CMS put out its 2026 participation numbers and the value-based crowd took a victory lap. 511 ACOs in the Medicare Shared Savings Program now. 12.6 million people in traditional Medicare attributed to them. The largest the program has ever been, up from 476 ACOs the year before. Add the most recent reconciled results on top of it, where ACOs earned around $4.1 billion in shared savings and saved Medicare roughly $2.5 billion in PY2024, and you'd think we finally turned the corner.
Here's the part nobody on the panel says out loud. Signing the contract is not the intervention. And most of these organizations signed a contract and changed almost nothing about what happens to a patient after they walk out of the office.
I spent years inside payers building these deals. I'll tell you what a lot of value-based care actually was. A financial modeling exercise. We shifted risk onto a health system, and if they happened to do well we cut them in, and if they did poorly they owed us money back. The spreadsheet changed. The benchmark changed. The care a diabetic got when she went home and got confused about her insulin did not change one bit. We dressed up risk transfer as transformation and called it a day.
The 2026 data quietly proves the point. Look at who actually earns. The high-needs tracks in ACO REACH, the ones built around the sickest, most complex Medicare patients, posted net savings rates north of 13 percent in 2023. Standard ACOs came in under 3 percent. I want you to sit with that gap. The same Medicare program, the same shared-savings math, and one group saved nearly five times the rate of the other. (Those figures come from reporting on the 2023 REACH results, so verify the exact numbers before you put them in front of a CFO.)
Avalere ran a similar cut and found the top MSSP performers looked a lot like those high-needs ACOs. They carried far more dually eligible patients and far more patients over 85 than the average. The pattern is not subtle. The money is in managing the people who get sick, get readmitted, and bounce back to the ED. It is not in the logo on the contract.
So why do most ACOs underperform? Because they treat the contract as the work. They attribute a panel, they run some quality reports, and they wait for a check. Health Affairs said it cleaner than I can. Most ACOs chase savings through administrative moves and light-touch outreach, and very few ever touch the real cost drivers, which are specialty spend, clinical standards, and acute care. Translation. They are not actually managing patients between visits. They are managing a measurement.
A physician sees a patient for twelve minutes and then that patient goes home for ninety days. That ninety days is where the cost lives. That is where the missed medication, the blood pressure nobody checked, the swelling that becomes a CHF admission all happen. A risk contract does not show up in that patient's kitchen. A nurse does. A care coordinator does. An outreach call that catches a problem on day forty instead of day eighty does. If you are not funding that work, your shared-savings number is a coin flip, and the data says the coin is weighted against you.
It gets more urgent. ACO REACH, the one model that actually carved out a dedicated track for complex high-needs patients, sunsets on December 31, 2026. The successor, ACO LEAD, does not start until January 1, 2027, and runs as a ten-year model. CMS also tightened the methodology for 2026, leaning less on generous regional benchmarks and capping risk-score growth, which is a polite way of saying the easy money from coding and favorable benchmarks is drying up. So the organizations that were earning on paper rather than on care are about to find out the paper does not pay anymore.
This is good news if you build real clinical infrastructure. It is very bad news if you bought a contract and a dashboard.
What do you actually do Monday morning? Stop counting covered lives. That number tells you nothing about whether you will earn. Pull your attributed panel and count something harder. How many of your complex patients, your duals, your 85-and-up, your CHF and uncontrolled diabetes and uncontrolled hypertension patients, got a real clinical touch in the last thirty days. Not a portal message nobody opened. An actual human contact that changed something. If that number is low, you already know how your reconciliation ends.
Then ask the uncomfortable question. Can your practice realistically staff and run that between-visit care itself, with your own nurses, on your EMR, while your physicians are already underwater? For most groups the honest answer is no, and that is fine. That is exactly the work that gets handed off. We take the risk on the codes so you don't carry it, we run the program end to end, and we get paid out of what actually gets billed and collected. The point is not who does it. The point is that somebody does it, because the contract sure won't.
I am not anti-value-based care. I think it is the right direction and the participation growth is real. What I am against is pretending that joining an ACO is the same as practicing differently. It isn't. The systems that win the next five years will be the ones that treated the contract as the easy part and the care redesign as the actual job.
The contract was never going to save the patient. You have to go to the home.
Sources:
