Executive Summary

• On January 30, 2023, the Centers for Medicare and Medicaid Services (CMS) released a final rule pertaining to the Risk Adjustment Data Validation (RADV) process, which is designed to catch payment errors and discrepancies in its risk adjustment in the Medicare Advantage (MA) program.
• The new RADV rule confirms CMS’s ability to extrapolate error rates beyond the enrollees in a sample population to an entire MA contract when determining how much money the agency can claw back from MA plan sponsors; this extrapolation methodology will be applied retroactively to contracts from the 2018 plan year and thereafter.
• The RADV rule also states that the error rate in fee-for-service (FFS) Medicare – a so-called FFS adjuster – will not be used to adjust MA payments.
• The RADV rule, which CMS estimates will allow it to recover $4.7 billion from MA plan sponsors for the years 2018–2023, does not specify the particular methodology or methodologies CMS will use, which may increase uncertainty among MA plan sponsors and weaken existing MA plan offerings.

Introduction

On January 30, the Centers for Medicare and Medicaid Services (CMS) released a final rule regarding major changes to the Risk Adjustment Data Validation (RADV) process in the Medicare Advantage (MA) program.  This rule has two key components: It confirms the ability of CMS to extrapolate error rates in diagnostic codes from a subset of MA beneficiaries across an entire MA contract (which is made up of multiple plans from a sponsor) for the purposes of clawing back improper payments –  and this extrapolation will be applied from 2018 on – and it states that CMS will not be using a so-called “fee-for-service (FFS) adjuster” when deciding how to risk-adjust payments. This rule, originally proposed in 2018 and continuously delayed until now, follows on the heels of recent accusations of, and investigations into, widespread and allegedly intentional “up-coding” by MA plan sponsors.

Medicare Advantage Payment Basics

To understand why the RADV rule matters and why it is controversial, it is important to understand how MA payments to plans are determined. A previous American Action Forum Insight has more details on how MA payments are made, but the basics are as follows: CMS sets a benchmark rate – the highest the agency will pay a plan sponsor per beneficiary before risk-adjustment (see below) – in a given coverage area. This benchmark is based on a percentage of the average expenditure on an FFS (i.e., traditional Medicare) beneficiary in the coverage area.

This benchmark for a given plan is then adjusted for quality scores, plan bids, and finally, in a process known as risk adjustment, the risk scores of beneficiaries in the plan, which are based on a beneficiary’s diagnoses from the prior year. The sicker the beneficiary, the greater the risk adjustment, and thus the higher the payment to the MA plan sponsor. In order to compensate for MA plans potentially over-diagnosing, whereby a plan may look for every possible diagnosis in order to make policyholders appear sicker than they truly are, and thus increase risk adjustment payments, a coding intensity adjustment is applied to reduce beneficiary risk scores.

Traditionally, when CMS would audit MA plans to look for payment errors and discrepancies in its risk adjustment, the agency would only audit a small portion of the records from only a few plan sponsors. If over- or under-payments were found, those mistakes were subsequently corrected by CMS, which either clawed back money from MA plans or paid MA plans more. These payment fixes were limited to only the enrollee level and not to entire plans or contracts.

Error Rate Extrapolation and Retroactivity

The new RADV rule confirms CMS’ ability to now extrapolate error rates across the whole of an MA contract, rather than at the enrollee level. On a practical level, the process will look like this: CMS will inspect a subset of MA enrollee data by looking at the diagnosis codes submitted by plan sponsors and comparing those codes to the beneficiaries’ actual medical records. The number of discrepancies between the plan-submitted diagnosis codes and beneficiaries’ medical records will determine the error rate. CMS will then apply that error rate to the MA contract when determining how much money to claw back from the plan sponsor (or return in payment, as the case may be).

Crucially, CMS does not specify the exact methodology it will use to determine the error rate. No specific sample sizes are listed – CMS merely notes two specific methodologies it has explored as possible candidates. The first of these methodologies was developed in 2012 and would investigate the records of up to 201 enrollees from each audited contract (MA contracts average around 36,000 enrollees), then extrapolate that error rate across the entire contract. The second methodology would explore the use of subcohorts of enrollees, potentially those with the same medical condition, and extrapolate the error rate of that subcohort across the contract. As specified in the CMS rule: “We are not adopting any particular statistical sampling methodology in this final rule. As previously noted, CMS will use statistically valid methods for sampling and extrapolation that we determine to be well-suited to a particular RADV audit.”

Part of this provision of the RADV rule also involves retroactively applying the extrapolation back to 2018. CMS had originally considered applying the extrapolation methodology all the way back to 2011 but decided against it to limit the number plan sponsor appeals the agency would have to sift through.

Fee-for-service Adjuster

As noted above, FFS claims influence MA benchmarks and thus payments to MA plan sponsors. Since at least the release of the 2012 extrapolation methodology, plan sponsors have asked MA to include an FFS adjuster in its RADV process to account for the potential influence of inaccurate FFS claims on MA payments. In the final RADV rule, CMS has opted not to apply an FFS adjuster to RADV audits. In the RADV rule, CMS claims that a previous study on the impact of FFS diagnostic errors on MA payments revealed little, if any, impact on systematic payment error in MA (plan sponsors dispute these findings). CMS also asserts in the RADV rule that an FFS adjuster has no bearing on the legal responsibility of MA plan sponsors to provide accurate documentation supported by the medical record, which is what the RADV process is meant to confirm.

Impact and Conclusion

By CMS’ estimate, the new RADV rule will lead to $4.7 billion in claw backs over the next decade, including from plan years 2018–2022. MA plan sponsors have fought CMS’ use of extrapolation over the last decade, and that fight was the cause of the RADV rule’s nearly five-year delay before implementation. MA plan sponsors insist the extrapolation and its retroactive application are illegal. They also hold that CMS has a legal obligation to use an FFS adjuster to bring about actuarial equivalence between MA and traditional Medicare. Two district court rulings have supported the MA plan sponsors’ position on actuarial equivalence, though one was overturned by the D.C. Circuit Court. Given the last decade of legal action over RADV policies, expect the new final RADV rule to be tied up in court for a while before implementation can begin.

CMS’ lack of specific proposed methodology for extrapolation is concerning and brings greater uncertainty to the MA program. CMS does not define what “statistically valid methods” are or explain why one audit would require a different method from another. The new RADV rule has the potential to lead to arbitrary decisions around methodology, and thus uncertainties for MA plan sponsors. This increased uncertainty may cause MA plan sponsors to limit plan offerings to beneficiaries or in some cases even pull out of the MA program entirely.