CMS Is Moving Ahead With RADV — Even as the Humana Case Continues

August 28 was an important day for Medicare Advantage Risk Adjustment Data Validation (RADV). CMS officially initiated the Payment Year (PY) 2024 RADV audits, notified the Medicare Advantage organizations selected for audit, and released the PY2024 Audit Methods & Instructions and list of audited contracts.¹

The same day also marked the closure of the PY2020 medical record submission window, while PY2021 RADV is already underway. ²

The timing of the PY2024 release is particularly noteworthy because an important legal challenge involving CMS's RADV methodology is still making its way through the courts. In September 2025, Humana successfully challenged CMS's 2023 RADV Final Rule in Humana Inc. v. Becerra. The federal district court vacated the rule, which established CMS's approach to extrapolation beginning with PY2018.³

The government appealed the decision, and the case, now Humana v. Kennedy, is scheduled for oral argument before the Fifth Circuit on October 5, 2026.⁴

CMS, however, is continuing to move forward with RADV while the appeal is pending. In fact, the PY2024 methodology released on August 28 directly addresses the issue of extrapolation. CMS acknowledges that the 2023 RADV Final Rule was vacated and is under appeal but says the PY2024 audits have still been designed to calculate extrapolated overpayments. What CMS has not yet decided is whether it will actually collect those extrapolated amounts or limit recovery to the overpayments associated with the sampled enrollees.⁵

For Medicare Advantage organizations, the difference between those two approaches could be significant. It also means plans are moving into a new round of RADV audits without yet knowing how one of the most consequential pieces of the financial methodology will ultimately be applied.

How We Got Here

The dispute between Humana and CMS goes back several years and centers in large part on the agency's decision to eliminate what is known as the Fee-for-Service (FFS) Adjuster.

The CMS-HCC risk adjustment model is calibrated using Medicare fee-for-service data, while RADV audits validate diagnoses submitted by Medicare Advantage organizations against medical record documentation. CMS had previously recognized that differences between those two environments could create a mismatch when audit findings were extrapolated. In 2012, the agency announced that it would apply an FFS Adjuster when calculating extrapolated RADV recoveries to account for differences between the medical-record documentation standard used in RADV and the FFS claims data used to develop the Part C risk adjustment model.⁶

CMS later changed course. In 2018, the agency proposed eliminating the FFS Adjuster, and that approach was finalized in the January 2023 RADV Final Rule. The rule also established that CMS would use extrapolation beginning with PY2018 RADV audits.⁷

Extrapolation significantly increases the potential financial impact of a RADV audit. Without it, CMS can recover overpayments associated with diagnoses that cannot be validated for members included in the audit sample. With extrapolation, CMS can use the findings from that sample to estimate overpayments across a much larger population. As a result, unsupported diagnoses identified in a relatively small sample can potentially translate into a much larger recovery.

Humana challenged the rule in federal court.

What Humana Won

Humana raised several issues with CMS's Final Rule, including the elimination of the FFS Adjuster, actuarial equivalence, retroactivity, and the process CMS followed when finalizing its policy. Ultimately, however, the district court did not need to decide all of those issues. Instead, its decision focused on the rulemaking process.

Under the Administrative Procedure Act, federal agencies generally must provide adequate notice of proposed regulatory changes and give interested parties a meaningful opportunity to comment. The court found that the rationale CMS relied upon in the 2023 Final Rule differed materially from the rationale the agency presented when it originally proposed eliminating the FFS Adjuster in 2018.

The court determined that the Final Rule was not a logical outgrowth of the Proposed Rule. On September 25, 2025, it granted Humana's motion for summary judgment and vacated and remanded the 2023 RADV Final Rule.³

It was a significant win for Humana, but the decision is important to understand in context. The court did not eliminate RADV, nor did it rule that CMS could never extrapolate RADV findings. Because the case was resolved on procedural grounds, the court also did not reach several of Humana's broader substantive challenges.³

CMS retains the authority to audit diagnoses submitted for Medicare Advantage risk adjustment and recover overpayments when those diagnoses cannot be supported. The unresolved question is how CMS can calculate recoveries beyond the members actually included in the audit sample.

That question is now part of the appeal before the Fifth Circuit.

CMS Is Moving Ahead With PY2024 RADV

While the Humana case continues, CMS has not slowed its RADV activity. The August 28 initiation of PY2024 is the latest example. CMS notified Medicare Advantage organizations with contracts selected for PY2024 RADV audits and released the PY2024 Audit Methods & Instructions, list of audited contracts, and updated RADV FAQs.¹

The timing also illustrates how much the operational demands of RADV are changing. PY2020 medical record submissions closed on August 28, PY2021 is already underway, and PY2024 has now been initiated. For plans involved in more than one payment year, different RADV audits may therefore be moving through different stages at the same time.²

That can create very practical challenges. A risk adjustment team may be completing submissions for one audit while managing retrieval, coding review and validation for another. Tracking, issue resolution, vendor oversight and leadership reporting can also overlap. Processes that worked when RADV audits were more spread out may need to be reconsidered as CMS moves payment years through the process more quickly.

At the same time, the PY2024 guidance introduces the additional uncertainty created by the Humana litigation. CMS specifically acknowledges in the new methodology that the 2023 RADV Final Rule was vacated by a federal district court and that the decision is under appeal. Despite that, CMS has designed the PY2024 audits to calculate extrapolated overpayments.⁵

The agency has not, however, decided whether those extrapolated amounts will ultimately be collected. CMS states that it may instead collect only the overpayments associated with sampled enrollees and that audited Medicare Advantage organizations will be notified of its decision later.⁵

That distinction matters. An audit recovery limited to sampled members represents one level of financial exposure. Applying the results of that sample across a larger population through extrapolation could represent something considerably different. Plans may therefore move through much of the PY2024 audit process before knowing which approach CMS will ultimately use.

PY2024 RADV Is Also More Targeted

Extrapolation is not the only important part of the new PY2024 methodology. CMS is also changing the size and composition of the audit sample.

For PY2024, sample sizes may include 35, 50, 100 or 200 enrollees, depending on the number of eligible members in the contract's sampling frame. This continues CMS's move away from the historical 35-enrollee RADV sample and could substantially increase the amount of medical record retrieval and review work required for some organizations.⁵

CMS is also using predictive models as part of the process for determining which members are eligible for the sampling frame. Among other criteria, members must fall within the top quartile of RADV-eligible enrollees based on one or both of CMS's PY2024 Medicare Advantage improper-payment prediction models. CMS describes these as members predicted to have the greatest reduction in risk score as a result of a RADV audit. ⁵

This is an important shift for risk adjustment teams. CMS is using its data and analytics to focus audit activity on members it believes have greater potential for payment error. Plans should consider whether their own internal auditing strategies are keeping pace.

What Should Plans Be Doing Now?

The outcome of Humana v. Kennedy may ultimately affect how CMS calculates and collects RADV recoveries, but it should not change the fundamentals of a health plan's risk adjustment compliance program. Every diagnosis submitted to CMS for risk-adjusted payment still needs to be accurate and supported by the medical record.

Plans should continue evaluating their data for areas that may create RADV exposure. That includes high-value HCCs, unusual provider patterns, diagnoses originating from specific programs or vendors, coding trends, deletion activity, and conditions that have historically presented documentation challenges. Random coding audits remain valuable, but CMS's increasing use of predictive analytics also supports a more targeted approach based on provider, HCC, diagnosis source, financial impact and other indicators of risk.

Medical record retrieval should be part of that discussion as well. A diagnosis may have been appropriately coded when it was originally submitted, but the plan still needs to be able to produce the supporting documentation when CMS requests it. With audits occurring years after the dates of service under review, plans should understand where their records are located, which providers present retrieval challenges, and whether their processes can reliably produce documentation within CMS timelines.

Vendor oversight is equally important. Medicare Advantage organizations may outsource coding, retrieval, analytics, in-home assessments and other risk adjustment functions, but responsibility for the data submitted to CMS remains with the plan. Organizations need visibility into what their vendors are finding and submitting, how coding accuracy is validated, whether identified errors are corrected, and whether the documentation supporting submitted diagnoses can withstand an audit.

Perhaps one of the most useful questions for plans to ask now is: If CMS used its analytics to identify our highest-risk members today, what would it find?

The answer can help shape a much more meaningful internal audit strategy than waiting to see which members eventually appear on a RADV Enrollee Data List.

The Financial Question Isn't Going Away

The Humana case is not only an issue for risk adjustment, legal and compliance teams. Finance and executive leadership should understand it as well because the difference between sample-member recoveries and extrapolated recoveries can be significant.

Until the legal questions are resolved and CMS announces how it intends to handle PY2024 recoveries, organizations may need to consider RADV exposure under both scenarios. Plans should understand what their potential exposure looks like if CMS collects only the overpayments identified in the sample and how that exposure could change if CMS ultimately applies extrapolation.

The uncertainty surrounding extrapolation does not change the part of the equation plans can control: the accuracy and supportability of the diagnoses submitted to CMS. The fewer unsupported diagnoses identified in the audit sample, the less there is to recover regardless of how CMS ultimately calculates the amount.

What Comes Next?

The government's appeal of the district court decision is now pending before the Fifth Circuit, with oral argument scheduled for October 5, 2026. The appellate court could affirm the district court's decision, reverse it, or issue a ruling that leaves additional questions for CMS and Medicare Advantage organizations to work through.⁴

In the meantime, CMS is clearly not putting RADV on hold. The agency initiated PY2024 on August 28, is increasing potential sample sizes, is using predictive models to identify members with greater potential for payment error,  and has designed the audits so that extrapolated overpayments can be calculated. What remains undecided is whether CMS will actually collect those extrapolated amounts.¹,⁵

That leaves Medicare Advantage organizations preparing for and responding to RADV while an important part of the potential financial outcome remains unresolved. Waiting for the courts to settle the issue is not a practical RADV strategy. Plans still need strong coding validation, documentation review, targeted auditing, medical record retrieval, data integrity and vendor oversight regardless of how Humana v. Kennedy is ultimately decided.

Humana's victory may help determine how CMS can calculate and collect future RADV overpayments. Health plans, however, still have considerable control over something more fundamental: whether the diagnoses CMS reviews are supported in the first place.

How Rebellis Can Help

With multiple RADV payment years moving through the process and CMS taking a more targeted approach to its audits, plans may need to take a broader look at how they manage RADV readiness.

Rebellis can assist with medical record retrieval oversight, coding and documentation validation, targeted and random audits, submission review, audit tracking and reporting, vendor oversight, and overall RADV project management. We can also help plans look beyond the immediate CMS sample to identify areas of potential risk across the broader risk adjustment program before those issues become audit findings.

The methodology and the legal questions surrounding RADV may continue to evolve, but the goal for health plans remains straightforward: submit accurate, well-supported data and be prepared to defend it when CMS asks.

Contact us

Sources

1. Centers for Medicare & Medicaid Services. RADV Announcements - CMS Initiates Payment Year 2024 RADV Audits.Published August 28, 2026.

2. Centers for Medicare & Medicaid Services. RADV Documents and Data. Current RADV audit materials for Payment Years 2020, 2021 and 2024.

3. U.S. District Court for the Northern District of Texas. Humana Inc. and Humana Benefit Plan of Texas, Inc. v. Becerra et al. Civil Action No. 4:23-cv-00909-O. Order granting Humana's motion for summary judgment and vacating and remanding the Final Rule. September 25, 2025.

4. U.S. Court of Appeals for the Fifth Circuit. Humana, Incorporated, et al. v. Robert Kennedy, et al. No. 25-11293. Oral argument scheduled October 5, 2026.

5. Centers for Medicare & Medicaid Services. Payment Year 2024 Medicare Advantage Risk Adjustment Data Validation Audit Methods & Instructions. Published August 28, 2026.

6. Centers for Medicare & Medicaid Services. Notice of Final Payment Error Calculation Methodology for Part C Medicare Advantage Risk Adjustment Data Validation Contract-Level Audits. February 24, 2012.

7. Centers for Medicare & Medicaid Services. Medicare Advantage Risk Adjustment Data Validation Final Rule (CMS-4185-F2) Fact Sheet. January 30, 2023.