What Is the CMS TEAM Model? The 30 Days After Surgery

The CMS TEAM model is a mandatory Medicare payment model that makes selected hospitals accountable for a patient’s care from surgery through the 30 days after discharge. TEAM stands for the Transforming Episode Accountability Model. According to the Centers for Medicare & Medicaid Services, it began on January 1, 2026, runs for five performance years, and covers people with Original Medicare who undergo one of five procedures: lower extremity joint replacement, surgical hip and femur fracture treatment, spinal fusion, coronary artery bypass graft, and major bowel procedures. CMS lists 716 participating hospitals.

Eddie Qureshi, founder and CEO of Rainfall Health, has built his company around the model. On Open Door Salon he told Lori Ellis what TEAM asks of hospitals, what it leaves open, and where he thinks AI belongs in those 30 days.

What does the TEAM model ask hospitals to do?

It asks them to treat a surgical episode as one accountable stretch of care rather than a procedure followed by a discharge. Qureshi described the two questions he thinks Medicare is answering with it:

How do you treat the patient as a whole and how do you provide regulations to do it safely and what does success look like, right?

He said that is where policy set by the government goes a long way, because it gives every participating hospital the same rules and the same definition of success.

What does the TEAM model leave to hospitals?

The means of getting there. Qureshi was direct that the goals arrive without the machinery to meet them:

It sets a set of ground rules and it defines success metrics. And now here’s the challenge is those are goals. To actually get there, you need operations, you need investment, you need infrastructure. And that’s been left open right now.

Rainfall Health narrowed its business to fill that gap. Qureshi said the company narrowed over seven years from rural health work with Indigenous communities and veterans to a single focus:

We work on managing episodes of care for Medicare patients. It’s incredibly specific.

Do hospitals in the TEAM model know they are in it?

Not always, by Qureshi’s account. He described a conversation, about two months before the recording, with a CFO at a participating hospital who had not heard of the model, and his reaction to it:

That’s really bad because you’re 8 months into actually operating on it.

At the other end, he said, some systems have assembled teams of around 30 people to work out their strategy. The spread matters for anyone selling into these hospitals, because the same policy produces very different starting points. How that narrow focus shortens Rainfall’s sales cycle with these hospitals is in our post on why hospital AI pilots fail to scale.

How can AI help patients in the 30 days after surgery?

By answering first and giving the patient one place to call. Lori Ellis described the experience many people know: home from surgery, calling a nurse every time a temperature moves. Qureshi said the application that gets the strongest response from both hospitals and patients is an AI care coordinator, which by his account speaks 76 languages:

That means that my grandmother could just pick it up and start speaking in the language that she’s familiar with and it could help her on the first ring.

He tied that to treating the patient as a consumer:

You’re not on hold, you’re not on a phone tree. You just start talking and somebody is there to help you.

The second benefit he described is a single point of entry. A patient in the middle of recovery often does not know whom to call:

Do I call my original, the surgeon who did my work? Do I call my family physician who helps me on the day-to-day? Like, who am I calling?

When does a human take over?

When the case needs one. Qureshi said Rainfall records each conversation in the same format as the electronic health record and translates it into the clinician’s preferred language, so a human care coordinator can intervene in real time. He estimated that most calls do not need that:

Ninety percent of the time you don’t need that additional intervention. You just need information. You just need advocacy.

The argument is about bandwidth. If the routine questions are handled, he said, the humans can give the complicated cases their full attention instead of feeling rushed. Where patient data goes matters as well: health information patients put into consumer AI apps usually falls outside HIPAA.

How long do Medicare models like TEAM take to arrive?

Years. Qureshi explained the timeline to the audience before describing the hardest decision of his career:

Sometimes they’re in the works for almost a decade before they come out.

Rainfall committed to TEAM well before it took effect. In his words, the company bet everything on it three years before we knew it was going to see the light of day. For a company whose market is defined by a payment rule, that is the planning horizon. Payment rules shape strategy across life sciences too, which is the subject of our earlier post on when a biotech should plan for reimbursement.

Companies that want to reach the leaders who follow this conversation can learn about sponsoring Open Door Salon.

This post is drawn from the recorded, on-the-record conversation between Lori Ellis and Eddie Qureshi on Open Door Salon.