CMS WISeR model: PT is cited, not reviewed

Since January 15, 2026, a handful of outpatient procedures billed to Original Medicare in six states have to clear an algorithm before Medicare pays for them. CMS calls the WISeR Model the first Innovation Center model in which “technology innovators are the only model participants,” and the first to pay for using those tools to check that payment follows Medicare’s coverage, documentation, payment and coding rules. It is scheduled to run for six performance years.

No therapy code is in it. Not one. What is in it is a short list of procedures whose federal documentation requirements name physical therapy on the page.

What is the WISeR model?

WISeR stands for Wasteful and Inappropriate Service Reduction. The CMS Innovation Center describes it as using “enhanced technologies, such as Artificial Intelligence (AI) and Machine Learning (ML), along with human clinical review” on a pre-selected set of services, and it runs from January 1, 2026 to December 31, 2031 in New Jersey, Ohio, Oklahoma, Texas, Arizona and Washington. It does not touch Medicare Advantage.

The participants are not health systems. They are six technology companies, one per assigned state: Cohere Health in Texas, Genzeon in New Jersey, Humata Health in Oklahoma, Innovaccer in Ohio, Virtix Health in Washington and Zyter in Arizona. CMS pays them “a percentage of the expenditures associated with averted wasteful, inappropriate care”, adjusted for performance measures that include provider experience.

A provider in a WISeR state who furnishes an in-scope service picks a lane: submit a prior authorization request, or skip it and take a post-service, pre-payment medical review instead. Determinations typically come back within three calendar days, two if the request qualifies as expedited.

Is physical therapy subject to WISeR prior authorization?

No. APTA states that “[p]hysical therapist services are currently not included under the WISeR model,” and version 7.0 of the CMS Provider and Supplier Operational Guide, dated July 24, 2026, backs that up: there is not a single 97000-series therapy code anywhere in its appendices.

What the guide does contain is a documentation checklist per service. Five of the thirteen live categories require proof that conservative care was tried, and CMS spells out what it means by conservative care.

The table cuts both ways. Wound care, sleep apnea and trigeminal neuralgia have nothing to do with an outpatient ortho clinic. And the knee category, the one CMS and its critics both reach for first, does not list physical therapy in its documentation requirements at all.

What actually changes for a PT clinic

The epidural steroid injection rules are the sharpest example, and CMS rewrote them in version 6.0 of the guide on April 24, 2026. To get an ESI affirmed in a WISeR state, the requesting physician must document “pain duration of at least 4 weeks and the inability to tolerate noninvasive conservative care OR medical documentation of failure to respond to 4 weeks of noninvasive conservative care.” The injection has to be performed “in conjunction with conservative treatments,” and CMS lists physical therapy second, after medication. Then a separate line: documentation “that the patient is part of an active rehabilitation program, home exercise program, or functional restoration program.” There is also a requirement that an objective pain scale or functional assessment be recorded at baseline and repeated at each follow-up on the same scale.

Read that as a therapist and the consequence is small but real. Your progress note and your objective measures are now evidence in a file somebody else submits, on a three-day clock, to a vendor paid out of savings. That does not make you responsible for their authorization. It does mean a note reading “tolerating well” is worth less to a referral partner in Ohio this year than it was last year, and that the same instrument at the eval, at each progress report and at discharge is the version that travels. Our outcome measures post covers which ones hold up; the medical necessity post covers what commercial reviewers want from the same note.

Does the evidence support PT as the alternative for knee arthroscopy?

For this specific knee procedure, yes, and the scope matters.

Medicare has not covered arthroscopic lavage alone for the osteoarthritic knee, or debridement for patients presenting with knee pain only, or lavage and debridement for severe osteoarthritis, since NCD 150.9 took effect on June 11, 2004. What stayed at local contractor discretion was debridement for patients without severe osteoarthritis who present with symptoms other than pain alone. So WISeR is not a new restriction on knee arthroscopy. It is automated checking of a 22-year-old one.

The trial record on this procedure is unusually clean, and most of it arrived after the NCD did. Moseley’s 2002 NEJM trial randomized 180 patients with knee osteoarthritis to arthroscopic debridement, arthroscopic lavage or placebo surgery, and found that “at no point did either of the intervention groups report less pain or better function than the placebo group.” Kirkley’s 2008 trial assigned 92 patients with moderate-to-severe knee osteoarthritis to surgery plus optimized physical and medical therapy and 86 to physical and medical therapy alone, and concluded that arthroscopic surgery “provides no additional benefit to optimized physical and medical therapy.” Thorlund’s 2015 BMJ meta-analysis of nine trials of arthroscopic surgery for the degenerative knee found a pain benefit of 2.4 mm on a 0 to 100 scale that was gone by one to two years, no significant benefit to physical function, and harms including symptomatic deep vein thrombosis at 4.13 events per 1000 procedures.

Hold the scope steady, though. Those trials are about the osteoarthritic and degenerative knee, not arthroscopy generally, and none of them speaks to acute injury, a locked knee or a young athlete. Nor does the famous head-to-head apply here: Katz’s 2013 METEOR trial randomized 351 patients with a meniscal tear and mild-to-moderate osteoarthritis to partial meniscectomy or a standardized physical-therapy regimen, and partial meniscectomy is not the procedure WISeR reviews. It found no significant difference in WOMAC function at six months. It also found that 30% of the patients assigned to physical therapy alone had the surgery within six months anyway.

Has anyone measured a referral shift?

No. There is no published dataset showing WISeR moving referrals toward physical therapy, or away from anything, and anyone selling you one is guessing. CMS has not released implementation data at all, which is the substance of a June 22, 2026 letter from 31 House members asking for affirmation rates, turnaround times, appeal and overturn rates and peer-to-peer wait times, state by state. That letter reports that “patients and providers in the affected states have reported care denials, long wait times, and increased administrative burden.”

It is contested on the record: the Senate rejected a motion to proceed to S.J.Res. 198, which would have disapproved the WISeR prior authorization rule, by 46 to 50 on July 16, 2026. CMS’s answer to the central objection is specific. Its FAQ says any recommendation against affirming coverage “will be made by an appropriately licensed human clinician, not a machine,” that peer-to-peer review is available, that resubmissions are unlimited, and that participants “will be financially penalized for inappropriate denials.”

What a practice can reasonably do

Put a date on anything you conclude from this. WISeR is a CMS Innovation Center demonstration running under Section 1115A authority, and it moves: seven versions of the operational guide since October 10, 2025, two services pulled out of the launch, and a standing note that services with broader coverage criteria are limited to select indications for now and “can be expanded in the future.”

If your clinic is in Arizona, New Jersey, Ohio, Oklahoma, Texas or Washington and you take referrals from an interventional pain or spine practice, the useful move is unglamorous: ask the referring office whether WISeR touches them, find out what they need from your notes, and send objective measures on the same scale every time. There is a compounding reason to bother. Since July 2026, WISeR participants automatically exempt individual providers from prior authorization at the NPI level once they submit at least 10 requests and clear an affirmation-rate threshold. Whoever holds the pen on that request is counting affirmations, and part of what earns one is documentation you wrote.

The rest of prior authorization has not changed, and for most practices that fight is still with commercial plans and Medicare Advantage, where therapy genuinely is the reviewed service.