In partnership with

Hospitalogists,

2026 has seen a marked shift in my conversations, where many are focusing on direct contracting and lots of innovation happening in the employer segment. Today I want to distill down the arguments and perspectives I’ve gathered along with providing a few angles of my own on where the future potential of this space lies, and whether managed care is dying, or if we’re just in another one of those healthcare cycles.

Also, quick favor to ask! This September, healthcare's finance and strategy crowd will descend on Las Vegas for Vizient Connections Summit.

I'm hosting a happy hour while I'm in town: drinks, conversation, and no forced networking energy. Which leads me to ask: are you going to Vizient Connections this year?

Let me know here.

I'm gathering interest before locking in the details. More to come!

Was this email forwarded to you?

Subscribe Here

Sponsored by Layer Health

Clinical registries were once just a reporting requirement. Now, as health systems evaluate clinical AI, they're emerging as one of the safest, most measurable proving grounds for adoption.

Registries combine clinically complex reasoning, standardized performance benchmarks, gold-standard training data, and clear operational ROI, making them an ideal environment to validate AI before expanding into more prospective clinical workflows. 

Layer Health's new registry guide explores what it takes to modernize registry operations with AI-assisted abstraction. 

Beyond demonstrating 60%+ reductions in abstraction time and 98%+ validated accuracy, AI helps to standardize abstraction practices, reduce variability between abstractors, and create a reusable clinical intelligence foundation for future use cases like case finding and quality gap identification.

See how proving AI in a high-value, low-risk workflow helps you build trust and scale clinical AI responsibly.

Get Your Copy

Employers Just Hit the Toll Booth Limit: A State of the Union on the $37,824 Problem

Somewhere between last year's renewal cycle and this one, employer-sponsored healthcare crossed over from "expensive line item" to "board-level emergency." Per the 2026 Milliman Medical Index, healthcare for a family of 4 on a typical employer plan now runs $37,824 a year, with per-person costs up 7.9%. This not-so-fun factoid represents the steepest non-COVID increase in over a decade. So, healthcare costs are rising, and accelerating.

Mercer pegs the 2026 employer increase at 6.7% (highest in 15 years, pushing average cost above $18,500 per employee), and that's after plan design changes; the raw number before offsets was closer to 9%. Meanwhile the sophisticated buyers are printing worse. John Zutter told me Lantern's own self-funded plan took a 20%+ increase this year, and across client summits representing about 60 of his largest accounts, average reported trend landed in the mid-teens.

• "Last year, people were surprised that it wasn't going to be six, it was going to be eight or nine. And then this year, they're saying we were expecting 15. And that's for some of the largest, most sophisticated employers in the country.”

Over on the individual market, enhanced subsidies expired at the end of 2025, net premium payments jumped 58% on average, deductibles spiked 37% to a record $3,786, and about 3M people dropped ACA coverage between the end of 2025 and February 2026. Underlying ACA premiums are set to rise another 14% into 2027 as the risk pool deteriorates. JP Morgan's own economic institute is watching small businesses drop coverage in real time because they can't absorb 10-20% increases on top of single-digit wage growth.

I've spent the past several months interviewing the people rebuilding this market and conducting research — to name a few: Nick Stefanizzi (Northwell Direct), Dan Mendelson (Morgan Health), John Zutter (Lantern), Ali Diab (Collective Health), Jason Ross (Privia) — and I want to present some findings along with a thesis around what’s emerging.

My thesis is, more or less, that direct contracting, price transparency, and consumerism, supercharged by AI-enabled navigation, will shape the next decade of benefit design. So we’ll walk through that, but I also think the bear case deserves real airtime, because healthcare has a long history of burying "employer revolutions" in shallow graves and healthcare tends to be cyclical in who the dominant party is at any given moment.

$37,824 buys you a Tesla Model Y. Every year.

Imagine buying a new Model Y each year, except the car never shows up. Healthcare has climbed from 10% of GDP to 20% in less than a generation, and on its current trajectory it's heading for 21% by 2034, or $9 Trillion. Worse, the burden keeps sliding toward employees. Milliman's 20-year data shows the employer share of total cost has fallen from 61% to 58% since 2005 while the employee premium contribution climbed from 21% to 27%. Mercer found 59% of employers will make cost-cutting plan changes in 2026, up from 48% in 2025 — and in most shops, "cost-cutting" means higher deductibles and copays, which is cost-shifting to the end consumer.

At the small end of the market, it's carnage. Eric Bricker told me on a recent episode that fully insured groups under 200 lives saw renewals of 30%, 40%, in some cases 80% this cycle, as carriers repriced commercial books to cover their MA wounds. Small employers now face a three-way fork: join a captive and go self-funded, punt employees to the individual market via ICHRA, or eat the renewal. A growing number are choosing the unmentioned door number 4 — dropping coverage and handing out stipends — which Mendelson called out as an unacceptable answer for the American workforce, and I agree with him.

• "Healthcare makes people feel dumb. And that is in part why they don't trust a lot of things." — Teira Gunlock

• "Anytime you hear a discussion about healthcare, just know that the priorities of the patient is actually at the bottom. Every other hand is in the till above the patient." — Dr. Eric Bricker

I often ask and wonder…'is there a breaking point with employer healthcare affordability? When do we reach this point?’ A 6.7% print in the surveys and a mid-teens print among sophisticated self-funded buyers is a lot to stomach. Pharmacy ran +14.8% on GLP-1s and specialty. With advanced therapeutics and consolidation, these dynamics won’t slow down unless policymakers bear their teeth. Trends like this force structural responses.

• "The path to healthcare reform is campaign finance reform." — Dr. Eric Bricker, quoting Donna Shalala

Toll booth economics: Diab's case against the MLR

If you want the most provocative diagnosis of why the system inflates, Collective Health’s Ali Diab's (in an upcoming podcast episode; stay tuned) is hard to beat: the MLR — the ACA provision everyone treats as consumer protection — is functionally a cost-plus contract.

If an insurer must spend 85 cents of every premium dollar on care, the only way to grow the 15 cents is to grow the dollar. He calls it "the $600 Navy submarine toilet seat problem all over again." Under that structure, carriers profit when unit costs and utilization rise, and every other stakeholder becomes what he calls a witting or unwitting accomplice: hospitals running thin margins take the extra dollar, drug makers take the extra dollar, and the toll booth collects its percentage on the way through.

Layer on his second beef — the "discount off the Neiman Marcus price," where networks and PBMs negotiate rebates off an inflated MSRP and then market the markdown as value — and you get a rubric where nobody in the payment chain wins by making the underlying thing cheaper. His personal knee surgery story is the whole system in miniature: a $340 coinsurance charge for a fixed-hinge knee brace that retails for $125 at Dick's, non-negotiable, because the brace was baked into an insurance bundle he couldn't unbundle. And this is the CEO of a health benefits platform!!

Anyone who's driven the Dallas North Tollway at rush hour understands the model: the road doesn't get faster, the toll just goes up.

I enjoyed Diab’s dialogue around this setup given his economic background. Stop insuring the whole enchilada where so many costs can hide or be veiled within verticalized bundles. Insurance exists for ruinous tail risk, and running preventive visits, generics, and birth control through an actuarial toll booth adds cost without adding protection. Get big enough, self-fund, buy stop-loss for the catastrophic layer, and contract for the rest. Agree or disagree, that logic is now driving real purchasing behavior, and it's why the self-insured trend line (high single digits) keeps beating the fully insured one (mid-to-high teens). It’s not too dissimilar from what Cuban has been saying on how healthcare should be financed - like car insurance, with cash pay on the rest.

The Waning Power of the Managed Care Industry

My favorite anecdote from this whole run of interviews was with Morgan Health and Dan Mendelson. Right after Jamie Dimon announced Morgan Health's mandate, Mendelson got a call from the CEO of one of the country's largest health plans. One question. "Are you friend or foe?" Mendelson replied, "That's up to you."

This exchange is the entire future of managed care compressed into few words. Because nuance obviously exists here (and both Mendelson and Zutter were careful with their phrasing) and plans aren't failing at everything. There’s a reason why this business model exists. These huge, scaled behemoths have negotiating power and the previously bundled nature of everything, while opaque, was easy for employers to work with. A one-stop shop. And members still want open access for the everyday 95%: the OB-GYN relationship, the behavioral health visit in a supply-constrained market, the ER on vacation without a surprise bill. Incumbents deliver that fine.

Where traditional networks are failing though, in Mendelson's own words, is the high-acuity layer, and the math there gets brutal. Zutter provided a cost composition breakdown for me on where things actually add up in the healthcare cost equation:

• planned procedures (~20% of spend),

• oncology (15-30%),

• other infusions (5-10%), and

• specialty pharmacy (~10%) stack to about 60% of total cost, each growing 1.5-3x the overall rate.

Meanwhile 1% of members drive a third of spend, 5% drive ~60%, and 10% drive 85%.

So in other words, employer healthcare is an acuity problem masquerading as a population problem, and broad PPO networks were never engineered for the 5%. So the disaggregated layer is getting built around the incumbents — Lantern on specialty steerage (12M lives, 1,000+ plan sponsors, ~70% compounded growth for a decade), Thyme Care on oncology, Centivo running narrow-network plans anchored to Baylor right here in Dallas.

Why am I considering these moves an inflection point rather than an insurrection? The smart plans are now bringing Lantern to their own clients, and CVS Aetna co-invested with Morgan Health in Thyme Care. As Mendelson put it, being a foe to employers and what they’re trying to accomplish is optional. Irrelevance is too, for now. But clinging to the old way of doing things in an anti-consumer, opaque financial engineering fashion is how you bury your own grave with your future customers and with Congress.

• "The traditional networks, the networks that are commonly deployed in the kind of traditional fee-for-service stuff, are failing. They're just not working." — Dan Mendelson

Express lanes: how direct contracting gets deployed in the wild

For readers who want the operating manual rather than the thesis, I have an upcoming conversation with Northwell Direct’s Nick Stefanizzi (subscribe here - we’re over 100 subs baby!), but I also had the chance to chat with him a few months ago to get more details on the direct deal. 170K union members migrated onto a high-performance network, closed in 75 days, bringing Northwell Direct past 310K covered lives with $100M+ proposed in union healthcare cost savings over 3 years.

I found 3 mechanics from that deal pretty notable.

• First, the pricing stack: a near-term rate concession from Northwell (around a 20% cut if memory serves), a contractual 5% YOY trend cap, benefit design that steers members with lower copays rather than forced switching, and clinical management on the back end that bends the underlying trajectory.

• Second, objective verification: Northwell let an independent third party run a regression on 32BJ's historical claims against proposed network rates, which validated the promised 20% savings before signature. Pre-deal, third-party re-adjudication is becoming table stakes, and any system unwilling to open the books will lose to one that will.

• Third, the frenemy network: around 20% of Northwell Direct's capacity is non-Northwell providers, because a network you can ship to members scattered from Staten Island to Connecticut beats a walled garden of just Northwell providers every time.

Note what Northwell refuses to do: file for insurance licenses, build a claims shop, run its own UM. Stefanizzi's "I'm not a health insurance company" stance is a disaggregation/unbundling of the plan, with TPAs handling adjudication and member service, stop-loss panels carrying risk, and PBMs on pharmacy. Lower capital, less regulatory drag. Mercy, Memorial Hermann, Geisinger, Baylor, and Intermountain are all experimenting in this arena as interest swells.

Unbundling the whole enchilada

Employer stack 2.0 is here, and it looks like a component, modular architecture. Self-funded employers are picking a chassis (an administrator like Collective Health, which Diab pitches as the Stripe of health benefits), then bolting on direct contracts, high-performance networks, specialty navigators, digital-first providers (Hinge, Lyra, Progyny), and direct pharmacy relationships.

About 40% of volume in the fastest-growing drug class now moves through direct channels like LillyDirect, and employers are starting to buy Keytruda-class drugs straight from manufacturers with performance guarantees attached. Even Medicare, as Diab pointed out, amounts to the largest direct contract in the country: AKA, here's the public rate, take it or leave it.

Unbundling is yielding directional results self-insured employers are looking for. Collective Health's book is running a 3.9% annualized trend over the trailing 24 months versus a ~9% self-insured average, which Diab attributes to real-time routing of members to the lowest-cost appropriate option across all the integrated components. A near-6-point trend delta, sustained, is the kind of number that reprices an entire market's expectations. Mercer's data says the market is following: 35% of large employers now offer at least 1 high-performance network plan, and 67% offer 3+ medical plan choices, up from 60% in 2023. Still, while this approach works for large employers, small and medium businesses continue to fall behind the cost 8-ball. I expect more solutions to be built for this segment in the near future.

Cell and gene therapy, at $1M-$10M per dose, is starting to get excluded from small-group products, which Mendelson called out for what it is: the moment insurance stops being insurance.

• "Insurance should be insurance and it's got to cover expensive stuff.”

And ICHRA — the full unbundling endgame — keeps stubbing its toe on narrow exchange networks and a risk pool that just shed its healthiest members. This dynamic is why group captive as an emerging option is incredibly compelling for SMBs:

• "If you join a captive and you pool all these smaller self-funded groups together, then you can go out to a reinsurance company and they'll be like, okay, we'll actually insure this larger pool of employers and we'll issue a policy that does not have a laser in it. And that gives the employers a lot more confidence when they go to a self-funding arrangement.”

• "If you're a fully insured employer, then you basically have one of two choices... You're either going to say we're going to join a captive and go self-funded, or we're just going to get out of the business totally and just go ICHRA. But you're not just going to stay in the fully insured world and take your 80% renewal."

TollTag for healthcare: AI finishes what the HMO started

The HMO movement of narrow networks and steerage didn't die in the 90s because the math was wrong. Really the math in concept was great, and the HMO model is one such model proven to lower costs given steerage. No, they died because the consumer experience was adversarial, and people hate having less choice. Managed care asked members to accept restriction with no intelligence attached, and members burned the model down. What's different now is that AI-enabled navigation can answer questions, route members before they ask, nudge the overdue colonoscopy, and personalize around what Zutter calls the fingerprint problem. Every member's circumstances are unique, which makes healthcare a natural fit for probabilistic tools and a rotten fit for the deterministic ones we've been using. Narrow network plus real intelligence plus carrot-based benefit design is a new, HMO-like, but more consumer-oriented approach.

And to be clear-eyed about it, AI cuts both ways. Milliman flagged AI as a healthcare cost factor for the first time in the MMI's history this year — hospitals deploying it to optimize billing capture while payers deploy it on adjudication. Mendelson is experiencing the bot wars firsthand:

• "We're seeing providers right now using AI to upcode services."

We are watching an arms race between AI incentivized to inflate claims versus the AI incentivized to steer around them, and benefit design is one battleground in the war.

Is Going Direct a Fad, or a New Paradigm?

As mentioned, healthcare stakeholders wax and wane in influence based on policies and existing tailwinds in any particular decade. Zutter volunteered the failure modes of centers of excellence 1.0 without prompting: too narrow in scope, invisible to members, and dependent on patients flying across the country after back surgery, which nobody wants to do. Lantern only scaled after abandoning that model.

Jason Ross from Privia, when I asked him about going direct, gave me the most measured operator's take: direct-to-employer is "very attractive in terms of the theory and the academic exercise," and then it breaks down in execution. Stated differently, if you’ve seen one contract, you’ve seen one contract.

In today’s direct game at the scaled provider organization level, you need employer-side employee density in a tight geography, an HR team sophisticated enough to run a custom network, and a benefits consultant whose incentives aren't pointed elsewhere. Privia, with ~6,000 providers and real market density, has stayed by-the-books on employer contracting for those exact reasons.

Ross also offered the cycles argument I’ve been alluding to. As someone who’s walked the walk in healthcare for quite some time, he's watched value-based care get declared inevitable and dead on alternating 5-year intervals since 2007.

• "If I had a dollar for every time I guessed wrong on when health systems would say we can't do this anymore, I'd be broke."

Managed care has absorbed every previous insurgency — HMO backlash, CDHPs, wellness, navigation 1.0 — by co-opting the parts that worked, and the co-option is already underway: plans reselling Lantern, CVS Aetna inside Thyme Care’s cap table. There's a plausible 2030 future where "direct contracting" is just a product line inside the BUCAs, trend moderates for 2 years, benefits teams declare victory, and the whole apparatus gets shelved until the next 20% print.

My Take: The Paradigm Shift is Here

Given the dynamics at play and the Great Personalization we’re seeing with agentic AI as an action layer, the direct movement is a distinct paradigm shift and not going away any time soon.

What's different from the HMO era and the CDHP era is the simultaneous arrival of

• the math (5% of members driving 60% of spend, at 20% trend),

• the waning popularity of the incumbent BUCA vertically integrated business model,

• tooling (AI navigation and the agentic action/intelligence layer) that makes steerage feel like consumer-oriented service, and

• capital (Morgan Health's $280M and other large employers with similar ambitions).

Prior "revolutions" had 1 or 2 tailwinds. This one culminates and converges on many forces.

For health system leaders looking for revenue diversification or growth strategies, you don't have to figure out the exact macro debate to act on this trend. Start experimenting today. This was one of Nick Stefanizzi’s main points at Northwell.

If direct contracting keeps compounding, a system with an autonomous direct-to-employer arm — its own P&L and talent, a Northwell-style pricing stack — owns a durable revenue line and a hedge against commercial rate compression. If managed care comes roaring back and absorbs the insurgents, that same infrastructure makes you the network partner every plan wants for its high-performance products, on better terms than a take-it-or-leave-it fee schedule. Flexibility and personalization in benefit design are the rare strategy that pays off in both states of the world: adaptability and revenue diversification, plus a front-row seat to the buyer relationship. And don’t forget that these investments and initiatives translate directly (pun intended) into better relationships with patients.

Informed members and consumers with real information vote with their feet and their wallets. And as information asymmetry decreases, this more informed patient presents distinct disadvantages for the opaque incumbents in our healthcare landscape.

Over at Collective, Diab has watched people drive 10 extra miles for an MRI once they saw the price difference. Healthcare is becoming a consumer experience whether the incumbents cooperate or don't, and the organizations building for an informed, steered, AI-assisted patient are building for the only future on the ballot. Everyone else is defending a toll booth on a road people are learning to route around.

Let’s go direct.

HOSPITALOGY TOP READS & RESOURCES

• Good overview from Rochester Regional's ambient scribe rollout shows what real adoption, not just deployment, actually takes. Read more from Becker’s.

• More data hasn't solved healthcare's coordination problem. Join clinical leaders from Bookmark Medical and Navina on July 15 for a webinar on how AI helps care teams turn fragmented information into meaningful action.*

• OpenEvidence launches medical AI copilot feature that grades medical evidence. Read more from Fierce Healthcare.

*This resource is brought to you by one of my brand partners who help make this newsletter possible!

MISCELLANEOUS MADDENINGS

Today I’m in the motherland (Austin, TX) and just finished recording of my first in-person, in-studio podcasts with the one and only Will Johnson from Gyde. More things to come! Stay tuned.

Thanks for the read! Let me know what you thought by replying back to this email.

— Blake