Introduction
In November 2022, the Centers for Medicare & Medicaid Services (CMS) issued a document further extending until February 1, 2023, the deadline for its publication of the long-awaited Medicare Advantage (MA) risk adjustment data validation (RADV) final rule. The proposed rule, originally published in November 2018, proposed regulations on the use of extrapolation and the application of a fee-for-service adjuster (FFS Adjuster) in the MA RADV program. Below, we provide a refresher on these issues in anticipation of the final rule.
Revisiting the 2018 Proposed Rule
Extrapolation
CMS currently has authority to conduct contract-level audits and recover any potential overpayments from MA plans based on member-level findings. These audits are known as “contract-level” audits, as they are done for a contract that an MA plan has with CMS to provide MA benefits (i.e., a plan may have multiple contracts with CMS with different member populations ). Extrapolation refers to the process of selecting a sample of members to audit in order to estimate recoveries for all the members in a contract. The current RADV process (see prior blog for more information on the RADV process) relies on manual medical record reviews, a burdensome and time consuming process. Conducting reviews on all MA enrollees would be impractical, if not impossible, in a timely, cost-effective manner. As such, sampling and extrapolation are necessary features of the overpayment recovery mechanism.
The proposed rule would allow extrapolation in RADV contract-level audits of MA plans starting with payment year 2011. At the time the proposed rule was written (in 2018), audits for payment years 2011-2013 had been conducted according to a previously published methodology in February 2012. However, the methodology did not go through formal rulemaking, which is now required before CMS can make changes to payment. Thus, the contract-level recoveries on those earlier audits and all subsequent audits are dependent on finalizing the methods in the proposed rule.
With the 2018 proposed rule, CMS is seeking comment on the approach for extrapolation already published in 2012, based on a contract-level sample (i.e., a representative sample of enrollees associated with a specific MA contract). CMS also solicits comments for alternative approaches, including an approach that would focus audits on a sub-cohort of MA plan enrollees (e.g., enrollees with a specific diagnosis). CMS suggests that the sub-cohort approach could allow for a smaller sample from each contract, while also allowing more plans to be sampled. Such an approach was utilized by CMS to conduct audits on payment years 2014 and 2015.
In the upcoming final rule, CMS will need to decide whether to finalize the methods proposed. If a proposed method is finalized, CMS will need to determine for which payment years the method will apply. For example, using extrapolated recoveries from the 2014 and 2015 audits already conducted raises a question of whether CMS will pursue extrapolated recoveries on earlier years (2011-2013) using the earlier methodology. Failure to finalize any of the proposed approaches will leave the utilization of extrapolation in the contract-level RADV audits in limbo for the foreseeable future.
FFS Adjuster
The proposed rule also readdresses whether to apply a fee-for-service (FFS) Adjuster to audit findings, proposing not to do so. This comes after a 2012 announcement where CMS stated it would apply a FFS Adjuster to extrapolated contract-level RADV audits after conducting additional research to determine the size of the adjustment.
According to the 2012 announcement, the FFS Adjuster was intended to account for potential systematic differences in the documentation standard used in RADV audits to determine a contract’s payment error from the documentation standard used in the traditional Medicare FFS program. This is important because FFS claims data are currently used to calibrate the risk factors used in the MA risk adjustment model. Consequently, any bias in the FFS data (relative to the MA RADV data) would permeate through to MA payments. At the time, CMS contended that the FFS Adjuster would calculate a permissible level of payment error (for example, a percentage of the total payments made on an MA contract in a given year) and limit RADV audit recovery to payment errors above that level.
Moreover, an actuarial analysis commissioned by the health plan industry focused on the notion that adjusting the RADV audit recoveries could help ensure actuarial equivalence between FFS and MA payments, which is required by law. This is based on the notion that actuarial equivalence may be compromised if the diagnosis documentation standards used in the FFS program (which is used to recalibrate the MA RA models) are different from those used to recoup overpayments in the MA program.
In a 2018 analysis of the FFS Adjuster (known as the “FFS Adjuster Study”), CMS suggested that errors in FFS claims data do not have any systematic effect on the risk scores calculated by the risk adjustment model. Therefore, CMS concluded that using FFS claims does not have any systematic effect on payments made to MA plans, leading CMS to propose excluding an FFS Adjuster in any final RADV payment error methodology. CMS also suggested in its analysis that if systemic biases were present, it would be more appropriate to address them in the original MA plan payments, rather than in audit recoveries.
The CMS analysis was criticized for various methodological issues. In response, CMS extended the comment period on the proposed RADV changes to allow more time for the industry to evaluate the analysis conducted and provide more comments. CMS made all of the data, programs, and empirical outputs underlying the study available for review. Thus, the RADV provisions introduced in the November 2018 proposed rule were left out of the subsequent final rule published in April 2019 on various technical changes to the MA program.
What Happens Next?
CMS is generally required to publish a final rule no later than three years after the proposed rule. However, as discussed in the November 2022 document, an extension due to “exceptional circumstances” is possible. Aside from the additional time for industry stakeholders to comment, the stated reason for the continued delays in this case includes a focus on the COVID-19 public health emergency. Assuming there are no further delays, the much-anticipated final rule will be published February 1, 2023, over 4 years from the original proposed rule publication. During that time, CMS has received many industry comments. There have also been court rulings related to the issue of actuarial equivalence and the FFS Adjuster, with the latest ruling suggesting that actuarial equivalence is not required for recovery (i.e., RADV) payments. These court rulings may provide foresight on CMS’s rationale and responses for the upcoming Final Rule.
If finalized as proposed, the rule would present MA plans with an audit landscape including full contract-level and sub-cohort-based RADV audits. Greater use of the sub-cohort-based audits would put a much greater proportion of the plans at risk of audit. Conversely, greater use of the contract-level audit would put fewer plans at risk, but a greater potential audit liability for any plan selected. The exclusion of the FFS Adjuster would mean greater financial liabilities in both audit scenarios. How CMS decides to apply extrapolation will also have a significant impact on the plans selected for audits completed on payment years 2011 through 2015. In the proposed rule, CMS restates its authority to perform retroactive recoveries. A total or partial decision to apply extrapolation on the earlier audits would have a significant impact on the audited plans. According to CMS, the extrapolated recovery estimate for payment years 2011 through 2013 alone is $650M.
Stay tuned for additional coverage of this issue in future blogs.