Program Overview
PROGRAM: WISeR — Wasteful and Inappropriate Service Reduction Model ADMINISTERING AGENCY: Centers for Medicare & Medicaid Services LAUNCH DATE: January 1, 2026 DURATION: Six years, through December 31, 2031 COVERAGE ZONE: Arizona, New Jersey, Ohio, Oklahoma, Texas, Washington BENEFICIARY POPULATION: 6.4 million fee-for-service Medicare beneficiaries in the model states PURPOSE: Reduction of waste, fraud, and abuse in fee-for-service Medicare through AI and machine learning prior authorization review VENDOR COMPENSATION: [REDACTED — see EFF v. CMS, filed March 25, 2026]
The WISeR model was announced by CMS Administrator Dr. Mehmet Oz, who stated that "CMS is committed to crushing fraud, waste, and abuse, and the WISeR Model will help root out waste in Original Medicare."
The Bureau notes the name. Wasteful and Inappropriate Service Reduction. The program's architects selected an acronym. The acronym they selected is WISeR. The Bureau notes the name and continues filing.
Vendor Selection and Compensation Structure
Six vendors were selected to administer prior authorization determinations across the six model states:
Cohere Health — Texas Genzeon Corporation — New Jersey Humata Health — Oklahoma Innovaccer — Ohio Virtix Health — Washington Zyter — Arizona
Vendor contracts: [REDACTED — CMS has not released these records. The EFF requested them. The EFF was not provided them. The EFF filed suit on March 25, 2026.]
AI model specifications: [REDACTED — CMS has not released these records.]
Accuracy and bias test records: [REDACTED — CMS has not released these records.]
Program evaluation documentation: [REDACTED — CMS has not released these records.]
What has been released is the compensation structure. According to reporting by Modern Healthcare, confirmed by KFF analysis and the text of the Democrats' proposed repeal legislation, WISeR vendors earn 10–20% of the savings generated when prior authorization requests are denied.
The Bureau will state the mechanism plainly. The vendor reviews a prior authorization request. If the vendor approves the request, Medicare pays for the service. The vendor earns nothing from the transaction. If the vendor denies the request, Medicare does not pay for the service. The money Medicare did not spend is classified as savings. The vendor receives between 10 and 20 percent of those savings.
There are two possible determinations. One of them generates revenue. The program cannot produce savings without producing denials. The vendors cannot produce revenue without producing savings. The Bureau notes that this sentence has only one logical direction.
BUREAU NOTE: The Bureau's records show that KFF describes the model as one that "rewards vendors, in part, based on the volume of care that they deny." The Bureau's experience suggests that "in part" is doing a great deal of administrative work in that sentence. The savings that fund the reward are generated exclusively by denials. The reward exists in no other configuration. The Bureau files this observation under: distinctions that hold legally and nowhere else.
The Official Position
CMS has addressed the incentive question directly. The WISeR Model Frequently Asked Questions states that contractors "will not be incentivized to deny claims, but to get the determination right."
The Bureau accepts this framing for the record.
The Bureau also notes that "the determination that generates savings" and "the determination that denies the claim" are the same determination. The determination that denies the claim is the only determination that gets the vendor paid. Getting the determination right, in the compensation architecture of WISeR, means arriving at the determination that maximizes the vendor's share of the savings generated by the denial of the claim.
The FAQ also states that vendors will be subject to quarterly accuracy audits.
Quarterly accuracy audit results: [REDACTED — CMS has not released these records.]
The program has been operational since January 1, 2026. The Bureau calculates that three audit periods have elapsed as of the date of this dispatch.
Relevant Prior Art
In 2022, Humana's post-acute care prior authorization denial rate was 16 times higher than its overall prior authorization denial rate across all service categories. The increase in post-acute care denials occurred in parallel with Humana's deployment of AI and predictive technology for prior authorization decisions. This is the finding of the Senate Permanent Subcommittee on Investigations, whose majority staff report — Refusal of Recovery: How Medicare Advantage Insurers Have Denied Patients Access to Post-Acute Care — was released on October 17, 2024.
The Senate investigated three insurers — UnitedHealth Group, Humana, and CVS/Aetna — and found that by 2022, they were collectively denying approximately 25% of all post-acute care prior authorization requests as AI deployment increased.
Humata Health, one of the six WISeR vendors, is financed by Optum Ventures (a UnitedHealth Group investment vehicle), Blue Venture Fund (Blue Cross Blue Shield), and Highmark Ventures. Virtix Health provides risk adjustment coding services to Medicare Advantage plans. Multiple WISeR vendor executives previously held senior roles at Elevance, Optum, Kaiser, Highmark, and HCSC.
The Senate investigated the parent companies of the firms that financed the contractors. The contractors were subsequently selected to administer Medicare prior authorization for original Medicare beneficiaries. The Bureau notes the continuity without further characterization.
BUREAU NOTE: A clarification for the record: the Bureau is not suggesting that Humata Health will replicate Humana's 2022 post-acute care denial rates in original Medicare. The Bureau is noting that Humata's investors are the entities the Senate investigated for those rates, that the compensation structure CMS deployed in original Medicare is structurally similar to the one that produced those rates, and that the documentation required to determine whether the rates are being replicated is [REDACTED].
Early Implementation Findings
The Center for Medicare Advocacy — citing Washington Post reporting from early 2026 — documents the following early findings from WISeR's first months of operation:
- Patients are experiencing pain while waiting for prior authorization approvals that have not arrived.
- Doctors in Ohio are considering stopping certain complex treatments because the AI model cannot evaluate them accurately enough to approve them.
- Portal failures, technical glitches, and communication breakdowns with vendors are documented.
- Approval rates under WISeR are lower than the already-contested approval rates in Medicare Advantage plans.
The Bureau notes that the program was designed to reduce inappropriate care. The Bureau notes that some of the care being delayed or denied appears, by physician and patient assessment, to be appropriate.
The Bureau further notes that the program does not distinguish between these two categories in its vendor compensation formula. A saved dollar is a saved dollar. The vendor's cut is the same regardless of whether the service denied was wasteful or warranted.
Outstanding Documentation Request
On March 25, 2026, the Electronic Frontier Foundation filed EFF v. CMS in federal court, seeking compelled disclosure of the records CMS declined to provide in response to a prior FOIA request. The requested records are:
- Agreements with software vendors participating in the WISeR model
- Records related to any tests for accuracy, bias, or hallucinations in vendors' technology
- Records related to any audits, monitoring, or evaluation of WISeR and participating vendors
EFF Director of AI and Access-to-Knowledge Legal Projects Kit Walsh stated that "tasking an algorithm with making determinations about treatment can create unwarranted — and even discriminatory — delays or denials of necessary medical care."
The Bureau notes that the EFF's request for transparency materials was filed eighty-four days after the program launched. The Bureau notes that as of the date of this dispatch, no records have been provided.
The program has been making prior authorization determinations, at scale, for 207,500 Medicare beneficiaries who use WISeR-targeted services, since January 1, 2026, under contracts whose terms are [REDACTED], using AI models whose specifications are [REDACTED], with accuracy records that are [REDACTED].
The waste reduction is ongoing. The revenue is accruing. The documentation will follow.
The Bureau of Appropriate Service Reduction, Claims Optimization Division, operates within the Bureau of Public Agreement. The Bureau wishes to advise all beneficiaries in the six model states that prior authorization requests should be submitted through the appropriate vendor portal. Determinations will be returned within 72 hours, or 48 hours for expedited requests. A gold-carding exemption for clinicians with consistent approval histories is planned for mid-2026. The Bureau's contract terms are not available for review. The Bureau's accuracy audits are not available for review. The Bureau's name, as noted, is WISeR.