 30+ years, Pacific Dental Services grew almost entirely de novo — opening new offices rather than acquiring them — and never took venture or private equity money, letting founder Steve Thorne chart his own course. The connective tissue between both models is ownership alignment. PDS’s “owner doc” structure puts the dentist’s name on the door with real equity, and for years the support team’s bonuses were tied directly to increasing owner-doc profit distributions. As Daniel Burke put it, that design avoided the “suits versus clinicians” dynamic that undermined earlier MSOs. Whether you roll up or build, the platforms that endure are the ones where the physicians have genuine skin in the game and local decision-making authority.
A subtle but revealing detail from US Heart and Vascular: the company deliberately avoids joint-venturing its ASCs with hospitals and health systems, even though the hospital outpatient department (HOPD) rate structure is “temporarily good” in such arrangements. The reason is strategic alignment with value-based care — HOPD rates tend to be much higher, and high site-of-service costs are a liability, not an asset, when you’re moving into risk-based contracting where you’re accountable for total cost.
This is the kind of decision that separates MSOs building toward genuine value-based care from those optimizing for near-term fee-for-service economics. Tesha Simpson’s Millennium Physician Group makes a parallel choice on the multi-specialty side: rather than operating as a loose network, its cardiologists, rheumatologists, GI, and urology specialists sit on the same technology stack within a tight geographic circle, actively helping primary care doctors manage conditions — a deliberately coordinated, patient-centered-medical-home design built for managing risk, not just capturing referrals. The strategic signal across both: the MSOs positioning for the next decade are architecting their site-of-service mix, their specialty integration, and their technology stack specifically around taking on risk, even when fee-for-service would pay more today.
Daniel Virnich — CEO, The Oncology Institute
Glen Tullman — CEO, Transcarent; founder of Livongo
Paul Teirstein — Chief of Cardiology, Scripps Clinic
Daniel Burke — Chief Enterprise Strategy Officer, PDS Health
Robbie Allen — CEO, US Heart and Vascular
What actually works, what’s getting commoditized, and the governance gap most organizations bought technology without closing
Beverly Malone — President and CEO of the National League for Nursing, former General Secretary of the UK Royal College of Nursing, and former President of the American Nurses Association — was direct that one in three nurses considered leaving the profession after COVID. Nurses remain the most trusted profession in Gallup polling, but the pipeline is capacity-constrained: nursing faculty positions are underpaid and unfilled, clinical training sites are limited.
Malone made a point that reframes the issue: nursing diversity improves outcomes for all patients, not just underrepresented ones. A workforce that doesn’t reflect the populations it serves is a clinical quality problem, not merely an organizational value. Her career underscores how durable these issues are — she worked on unsafe staffing and mandatory overtime during the Clinton era (1996-2000), and the same friction (recent nurse strikes across large systems) persists. The organizations gaining ground are doing unglamorous things: meaningfully reducing documentation burden, building internal career pathways, and staffing so nurses practice at the top of their license.
Rahul Gupta, who served as the national Drug Czar (Director of ONDCP) for three and a half years, brought numbers that reframe the scale of the crisis: when he started in 2021, overdose deaths were rising 31% year-over-year and reached roughly 112,000 per year. Beneath that — what he called the tip of the iceberg — about 49 million Americans, one in seven, suffer from an addiction of some kind.
His policy response centered on treating addiction as a chronic disease rather than a moral failing: making naloxone widely accessible, expanding treatment, and treating fentanyl as a national security issue manufactured through irregular channels. The workforce implication is significant and under-discussed — addiction at this scale strains every part of the delivery system, from EDs to primary care to behavioral health, and the treatment capacity to meet 49 million people simply doesn’t exist. Gupta’s framing — “you learn from your dead so you can help prevent that from happening” — captures the urgency that the aggregate statistics obscure.
Graduate medical education funding was effectively capped by Congress in the early 1990s. The result is a structural ceiling on physician supply that medical school expansion cannot overcome if residency slots don’t keep pace — the bottleneck is post-graduate training capacity, not interest in medicine. Paul Teirstein, who runs training programs at Scripps, described the absurdity: expanding a program comes down to finding grateful patients to donate money, because residents must be paid, and they work brutal hours for under $100,000 a year.
Expanding residency funding requires congressional action — and the political will to act on a crisis that’s visible in four-month wait times but abstract in political-urgency terms. The near-term implication for every healthcare organization: the shortage will not resolve on its own. The organizations not actively building workforce pipelines, residency partnerships, and administrative-burden reduction are betting on a recovery that the structural math does not support.
The workforce crisis is most acute, and most consequential, in behavioral health. Mark Frank of SonderMind described a specific failure point that captures the whole problem: when a patient is discharged from an inpatient psychiatric setting, they can’t safely leave until a first outpatient appointment is confirmed — but the outpatient capacity to provide that appointment often doesn’t exist. The gap between inpatient stabilization and outpatient follow-up is where patients fall, relapse, and return.
SonderMind’s response is a workforce-multiplication strategy: AI-driven matching that connects patients to the right therapist by specialty, approach, insurance, and availability, plus payer-direct partnerships that put the service in front of an entire member population before they reach crisis. The broader lesson for any workforce-constrained specialty is that you cannot hire your way out of a shortage this large; you have to use technology to extract more effective capacity from the clinicians who exist and to reach patients earlier, when intervention is cheaper and more effective. Reaching a population before crisis isn’t just better clinically — it’s the only math that works when supply is fixed and demand is exploding.
The clearest evidence of how heavy administrative burden has become is what physicians will give up to escape it. Tesha Simpson described how Millennium Physician Group was founded in the early 2000s on exactly this premise: be the solution for independent physicians who didn’t want to affiliate with a health system but were drowning in the administrative demands of payers and government. The pitch was simple — partner with us, and we’ll remove the administrative burden so you can practice at the top of your license.
That value proposition has only grown more powerful. As Simpson put it, nearly every physician and advanced practitioner, in primary care or specialty, would say they experience significant administrative burden just to meet payer and regulatory demands. The strategic insight for any organization competing for clinical talent is that administrative-burden relief is not a perk — it is increasingly the deciding factor in where physicians choose to practice. The MSOs and groups winning the recruitment war are the ones that can credibly promise, and deliver, a practice environment where the clinician spends their time on patients rather than paperwork. This is also where AI’s workforce impact is most direct: every hour of documentation, prior authorization, and inbox management that technology removes is an hour returned to clinical work and a reason for a physician to stay.
Beverly Malone — President & CEO, National League for Nursing
Rahul Gupta — former Director, ONDCP (national Drug Czar)
Paul Teirstein — Chief of Cardiology, Scripps Clinic; residency program director
Tesha Simpson — Millennium Physician Group
India’s 400,000 daily teleconsultations, Pakistan’s two-problems-one-solution model, and what the US gets right by accident
The series deliberately reached beyond the US bubble. The global perspectives surfaced a specific and underappreciated point: the US system’s genuine strengths — innovation infrastructure, value-based contracting, health technology — are what it exports. Its weaknesses are largely internal. Meanwhile, other countries have executed at a scale the US debates endlessly but rarely achieves.
Rajendra Pratap Gupta, an architect of India’s National Health Policy 2017, described execution at a scale that reframes US telemedicine debates. eSanjeevani conducts 300,000-400,000 telemedicine consultations per day — free, government-funded — and has completed more than 425 million teleconsultations. India also launched PMJAY, the world’s largest insurance scheme covering 500 million people, and a National Digital Health Mission that has generated 825 million health IDs. No country has digitized health records at that scale before.
The contrast with US policy is instructive: a developing nation found the political will to make telemedicine free, permanent, and population-scale, while the US — despite COVID proving telemedicine works — still operates under a temporary statutory framework that hasn’t been renewed. The political economies differ completely, but the demonstration matters: population-scale digital health is technically and organizationally achievable. The question for any system is will, not feasibility.
Priya Basu, who leads the World Bank’s Pandemic Fund, gave the sobering global picture: 4.5 billion of the world’s 8 billion people still lack access to basic health services they need. Child mortality under five has dropped by more than half over 20 years despite more births — real progress — but since 2015 progress on major indicators has stalled, and COVID reversed life expectancy in many countries for the first time in a generation.
Basu also surfaced the structural financing gap COVID exposed: when a pandemic hits and you need $2 billion in 90 days to shore up health capacity across dozens of low-income countries, neither the WHO (standards and surveillance) nor the World Bank (long-term development lending) was built to move that fast — which is the gap the Pandemic Fund fills. And she reframed a challenge every wealthy country faces: each must effectively build two health systems, one for the population it has and one for the population arriving. With 2 billion people living in conflict-affected countries and 61 conflicts in a single recent year, that second system is not optional.
Mark Bertolini — CEO of Oscar Health, former CEO of Aetna — pushed back on the reflexive framing of US healthcare by what it gets wrong. The US private healthcare sector alone is $3 trillion; the government piece adds $2 trillion; total US healthcare spend is larger than the entire GDP of 190-195 countries. That scale produces more health technology investment than any other country by orders of magnitude, and more value-based care than anywhere else — there is essentially no value-based care in Europe, which remains entirely fee-for-service.
Bertolini called the life-expectancy comparison “lazy”: the US has 350 million people with a 40% obesity rate, not 10 million in New Zealand with a 20% rate. When US counties are segmented by health behaviors, the healthiest county populations match or exceed the world’s longest life expectancies, including Japan’s. The aggregate is dragged down by obesity, geographic and income disparity, and violence — none primarily healthcare-delivery problems. The honest read: the US doesn’t have a delivery problem at the top of the market. It has a prevention, access, and equity problem.
Saeed Khurram (Dr. Sara) built the largest telehealth platform in Pakistan around an elegant insight: Pakistan has a severe physician shortage, and simultaneously a large population of female physicians who can’t leave home to practice due to family obligations and social norms. Sehat Kahani connects those two populations — home-based women physicians with the hundreds of millions of underserved patients who have no realistic alternative.
The model is powerful precisely because it solves a supply problem and an access problem at once, without new infrastructure, capital, or policy reform. The supply already existed — it was just geographically and socially inaccessible — and the platform made it accessible. For any health system facing both workforce constraints and access gaps, the conceptual frame transfers directly: identify supply that exists but is locked away, and build the connection layer. The same theme echoed in Jorge Fernandez’s Hospital Clínic de Barcelona, where 120 years in a constrained facility forced an innovation culture that produces clinical gains using near-free tools — VR headsets for patient anxiety, gamified rehab in hallways — because the people defining the problem also define the solution.
Priya Basu’s Pandemic Fund offers a financing lesson that extends well beyond pandemics. The fund has raised $3 billion in donor commitments from 25 governments and three philanthropies — but its real impact comes from catalytic leverage. For every dollar it allocates, it mobilizes roughly seven more in co-financing from multilateral development banks and recipient countries themselves. The result: a portfolio approaching $7 billion in projects across 75 countries, built on a fraction of that in direct grants.
This catalytic model — using a relatively small pool of flexible capital to unlock far larger committed financing — is the mechanism that makes global health investment tractable at scale. Basu’s operational insight is equally sharp: in a pandemic, the binding constraint is surge capacity. You suddenly need orders of magnitude more nurses, clinicians, and technicians to handle a wave of sick patients, and that capacity has to be built before the crisis, not during it. The lesson for any health system is that preparedness is an infrastructure investment that looks wasteful right up until the moment it’s the only thing that matters.
Jorge Fernandez offered a clear-eyed window into how a high-performing European system actually works. In Catalonia, hospitals are predominantly public — 70 hospitals serving an 8-million population, with a mix of fully public, publicly financed but independently governed, and private institutions. Most citizens use the public system for anything serious (a cancer, a major diagnosis) and a private supplemental plan for speed and convenience. Roughly 20-25% of Spaniards carry this “double coverage.”
The price difference is staggering and worth sitting with: what a Spaniard pays for a year of private supplemental coverage is roughly what an American pays for a month. That affordability comes with trade-offs the US has historically refused — government-owned hospitals, government-employed physicians, and elected officials who can remove hospital CEOs. The honest comparison isn’t “Europe is better” or “the US is better”; it’s that each system has made different choices about the trade-off between universal access, cost, and the consumer-grade speed and choice the US market prizes. Understanding which trade-offs you’re actually making — rather than pretending they don’t exist — is the start of any serious reform conversation.
Rajendra Pratap Gupta — architect of India's National Health Policy 2017
Priya Basu — Executive Head, the Pandemic Fund (World Bank)
Mark Bertolini — CEO, Oscar Health; former CEO, Aetna
Jorge Fernandez — Chief Innovation Officer, Hospital Clínic de Barcelona
Francesca Colombo — Head of the Health Division, OECD
The revenue cycle is where operational dysfunction becomes measurable. After a decade of AI analytics applied to front-desk calls and patient financial data, specific numbers have emerged that most organizations have never seen — because most don’t have the infrastructure to see them. These figures are the starting point for an honest conversation about what front-office optimization is actually worth.
Amol Nirgudkar’s Patient Prism has analyzed millions of patient phone calls since 2015, surfacing a number most organizations have never quantified: roughly 50% of inbound calls from patients actively trying to schedule care do not result in a booked appointment. That’s half of all motivated, interested patients failing to convert into care.
The causes fall into three identifiable categories: the front-desk or call-center agent fails to convert an interested patient (a sales/communication failure); marketing drove a patient who isn’t a good fit (a lead-quality failure); or the practice has a real operational constraint — no capacity, wrong insurance, service not offered (an operations failure). Most organizations assume the problem is the first category; the data shows all three matter, varying by location, agent, channel, and service line. For PE-backed MSOs and DSOs, the financial implication is quantifiable: organizations using AI call analytics to fix conversion failures report 15x to 45x ROI within the first 90 days, measured purely from recovered revenue — and PE operating partners verify it by adjudicating the recoveries directly in the database.
Nirgudkar exposed a false premise at the foundation of healthcare digital marketing: Google and Meta count any phone call over 20-30 seconds as a “conversion.” But a 25-second call where a patient asks about hours and hangs up is not a booked appointment. Healthcare organizations are allocating marketing budget across Instagram, Google, Facebook, TV, and radio based on metrics that bear no relationship to actual patient acquisition.
The result is systematic misallocation — channels generating calls but not appointments get budget that should flow to channels generating actual patients, and high-value service lines (implants, complex procedures, higher-lifetime-value patients) are invisible without appointment-level attribution. AI call analytics that connect marketing channel to booked appointment, and feed validated booking signals back to ad platforms as optimization targets, give the algorithm a real signal to optimize toward. The compounding effect: lower cost per acquisition, higher lead quality, and budget allocated by what produces patients rather than what produces calls. For groups spending meaningfully on patient acquisition, this is a dollar-for-dollar improvement.
Itzik Cohen’s Payzen addresses a number that explains a great deal of downstream dysfunction: roughly 70% of patients cannot pay their medical bills in full. The effects extend far past the revenue cycle — patients delay or avoid care out of financial anxiety, leave the ED before treatment, don’t fill prescriptions, skip follow-ups. The financial problem is a clinical problem.
AI-powered payment plans that assess ability to pay and offer personalized, affordable installments at the point of service — rather than in collections — increase collection rates by 40%+. But the more important outcome is access: patients stop avoiding care because they’re afraid of the bill. The same RCM logic appears at scale in Daniel Virnich’s Oncology Institute, which handles 500,000+ patient calls per month and identifies prior authorization — “protocolized, manual, labor intensive” — as its highest-ROI AI target, precisely because a prior-auth delay in oncology means a cancer patient waiting. Across these conversations, the highest-value RCM AI targets are consistently the most protocol-driven, highest-volume tasks — not the complex clinical judgments.
Glen Tullman, who founded Livongo and now leads Transcarent, made a point that reframes how to evaluate any healthcare investment or initiative: at Livongo, the company delivered higher quality, a better patient experience, and lower cost — all three at once. As Tullman noted, “no one had ever done those three things together,” which is precisely why Livongo commanded the valuation it did. The reason it worked was predictive analytics — what we’d now call AI — that could tell what was happening in a diabetic patient’s body before they even knew it.
The reason this matters for revenue cycle and patient finance is that most healthcare “savings” come by sacrificing one of the three — cutting cost by degrading experience, or improving experience by raising cost. The genuinely valuable innovations are the ones that move all three in the same direction, and they almost always require technology to do it. When evaluating an RCM tool, a financing platform, or a care model, the question that separates real value from incremental tinkering is: does this improve quality, experience, AND cost together, or is it just shifting the burden from one to another? Tullman’s other warning applies directly here — friction lives in the middle, where ‘the middle keeps getting bigger and bigger,’ and that administrative middle is exactly where cost accumulates and patient experience dies.
Itzik Cohen of Payzen put a precise number on the structural shift that reshaped revenue cycle management. Thirty years ago, roughly 5% of a provider’s revenue came directly from patients; today it’s about 25%. The cause is the steady migration of payment responsibility from payers to patients through rising deductibles and cost-sharing — the patient, in Cohen’s framing, has effectively become the payer.
The operational implication is enormous and underappreciated. A provider used to collect 95% of revenue by dealing with a handful of commercial payers and Medicare — “five calls to make, really.” Now a quarter of revenue depends on collecting from thousands of individual patients, each with different financial circumstances and most unable to pay in full. That’s not a marginal change to the collections function; it’s a wholesale transformation of how providers get paid, and most billing operations were built for the old world. Cohen’s path to this insight is itself telling: at the marketplace lender Prosper, medical bills were one of the fastest-growing reasons people took personal loans — a market signal that the patient-financing problem was metastasizing well before most of healthcare noticed. The non-elective side, driven by deductibles, turned out to be a far bigger problem than the elective procedures (dental, plastic surgery, fertility) that financing companies originally targeted.
Noam Levey of KFF — who covered the crafting of the ACA from inside Congress and has reported on healthcare for 17 years — produced the landmark investigation quantifying America’s medical debt crisis: roughly 100 million Americans carry debt from medical or dental care. That number reframes the patient-financing conversation in this chapter. The 70% of patients who can’t pay their bills in full and the shift of provider revenue from payers to patients aren’t operational quirks — they are the mechanics behind a debt burden touching four in ten adults.
Levey’s longer view is also worth holding: he still considers the ACA the most transformational healthcare story of the past two decades — bigger than meaningful use, transparency rules, or Medicare drug negotiation — precisely because it restructured who bears cost. The unfinished part of that restructuring is what the medical debt numbers reveal: coverage expanded, but cost exposure migrated to deductibles and cost-sharing, turning insured patients into debtors. For providers and MSOs, that means patient financial experience isn’t a billing-department concern — it’s the front line of both collections performance and community trust.
Amol Nirgudkar — founder & CEO, Patient Prism
Daniel Virnich — CEO, The Oncology Institute (500,000+ calls/month)
Itzik Cohen — CEO, Payzen
Glen Tullman — CEO, Transcarent; founder of Livongo
Craig Kurzweil — Chief Data & Analytics Officer, UnitedHealthcare
Cleaning out his house after leaving Aetna, the former CEO (and current CEO of Oscar Health) found a box of overhead projector slides from 1991: virtual healthcare, social determinants of health, the human genome project. He looked at them and thought, “my God, we have come so little in so long.” Ideas that made people look at him “like I had three heads” three decades ago were still unresolved in 2025.
The observation is less about healthcare’s slowness than about what sustains a leader’s relevance across thirty years in a slow-moving, entrenched industry: a consistent long-horizon thesis, held through the market’s repeated failures to catch up. Bertolini’s career — Cigna to a startup to Aetna to Oscar — is built on the same thesis he had in 1991; he just found different vehicles for it. The uncomfortable lesson for executives: if your view of where healthcare is going has shifted materially in five years because of regulatory or investor pressure, you may have a position, not a vision. Positions change with conditions. Visions don’t.
Richard Zane, Chief Medical and Innovation Officer at UC Health (14 hospitals, $9 billion in revenue), runs a deceptively simple process. Identify a problem statement from the people doing the work — frontline clinicians and staff, not executives. Validate that the problem actually exists as described. Recognize that if it’s a real problem at UC Health, it’s probably a real problem at the other 2,000 systems in the country — so check whether anyone has already solved it. Then decide whether it’s solvable, because some problems aren’t.
The key move is in the third step: most health systems reinvent rather than borrow, wasting time and money on solutions that already exist. As Zane noted, every patient has 23 pairs of chromosomes, two lungs, and four heart chambers regardless of geography — the clinical problems are largely the same, so the solutions should travel. That willingness to adopt rather than invent — to admit your organization isn’t uniquely positioned to solve every problem — dramatically shortens time-to-implementation and raises success rates. Jason Suzuka, CDO at Bon Secours, reinforced the cultural prerequisite: put a clinician on your innovation team, because frontline relationships and credibility are what get innovations actually implemented and scaled.
Sachin Jain, CEO of SCAN Health Plan — the largest not-for-profit Medicare Advantage plan in the US, with roughly 440,000 members — made a strategic bet to expand from payer into care delivery (home care, primary care) for senior members. The logic was sound: as a payer you have the financial motivation to keep members healthy, so controlling delivery is the logical next step.
The reality was humbling. Jain’s description was unusually candid: the expansion was “really, really hard” and “has not gone as smoothly as we thought.” There is, he said, an “if you build it, they will come” mentality with clinical services that simply doesn’t hold — members don’t automatically use new care assets just because they’re available and covered. Clinical network effects, patient trust, geographic convenience, and the inertia of existing relationships all matter enormously. His broader posture is the real lesson: “this industry is broken beyond repair — there’s plenty to learn still.” The leaders worth listening to combine ambition with the humility to name what isn’t working, even in their own organizations.
Inside the same Oscar Health conversation was one of the sharpest leadership stories in the series. Bertolini described Aetna — a 167-year-old company — as “stuck in the way we always do things.” He kept hearing “they told me we can’t do that,” so he started asking to meet “they.” “Bring they up here — I’d really like to meet them.” The point was that the invisible, diffuse “they” — the unwritten rules, the assumed constraints, the institutional inertia — is what actually blocks change, and naming it is the first step to dismantling it.
What turned Aetna’s culture in his final three years wasn’t strategy; it was trust. Drawing on Dov Seidman’s book How, Bertolini built what he called a trust model: “trust is like love — you have to give it to get it.” Once the organization believed leadership trusted them, ordinary people started doing extraordinary things. The contrast he drew was pointed: when the business reverted to “financial engineering more than actual product engineering” — chasing earnings flow over building things that improve care — the energy drained out. For executives running consolidation plays and transformations, the lesson is that culture and trust are not soft accompaniments to the financial model; they are the thing that determines whether the financial model actually delivers.
Beyond his innovation filter, Richard Zane articulated a disciplined model for how a health system should engage with companies — one that resolves the perennial build-versus-buy-versus-partner question. Once UC Health validates a real, unsolved, solvable problem and scans the market for who’s working on it, the ideal partnership does three specific things: the system commits to being a customer of whatever gets built, a co-developer of the solution, and a participant in the value creation. Not just a pilot site. Not just a logo on a slide. All three.
This structure aligns incentives in a way that one-off vendor pilots never do — the health system has skin in the outcome, the company gets a genuine development partner with real clinical context, and both share in the upside. Zane paired this with a posture worth emulating: as Chief Medical Officer, he refuses to be “the gatekeeper for anything innovative.” When frontline people have an innovative solution, they don’t need his permission to pursue it — they may need his help, his endorsement, or access to resources, but not his approval. That distinction between enabling innovation and controlling it is what separates innovation functions that accelerate change from those that quietly strangle it.
David Lubarsky, CEO of Westchester Medical Center Health Network and former CEO of UC Davis Health, described a positioning strategy that applies far beyond academic medical centers. Westchester receives nearly 14,000 inpatient transfers a year — more than almost any hospital in the United States, and more than double what UC Davis received in a comparable geography. The reason is that it is the only tertiary center capable of advanced, complex care across a 6,800-square-mile region between New York City and Albany.
The strategic insight is that even Westchester’s competitors send it patients — because those community hospitals don’t have the capability, and the alternative is transporting a critically ill patient two hours into Manhattan. By being genuinely indispensable for the most complex care in its region, Westchester turned competitors into referral sources. The broader leadership lesson is about choosing where to be irreplaceable. In a consolidating, cost-pressured market, the durable position isn’t being adequate at everything — it’s being the one organization in your geography that can do the thing nobody else can, such that even rivals route their hardest cases to you. That requires the discipline to invest deeply in genuine differentiation rather than spreading thin across commoditized services.
Mark Bertolini — CEO, Oscar Health; former CEO, Aetna
Richard Zane — Chief Medical & Innovation Officer, UC Health
Sachin Jain — CEO, SCAN Health Plan
David Lubarsky — CEO, Westchester Medical Center Health Network; former CEO, UC Davis Health
The revenue cycle is where operational dysfunction becomes measurable. After a decade of AI analytics applied to front-desk calls and patient financial data, specific numbers have emerged that most organizations have never seen — because most don’t have the infrastructure to see them. These figures are the starting point for an honest conversation about what front-office optimization is actually worth.
Patrik Feeney left a career in private equity — he co-founded Lateral Investment Management and scaled it from zero to $700 million in assets — to build Pearl Medical, an outpatient cardiovascular platform, after CMS approved minimally invasive cardiac procedures to move outside hospital walls. His most expensive lesson wasn’t clinical or operational. It was contracting: the largest managed care contract in his first region, which he conservatively budgeted nine months to procure, took three years. Three years — for a physician who had been in the market for twenty years with established relationships and leadership positions. As Feeney put it, incumbents “will fight it tooth and nail.”
When the contract finally came through, it doubled Pearl’s volume — proof of how much value was locked behind that single negotiation. The generalizable lesson for anyone building an outpatient or site-of-service-shift model: the constraint isn’t demand, and it often isn’t clinical capability. It’s the payer relationships and network access that incumbents control, and displacement timelines should be modeled in years, not quarters. Pearl survived that period and now runs north of $60 million in revenue across 12 states — but the plan that got them there had to absorb a three-year contracting delay that would have killed a thinner-capitalized business.
Richard Zane, Chief Medical and Innovation Officer at UC Health (14 hospitals, $9 billion in revenue), runs a deceptively simple process. Identify a problem statement from the people doing the work — frontline clinicians and staff, not executives. Validate that the problem actually exists as described. Recognize that if it’s a real problem at UC Health, it’s probably a real problem at the other 2,000 systems in the country — so check whether anyone has already solved it. Then decide whether it’s solvable, because some problems aren’t.
The key move is in the third step: most health systems reinvent rather than borrow, wasting time and money on solutions that already exist. As Zane noted, every patient has 23 pairs of chromosomes, two lungs, and four heart chambers regardless of geography — the clinical problems are largely the same, so the solutions should travel. That willingness to adopt rather than invent — to admit your organization isn’t uniquely positioned to solve every problem — dramatically shortens time-to-implementation and raises success rates. Jason Suzuka, CDO at Bon Secours, reinforced the cultural prerequisite: put a clinician on your innovation team, because frontline relationships and credibility are what get innovations actually implemented and scaled.
Ira Kornbluth built Clearway Pain Solutions (formerly Smart Pain / KureSmart) into the largest pain management MSO in the country — through a 2017 private equity transaction and a recap less than two years later at what was then the highest multiple ever in the pain space. Asked which function changed most across his scaling journey, his answer surprised the room: compliance. He hired a dedicated compliance officer just four or five years into the business — early for the industry, and most practices Clearway evaluates for acquisition today still don’t have one.
His articulation is worth memorizing: “Compliance oftentimes doesn’t matter until it matters. And when it matters, compliance is the primary thing everyone focuses on.” In pain management — where opioid scrutiny is existential — Clearway’s compliance leader is a former COO of a large pain practice, positioned at the intersection of operations and compliance rather than in a policing silo. A proactive external review of hundreds of charts recently rated Clearway’s program the best the reviewers had seen in years. The strategic point extends to every regulated specialty: compliance done well is not a cost center — it’s risk mitigation, brand integrity, transaction readiness, and a genuine competitive asset in payer and PE negotiations. Notably, every physician from Kornbluth’s original transaction is still with the business, save one.
Lew Little has run scaled healthcare services businesses three times — Covenant Care, Harden Healthcare, and now Giving Home Health Care. His distillation of what multi-state, multi-site scaling requires is a three-item list, in order: get the culture right; get the right people in the right seats on the bus; and — the most painful — get processes in place that are tested and repeatable, so the organization isn’t taking “heroic action” every time something needs to get done.
Two of his hard-won specifics deserve attention. First, on clinician partners and first impressions: he cited a physician who complained bitterly about an acquisition experience — from seven years earlier. “They never forget.” And partners have what he calls selective retention: the farther you get from the deal date and the cash payment, the cloudier their memory of the benefits becomes. Underpromise and overdeliver, and in the spirit of John Wooden: “be quick, but don’t hurry.” Second, on governance: the right private equity partner is one of the most critical decisions a CEO makes, because “the money is not all similar” — KKR brought genuine operational resources through Capstone, while a misaligned board (five stacked rounds of preferred stock, common holders effectively out of the money) nearly paralyzed a company. His fix, and his advice to any CEO struggling with a board: bring in an independent director. “It’s a great lifeline for CEOs.”
Kris Shepherd has spent 18+ years building physician partnerships at what is now Advocate Health — 67 hospitals, over 1,000 sites, 150,000 employees, $30 billion in revenue, the fourth-largest health system in the country. His portfolio is a living taxonomy of every partnership model available to a health system: outright acquisition (Sanger Heart & Vascular, acquired 20+ years ago, now grown into a national service line); organic de novo builds (One Health Primary Care, roughly 200 physicians built from scratch); three-party private equity partnerships (Charlotte Radiology becoming the cornerstone of US Radiology alongside Welsh Carson); and joint ventures (SCA as their largest ASC JV partner, Surgery Partners in two markets — while Advocate simultaneously builds its own ASC management company).
Two observations from his experience matter for anyone on the other side of the table. First, three-party partnerships with large PE firms are the hardest to manage — “it’s tough to convince a large private equity group to do things the Advocate way” — because the incentive alignment is never fully clean. Second, clinical gravity outlasts corporate structure: Tryon Medical Partners left Atrium in 2018 and later took TPG investment, yet still predominantly uses Advocate’s specialty network years later. The relationships between physicians and referral networks persist long after the org chart changes — which is both a consolation when groups leave and a warning about how long it takes to truly displace an incumbent system.
Aaron Carroll, CEO of AcademyHealth, delivered the most intellectually honest framing of US healthcare costs in the masterclass series. Everyone wants a single villain — greedy insurers, overpaid CEOs, pharma, hospital monopolies. The evidence says: it’s everything. Physicians earn more relative to average citizens than in peer countries. So do nurses. Pharma makes more. Hospitals make more. Insurers make more. And the single largest overspending category isn’t any of the popular villains — it’s outpatient care. Eliminate all insurance overhead entirely and the US remains the most expensive system in the world. Zero out all pharmaceutical spending: still the most expensive. Refuse to pay physicians at all: still the most expensive.
The implication he draws is the one politicians refuse to say aloud: there are no solutions to US healthcare prices that involve no trade-offs. Coverage expansions require enormous subsidy dollars; removing subsidies exposes people to true prices (the 10-40% premium increases now hitting the market as ACA subsidies expire); and international comparisons cut both ways — the UK’s universal coverage comes with an inverse relationship between coverage breadth and actual supply, measured in wait times. For executives, the takeaway isn’t cynicism; it’s clarity. Any strategy premised on someone else’s margin being the fixable problem is built on sand. Durable strategies are built on genuinely delivering more value per dollar — because external price pressure, when it comes, will hit everyone.
Patrik Feeney — founder & CEO, Pearl Medical (fmr co-founder, Lateral Investment Management)
Ira Kornbluth — co-founder, Clearway Pain Solutions (largest pain MSO in the US)
Lew Little — CEO, Giving Home Health Care (fmr Harden Healthcare, Covenant Care)
Kris Shepherd — SVP of ASC Strategy & Physician Partnerships, Advocate Health ($30B, 67 hospitals)
Connect with us to explore how our trusted solutions can help you reach your goals. Fill out the form below, and a member of our team will follow up with personalized details for your organization.
