Last Friday we talked about prospective payment, the model where the payer decides in advance what an episode of care is worth and the hospital gets that amount whether they spend more or less. It was a revolution in 1983 because it broke the link between volume and revenue at the hospital level.
This week we’re talking about the model that breaks that link even more completely. Capitated payment.
Under capitation, the payer hands the provider a fixed amount of money per patient per month and says, in effect, here is your budget. Take care of this person. If they stay healthy, you keep what you don’t spend. If they get sick, that comes out of your pocket.
It is the most radical departure from fee for service that American healthcare has ever attempted. And like a lot of things in American healthcare, it started with a desert, a dam, and a man who needed to keep his workers alive.
What capitated payment actually is
The word “capitation” comes from the Latin caput, meaning head. A capitated payment is a payment per head. Per person. Per member per month, in the language of modern contracts. PMPM, if you’ve ever sat in a meeting with health plan executives.
Here is the mechanism. A health plan contracts with a provider organization, usually a group of physicians or an integrated delivery system, and agrees to pay that organization a set amount each month for every patient enrolled with them. Maybe it’s $400 a month for a healthy 35 year old. Maybe it’s $1,800 a month for a 78 year old with diabetes and heart failure. The amount is risk adjusted based on how sick the population is expected to be. But once the check is cut, the provider owns the risk.
If the patient never comes in all year, the provider keeps every dollar. If the patient ends up in the ICU for three weeks, the provider eats the loss.
This flips the entire incentive structure of American medicine. Under fee for service, the provider makes money when the patient gets sick. Under capitation, the provider makes money when the patient stays well. Suddenly preventive care, smoking cessation, blood pressure control, diabetes management, all the things that fee for service treats as charity work, become the things that keep the lights on.
That is the theory, anyway. The practice is more complicated.
Where it started: Kaiser Permanente and the birth of the HMO
The story of capitated payment in America really begins in 1933, in the Mojave Desert, with a young surgeon named Sidney Garfield.
Garfield had built a small hospital outside Desert Center, California, to care for the thousands of workers building the Colorado River Aqueduct. He was going broke. Insurance companies were slow to pay. Some workers had no coverage at all. Garfield was treating people for free because he could not bring himself to turn them away.
An insurance agent named Harold Hatch proposed a solution. Instead of paying Garfield after the fact for each treatment, what if the insurance companies paid him a fixed amount up front, every day, for every worker covered? Five cents per worker per day. Workers could kick in another nickel out of their own paychecks for non work related illness.
It was the first capitated payment arrangement in American medicine. And it worked. Garfield’s hospital stabilized. He started focusing on prevention, on safety education, on keeping workers from getting hurt in the first place, because every injury he prevented was money he kept.
A few years later, the industrialist Henry J. Kaiser recruited Garfield to set up a similar plan for the workers building the Grand Coulee Dam, then again for 30,000 shipyard workers in Richmond, California, during World War II. When the war ended, Kaiser and Garfield made a decision that would change American medicine. They opened the plan to the public.
On July 21, 1945, the Permanente Health Plan, soon to be renamed Kaiser Permanente, became available to anyone who wanted to enroll. By 1958, membership had grown to over 600,000 people.
What Kaiser had built was the prototype for what would later be called a Health Maintenance Organization, or HMO. The defining features were already there. A defined population of enrolled members. A fixed prepaid amount per member. An integrated group of physicians who worked exclusively for the plan. Hospitals owned by the same organization that owned the insurance. Everyone on the same side of the ledger.
The term HMO itself didn’t exist yet. It was coined in the early 1970s by a Minnesota physician named Paul Ellwood, who pitched the concept to the Nixon administration as a way to control runaway healthcare costs. Nixon, who had been impressed by the Kaiser model, embraced it. The HMO Act of 1973 provided federal grants and loans to encourage the formation of HMOs and required employers of a certain size to offer one as an option.
The HMO boom of the 1980s and 1990s followed. By the late 1990s, more than 80 million Americans were enrolled in some form of HMO. The model that started with five cents a day in the Mojave Desert had become a dominant force in American healthcare.
It also became deeply unpopular.
The HMO backlash
The problem with capitated payment is that the same incentive that rewards prevention also rewards under treatment.
If a provider keeps the difference between the capitated payment and what they actually spend, they have a financial reason to spend less. Sometimes that means more flu shots and better diabetes management. Sometimes it means denying a referral to a specialist, refusing an MRI, or making a patient jump through hoops to get a procedure they actually need.
By the late 1990s, the public had had enough. Stories of HMOs denying care became a staple of news coverage. The phrase “drive through delivery,” referring to insurers limiting maternity stays to 24 hours, entered the political vocabulary. Movies and television portrayed HMO administrators as villains. In 1997, the actress Helen Hunt got an extended ovation in a movie theater for a single line about her HMO, delivered with profanity I won’t reproduce here.
The backlash worked. Many HMOs loosened their restrictions, expanded their networks, and shifted toward a more flexible model called the Preferred Provider Organization, or PPO, which paid providers on a discounted fee for service basis rather than capitating them. Pure capitation declined. By the early 2000s, the HMO as Kaiser had originally conceived it was, outside of a few markets like California and the Pacific Northwest, in retreat.
But the underlying idea, that paying per person rather than per service might be a better way to align incentives, did not go away. It just went looking for a new vehicle.
The Affordable Care Act and the voluntary ACO
When the Affordable Care Act was being drafted in 2009 and 2010, the people writing it had a problem. They knew fee for service was driving costs up. They knew prospective payment had helped contain hospital costs but had not done much for the rest of the system. They wanted to push providers toward something that looked more like capitation, where the financial reward came from keeping people healthy rather than treating them sick.
But they also knew the political third rail they were walking near. The HMO backlash was less than fifteen years in the rearview mirror. They could not propose anything that looked like forcing patients into a closed network or capitating providers in a way that gave them a financial incentive to deny care.
Their answer was the Accountable Care Organization, or ACO.
The ACO is, in essence, a softer version of capitation. The model that became the Medicare Shared Savings Program, created by Section 3022 of the ACA in 2010 and launched in 2012, works like this. A group of doctors and hospitals voluntarily agrees to be collectively accountable for the cost and quality of care for a defined population of traditional Medicare beneficiaries. Medicare sets a spending benchmark for that population based on historical costs. If the ACO comes in under the benchmark and meets quality standards, they share in the savings. If they come in over the benchmark, depending on the track they’ve chosen, they may share in the losses.
Crucially, patients are not locked in. A Medicare beneficiary assigned to an ACO can still see any doctor who accepts Medicare. They cannot be denied a referral. They cannot be told their specialist visit isn’t covered. The ACO bears the financial risk for the patient’s total cost of care, but the patient retains full freedom of choice.
This is what makes the ACO voluntary in two senses. The providers choose to participate. And the patients are never required to stay in network.
It is capitation with the sharp edges sanded down. The provider still has a financial reason to keep patients healthy, to coordinate care, to avoid unnecessary admissions and duplicate testing. But the provider cannot stop a patient from seeking care elsewhere, which means the most aggressive forms of cost cutting that gave HMOs their bad name are not available.
How it’s going
Fifteen years in, the ACO model has grown into something substantial. As of 2025, the Medicare Shared Savings Program included 476 ACOs, with more than 655,000 healthcare providers, caring for over 11.2 million traditional Medicare beneficiaries. CMS reported that for performance year 2024, ACOs generated $4.1 billion in performance payments to themselves and saved Medicare $2.5 billion against benchmarks, the highest savings in the program’s history.
CMS has set a goal of having every traditional Medicare beneficiary in an accountable care relationship by 2030. As of January 2025, they were at 53.4 percent.
The results have been better than the HMO experiment, but more modest than the most enthusiastic proponents had hoped. Savings are real but not enormous. Some ACOs do well, particularly those led by primary care physicians. Hospital led ACOs have generally done worse, in part because hospitals make more money when their beds are full, and an ACO that is actually keeping people out of the hospital is cannibalizing its own revenue. Attrition is significant. The total number of ACOs peaked at 548 in 2018 and has since declined.
But the model has survived three administrations and a deeply polarized political environment, which is more than can be said for many ACA provisions. And the underlying concept, that providers should bear at least some of the financial risk for the populations they serve, has spread far beyond Medicare. Commercial insurers have built their own ACO contracts. Medicaid programs in many states have adopted similar arrangements. Medicare Advantage, the privately administered version of Medicare, is essentially a fully capitated program covering more than half of all Medicare beneficiaries today.
Why this matters
Capitated payment is the most ambitious idea in American healthcare. It is also the most dangerous. Done well, it builds organizations like Kaiser Permanente, which consistently rank among the highest quality and most efficient health systems in the country. Done poorly, it builds the kind of HMO that became a punchline in 1990s movies, the kind that denied care to make a margin.
The ACO is the country’s best attempt so far to capture the upside of capitation while avoiding its worst failure models. By keeping participation voluntary and preserving patient choice, it tries to harness the incentive to keep people healthy without unleashing the incentive to deny them care. Whether that compromise produces enough savings and enough quality improvement to bend the cost curve is still an open question. The early results are promising but not transformative.
What is clear is that capitation, in some form, is where American healthcare is heading. The fee for service world we grew up with is slowly being dismantled. What replaces it will look something like the model Sidney Garfield invented in the desert in 1933, adapted for an era that has learned, sometimes painfully, what can go wrong when you pay providers to spend less.
Next Friday we’ll talk about pay for performance, the model that tries to bolt quality measurement onto whatever payment system you happen to be using.