What 15 years at Kaiser Permanente taught me about the healthcare system we’re at risk of losing.

He was 32, and I described him as “previously healthy” in my note, which is what we say when you’ve just had something really bad happen to you. I was pretty sure he’d had his first seizure a few hours ago. By the time I saw him he was back to normal, and all the tests I’d done in the ER were normal, too. This is, of course, a good outcome, but it always leaves a big, unanswered question: why did he have a seizure in the first place?

He would eventually need an EEG, as well as an MRI to rule out a small brain tumor, but none of that needed to be done tonight, at midnight. At most hospitals, I would have just admitted him, because there was no reliable way to get it done quickly as an outpatient. He’d have spent a night or two in a bed he didn’t need, exposed to all the “fun” things that can happen to people in hospitals, paying hospital prices for something we could do tomorrow in the daylight. At Kaiser Permanente, I sent him home. I knew the scan, and the EEG, and the neurology followup would happen.

And this is the kind of medicine I’m afraid we’re about to lose: better for the patient, more organized care, and less wasteful of the money everyone pays for. But no rug will be pulled out from under Kaiser, or Stanford, or the Cleveland Clinic. It will be a slow, persistent leak, because the healthy people will quietly stop paying for it and find other ways to get their medical care, their way.

After three essays about changing my work and leaving clinical practice, I should probably explain something: I still desperately believe in the model I left a few months ago. I wouldn’t have stayed there as long as I did if I didn’t. That might seem contradictory, but part of what made leaving Kaiser and The Permanente Medical Group difficult was knowing what the system allowed me to do for patients. (And how easy it was for me to get care, too.) So let me tell you why I’m worried we could lose it all.

For the past 100 years, “The Healthcare System” was your only choice. If you needed care you saw a doctor, and that doctor billed your insurance. Even disruptors played by those rules. Sure, One Medical put a nicer front door on primary care. Oak Street Health built a better model for Medicare Advantage patients. But both were still doctors’ offices, still billing the traditional health insurance system. You could switch plans, but you could never really leave. American healthcare was the Hotel California.

Now, and for the first time at scale, healthy Americans have a real choice that has nothing to do with picking a doctor or a plan: it’s whether to stay in the traditional healthcare system at all. Healthcare is getting unbundled, just like cable television. You’ve got options like Counsel Health or Doctronic: AI-enabled or AI-first, chat-based medical care. You’ve got direct primary and even direct specialty care, subscription-based and cash-pay, opting out of insurance and its bureaucracy entirely. Your GLP-1 can ship directly from its manufacturer. Mark Cuban's Cost Plus Wellness wants employers to contract directly with hospitals or medical groups, cutting out the insurance middleman entirely. A bunch of companies will run fifty dollars’-worth of labs for you for $500. “Consumers” can buy the routine care they want directly: faster, cheaper, often better.

But routine care leaving isn’t what worries me. (Arguably, that’s how it should have worked all along.) It's the fact that routine care is also bundled with a $10,000 annual insurance premium, and increasingly, healthy people have reasons to question whether they need to keep paying for the whole package.

If a subscription practice and a few direct-pay services handle most of what healthy people want and need, why keep paying ten grand a year for the whole package? The dangerous part isn’t buying your routine care somewhere else; it’s when healthier people or employers can also move into cheaper coverage priced around their lower expected costs, leaving everyone else in a more expensive pool. Employers with healthy workforces will ask the same question.

What would we lose along with it? Go back to my seizure patient. His MRI happened the next day because a system made sure it did — a system paid to take care of the whole person, not to bill for one scan at a time. That’s one of the things those premiums help support: not just insurance against a big bill, but an organization capable of taking responsibility when something goes wrong. That capability needs a stable way to be paid for, whether it comes from today’s premiums or something better. Robert Pearl, M.D. understood that better than anyone.

The $10,000 Visit

Dr. Pearl was The Permanente Medical Group’s CEO when I started there in 2011, and he became famous for harping on what he called the “$10,000 Visit.” A young, commercially insured, healthy member might seek care once a year from us — maybe for eyeglasses, a rash, or an annual exam — and that single appointment was their entire experience of Kaiser Permanente. It also represented all that the patient asked of us, annually, despite their employer paying thousands of dollars in premiums on their behalf.

Those premiums did two jobs. They paid for that member’s own care, and they quietly paid for someone else’s: the chemo, the NICU stay, the trauma bay staffed at 3am. That second job is the whole point of insurance. Nobody advertises it, but every hospital’s budget depends on it. We were, quietly, asking those members to help pay for a LOT of care they didn’t personally need.

Dr. Pearl’s obvious conclusion was that we’d better make that visit count for that healthy member. We should make it easy for them to get an appointment, and give them a reason to be glad they belonged to KP. In return for everything we were asking of them, we needed to do a good job with the relatively little they asked of us. (While this seems obvious enough, you have to remember that healthcare operates on triage. And since doctors mostly see really, really sick people all day long, medical culture often under-appreciates the needs of the healthy. Myself included, both as a doctor and a healthy person.)

Back then, Dr. Pearl was worried about losing that member to UCSF or Stanford down the street, or to Cigna or United, and with them years of relatively predictable premium revenue. But now enter telemedicine, automation, and AI. Increasingly, someone else is happy to offer the healthy a much more convenient, much more affordable option than health systems can. In 2026, I’m worried about the $10,000 visit because those healthy members might not just leave Kaiser for another insurer. They may find ways to pay much less into the system altogether.

And that would be a real loss, because I personally think systems like Kaiser Permanente are about as close as America has come to building the healthcare system we’ve spent decades saying we want. A 2018 study in the American Journal of Medicine found that by 2015, heart disease mortality among Kaiser Northern California members was 30% lower than the national average. None of that came from smarter doctors (it has plenty). I didn’t become a fundamentally different physician when I left Stanford for TPMG. What changed was whether that MRI got scheduled, whether the follow-up I’d arranged actually happened, and whether I could practice a type of medicine that focused on using our members’ dollars fairly and wisely.

Why Healthy People Are Fed Up

There’s an old saying: when elephants fight, it’s the grass that suffers. In American healthcare, the elephants are health systems and health insurers, fighting over the almighty Healthcare Dollar while patients get slow, fragmented, expensive care. And in 2026, everyone’s tired of being the grass. And now that they have somewhere else to go, the elephants should worry less about winning the fight and more about why anyone would keep paying to be trampled on.

Patients don’t need to understand the fight to notice when their doctor is distracted, their appointment feels rushed, their referral takes weeks (or months), and the next available slot is in December.

Kaiser has always been focused on keeping patients healthy: prevention is the goal. “We win when patients win.” KP doesn’t need a billing code to make preventing a stroke or avoiding an unnecessary hospitalization worthwhile. That’s literally the point of the model.

But Medicare Advantage and the ACA marketplaces adjust payment based on how sick their members are. That process, called risk adjustment, was designed in part to discourage plans from cherry-picking healthy patients. Makes sense, right? Unfortunately, it’s also created a powerful incentive to document every diagnosis, even when it makes little difference to the care of the patient.

The ACA marketplaces and Medicare Advantage do this differently, but the incentive is similar: if your patients look healthier on paper because you're not documenting every diagnosis, your organization can lose a lot of money. (Billions, with a B.) So even a system built to prevent illness has to spend time proving how much “illness” its members have. Aortic atherosclerosis. Senile purpura. Sure, these are real diagnoses, but they often make little difference to what I’m doing for the patient in front of me. And much of that work lands on the same doctors whose time patients were already struggling to get.

Here’s the irony I see: the machinery built to keep plans from cherry-picking healthy patients has become part of the paperwork driving healthy patients away.

And risk adjustment is just one example. Prior authorizations, billing rules, quality-measure reporting, compliance attestations: each is a commitment the system makes to someone other than the patient, and each one is paid for in clinician time. Paperwork, even at Kaiser Permanente, has gotten worse — some of it thanks to programs meant to protect patients from worse care.

At the same time, the population got older and sicker and needed more care. When I arrived at my Kaiser hospital in 2012, on some overnight shifts one ER doctor could actually lie down for an hour while the other ran the department. A nap! In the emergency department! During a shift! By 2013 or 2014, that was over. Nobody banned naps. The department just got too busy.

Add it up and you get a system with no slack in it. And when any system is operating at capacity, small surges create long waits and more expensive care. As I wrote in my previous essay, patients could board in the ER for literal days because there was nowhere safe for them to go. The rest of the hospital wasn’t playing Candy Crush while we struggled. They were drowning, too.

And here’s where those healthy people finally come into the story: triage appropriately pushes them to the back of the line: the stroke before the sprained ankle, the suspected cancer before the itchy rash. More cruel irony:

  • The people paying the most for the system…
  • Who use it the least…
  • Also get to wait the longest to access it.

What a lousy way to show them why they should keep paying $10,000 into the system.

This is what health systems and insurers keep missing: we're giving healthy people every reason to walk away because the system spends its scarcest resource — physician time — proving things to payers and regulators and pushing paper while patients just wait. Healthy and sick alike!

Doctors have somewhere else to go, too. A national study counted 83% more concierge and direct primary care sites in 2023 than in 2018. In Hint Health’s own data, the number of direct primary care clinicians rose 555% from 2017 to 2025. And last week, CMS announced they’d cut the base rate for most doctors by 2.31% starting January 1, 2027. After years of being asked to document senile purpura, I can see why some doctors might stop working with insurance companies and health systems. Why not just be another consumer subscription like Netflix? But then who’s going to see the Medicare patients who can’t afford a membership fee?

Why the Math Is Getting Harder to Ignore

Most young people mostly need good primary care. I’ll use myself as the example. I’m 46 (young, I swear), and in 20 years I think I’ve needed 20 “beyond primary care” appointments, mostly sports medicine and eye care. The services I personally use in a typical year could plausibly be purchased for a modest fee or occasional direct-pay visits. It wouldn’t replace insurance against cancer, a major accident, or a prolonged ICU stay. But compare my ordinary annual needs with the 2026 average employer-sponsored premiums reported by KFF, which were $9,804 for individual coverage and $28,303 for family coverage. I’m not anywhere near using $10,000 of healthcare every year.

Now imagine being even younger and healthier: 27 years old, paying rent, carrying student loans, trying to decide whether you can afford kids, and learning how much your employer pays for healthcare benefits that can’t reliably get you a timely appointment. Younger people already have plenty of reasons to be suspicious of arrangements that seem to work better for older generations. Are we really going to assume that the youth who came up with “OK, Boomer” will be eager to preserve the insurance model out of appreciation for “actuarial risk pooling” when other options are fast, transparent, easy to use, and maybe even higher quality? Who are we kidding?

I don’t think healthy people are selfish for asking the question. The bigger threat is what happens if employers start asking whether they should purchase the traditional package for employees who are telling them they don’t want it.

Employers have spent decades watching healthcare costs rise and negotiating with insurers and health systems over the price of increasingly expensive benefits. Now imagine a business with a largely healthy workforce looking at a $28,000 average family premium and realizing employees mainly want timely primary care, mental health, physical therapy, eye care, a little sports medicine, and decent protection against serious illness. Why not assemble a package that provides those services directly, and pair it with meaningful insurance against truly large medical expenses? If that delivers better care for less money while preserving protection for sick employees, wonderful.

Employers are already reconsidering the package. In September 2026, Becker’s reported an eHealth survey in which 73% of small and mid-size employers currently offering group coverage said they were considering dropping it for 2027. Considering isn’t doing. But that’s not a number anyone should ignore. And some departures replace group coverage with money for employees to buy individual policies, as Bloomberg reported in August.

The more specific danger is that healthier groups find a way to pay according to their own lower expected costs while sicker groups remain in coverage that shares costs more broadly. Then the remaining pool gets more expensive, premiums rise, and the next-healthiest group heads for the door.

Records Aren’t Responsibility

As care gets unbundled, better records and AI-generated summaries will also make it easier to see doctors in different systems. Patients will have way less incentive to get their primary AND specialty care “under one roof” when all the records can be shared. I can see a UCSF primary care doctor and a Stanford cardiologist, go to Sutter’s ER and John Muir’s urgent care, and see a direct specialty care gastroenterologist.

Having all my records available makes that easier. It doesn’t tell me who’s responsible for what happens next. Every one of those doctors could read my seizure patient’s chart. Someone still has to get his MRI scheduled tomorrow. Someone still has to own what happens next, and someone has to pay for that responsibility.

Who Gets Left Behind

Here’s the part I worry about most: a company can make one piece of healthcare cheaper without taking responsibility for the rest. But the rest still has to exist when we need it.

The crucial difference to me is ethical obligation. As an emergency physician, I can’t look at a frail 85-year-old with heart failure, diabetes, and dementia and say, “Sorry, you’re not our demographic.” I take care of the patient who arrives. An integrated system like Kaiser Permanente has a broader responsibility for its covered population, including the members who become extraordinarily expensive.

But a company offering a narrow healthcare service generally can define the patients and problems it serves, and leave everything else to someone else. That’s not necessarily wrong. Some of these companies are genuinely making care better and cheaper. But somebody still has to take care of the patients and problems they don't.

And somebody has to pay for it. Commercial insurance often pays hospitals far more than Medicare does, and those commercial payments help fund the expensive services we all depend on: the ICU, the OR, and the neurosurgeon on call tonight. If new arrangements squeeze out waste, great. But if they also pull money away from hospitals without replacing the funding for those services, something eventually has to give.

Maybe taxpayers should pay more directly to keep those services available. I’m open to that. But if the money that paid for that readiness walks out and nothing replaces it, the readiness goes too. We just won’t find out until we need it.

To be clear here, I think the existing system does need a shake up. And costs also need to come down. I have zero interest in defending a $1,000 bill for a bandaid and Tylenol. Some of this disruption may be exactly what healthcare needs. I want us to recognize what’s worth keeping while we tear out what isn’t.

What I Want to Survive

At Kaiser Permanente and The Permanente Medical Group, where I practiced for 15 years, I worked with people who took prevention, appropriate restraint, and responsibility for very sick patients seriously. It wasn’t perfect, and I have plenty of opinions about its bureaucracy. But systems that want to do things right are worth preserving, even if we eventually pay for it differently. Nobody chooses to be sick. I want the systems that care for the sick — and those values — to succeed.

But the system can’t expect healthy patients to keep financing that care while making their own care so difficult to get. Some of the solution is policy: we have to decide how to finance the readiness and responsibilities everyone depends on. Some of it is much simpler, and already within health systems’ control:

  • Make the one visit easy to get. If our healthiest members can’t get seen for a rash without a three-month wait, someone else will see them in fifteen minutes, and eventually, their annual premium will go with them.
  • Give doctors their time back. Every hour spent documenting senile purpura is an hour not spent on the patient deciding whether we’re worth it. The same tools that make it easy to leave — ambient documentation, automated coding, AI that handles the prior auth — can give that time back, if incumbents use them for that instead of squeezing more diagnoses out of every visit.
  • Tell people what they’re paying for. Most healthy members have no idea their premium keeps the cath lab staffed at 3am. Some might keep paying if they knew. None will if all they ever see is the wait. And if the honest answer is that readiness costs more than healthy members will pay, that’s a conversation for policymakers and taxpayers — not a cost to keep hiding in the premium.
  • Offer care people actually want. Mental health, ortho/sports care, eye care, and easy access to primary and urgent care matter to younger members. As they start families, so do obstetrics, fertility, and pediatrics. The services aren’t exotic. The challenge is making them as easy to use as the apps.

Even traditional fee-for-service systems are starting to recognize that access is a problem for the entire organization. I know the chair of a large department at an academic medical center who was told the next available appointment at her own institution was nine months away. Nine months for a VIP! If that’s what happens to a department chair, imagine everyone else. Some of these organizations are starting to act more like systems that have to keep patients healthy and connected to care, even if their financial incentives haven’t caught up. Meanwhile, KP Northern California has made big strides on convenience and responsiveness.

For example, TPMG launched a program called RESET a few years ago. A study found lower burnout and higher professional fulfillment among participating physicians; according to my friends in primary care, it has also been transformative for access, scheduling, and autonomy. That’s the kind of change I want to see more of: not just making healthcare cheaper, but making the systems responsible for caring for us actually work better.

The value of that system became obvious to my seizure patient at midnight, after his first seizure. It had been much less obvious when he was healthy. We can’t expect people to keep helping finance that capacity if their ordinary experience of the system is a six-month wait. “Because we need your money” isn’t a persuasive answer.

I want Kaiser Permanente and systems like it to survive because I know what I could do for patients when the system behind me worked. Nobody chooses to have their first seizure, or get cancer, or wind up in an ICU. But we all depend on someone being ready when those things do happen. I want that seizure patient to get his MRI tomorrow. And I want to help the healthy person stay healthy AND get them an appointment for their rash before they decide the whole arrangement is a ripoff. All of us deserve a system that works.

Ten thousand bucks a year! We'd better make it count.


This is somehow part 4 in this series (one more to go!) on my reflections in healthcare. If you're just joining us, you've missed some zingers.

Part 1: Back to Where I Started... 21 Years Later

Part 2: What Do You Mean by Meaning?

Part 3: Always Open. Already Full.

You can check out more of my writing here on LinkedIn by following me, or on drgrahamwalker.com.