WISeR

For sixty years, one of the defining structural distinctions between Original Medicare and Medicare Advantage was prior authorization. MA plans used it routinely: 99 percent of MA enrollees are in plans requiring PA for some services, and in 2023 MA plans made 1.8 prior authorization determinations per enrolled beneficiary. Traditional Medicare did not. In FY2023, the existing Medicare FFS prior authorization programs reviewed 3.1 million claims, representing less than one percent of the 1.2 billion total Part A and B claims processed that year. The asymmetry was so pronounced that it served as a foundational argument for choosing Original Medicare over MA — the argument that a beneficiary’s doctor, not an insurance plan, would make treatment decisions without a middleman.

On January 1, 2026, that asymmetry narrowed. Not eliminated, and not reversed, but the boundary moved in a direction it has never moved before.

The Wasteful and Inappropriate Service Reduction Model, announced June 27, 2025, and launched on schedule, introduced AI-powered prior authorization and prepayment medical review into Original Medicare fee-for-service for the first time at any material scale. Six states, four MAC jurisdictions, a defined list of services historically associated with fraud, waste, and low clinical value, and a structural design that makes provider participation functionally mandatory even though the model’s official classification is voluntary. WISeR is simultaneously a fraud prevention initiative, a test of AI-assisted clinical review infrastructure, and a proof-of-concept for whether prior authorization can be imported into FFS Medicare without generating the access harm that has dogged its use in MA.

Whether it achieves all three purposes, or trades success on one against failure on the others, is the question that will define its legacy.

What WISeR Tests

#

WISeR’s formal description is precise: a six-year CMMI model running from January 1, 2026 through December 31, 2031, structured in two three-year agreement periods, operating in six states across four MAC jurisdictions, in which CMS contracts with technology companies to implement AI-assisted prior authorization and prepayment medical review for a defined list of services that CMS characterizes as vulnerable to waste, fraud, and abuse.

The model tests four distinct things simultaneously. The first is whether AI and machine learning tools can conduct clinically accurate medical necessity determinations at the speed and scale required for a functional PA system in FFS Medicare. The second is whether a novel contractor payment structure, compensating vendors based on a share of averted expenditures rather than a traditional acquisition fee, creates the right incentives or the wrong ones. The third is whether a gold-carding mechanism, planned for mid-2026, can reward high-compliance providers with reduced administrative burden and whether that mechanism can be scaled. The fourth is whether the political and legal durability of PA in FFS Medicare is sufficient to support expansion if the first three tests succeed.

All four are genuinely uncertain. The model’s design encodes assumptions about each that are contestable, and the contestation is already active.

The Service List: Skin Substitutes and the Spending Surge

#

CMS selected WISeR’s target services based on three criteria: the services are non-emergent, they are subject to prior authorization under MA plans and commercial insurers with existing publicly available coverage requirements, and they have documented histories of waste, fraud, or low-value billing patterns in FFS Medicare.

The full list for the 2026 performance year includes electrical nerve stimulator implants, sacral nerve stimulation for urinary incontinence, phrenic nerve stimulators, induced lesions of nerve pathways, epidural steroid injections for pain management, percutaneous vertebral augmentation for vertebral compression fractures, cervical fusion, knee arthroscopy for osteoarthritis, hypoglossal nerve stimulation for obstructive sleep apnea, incontinence control devices, and skin and tissue substitutes. Two services originally scheduled for 2026 inclusion were delayed: deep brain stimulation and percutaneous image-guided lumbar decompression for spinal stenosis. Those two represented less than one percent of WISeR service spending and were removed because each had clinical study approvals or NCD complexities that required additional specification time before the review infrastructure could be deployed.

The skin substitute category deserves separate treatment, because its spending trajectory is unlike anything else in the WISeR service list. According to KFF analysis of CMS claims data, traditional Medicare spending on skin and tissue substitutes grew from $509.6 million in 2019 to $10.3 billion in 2024, a nearly 2,000 percent increase in five years. That single category accounted for 83.4 percent of all WISeR service spending in 2024. In the first half of 2025, skin substitute spending was nearly 3,000 percent higher than in the comparable 2019 period. The OIG has documented unusual billing patterns, inadequate pricing transparency from manufacturers, and multiple fraud instances in this category. The MedPAC figure of approximately $6 billion in annual FFS Medicare spending on services with little or no clinical benefit, cited at multiple points in the WISeR model documentation, understates the skin substitute problem by the time the model launched. CMS’s own claims data had already shown the 2,000 percent growth trajectory before the January 2026 launch date.

The clinical case for skin substitutes is real in legitimate wound care: chronic, non-healing lower extremity wounds in diabetic patients, venous ulcers, and pressure injuries can benefit from bioengineered skin products when conservative wound management has failed and the patient meets specific criteria. The billing abuse that drove the spending surge involves applying these products, which have very high per-unit billing values, to wounds that do not meet coverage criteria, billing at the maximum reimbursable quantity regardless of actual product use, and in some cases billing for application without documentation of the wound characteristics required under the applicable Local Coverage Determinations.

Prior authorization is the correct tool for this specific problem. The question is whether it is adequately precise — whether the WISeR review infrastructure can accurately distinguish covered from non-covered applications, and whether it can do so at the volume skin substitute claims represent, without creating delays that harm the patients with legitimate wound care needs.

The Contractor Structure: Who Does the Review and How They Are Paid

#

WISeR’s most structurally novel feature is that the model’s participants are not health care providers or insurers. They are technology companies. This is the first CMMI model in which technology companies are the only model participants. CMS selected vendors through a competitive application process that closed July 25, 2025. Participants were announced November 6, 2025, with each assigned to a specific MAC jurisdiction or geographic region within the six WISeR states.

The contractor compensation structure is what generated the most significant stakeholder concern before and after launch. Participants are paid a percentage of demonstrated savings from avoided payments for WISeR-selected services. That percentage is adjusted based on the participant’s performance on process measures related to provider experience, but the baseline compensation mechanism is averted expenditures. A contractor that determines more services are not medically necessary receives more revenue. A contractor that affirms more services receives less.

The American Hospital Association, in its October 2025 comment letter, drew an explicit analogy to the MultiPlan controversy: the concern that a compensation model tied to denial volume creates financial incentives for inappropriate restrictions that are structurally similar to the incentives critics have identified in other health care utilization management contexts. The AHA recommended a flat-fee compensation structure to remove the financial motivation for restricting beneficiary access.

CMS designed guardrails to address this concern. Contractors are required to use AI as a process tool, not a final arbiter. When a service is not affirmed through the prior authorization pathway, a human clinician with relevant clinical expertise must review the determination before the denial is finalized. Audits of contractor performance are built into the model design. And the provider experience performance measures, which adjust the averted-expenditure percentage, create a countervailing incentive: contractors that generate excessive provider friction or inappropriate denials face payment adjustments.

Whether these guardrails are sufficient is a live empirical question. The concern is not that CMS failed to attempt mitigation. The concern is whether the mitigation is structurally robust against a compensation architecture that, at its foundation, pays contractors more for saying no.

CMMI Director Sutton’s October 2025 webinar framing was direct on this point: contractors are incentivized to “get the determination right,” not to deny. The coverage determinations that generate savings are specifically those in which the service did not meet Medicare’s existing coverage criteria. A medically necessary service that meets NCD and LCD requirements should be affirmed, and the contractor’s process measures track affirmation accuracy alongside denial rates. The design intent is clear. The question of whether the intent survives implementation at scale is what the six-year model test is designed to answer.

The Three Pathways: Voluntary Classification, Mandatory Reality

#

WISeR is formally classified as a voluntary model. The classification is technically accurate and operationally misleading.

All Medicare-enrolled providers and suppliers in the six WISeR states are affected by the model for the services on the WISeR list. A provider in an affected state who does not submit a prior authorization request before furnishing a WISeR-listed service does not escape review. Instead, the claim is automatically routed to prepayment medical review conducted by the WISeR participant, with payment held pending completion of that review. The practical effect is that every provider in the WISeR states furnishing covered services on the WISeR list faces a choice between three pathways — none of which is simply billing and getting paid as before.

The first pathway is direct prior authorization to the WISeR technology participant. The provider submits documentation before furnishing the service. The participant reviews it, makes a determination, and issues an affirmation or non-affirmation within 72 hours, 48 hours for expedited cases. An affirmed service can proceed. A non-affirmed service triggers provider notification and, if the provider proceeds, a required Advanced Beneficiary Notice of Non-Coverage to the beneficiary before the service is furnished. All existing appeal rights are preserved: providers and beneficiaries retain the right to appeal denied claims through Medicare’s administrative appeals process.

The second pathway is submission through the MAC, which routes the request to the WISeR participant. This pathway exists to minimize disruption for providers whose existing workflows involve MAC communication, and to preserve the MAC’s role in the jurisdictional claims processing infrastructure.

The third pathway is doing nothing — furnishing the service and submitting a claim without a prior authorization request. That claim goes to prepayment review, not to the WISeR participant directly, but the review is conducted under the WISeR framework before payment is released. The prepayment review pathway is slower and operationally more disruptive than the prior authorization pathway; it is the functionally punitive option for providers who decline to engage with the PA system.

The “voluntary” classification reflects the fact that providers can technically choose not to seek prior authorization. They cannot choose to avoid the review.

Geographic Footprint: Why These States

#

The six WISeR states, New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington, were selected based on a combination of criteria: comparison feasibility for the evaluation design, service volume for the WISeR-listed services, geographic diversity across different practice environments, and MAC jurisdiction structure.

The selection spans four MAC jurisdictions: JL (New Jersey), J15 (Ohio), JH (Oklahoma and Texas), and JF (Arizona and Washington). This multi-jurisdictional structure ensures that the model tests contractor performance across different MAC administrative environments, different regional claims processing systems, and different LCD landscapes. Skin substitute LCDs, for example, vary across MAC jurisdictions; WISeR’s applicability to skin substitutes is limited to jurisdictions with active LCDs in place.

Together, the six states cover nearly one in five Medicare FFS beneficiaries — a footprint large enough to generate statistically robust evaluation data and large enough that the model’s operational performance will have visible beneficiary and provider effects before the first formal evaluation report is published.

States not included in WISeR are affected in a different way: providers in those states are watching carefully, because CMS has stated it will evaluate WISeR’s outcomes before considering geographic expansion. If the model achieves certifiable savings without generating the access harm critics predicted, the case for national expansion is straightforward. If it generates savings but also generates elevated inappropriate denial rates, complaint volume, and beneficiary harm, the expansion case is complicated by a quality problem that may not be easily fixed through administrative adjustment.

AI in the Review Process

#

The role of AI and machine learning in WISeR is operationally significant and deliberately bounded. CMS describes AI as serving “primarily as a process tool to streamline approvals.” The AI layer is intended to automate the documentation matching and coverage criteria verification steps that make manual PA review slow and labor-intensive, not to make autonomous clinical judgments that are not subject to human review.

CMS cited evidence that AI has the potential to automate between 50 and 75 percent of the manual work involved in processing prior authorizations, including prior experience with the CMS prior authorization program for certain hospital outpatient department services, where targeted PA contributed to a decline from $51 million in the second half of 2019 to $35 million in the second half of 2023 in spending on specified OPD services.

The model does not specify which AI tools or methodologies participants must use, beyond the requirement that they have FedRAMP-certified workflows, comply with FISMA regulations and CMS information security frameworks, and enter into HIPAA business associate agreements with CMS and the MACs. This gives participating technology companies flexibility in their AI implementation designs, which is appropriate for an exploratory test, and also means that different participants may deploy materially different AI architectures across their assigned jurisdictions. The evaluation design will need to account for AI methodology variation as a confounding variable in any comparison of contractor performance across jurisdictions.

When AI flags a service as potentially not meeting coverage criteria and a human clinician review confirms the non-affirmation, the clinician must have relevant clinical expertise in the service category. A determination on a spinal cord stimulator implant must involve a clinician with relevant neurological or interventional pain expertise. This requirement is not merely aspirational. It is a response to the documented pattern in MA PA, where generalist reviewers make clinical determinations on procedures requiring specialty expertise, generating inappropriate denials that are then overturned on appeal at rates that the OIG found alarming — a 2018 HHS OIG report found that 75 percent of denied MA PA requests were overturned on appeal.

The Gold-Carding Mechanism

#

CMMI committed to piloting a gold-carding feature for WISeR by mid-2026. The gold card, as the model defines it, exempts providers with demonstrated records of compliance and consistent PA approval histories from the prior authorization or prepayment review requirement for a defined period.

The gold-carding concept is not new in health insurance administration. Several MA plans and commercial insurers have implemented gold card programs under various designs, and at least one state, Texas, has enacted legislation requiring MA plans operating in the state to offer gold-carding to qualifying physicians. The WISeR implementation will be the first test of gold-carding at a program-wide scale in FFS Medicare.

The policy logic is straightforward: if the purpose of WISeR’s PA requirement is to identify the subset of providers and suppliers whose billing patterns reflect inappropriate utilization, and if a provider’s history of approval demonstrates that their practice patterns consistently meet coverage criteria, then subjecting that provider to the same administrative burden as high-risk billers generates costs without proportionate benefit. Gold-carding concentrates administrative friction on the high-risk tail while releasing the compliant majority from ongoing review.

The design questions that will determine the gold card’s success are not conceptual but operational. What is the evaluation period before a provider qualifies? How many approvals and over what timeframe constitutes a “consistent approval history”? What is the gold card’s duration before requalification is required? What happens when a previously gold-carded provider receives a denial — is the card suspended, revoked, or subject to graduated response? None of these parameters has been publicly specified for the WISeR mid-2026 rollout. The specifics will be published as implementation guidance.

Provider Revenue Risk

#

For specialists whose practices include WISeR-listed services in the six pilot states, the model’s revenue implications are not theoretical.

Skin substitute specialists face the most concentrated risk. The wound care practices, wound care product distributors, and podiatry practices that drove the skin substitute billing surge from $509 million to $10 billion in five years are precisely the population WISeR is designed to target. Providers with legitimate, documentation-supported skin substitute practices should be affirmed through the PA pathway. Providers whose billing patterns reflect the fraud and abuse the OIG documented will face non-affirmation and, if they proceed without authorization, post-service claim denials with no payment.

The revenue risk for compliant providers is administrative rather than clinical. The PA workflow adds time and documentation burden to services that were previously billed without review. A wound care physician who was submitting skin substitute claims and receiving payment within two weeks now submits PA requests, waits up to 72 hours for determination, receives an affirmation, schedules the service, and then bills. The clinical outcome is unchanged. The cash flow timing is different. The administrative staffing requirement is higher.

For interventional pain specialists, neurosurgeons performing cervical fusions, and implanting physicians for nerve stimulators and hypoglossal stimulators, the coverage criteria verification that WISeR requires is the central issue. Coverage policies for these services are detailed, condition-specific, and require documentation of prior treatment failure, specific diagnostic findings, and in some cases multidisciplinary evaluation. A practice that maintains comprehensive clinical documentation should be affirmed. A practice that has been billing based on physician attestation without the supporting clinical record will face non-affirmation.

The AHA’s concern is that even compliant practices will face denials “in the margins” — determinations where the clinical documentation is ambiguous relative to coverage criteria — and that the averted-expenditure compensation structure incentivizes contractors to resolve ambiguous cases toward non-affirmation rather than affirmation. This concern is grounded in the documented MA PA experience, where marginal cases have historically been resolved in plans’ financial interest rather than beneficiaries’ clinical interest. Whether WISeR’s design, specifically the human clinician review requirement and the provider experience performance measures, effectively neutralizes this incentive is the central empirical question the model will answer.

The Congressional Intervention That Did Not Happen

#

In September 2025, the House Appropriations Committee approved an amendment to the FY2026 government funding bill that would have prohibited spending for the implementation of WISeR and, notably, for any future CMMI model testing prior authorization in traditional Medicare. The amendment passed committee. It was not included in the Consolidated Appropriations Act of 2026, signed into law in February 2026.

The appropriations process outcome preserved WISeR’s operational continuity. But the committee vote is a meaningful signal. There is active congressional concern — not limited to one party, based on the breadth of stakeholder opposition — about the implications of PA in FFS Medicare, the contractor compensation structure, and the precedent that a successful WISeR expansion would set for the traditional Medicare benefit design.

The separate July 2025 letter signed by House Democratic members to Administrator Oz, questioning the averted-expenditure contractor payment structure and the adequacy of safeguards against inappropriate denials, reflects concern that crosses the political boundary typically associated with health care oversight disputes. The operational design of WISeR is under sustained congressional scrutiny in a way that most CMMI models are not.

If the model’s first-year evaluation data shows material rates of inappropriate non-affirmations, the appropriations vehicle will be available again in FY2027 to revisit the funding prohibition. CMS’s ability to defend and expand WISeR depends not only on whether it generates savings but on whether those savings are generated without the access harm that the OIG has documented in MA PA and that congressional critics specifically referenced in their objections.

The MA-FFS Convergence Signal

#

CMMI has been explicit about WISeR’s broader implication: the model tests whether it is possible to move toward a more uniform set of administrative rules across Medicare, so that providers are accountable to the same documentation and appropriateness standards regardless of whether they treat a beneficiary in MA or in FFS.

That framing connects WISeR to the convergence argument in MCR-0.2: the structural distinction between MA and FFS is eroding. MA is being pressured from above, through risk adjustment reform, near-flat rate increases, and benefit design scrutiny, in ways that reduce its financial and administrative advantages. FFS is being changed from below, through WISeR’s PA introduction and ACO accountability requirements, in ways that reduce its unmanaged-care character. The two systems are moving toward each other.

WISeR’s convergence contribution is narrow: it addresses a specific list of services in six states over six years. It does not impose MA-style broad-spectrum PA on FFS Medicare. The contrast in PA rates — 1.8 determinations per MA beneficiary versus the new FFS floor that WISeR creates in six states for defined services — will remain enormous even after WISeR is fully operational. But the direction of movement is established. CMS used CMMI authority to waive the statutory limits on PA in Medicare that would otherwise prevent this. If WISeR certifies savings, the waiver authority scales. The convergence continues.

Appeal Rights and Beneficiary Protections

#

Medicare’s administrative appeals system applies to WISeR determinations. A beneficiary or provider who disagrees with a non-affirmation decision can appeal through the standard five-level Medicare appeals process: redetermination by the WISeR participant, reconsideration by a Qualified Independent Contractor, Administrative Law Judge hearing, Medicare Appeals Council review, and federal district court review.

If a service is non-affirmed and the provider proceeds with furnishing it, the provider must issue an Advanced Beneficiary Notice of Non-Coverage before the service is delivered. The ABN informs the beneficiary that Medicare may not pay for the service and that the beneficiary may be responsible for the full cost. The beneficiary then has a documented choice: accept the potential financial liability and receive the service, or decline the service based on the coverage uncertainty. This notice requirement is both a patient protection mechanism and an administrative safeguard: it ensures that beneficiaries are not unknowingly exposed to financial liability for services that were not affirmed.

The access implication of the ABN requirement is real and not fully resolved. A patient for whom a WISeR-covered service has been recommended by their physician, and for whom the prior authorization request was not affirmed, must decide whether to absorb what could be substantial out-of-pocket costs while the appeal process resolves. For beneficiaries without Medigap coverage, the 20 percent Part B coinsurance on a skin substitute application or a nerve stimulator implant could represent significant financial exposure. The ABN protects the beneficiary’s legal rights. It does not protect their financial interest if the service is ultimately appropriate but the initial PA determination was wrong.

What WISeR Tells Us About What Comes Next

#

WISeR is explicitly described by CMS as a potential “blueprint for broader prior authorization reforms across Medicare, as well as potentially the private sector.” That framing is not incidental. The model’s design choices — AI-assisted review, human clinician oversight for denials, averted-expenditure contractor payment with process measure adjustments, gold-carding for compliant providers, a defined service list tied to existing coverage criteria — were made with replication in mind.

If the six-year evaluation shows savings certification is achievable without documented access harm, CMS has a certifiable model it can propose expanding to additional states or service categories through rulemaking. The contractor infrastructure that WISeR builds — six technology companies with FedRAMP-compliant AI review systems, MAC integration protocols, and beneficiary notification workflows — does not dissolve at the end of 2031. It becomes the deployment foundation for a scaled FFS PA system.

The policy horizon beyond WISeR depends on the evaluation data. Additional service categories are already being analyzed; the model’s RFA language explicitly anticipated that CMS would monitor utilization trends and consider service list expansion within the model’s existing period. The geographic expansion question is procedurally simpler than service list expansion: adding states requires a model modification, not a new rulemaking cycle.

The larger question is whether WISeR’s political sustainability holds through the six-year evaluation period. The appropriations threat was real and was only resolved by exclusion from the final spending bill, not by congressional endorsement. A successor appropriations vehicle, a change in CMS leadership, or a significant documented access harm event could interrupt the model before its evaluation is complete.

The structural argument for WISeR, and for PA in FFS Medicare generally, rests on a simple premise: the Medicare FFS payment system has no mechanism for prospective clinical review of services with documented overuse and fraud patterns, and the absence of such a mechanism costs the program billions annually while exposing beneficiaries to medically unnecessary procedures. That premise is empirically well-supported. The implementation challenge is building a review system that is accurate, timely, clinically rigorous, and immune to the financial incentives that have corrupted PA in the MA context. Whether WISeR’s design accomplishes that will be apparent in its first two years of evaluation data. The model’s future, and the future of PA in FFS Medicare, depends on the answer.

Related Reading

#

MCR-03_02 The Prior Authorization Divide: WISeR (FFS) vs. MA Plans MCR-06_11 Clinical Decision Support and the WISeR Vendor Ecosystem MCR-05_01 The Provider’s New Reality: Revenue, Authorization, and Accountability MCR-11_04 Arizona and Nevada: Sun Belt Medicare in the WISeR Era