The Court: .... The definition of identified doesn't mean knew, it means knew or with reasonable diligence should have known or maybe didn't care to look.
[Counsel]: Yes, your Honor.
The Court: That's all negligence.
[Counsel]: It bears some similarities to negligence, your Honor.
The Court: Right. So it's not a knowledge based thing?
[Counsel]: Not as it has been interpreted in the overpayment rule.
Hearing Tr. at 34-36.
Most critically for the present challenge, the 2014 Overpayment Rule did not adopt something like an "FFS Adjuster" to recognize that the sources of data are not compatible, i.e. , unaudited traditional Medicare records to determine payments to Medicare Advantage insurers and audited medical charts to determine overpayments. UnitedHealth argues that the 2014 Overpayment Rule thus fails to ensure "actuarial equivalence" between CMS's own costs and what CMS pays Medicare Advantage insurers to provide the same coverage. Rather, it subjects the insurers to a more searching form of scrutiny than CMS applies to its own enrollee data, thus resulting in a false appearance of better health among Medicare Advantage beneficiaries compared to traditional Medicare participants and systemic underpayments for healthcare costs to Medicare Advantage insurers. UnitedHealth also argues that the "negligence" standard of liability imposed by the 2014 Overpayment Rule constitutes an unlawful departure from the standard for liability under the False Claims Act.
III. LEGAL STANDARD
Summary judgment is available when "the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a). A fact is material if it is capable of affecting the outcome of litigation. Anderson v. Liberty Lobby, Inc. ,
Summary judgment is the proper stage for determining whether, as a matter of law, an agency action is supported by the administrative record and is consistent with the Administrative Procedure Act (APA). Richards v. INS ,
A. Statutory Requirement of "Actuarial Equivalence"
The statutory provision at issue states that "the Secretary shall adjust the payment amount" of fixed monthly payments to Medicare Advantage insurers "for such risk factors as age, disability status, gender, institutional status, and such other factors as the Secretary determines to be appropriate, including adjustment for health status ... so as to ensure actuarial equivalence." 42 U.S.C. § 1395w-23(a)(1)(C)(i). A traditional rule of statutory interpretation renders the use of "shall" a mandatory obligation. See Anglers Conserv. Network v. Pritzker ,
UnitedHealth argues that the 2014 Overpayment Rule violates the statutory mandate of "actuarial equivalence." CMS responds that Medicare Advantage insurers are paid "a sum equal to the cost that CMS would expect to bear in providing traditional Medicare benefits to a given beneficiary" and there is thus "equivalence between an expected cost, on the one hand, and a known payment, on the other." CMS Mot. at 36.
In its briefs, CMS fails adequately to address the actuarial problem posed by the 2014 Overpayment Rule because of the different data sources on which it rests; the same actuarial problem was recognized and mitigated by CMS in 2012 with the FFS Adjuster for RADV audits but, surprisingly, omitted in 2014. The record is clear that payments for care under traditional Medicare and Medicare Advantage are both set annually based on costs from unaudited traditional Medicare records, but the 2014 Overpayment Rule systemically devalues payments to Medicare Advantage insurers by measuring "overpayments" based on audited patient records. This distinction makes an actuarial difference.
In plain English, doctors treating patients under traditional Medicare bill CMS by the procedure involved and not by diagnosis code(s). While the doctors are required to enter diagnosis codes, that information is irrelevant to payment. As far as the record reveals, the diagnosis codes in traditional Medicare are never verified because they do not matter to payment. "[T]he risk adjustment model is built on unaudited data about traditional, fee-for-service Medicare beneficiaries, which must contain errors." CMS Mot. at 37. However, those very same diagnosis codes are presumed to have been accurate when CMS inputs all the data concerning beneficiaries of traditional Medicare into its regression model, which ultimately computes a value for each diagnosis. In consequence, the rates at which CMS pays Medicare Advantage insurers are based on flawed data across the millions of people in traditional Medicare. Yet the 2014 Overpayment Rule ignores those flaws when defining an "overpayment."
It is critical to appreciate that CMS does not claim that it audits traditional Medicare patient records; to the contrary, it accepts their diagnosis codes as given. See CMS Mot. at 7 (agreeing that traditional Medicare payments to doctors "depend only on the services ... and not in any way on the diagnoses submitted"). It is also critical to appreciate that CMS does not show more errors or fraud in the charts of Medicare Advantage beneficiaries than in the charts of traditional Medicare beneficiaries. But the effect of the 2014 Overpayment Rule, without some kind of adjustment, is that Medicare Advantage insurers will be paid less to provide
Neither party cites, and the Court has not located, any case in which a court has defined the precise meaning of "actuarial equivalence" as used in 42 U.S.C. § 1395w-23(a)(1)(C)(i). Congress used the same language in the Employee Retirement Income Security Act (ERISA),
The term also appears in the Medicare Part D statute, which provides that certain prescription-drug coverage is subject to an "actuarial equivalence requirement" that is described in implementing regulations as "a state of equivalent value demonstrated through the use of generally accepted actuarial principles and in accordance with ... CMS actuarial guidelines."
Based on these references to actuarial equivalence, CMS argues that the term "means to equate either an expected value with a known value (as in the case of an annuity and a lump sum payment) or two expected values (as in the case of benefit plans)."
Of particular assistance here, the D.C. Circuit specifically noted that two figures are actuarially equivalent only when they
CMS is the insurer for traditional Medicare. Under the 2014 Overpayment Rule, however, the "expected cost" to the government insurer for traditional Medicare, i.e. , CMS, would be less than the "expected cost" to a private insurance company offering Medicare Advantage coverage. The problem would immediately arise when a Medicare Advantage insurer found its payments from CMS lower than traditional Medicare payments for comparable patients, due to reductions for any "overpayments" as defined by the 2014 Overpayment Rule. The use of unaudited CMS data, with its known and unknown errors, to set the rates by which Medicare Advantage insurers are paid and then the use of audited data to define "overpayments" will lead to this result. See Academy of Actuaries Comment at AR5235 ("An underlying principle of risk-adjustment systems is that there needs to be consistency in the way the model was developed and how it is used. The [model's] risk-adjustment factors were developed with FFS data that, to the best of our knowledge, were not validated or audited for accuracy.").
RADV audits, of course, are conducted for the same purpose as the 2014 Overpayment Rule: to identify those claims for medical care that are not supported by medical diagnoses. In the context of an RADV audit, a contract-wide "error rate" is extrapolated from a sample and extended to an entire contract; a Medicare Advantage insurer may be required to return monies to CMS based on the extrapolated error rate. In that context, CMS heeded the advice of actuaries and adopted the FFS Adjuster to achieve actuarial equivalence between Medicare Advantage and traditional Medicare. Under an RADV audit, therefore, an "overpayment" is shown when, and only when, the error rate for a Medicare Advantage contract is greater than the CMS error rate. See RADV Final Methodology at AR5314 ("[T]o determine the final payment recovery amount, CMS will apply a Fee-for-Service Adjuster ... as an offset to the preliminary recovery amount.").
The base rate for the "average Medicare beneficiary" and specific rates for diagnosis codes are determined using unverified CMS data. From this uncontested fact, UnitedHealth argues that relying on audited data to identify alleged overpayments to Medicare Advantage insurers is actuarially unsound and violates the statute. It contends that the statutory mandate of actuarial equivalence requires CMS to use the "same methodology" for each. See 42 U.S.C. § 1395w-23(b)(4)(D). According to the argument, CMS cannot subject the diagnosis codes underlying Medicare Advantage payments to a different level of scrutiny than it applies to its own payments under traditional Medicare without impermissibly skewing the calculus: by doing so, it ensures that there will not be actuarial equivalence between traditional Medicare payments and Medicare Advantage payments for comparable patients.
CMS fails to respond adequately. The agency has been explicit that the 2014 Overpayment Rule requires "proactive compliance activities" and other measures to ensure that overpayments, defined as any unsupported diagnosis, are identified and repaid promptly.
B. Statutory Requirement of "Same Methodology"
UnitedHealth argues that the 2014 Overpayment Rule violates other statutory requirements as well. In computing expenditures for traditional Medicare (information that determines patient risk scores and Medicare Advantage payment rates), CMS must "us[e] the same methodology as is expected to be applied in making payments" to Medicare Advantage plans. 42 U.S.C. § 1395w-23(b)(4)(D). UnitedHealth insists that CMS fails to comply with this mandate because the "methodology" applied in "making payments" to the insurers involves reconciliation based strictly on audited diagnosis codes for Medicare Advantage patients, in sharp contrast to unverified diagnosis codes for traditional Medicare patients from which payment rates were set. The argument also raises the question of the meaning of "applicable reconciliation" contemplated by the statute. Id. § 1320a-7k(d)(4)(B). The logic of the earlier discussion of "actuarial equivalence" commands the results here.
For present purposes, the fly in the ointment is that CMS recognized the actuarial need to apply an FFS Adjuster to the RADV audit program because of its failure, as proposed, to maintain actuarial equivalence in payments between traditional Medicare and Medicare Advantage but CMS refused to maintain such actuarial equivalence in the 2014 Overpayment Rule. Yet without some adjustment, the entire Rule would fail. Whether analyzed as a direct question of the statutory requirement of actuarial equivalence or an indirect question of the requirements of explicit statutory language concerning "same methodology," the result is the same: the 2014 Overpayment Rule fails to recognize a crucial data mismatch and, without correction, it fails to satisfy 42 U.S.C. § 1395w-23(b)(4)(D).
C. Arbitrary and Capricious
It is established law that an agency must provide a legitimate reason for departing from or rejecting a previous rule. See Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm Mutual Auto. Ins. Co. ,
The first, most recent, and most apt is the stated rationale on which CMS ultimately included the FFS Adjuster in the RADV audit process, as explained in the official notice of the methodology CMS would use to extrapolate payment errors to a contract-wide error rate. See RADV Final Methodology at AR5311-15. After notice and comment on the proposed audit process, including from the American Academy of Actuaries, CMS explained:
CMS will apply a Fee-for-Service Adjuster (FFS Adjuster) amount as an offset to the preliminary recovery amount.... The FFS adjuster accounts for the fact that the documentation standard used in RADV audits to determine a contract's payment error (medical records) is different from the documentation standard used to develop the [MA] risk-adjustment model (FFS claims) . The actual amount of the adjuster will be calculated by CMS based on a RADV-like review of records submitted to support FFS claims data.
RADV Final Methodology at AR5314-15 (emphasis added). 42 U.S.C. § 1395w-23(b)(4)(D). (At oral argument, counsel for CMS stated that the anticipated audit, whose goal is to "publish[ ] a finalized FFS adjuster," is not concluded. See Hearing Tr. at 31-32.) UnitedHealth urges the Court to find that this CMS explanation of the need for an FFS Adjuster for audits, due to the different data sources from which pay rates and error rates are determined, is a singular and strong demonstration of the inadequacies of the 2014 Overpayment Rule, which is based on the same dissimilar data sources but lacks such an adjustment.
Second, UnitedHealth points to two notices from CMS that recognized the differences in data for traditional Medicare and Medicare Advantage healthcare coverage. It notes the CMS rationale for applying a "Coding Intensity Adjustment" to Medicare Advantage insurers. Medicare Advantage plans contain more diagnosis codes than does traditional Medicare, which could lead to overpayments relative to traditional Medicare costs for the same patient. CMS implemented a Coding Intensity Adjustment to adjust for the higher prevalence of diagnosis codes in Medicare Advantage plans. When it did so, CMS emphasized that it was concerned about the imbalance in the number of diagnosis codes between traditional Medicare and Medicare Advantage and not "improper coding." Advance Notice of Methodological Changes for CY 2009 Parts C and D Rates and Policies (Feb. 22, 2008) (2009 Advance Notice) at AR4231 ("We do not assume that the coding pattern differences that we found in our study are the result of improper coding.... However, because MA coding patterns differ from FFS coding patterns, the normalization factor (which is calculated based on FFS coding) does not currently adjust for these different coding patterns."). In addition, UnitedHealth points to a CMS 2010 rate announcement for Medicare Advantage plans which recognized that because "MA payment methodology is based on fee-for-service payments" by traditional Medicare, such "plans must code the way Medicare Part A and B providers do in order for risk adjustments to be valid." Announcement of CY 2010 Parts C and D Rates and Policies (Apr. 6, 2009) at AR4335.
CMS dismisses these earlier statements as only "varied comments about the purpose of the coding difference adjuster, made in an effort to explain why insurers' search for every supportable diagnosis would lead to overpayment." CMS Mot. at 35. It insists that the agency "has always understood a certification of the 'accuracy' and 'truthfulness' of risk adjustment data to require that any reported diagnosis be substantiated" by underlying records.
The CMS argument does not misstate its regulations but misses the point. UnitedHealth does not contend that Medicare Advantage insurers should be permitted knowingly or recklessly to bill CMS for erroneous diagnosis codes. Instead, it argues that the Medicare statute requires CMS to pay for the healthcare of Medicare Advantage beneficiaries in the same manner, and by the same standards, by which CMS pays for traditional Medicare beneficiaries. That means, for the millions of Americans covered by Medicare and Medicare Advantage, that there are error rates; UnitedHealth argues that it should not be subject to lesser payments, False Claims Act liability, or debarment for errors over these huge populations that are fewer than those errors made by CMS itself.
CMS fails to address the central issue here. The question is whether the documents cited by UnitedHealth constitute an agency policy or position from which the 2014 Overpayment Rule deviated without a reasoned explanation. More specifically, UnitedHealth argues that the analysis in the RADV Final Methodology constituted an agency decision or policy that recognized the necessity of an FFS Adjuster-type procedure to account for discrepancies between the documentation for setting payments to Medicare Advantage insurers and that used for determining whether an "overpayment" had occurred. As to this argument, CMS is essentially silent.
Agency policies and practices may take many forms and still be sufficiently established so that any change in the policy must be explained. Republic Airline Inc. v. U.S. Dept. of Transp. ,
This Court comes to the same conclusion. Having recognized that actuarial equivalence, mandated by statute, required an FFS Adjuster for purposes of defining overpayments because of dissimilar data for RADV audits, CMS provides no legitimate
D. False Claims Act Liability
1. Negligence Standard
UnitedHealth further complains that the 2014 Overpayment Rule unlawfully imposes a negligence standard on Medicare Advantage insurers to identify and report "overpayments," which is inconsistent with the standards of the False Claims Act to which it would otherwise align enforcement. CMS objects, contending that the standard adopted in the 2014 Overpayment Rule, including its requirement of "reasonable diligence," is indistinguishable from the CMS 2000 Rule that required Medicare Advantage insurers to certify to the accuracy of risk adjustment data. See Medicare + Choice Program,
Back to basics. The ACA requires that "[a]n overpayment must be reported and returned" within "60 days after the date on which the overpayment was identified." 42 U.S.C. § 1320a-7k(d)(2). The 2014 Overpayment Rule provides: "The MA organization has identified an overpayment when the MA organization has determined, or should have determined through the exercise of reasonable diligence, that the MA organization has received an overpayment."
In contrast, the False Claims Act-which the ACA refers to for enforcement, see 42 U.S.C. § 1320a-7k(d)(3) -imposes liability for erroneous ("false") claims for payment submitted to the government that are submitted "knowingly." "Knowingly" is a term of art defined in the FCA to include false information about which a person "has actual knowledge," "acts in deliberate ignorance of the truth or falsity of the information," or "acts in reckless disregard of the truth or falsity of the information."
"Congress clearly had no intention to turn the FCA, a law designed to punish and deter fraud, into a vehicle for either 'punish[ing] honest mistakes or incorrect claims submitted through mere negligence' or imposing 'a burdensome obligation' ... rather than a 'limited duty to inquire.' " United States v. Sci. Applications Int'l Corp. ,
2. Definition of "Identified"
UnitedHealth also notes that the proposal for the 2014 Overpayment Rule stated that a Medicare Advantage insurer would have "identified" an overpayment when "it has actual knowledge of the existence of the overpayment or acts in reckless disregard or deliberate ignorance of the existence of the overpayment." 2014 Proposed Rule at 1997 (AR81). However, the final 2014 Overpayment Rule stated that a Medicare Advantage insurer would have "identified" an overpayment when "it has determined, or should have determined through the exercise of reasonable diligence, that the MA organization has received an overpayment."
A regulation "violates the APA, if it is not a 'logical outgrowth' of the agency's proposed regulations." Ass'n of Private Sector Colleges and Univs. v. Duncan ,
IV. CONCLUSION
For the foregoing reasons, the Court will grant UnitedHealth's Motion for Summary Judgment, Dkt. 47; deny CMS's Cross-Motion for Summary Judgment, Dkt. 57; and vacate the 2014 Overpayment Rule. A memorializing Order accompanies this Opinion.
Notes
See Pls.' Mot. for Summary J. (United Mot.) [Dkt. 47]; CMS Mot.; Mem. in Opp'n to Mot. for Summ. J. [Dkt. 58]; Pls.' Mem. in Opp'n to Cross-Mot. for Summ. J. (Pls.' Opp'n & Reply) [Dkt. 60]; Reply to Opp'n to Mot. for Summ. J. [Dkt. 61]; Defs.' Reply to Opp'n to Cross-Mot. for Summ. J. (Defs.' Reply) [Dkt. 64].
The parties argue about the validity of CMS risk factors and risk scores, which, as stated, form the basis for (unaudited) CMS payments to traditional Medicare beneficiaries and payments to Medicare Advantage plans (subject to RADV audits and to the 2014 Overpayment Rule). Going back to these basics and redefining all the risk factors and all the diagnostic codes to account, within that structure, for actuarial equivalence may be the preferred approach but the very heart quakes at the thought, if one or more actuarially-sound "adjusters" might resolve the obvious dissonance in the 2014 Overpayment Rule.
UnitedHealth further urges the Court to find that "it is inherently arbitrary and irrational to calibrate a payment model using one type of data and then operate the model using a different type of data." United Mot. at 32. As discussed above, the Court recognizes and gives substantial weight to the American Academy of Actuaries' analysis of why it is actuarially unsound to "apply the risk-adjustment model in a way that is inconsistent with the way it was developed." Academy of Actuaries Comment at AR5235. Further, " 'unexplained departure from prior agency determinations' is inherently arbitrary and capricious." Nat'l Treasury Emps. Union v. Fed. Labor Relations Auth. ,
The ACA does not use the term "knowingly" but defines it by cross-reference to the FCA. See 42 U.S.C. § 1320a-7k(d)(4)(A) ("The terms 'knowing' and 'knowingly' have the meaning given those terms in section 3729(b) of Title 31.").
