Introduction
On August 13, 2021, the United States Court of Appeals for the District of Columbia Circuit issued a decision in UnitedHealthcare Insurance Co., et al. v. Becerra, No. 18-5326, affirming efforts by the Centers for Medicare and Medicaid Services (CMS) to recover overpayments made to Medicare Advantage (MA) organizations.
Notably, the unanimous three-judge panel of the court acknowledged a long-running methodological dispute between CMS and insurers over the calculation of improper payments within the MA Risk Adjustment Data Validation (RADV) program. In a challenge to the Overpayment Rule, 42 C.F.R. § 422.326, UHC had argued that CMS must apply a “Fee-for-Service Adjuster” to payment recoveries that would offset inaccuracies in medical claims data used to calibrate the MA risk adjustment model—and thereby reduce the amount insurers reimburse to CMS. The court rejected both the statutory and empirical bases of UHC’s arguments.
The D.C. Circuit precedent strengthens CMS’s hand as the agency nears a November 1, 2021 deadline for finalizing its RADV audit payment error calculation regulation. Although MA organization comments have opposed any audit rule that lacks a Fee-for-Service (FFS) Adjuster (as the agency has proposed), CMS will be able to argue that the logic of the Becerra decision extends to RADV audit activity.
Overpayment Rule: The UHC Case
As finalized in 2014, the Overpayment Rule requires the return of any overpayments identified by an MA organization. § 422.326(c). An “overpayment” results from any diagnostic code submitted by an MA organization to CMS that was not supported in a patient’s medical records. 79 Fed. Reg.29,844, 29,921 (May 23, 2014).
UHC argued that this requirement conflicted with the methodology used to determine MA payment rates. CMS uses claims from traditional Medicare to determine the expected costs of treating patients with particular demographic and health characteristics. Those expected costs are then used in calculating payments made monthly to MA organizations. However, claims submitted in traditional Medicare (i.e., Parts A and B) are expected to contain some errors, such as unsupported diagnosis codes. By contrast, when an MA organization reports and reimburses CMS for overpayments, unsupported diagnostic codes are removed from diagnoses submissions justifying its initial payments. Accordingly, UHC argued that its payments for the accurate diagnoses would be based on costs calibrated using inaccurate claims, and that this relationship caused the plan to be underpaid.
In a significant legal victory for MA insurers, a federal district court judge awarded summary judgment to UHC and struck down the Overpayment Rule in September 2018. UnitedHealthcare Insurance Co., et al. v. Becerra, 330 F. Supp. 3d 173, 192 (D.D.C. 2018). Accepting UHC’s arguments, the district court held that the Overpayment Rule violated the Medicare Act’s requirements of “actuarial equivalence,” (42 U.S.C. § 1395w-23(a)(1)(C)(i)) and “same methodology” (42 U.S.C. § 1395w-23(b)(4)(D)). 330 F. Supp. 3d at 187. The court also held the regulation was inconsistent with CMS’s purported conclusion that a “Fee-for-Service Adjuster” was required to render the traditional Medicare and MA data sets equivalent in the context of RADV audits. Id. at 187, 189-90.
The D.C. Circuit’s recent decision reverses the 2018 judgment, allowing enforcement of the Overpayment Rule.
The appellate court noted that it had never “previously decided any case involving ‘actuarial equivalence’” in the MA program. Slip op. at 29. It interpreted that provision narrowly: the “actuarial-equivalence requirement is not broadly applicable, but instead limited to the specified context of CMS’s calculation and disbursement of monthly payments in the first instance.” Slip op. at 32 (emphasis added). Similarly, the court held that the “same methodology” provision of the Medicare Act did not impact either the operation of the CMS-HCC risk adjustment model or the Overpayment Rule. Id. Slip op. at 46.
The court ventured beyond interpretation of the Medicare Act’s text, opining that even if the actuarial equivalence requirement applied to the Overpayment Rule, UHC had not provided any logical or empirical basis to conclude that errors in traditional Medicare claims would inevitably penalize MA organizations who reported overpayments. However, the court noted that its analysis was limited to the Overpayment Rule context where insurers identify some, but not all, underreported codes. Slip op. at 44.
UHC also challenged the standard which the Overpayment Rule applied to insurers. As finalized, the rule applies to any instance where an “MA organization has determined, or should have determined through the exercise of reasonable diligence, that the MA organization has received an overpayment.” 42 C.F.R. § 422.326(c) (emphasis added). The district court agreed that this definition—effectively holding plans to a negligence standard—was impermissibly broad. Instead, the court held that the Overpayment Rule should apply at most to cases where insurers have actual knowledge or act with reckless disregard or deliberate ignorance of an overpayment. The agency did not challenge this holding, and accordingly it remains in force.
RADV Audits
A significant question remains following the UHC decision: how will courts treat recoveries in a context where all unsupported codes are removed from a MA organization’s data? This is potentially the situation with MA RADV audits.
In an outstanding Notice of Proposed Rulemaking (NPRM), NPRM-4185-P, CMS has published draft regulations that would allow it to recover overpayments based on auditing selected medical records from an MA plan and then extrapolating its findings across a particular contract[1]—in effect calculating payments for an MA data set without unsupported diagnosis codes. Further, CMS has published an empirical study finding that inaccurate claims data from traditional Medicare have no systematic effect on the MA payments—if anything, CMS believes any errors have an average impact of less than one percent on payment recoveries, benefiting MA plans. Accordingly, in its rulemaking CMS proposed not to apply any FFS Adjuster when calculating RADV audit findings. The agency has until November 2021 to finalize its proposed rule.
If CMS issues a final RADV rule with extrapolated recoveries and no FFS Adjuster, any litigation challenging the audit methodology is likely to revisit these empirical disputes in greater depth. The Becerra court distinguished Overpayment Rule recoveries from circumstances where “a Medicare Advantage insurer’s data will consist of only supported [diagnosis] codes,” like in RADV audits with extrapolated, contract-wide recoveries. Slip op. at 44. However, in practice, the circuit court’s reasoning applies equally in the RADV context: the court observed that, contrary to UHC’s premises, the fee-for-service model of Part B “would seem to tend toward underreporting, not overreporting, of diagnoses within traditional Medicare”—driving MA rates higher—and that there was no reason to expect that unsupported diagnoses in traditional Medicare claims were “materially analogous” to the codes that caused MA overpayments such that the regulation would depress an insurer’s payment. Slip op. at 39. CMS will likely have a strong argument that a decision not to apply a Fee-for-Service Adjuster to extrapolated recoveries is an appropriate exercise of its administrative discretion.
In addition, the D.C. Circuit court’s textual analysis will also loom over potential challenges to a final RADV rule. Future courts will likely seek to address the threshold question of whether the “actuarial equivalence” and “same methodology” requirements of the Medicare Act apply to RADV audit recoveries, which have a different statutory basis than the Overpayment Rule.[2]
The outcome of any litigation could have a significant dollar impact on MA participants. CMS has identified very large historical payment errors in MA, and has estimated $435 million in annual payment recoveries from RADV audits using extrapolated payment error calculations and no FFS Adjuster. 83 Fed. Reg. at 55,067.
Insurers should also note the methodology applicable to a particular RADV audit year. CMS has advised that it intends to seek extrapolated recoveries for MA payment year 2011 and all subsequent payment years. However, the agency acknowledged some doubt as to whether extrapolating recoveries prior to PY 2014 would require authority to act retroactively (as audits for payment years 2011-2013 had been performed prior to publication of the NPRM). 83 Fed. Reg. at 55,039. Should CMS not seek extrapolated recoveries for any particular payment years, each audited MA plan would be left with a data sample containing some, but not all, of the originally unsupported diagnoses submissions–resembling those in the Overpayment Rule context. Accordingly, under the Becerra precedent, courts should not require application of a FFS Adjuster to unextrapolated recoveries. By contrast, payment years subject to an extrapolated payment error calculation methodology still present an open question. Although, as discussed above, CMS will have a powerful basis to argue that under the Becerra court’s reasoning, no FFS Adjuster should be required for extrapolated recoveries, either.
Whistleblower Actions
The Becerra decision may also impact MA insurers and health care providers facing whistleblower lawsuits. Although some defendants in these suits have raised defenses based on the “actuarial equivalence” language of the Medicare statute, the D.C. Circuit’s decision will present a formidable obstacle, at least, to these arguments moving forward.
At the same time, defendants can still point to Becerra as affirming that they cannot be held liable for overpayments that they “should have determined through the exercise of reasonable diligence.” Instead, relators and the government must continue to meet a higher standard of proof: either actual knowledge or reckless disregard or deliberate ignorance. Insurers and providers participating in Medicare Advantage should be careful to ensure their medical records systems and submissions adhere to these standards and that they report overpayments accurately.
Conclusion
Pending any further appeals,[3] the Becerra decision should carry great weight in the MA space moving forward. With the Overpayment Rule again enforceable, MA organizations must be careful to report unsupported medical diagnoses identified in medical record reviews. Plans defending whistleblower litigation or preparing to report overpayments must be prepared to acknowledge that principles of “actuarial equivalence” are unlikely to provide a strong defense.
Similarly, unless any insurers who challenge the anticipated RADV audit rule successfully distinguish the D.C. Circuit’s new precedent, plans must be prepared to dispute or reimburse RADV audit recoveries without application of an FFS Adjuster.
[1] CMS has previously described auditing a sample of 201 enrollees from a particular plan, divided into three groups of 67 enrollees representing low, moderate, and high risk. Contract-wide payment error would be calculated from these samples. In its 2018 NPRM, the agency proposed that it might also sample enrollees representing a particular “sub-cohort or sub-cohorts” of the contract, representing a particular hierarchical condition category (HCC) or set of HCCs. Payment errors would then be calculated and extrapolated across the respective sub-cohorts. 83 Fed. Reg. 54,982, 55,038-55,039 (Nov. 1, 2018).
[2] CMS developed the RADV audit program to comply with the Improper Payments Information Act of 2002, as amended by the Improper Payments Elimination and Recovery Act of 2010 and Improper Payments Elimination and Recovery Improvement Act of 2012.
[3] UHC has not yet challenged the August 13, 2021 Becerra decision. Although UHC may still request a rehearing or en banc proceedings before the D.C. Circuit, or petition for Supreme Court review, courts would have discretion to accept or decline review.