444 F. Supp. 3d 1010 (N.D. Cal. 2020)

From Casetext: Smarter Legal Research

United States ex rel. Ormsby v. Sutter Health

United States District Court, N.D. California, San Francisco Division.

Mar 16, 2020

444 F. Supp. 3d 1010 (N.D. Cal. 2020)

Opinion

Case No. 15-cv-01062-LB

2020-03-16

UNITED STATES of America EX REL. Kathy ORMSBY, Plaintiff, v. SUTTER HEALTH, et al., Defendants.

Olga Yevtukhova, Albert Thomas Morris, United States Department of Justice, Lyle Gruby, Washington, DC Benjamin Joseph Wolinsky, U.S. Attorney's Office Northern District of California, Kimberly Friday, U.S. Attorney's Office, Jeffrey Farley Keller, Kathleen R. Scanlan, Keller Grover LLP, Sarah Pascal Alexander, Constantine Cannon LLP, San Francisco, CA, Gordon Schnell Hamsa Mahendranathan, Constantine Cannon LLP, New York, NY, Mark Allen Kleiman, Pooja Rajaram, Santa Monica, CA, for Plaintiff. Jason Masashi Ohta, Katherine Ann Lauer, Amy Elizabeth Hargreaves, Steven Mark Bauer, Latham & Watkins LLP, San Francisco, CA, Spencer Kohl Turnbull, Latham Watkins LLP, Los Angeles, CA, for Defendants.

LAUREL BEELER, United States Magistrate Judge

Jason Masashi Ohta, Katherine Ann Lauer, Amy Elizabeth Hargreaves, Steven Mark Bauer, Latham & Watkins LLP, San Francisco, CA, Spencer Kohl Turnbull, Latham Watkins LLP, Los Angeles, CA, for Defendants.

ORDER DENYING DEFENDANTS' MOTIONS TO DISMISS

Re: ECF Nos. 66, 68, 81

LAUREL BEELER, United States Magistrate Judge

INTRODUCTION...1019

STATEMENT...1021

1. Medicare Advantage Generally...1021

2. The Government's Allegations Regarding Diagnosis Codes...1025

3. The Government's Allegations Against Sutter and PAMF...1026

3.1 Sutter's and PAMF's "RAF Campaign"...1027

3.1.1 Characterizing conditions as "chronic"...1027

3.1.2 Using a "pit crew" to add diagnosis codes to patient medical records...1028

3.1.3 Pressuring physicians to add diagnosis codes to patient medical records...1028

3.1.4 Pre-populating patient medical records with diagnosis codes...1031

3.1.5 Adding diagnosis codes as addendums to patient medical records...1032

3.2 "Red Flags" Regarding Coding...1033

3.2.1 Audits by MA Organizations...1033

3.2.2 Lack of compliance and training programs...1034

3.2.3 Internal reviews and audits by Sutter and PAMF...1035

3.3 Knowingly Ignoring Red Flags and Actual Notice of False Claims...1040

3.4 Examples of False Diagnosis Codes and Estimate of Total Overpayments...1045

4. The Relator's Allegations Against Sutter Regarding Its Non-PAMF Affiliates...1048

THE FALSE CLAIMS ACT...1054

1. Direct-FCA Claims...1054

2. Reverse-FCA Claims...1055

STANDARD OF REVIEW...1057

ANALYSIS...1057

1. The "Actuarial Equivalence" Argument Is Not a Valid Defense...1058

1.1 The Premise — That CMS Is Violating "Actuarial Equivalence" — Is Not Sound...1060

1.1.1 CMS's Medicare Advantage audit process and the "2014 Overpayment Rule"...1060

1.1.1.1 Audits of MA Organizations...1060

1.1.1.2 The "2014 Overpayment Rule"...1061

1.1.2 UnitedHealthcare 's invalidation of the 2014 Overpayment Rule...1062

1.1.3 UnitedHealthcare does not compel the conclusion that there is no actuarial equivalence here...1064

1.2 "Actuarial Equivalence" Is Not a Defense to an FCA Claim...1067

1.2.1 MA Participants cannot retain payments predicated on false diagnosis codes...1067

1.2.2 The "actuarial equivalence" requirement does not vitiate scienter...1071

2. The Relator Can Pursue Her Sutter-Wide Claims...1073

3. The Plaintiffs Sufficiently Plead Their Claims...1077

3.1 Reverse-FCA Claims...1077

3.1.1 Concealing or avoiding an obligation to pay the government...1077

3.1.2 Scienter...1081

3.2 Direct-FCA Claims...1082

3.2.1 False claims, or false records and statements...1082

3.2.2 Scienter...1083

3.2.3 Materiality and causation...1085

3.3 Common-Law Claims...1086

CONCLUSION...1086

INTRODUCTION

The government and relator Kathy Ormsby sued Sutter Health and its affiliate Palo Alto Medical Foundation ("PAMF") for violations of the False Claims Act ("FCA"), 31 U.S.C. §§ 3729 – 3733, and for common-law claims of payment by mistake and unjust enrichment. Sutter and PAMF control hospitals and physician foundations throughout California. The plaintiffs allege that Sutter and PAMF knowingly submitted thousands of false claims relating to the Medicare Part C Program, known as Medicare Advantage, and knowingly concealed and avoided their obligations to return Medicare Advantage overpayments that they received.

Under the Medicare Advantage program, the federal agency that administers the Medicare program — the Centers for Medicare and Medicaid Services ("CMS") — contracts with private health-insurance companies (known as "Medicare Advantage Organizations" or "MA Organizations") that operate health-insurance plans (known as "Medicare Advantage Plans" or "MA Plans") that cover Medicare beneficiaries. MA Organizations in turn contract with medical providers such as Sutter and PAMF for healthcare services (e.g., doctor's visits, hospitalizations, etc.) for the beneficiaries enrolled in the MA Plan. CMS pays MA Organizations a capitated (fixed) amount for each beneficiary enrolled in their MA Plans. MA Organizations share those payments with their contracted medical providers.

The amount that CMS pays for a given beneficiary depends in large part on the beneficiary's health status. Broadly speaking, CMS pays higher rates for sicker beneficiaries and lower rates for healthier beneficiaries, out of a recognition that sicker beneficiaries likely will need more care.

CMS relies on "diagnosis codes" to calculate beneficiaries' health statuses. Every disease, injury, infection, and symptom has its own diagnosis code. Medical providers such as Sutter and PAMF enter diagnosis codes into beneficiaries' medical records after "physician-patient encounters" (a doctor's physical examination of a patient). Medical providers submit the diagnosis codes to MA Organizations, which submit them to CMS. CMS uses some of the diagnosis codes (such as codes for major, severe, or chronic illnesses ) — referred to as "risk-adjusting diagnosis codes" — in a risk-adjustment model to calculate a risk score for each beneficiary. The diagnosis codes that medical providers submit are the only factors that CMS uses to determine a beneficiary's health status to calculate the beneficiary's risk score and thus to calculate how much CMS will pay for that beneficiary.

As the Ninth Circuit has recognized, participants in the Medicare Advantage program (MA Organizations and medical providers) (collectively, "MA Participants") have an incentive to over-report diagnosis codes in order to raise beneficiary risk scores and, in turn, increase the amount that CMS pays them. United States ex rel. Silingo v. WellPoint, Inc. , 904 F.3d 667, 672 (9th Cir. 2018) ; United States v. United Healthcare Ins. Co. , 848 F.3d 1161, 1168 (9th Cir. 2016) ( Swoben ).

The government alleges that Sutter and PAMF maximized the number of risk-adjusting diagnosis codes that they reported (through MA Organizations to CMS) in order to increase the payments that CMS paid to the MA Organizations and, ultimately, to Sutter and PAMF. Among other things, Sutter and PAMF (1) pre-populated Medicare Advantage beneficiaries' medical records with diagnosis codes before physicians saw or diagnosed the beneficiaries, (2) had non-physician "coders" review Medicare Advantage beneficiaries' medical records and retroactively add codes that the physicians supposedly "missed" or change the physicians' codes to codes for more severe conditions, and (3) knowingly submitted unsupported diagnosis codes to CMS and prohibited their internal coders/auditors from deleting unsupported diagnosis codes. Relator Kathy Ormsby, who worked as PAMF's Risk-Adjustment Project Manager from 2013 to 2015, conducted internal reviews and audits, found that large percentages of the risk-adjusting diagnosis codes that Sutter and PAMF submitted were false, and reported her findings to Sutter and PAMF management. The government alleges Sutter and PAMF were deliberately ignorant or reckless about their submitting false diagnosis codes and retaining payments predicated on false diagnosis codes and did nothing to fix the problem. The government alleges that Sutter and PAMF violated the FCA by (1) submitting false risk-adjusting diagnosis codes to CMS and (2) failing to return payments predicated on false diagnosis codes. These allegations also are the predicates for the government's common-law claims for payment by mistake and unjust enrichment.

Ms. Ormsby's complaint is broader than the government's complaint. The government's complaint is limited to PAMF (and Sutter's actions in connection with PAMF). Ms. Ormsby alleges that Sutter committed similar violations of the FCA through its other affiliates.

Sutter and PAMF move to dismiss both complaints under Federal Rules of Civil Procedure 12(b)(6) and 9(b). Sutter's and PAMF's main argument is that the FCA claims fail because the plaintiffs have not alleged a sufficient threshold diagnosis-code error rate. The argument is predicated on a statute — 42 U.S.C. § 1395w-23(a)(1)(C)(i) — that provides that CMS must adjust Medicare Advantage payments to ensure "actuarial equivalence" with traditional Medicare. Sutter and PAMF maintain that (1) traditional Medicare providers also submit false or unsupported diagnosis codes and (2) under the principle of "actuarial equivalence," Medicare Advantage providers that submit false or unsupported diagnosis codes are not overpaid unless their diagnosis-code "error rate" exceeds the "error rate" of traditional Medicare providers. Sutter and PAMF contend that because the plaintiffs have not alleged that their "error rate" exceeds the traditional-Medicare "error rate," they have not sufficiently pleaded a claim. Sutter also argues that where, as here, the government has intervened in an FCA case, a relator like Ms. Ormsby cannot independently pursue claims that exceed the scope of the government's intervention and that the court therefore should dismiss her claims regarding Sutter's non-PAMF affiliates. Sutter and PAMF also move to dismiss the claims on the ground that the plaintiffs did not sufficiently allege facts establishing their knowing failure to return overpayments or their knowing submission of false claims.

The court denies the motions to dismiss. First, the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i) is not a defense to an FCA claim, does not entitle MA Participants to submit unsupported diagnosis codes (or to keep and not return and report payments predicated on unsupported diagnosis codes), and does not require an FCA plaintiff to allege that an MA Participant's "error rate" exceeds the traditional-Medicare "error rate" to plead a claim. Second, the government's intervention with respect to Sutter and PAMF does not bar Ms. Ormsby from pursuing her claims regarding Sutter's non-PAMF affiliates. Third, Sutter's and PAMF's other challenges to the sufficiency of the plaintiffs' pleadings fail.

As explained below, traditional Medicare providers are paid under a fee-for-service model and are not paid based on the diagnosis codes they submit. A traditional Medicare provider that submits an unsupported diagnosis code does not cause CMS to pay out any additional money, whereas a Medicare Advantage provider that submits an unsupported diagnosis code does.

STATEMENT

1. Medicare Advantage Generally

In recent cases, the Ninth Circuit has discussed the Medicare Advantage program in the context of FCA claims. Silingo , 904 F.3d 667 ; Swoben , 848 F.3d 1161.

"Medicare Advantage is a modern adaptation of the momentous 1960s-era [Medicare] program." Silingo , 904 F.3d at 672. "Traditional Medicare uses a fee-for-service payment model, whereby the more services physicians perform, the more money they earn." Id. "After Medicare was enacted, however, experts came to realize that this payment structure encourages healthcare providers to order more tests and procedures than medically necessary." Id. (citing Thomas L. Greaney, Medicare Advantage, Accountable Care Organizations, and Traditional Medicare: Synchronization or Collision? , 15 Yale J. Health Pol'y, L. & Ethics 37, 38, 41 (2015) ). "Medicare Advantage seeks to improve the quality of care while safeguarding the public fisc by employing a ‘capitation’ payment system." Id.

Accord Gov't Compl. – ECF No. 41 at 2 (¶ 1), 10 (¶ 27). Citations refer to material in the Electronic Case File ("ECF"); pinpoint citations are to the ECF-generated page numbers at the top of documents.

Accord id. at 3 (¶ 3), 11 (¶ 30).

"Medicare beneficiaries have the option of receiving benefits through private health plans as an alternative to the traditional fee-for-service Medicare program." Swoben , 848 F.3d at 1167. "Under this option, known as Medicare Advantage or Medicare Part C, the government pays Medicare Advantage organizations a capitated (per enrollee) amount to provide medical benefits." Id. "Capitation means an amount is paid per person." Silingo , 904 F.3d at 672 (citing Capitation, Black's Law Dictionary (10th ed. 2014)). "Under Medicare Advantage's capitation system, private health insurance organizations provide Medicare benefits in exchange for a fixed monthly fee per person enrolled in the program — regardless of actual healthcare usage." Id. "These organizations pocket for themselves or pay out to their enrollees' providers the difference between their capitation revenue and their enrollees' medical expenses, creating an incentive for the organizations to rein in costs." Id. (citing Patricia A. Davis et al., Cong. Research Serv., R40425, Medicare Primer 20 (2017), https://fas.org/sgp/crs/misc/R40425.pdf (last visited Mar. 16, 2020)).

Accord id. at 2 (¶ 1), 10 (¶ 28).

Accord id. at 3 (¶ 3), 11 (¶ 30).

Accord id. at 11 (¶ 30).

Accord id.

"The government adjusts the monthly payments to Medicare Advantage organizations to reflect the health status of their enrollees." Swoben , 848 F.3d at 1167 (citing 42 U.S.C. § 1395w-23(a)(1)(C)(i), (a)(3) ; 42 C.F.R. § 422.308(c)(2) ). "This ensures Medicare Advantage ‘organizations are paid appropriately for their plan enrollees (that is, less for healthier enrollees and more for less healthy enrollees).’ " Id. (quoting Establishment of the Medicare Advantage Program , 70 Fed. Reg. 4588, 4657 (Jan. 28, 2005) ).

Accord id. at 3 (¶ 4), 11 (¶ 30).

Accord id.

"Medicare Advantage organizations obtain diagnosis codes from healthcare providers after these providers have had medical visits with plan enrollees." Silingo , 904 F.3d at 672 (citing Ctrs. for Medicare and Medicaid Servs., Pub. No. 100-16, Medicare Managed Care Manual , ch. 7, § 40 (2014), https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/mc86c07.pdf (last visited Mar. 16, 2020)). "Physicians and other health care providers submit diagnosis codes to the Medicare Advantage organizations[.]" Swoben , 848 F.3d at 1167 (citing Medicare Program; Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Drug Benefit Programs , 74 Fed. Reg. 54,634, 54,674 (Oct. 22, 2009) ). "In turn, Medicare Advantage organizations report the diagnosis codes that they receive to the Centers for Medicare and Medicaid Services (‘CMS’) for use in the risk adjustment model that is the key to calculation of capitation rates." Silingo , 904 F.3d at 672 (citing Medicare Managed Care Manual , ch. 7, § 40); accord Swoben , 848 F.3d at 1167. "The risk adjustment model deems a Medicare Advantage enrollee to be as healthy as the average Medicare beneficiary unless CMS receives updated diagnosis codes for the enrollee every year." Silingo , 904 F.3d at 672 (citing Medicare Managed Care Manual , ch. 7, §§ 20, 70, 70.2.5, 120.2.4). "These diagnosis codes contribute to an enrollee's risk score, which is used to adjust a base payment rate." Swoben , 848 F.3d at 1167–68 (citing 74 Fed. Reg. at 54,674 ).

Accord id. at 3–4 (¶¶ 4–5), 11–13 (¶¶ 30–35).

Accord id. at 3–4 (¶¶ 4–5), 13 (¶ 35).

Accord id. at 3–4 (¶¶ 4–5), 13 (¶¶ 34–35).

Accord id. at 13 (¶ 34).

Accord id. at 3–4 (¶¶ 4–5), 11–13 (¶¶ 30–35).

"Unfortunately, human nature being what it is, Medicare Advantage organizations also have some incentive to improperly inflate their enrollees' capitation rates, if these organizations fall prey to greed." Silingo , 904 F.3d at 672. "[T]here is an incentive for Medicare Advantage organizations to potentially over-report diagnoses so that they can increase their payment[.]" Swoben , 848 F.3d at 1168 (internal brackets omitted).

Accord id. at 4–5 (¶ 6).

Accord id.

"With data for millions of people being submitted each year, CMS is unable to confirm diagnoses before calculating capitation rates." Silingo , 904 F.3d at 672. "Instead, the agency accepts the diagnoses as submitted, and then audits some of the self-reported data a few years later to ensure that they are adequately supported by medical documentation." Id. at 672–73 (citing 42 C.F.R. §§ 422.310(e), 422.311 ; Contract Year 2015 Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Drug Benefit Programs , 79 Fed. Reg. 1918, 2001 (Jan. 10, 2014) ). "These audits have revealed excess payments for unsupported diagnoses steadily increasing over the last decade, reaching an estimated $16.2 billion — nearly ten cents of every dollar paid to Medicare Advantage organizations — in 2016 alone." Id. at 673 (emphasis in original) (citing James Cosgrove, U.S. Gov't Accountability Office, GAO-17-761T, Medicare Advantage Program Integrity: CMS's Efforts to Ensure Proper Payments and Identify and Recover Improper Payments 1 (2017), https://www.gao.gov/assets/690/685934.pdf (last visited Mar. 16, 2020); James Cosgrove, U.S. Gov't Accountability Office, GAO-13-206, Medicare Advantage: Substantial Excess Payments Underscore Need for CMS to Improve Accuracy of Risk Score Adjustments 9–10 (2013), https://www.gao.gov/assets/660/651712.pdf (last visited Mar. 16, 2020)).

"To combat the ‘incentive for Medicare Advantage organizations to potentially over-report diagnoses,’ Medicare regulations require risk adjustment data to be produced according to certain best practices." Id. (internal brackets omitted) (citing 79 Fed. Reg. at 2001 ). "Every diagnosis code submitted to CMS must be based on a ‘face-to-face’ visit that is documented in the medical record." Id. (citing Medicare Managed Care Manual , ch. 7, §§ 40, 120.1.1). "Medical records must be validated by qualifying ‘physician/practitioner signatures and credentials.’ " Id. (citing Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Drug Benefit Programs , 75 Fed. Reg. 19,678, 19,743 (Apr. 15, 2010) ). "Further, electronic medical records must meet special signature requirements and use software that is ‘protected against modification.’ " Id. (citing Ctrs. for Medicare and Medicaid Servs., Pub. No. 100-08, Medicare Program Integrity Manual , ch. 3, § 3.3.2.4 (2018), https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/PIM83c03.pdf (last visited Mar. 16, 2020)).

Accord id. at 10–11 (¶ 29).

Accord id. at 12 (¶ 32).

Accord id.

"As a further bulwark against fraud, Medicare Advantage organizations must certify the accuracy, completeness and truthfulness of the data they provide to CMS, including risk adjustment data, as a condition to receiving payment[.]" Swoben , 848 F.3d at 1168 (citing 42 C.F.R. § 422.504(l ) ); accord Silingo , 904 F.3d at 673 ("[I]t is an express condition of payment that a Medicare Advantage organization ‘certify (based on best knowledge, information, and belief) that the [risk adjustment] data it submits ... are accurate, complete, and truthful.’ ") (brackets and ellipsis in original) (quoting 42 C.F.R. § 422.504(l )(2) ). "The organization also is required to ‘adopt and implement an effective compliance program, which must include measures that prevent, detect, and correct non-compliance with CMS'[s] program requirements,’ such as written standards of conduct, the designation of a compliance officer, and other listed minimum requirements." Silingo , 904 F.3d at 673 (internal brackets omitted) (quoting 42 C.F.R. § 422.503(b)(4)(vi) ). Among other things, the compliance programs must "includ[e] ‘procedures for internal monitoring and auditing’ and for ‘ensuring prompt responses to detected offenses.’ " Swoben , 848 F.3d at 1174 (internal brackets omitted) (quoting 42 C.F.R. § 422.503(b)(4)(vi), (vi)(F), (vi)(G) ). Similarly, medical providers such as Sutter and PAMF that contract with MA Organizations must certify that the data (including diagnosis codes) they submit are "accurate, complete, and truthful." 42 C.F.R. § 422.504(l )(3). They also are subject to the MA Organization's compliance and training requirements, 42 C.F.R. § 422.503(b)(4)(vi)(C), and must "comply with all Medicare laws, regulations and CMS instructions," 42 C.F.R. § 422.504(i)(4)(v). "The importance of accurate data certifications and effective compliance programs is obvious: if enrollee diagnoses are overstated, then the capitation payments to Medicare Advantage organizations will be improperly inflated." Silingo , 904 F.3d at 673.

Accord id. at 10–11 (¶ 29).

Accord id.

Accord id. More specifically, Medicare regulations define the concepts of "first tier entities," which "means any party that enters into a written arrangement, acceptable to CMS, with an MA organization or applicant to provide administrative services or health care services for a Medicare eligible individual under the MA program," and "related entities," which "means any entity that is related to the MA organization by common ownership or control and (1) Performs some of the MA organization's management functions under contract or delegation; (2) Furnishes services to Medicare enrollees under an oral or written agreement; or (3) Leases real property or sells materials to the MA organization at a cost of more than $2,500 during a contract period." 42 C.F.R. §§ 422.2, 422.500. Medicare regulations impose various requirements with respect to "first tier entities" and "related entities." See, e.g. , 42 C.F.R. §§ 422.503(b)(4)(vi)(C), 422.504(i)(4)(v), (l )(3). The government alleges (and, for the purposes of their motions to dismiss, Sutter and PAMF do not deny) that Sutter and PAMF are "first tier entities" and "related entities." Gov't Compl. – ECF No. 41 at 10 (¶ 29).

Accord id. at 5 (¶ 6).

"The Medicare Advantage capitation payment system is subject to the False Claims Act." Id.

Accord id. at 8 (¶ 19).

2. The Government's Allegations Regarding Diagnosis Codes

In its complaint, the government alleges the following about diagnosis codes and their role in the Medicare Advantage program.

Under Medicare Advantage, CMS pays each MA Organization a monthly fixed, capitated, per-beneficiary amount, adjusted by the expected risk of each beneficiary. The baseline payment is set each year through a bidding process. Each MA Plan, through an MA Organization, submits a bid amount. CMS compares these bid amounts to a benchmark that it sets through a statutory formula. CMS then adjusts the baseline payment for each beneficiary based on that beneficiary's (1) demographic factors, such as age and gender, and (2) health status. Specifically, CMS uses a risk-adjustment model called the Hierarchical Conditions Category ("HCC") model that takes into account a beneficiary's demographic factors and health status and generates a numerical risk score, sometimes referred to as the "Risk-Adjustment Factor" ("RAF"), for the beneficiary. This risk score is a multiplier that is applied to the baseline payment to the MA Organization (e.g., a beneficiary's risk score of 1.2 means that CMS pays out 1.2 times the baseline payment for that beneficiary's care).

Id. at 11 (¶ 30).

Id.

Id. (citing 42 U.S.C. § 1395w-23(a)(1)(C) ).

Id. at 11–12 (¶¶ 30–31) (citing 42 U.S.C. § 1395w-23(a)(1)(G) ; 42 C.F.R. § 422.308(e) ).

Id. at 11 (¶ 30).

With respect to health status, CMS's HCC model relies on diagnosis codes documented by treating physicians during office visits and hospital outpatient and inpatient stays. Medical providers (e.g., physicians, or organizations that employ physicians such as Sutter and PAMF) submit the diagnosis codes to CMS, which uses the codes to determine a beneficiary's health status in order to calculate the beneficiary's risk score. These diagnosis codes, as reported by medical providers, are the only factors that CMS uses to determine a beneficiary's health status. Medicare regulations and guidance are clear that CMS relies on these diagnosis codes to make accurate payments for each beneficiary enrolled in the Medicare Advantage program.

Id. at 11–12 (¶ 31).

Id.

Id. at 12 (¶ 31).

Id. at 13–14 (¶ 36) (" ‘Accurate risk-adjusted payments rely on the diagnosis coding derived from the member's medical record.’ ") (quoting Ctrs. for Medicare and Medicaid Servs., 2013 National Technical Assistance Risk Adjustment 101 Participant Guide § 4.2 (2013) and citing 42 C.F.R. § 422.504(l )(3) ).

The higher a beneficiary's risk score, the higher the payments that CMS makes to MA Organizations for that beneficiary. When medical providers submit more risk-adjusting diagnosis codes, they increase the amount that CMS pays. Conversely, when providers delete risk-adjusting diagnosis codes (such as erroneous, invalid, unsupported, or otherwise false codes), they reduce the amount that CMS pays. Given the importance of accurate information, CMS requires MA Organization executives to certify that the diagnosis codes (and other patient data) that they submit to CMS are true and accurate. CMS requires these signed certifications as a condition of payment. If a subcontractor or related entity (such as a medical provider operating under a contract with the MA Organization) generates the data, the subcontractor or related entity also must certify that its diagnosis codes and data are true and accurate. Additionally, CMS audits MA Organizations, and MA Organizations in turn audit medical providers, regarding the accuracy of their diagnosis coding. If a medical provider submits erroneous risk-adjusting diagnosis codes (whether manually, through an automated "sweep" system, or otherwise), then CMS requires the return of any overpayments predicated on those codes.

Id. at 13 (¶ 34).

Id. at 5 (¶ 6), 13 (¶ 34).

Id. at 14 (¶ 37).

Id. at 44 (¶ 126).

Id.

Id. (citing 42 C.F.R. § 422.504(l )(3) ).

Id. (¶ 127).

Id. (citing Medicare Managed Care Manual , ch. 7, § 40; Swoben , 848 F.3d at 1176–77 & n.8 ).

MA Organizations and medical providers are able to delete diagnosis codes previously submitted to CMS that are erroneous, invalid, unsupported, or otherwise false. If an MA Organization or medical provider deletes or withdraws a diagnosis code, CMS's electronic-processing system automatically recalculates the respective payment amount, which is the first step in CMS's process to recoup the payment associated with the deleted or withdrawn diagnosis code.

Id. at 14 (¶ 37).

Id. at 9 (¶ 23), 14 (¶ 37).

3. The Government's Allegations Against Sutter and PAMF

Sutter, through its provider affiliates, including PAMF, furnishes healthcare services to thousands of Medicare Advantage beneficiaries under at least 10 MA Plans managed by three MA Organizations: UnitedHealth, Health Net, and Humana. Under their agreements with the MA Organizations, Sutter and PAMF submit patient-encounter data, including diagnosis codes, to the MA Organizations for their MA Plan beneficiaries. The MA Organizations then submit these diagnosis codes to CMS.

Id. at 3–4 (¶ 5). The government refers to the UnitedHealth entity as "United Healthcare Group" in one paragraph of its complaint, id. , but otherwise refers to it as "United Health Group," see, e.g. , id. at 22 (¶ 59). Ms. Ormsby refers to it in her complaint as "UnitedHealth Group Inc." Relator First Amend. Compl. ("FAC") – ECF No. 52 at 12 (¶ 39). This order refers to it as "UnitedHealth."

Gov't Compl. – ECF No. 41 at 4 (¶ 5), 13 (¶ 35).

Id.

The MA Organizations' agreements with Sutter and PAMF provide that they will pay Sutter and PAMF a set percentage of the payments they receive from CMS. Additionally, Sutter and PAMF have "gainsharing" agreements with the MA Organizations whereby Sutter and PAMF receive incentive payments based in whole or in part on the total revenues that the MA Organizations receive for beneficiaries that Sutter and PAMF treat. Consequently, when Sutter and PAMF submit more risk-adjusting diagnosis codes, Sutter and PAMF cause CMS to make higher payments to Sutter and PAMF.

Id. at 4–5 (¶ 6), 13 (¶ 34).

Id. at 4–5 (¶ 6).

Id. at 5 (¶ 6), 13 (¶ 34).

3.1 Sutter's and PAMF's "RAF Campaign"

"Beginning no later than 2010, Sutter and PAMF began a campaign to increase the number of risk-adjusting diagnosis codes for its MA patients, in order to generate revenue and maximize reimbursement from CMS. This effort became known as the RAF Campaign."

Id. at 16 (¶ 44).

In February 2012, Dr. Jeffrey Burnich (a Sutter Senior Vice President and Executive Officer) told Suzy Cliff (PAMF's Vice President of Revenue Cycle) and other members of management that PAMF was "leaving millions of dollars on the table" from "sub-par coding." A couple of days later, a member of Sutter leadership told PAMF to "identify a PAMF operational director to work with them in improving our RAF scoring/coding on our Medicare Advantage patients." Sutter and PAMF tasked Nancy McGinnis (Sutter's RAF Director) "to lead the efforts to improve Sutter Health's RAF scores."

Id. at 17 (¶ 46).

Id.

Id. (¶ 44).

By November 2012, Sutter and PAMF formalized the RAF Campaign, calling it the "Risk Adjusted Factor Project." The Project's goal was "to reach a 28% improvement in the HCC performance" for its MA Plan patients. By late 2012, Sutter and PAMF outlined new steps in the Project, including approving the hiring of a "Project Manager" to coordinate the Project and a "Database Analyst" to track the diagnosis-coding performance of network physicians.

Id. at 18 (¶ 48).

Id.

As part of the RAF Campaign, Sutter and PAMF identified so-called "Physician Champions" to "act as a liaison between the coding team and the physicians" on the theory that physicians would be more likely to accept diagnosis-coding guidance from other physicians. Dr. Veko Vahamaki, PAMF's Lead RAF Physician Champion, supervised the champions at PAMF's four divisions in Alameda (Dr. Amy Lin), Camino (Dr. Graham Dresden), Palo Alto (Dr. Anita Gupta), and Santa Cruz (Dr. Susan Schaefer).

Id. at 17 (¶ 45).

Id.

Sutter and PAMF engaged in activities to increase the risk-adjusting diagnosis codes they reported to the MA Organizations and CMS — and thereby increase their revenue — including the following.

3.1.1 Characterizing conditions as "chronic"

A medical provider's labeling certain diagnoses "chronic" (as opposed to "acute") allows the provider to add risk-adjusting diagnosis codes to a Medicare Advantage beneficiary's medical record and thereby increase the money that CMS pays for the beneficiary.

Id. at 32 (¶ 93).

In January 2012, Dr. Steven Lane (a physician in PAMF's network and PAMF's Electronic Health Record Ambulatory Physician Director) wrote an email with the subject line "HCC codes: more to consider as chronic?" Dr. Lane wrote, "Over the past year or two ... increasing attention has been focused on the importance of appropriately identifying and coding HCC diagnoses to improve RAF scores and Medicare managed care reimbursement ...."

Id. at 17 (¶ 44).

Id. (ellipsis in complaint).

In early February 2012, Greta Fees (Sutter's Director of Coding, Documentation, and Data Quality) expressed concern about "the added descriptive of chronic to the diagnosis code descriptions" for leukemia, bronchitis, and asthma, among other conditions. Ms. Fees believed that these diagnoses clinically related to "acute" rather than "chronic conditions" and explained that her research did "not support adding the descriptive term of chronic ... as that would change the definition, intent and possibly use of the code." Ms. Fees related these concerns to Dr. Lane and to Dr. Meg Durbin (a PAMF Regional Medical Director, Managed Care), who both wanted to add the "chronic" label to these diagnoses. Drs. Lane and Durbin pushed back in support of the "chronic" designation, tried to pressure Ms. Fees to accept their analyses, and claimed that they were "approaching a consensus" with Ms. Fees despite her continued disagreement with them.

Id. at 32 (¶ 93).

Id. (ellipsis in complaint).

Id.

3.1.2 Using a "pit crew" to add diagnosis codes to patient medical records

Sutter and PAMF maintained a team of non-physician "coders" whose function, among others, was to audit the accuracy of PAMF's diagnosis coding and medical-record documentation.

Id. at 7 (¶ 13).

In November 2012, Dr. Jeffrey Brown (PAMF's Associate Director for Managed Care) approved coders adding risk-adjusting diagnosis codes to patient medical records that physicians missed during their patient visits. Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion) called this a "pit crew plan" and believed that it "would significantly help with the RAF efforts."

Id. at 18 (¶ 49).

Id.

Dr. Christopher Jaeger (a Sutter Vice President and Sutter's Chief Medical Informatics Officer) expressed concern that "having a coder change an entry that I purposefully enter that has clinical meaning to me/others ... seems like a dangerous step." In response, Dr. Vahamaki defended the coder approach and forwarded his exchange with Dr. Jaeger to Julie Cheung (Sutter's RAF Program Manager).

Id. (ellipsis in complaint).

Id.

Another doctor, Dr. Douglas Tucker, complained to Dr. Vahamaki and others in PAMF's management about a coder changing a patient's diagnosis, stating that "changing a diagnosis from acute bronchitis to pneumonia is not a simple or unimportant change" and "it is so obviously unethical."

Id. at 33 (¶ 96).

3.1.3 Pressuring physicians to add diagnosis codes to patient medical records

By late 2012, Sutter asked coders in all of its affiliates, including PAMF, to schedule annual "Medicare Wellness Exams" for Medicare Advantage patients lacking any risk-adjusting diagnosis codes. The government alleges that Sutter scheduled these exams "to ensure the capture of every possible code that could increase CMS's payments." Sutter tracked the success of each affiliate, including PAMF, in scheduling the Medicare Wellness Exams and rewarded meeting a goal of 75% annual wellness visits with a 1% upside bonus at the group level. The government alleges that "Sutter and PAMF understood that ‘capturing more wellness exams’ increased the capture of risk-adjusting diagnosis codes and thus increased revenue." One doctor, Dr. Heather Linebarger, complained to Dr. Vahamaki and others that "I have serious questions about the new policy of booking in Medicare Advantage patients to review all HCC codes ... This represents a waste of time for the patient and a loss of appointment and worsening of access for me."

Id. at 18 (¶ 48).

Id.

Id. at 34 (¶ 96) (ellipsis in complaint).

Also in late 2012, some PAMF physicians began to receive "HCC/RAF cheat sheets" that identified risk-adjusting diagnosis codes that purportedly were common to many Medicare Advantage patients (such as diabetes). "The cheat sheets were used to pressure physicians to add these codes into the patient's electronic medical records even during encounters focusing on other patient healthcare problems."

Id. at 19 (¶ 50).

Id.

Sutter-affiliated physicians, including those at PAMF, received (through Sutter's electronic-medical-record system) a customized "Problem List" for their MA Plan patients. A "Problem List" is a list of health problems with corresponding diagnosis codes and can be used as a high-level summary of a patient's past health problems. At Sutter and PAMF management's direction, coders or Physician Champions pre-populated the Problem Lists with "lucrative" diagnosis codes and had the Problem Lists auto-flag these codes with "a red pushpin icon" that served as a "visual reminder" for physicians to examine patients with those codes in mind. To document a diagnosis code, the physician needed only to electronically move the diagnosis code from the Problem List to the patient-encounter part of the patient's electronic medical record. Sometimes, disputes arose over the pre-population of the Problem Lists. An example is when physicians or PAMF employees did not believe that patient diagnoses qualified as "chronic" (a label that permits the addition of risk-adjusting diagnosis codes and increases the amount that CMS pays), and management disagreed and often "err[ed] on the side of including the [diagnoses] as chronic."

Id. (¶ 51).

Id.

Id. at 19 (¶ 51) (brackets in complaint), 32 (¶ 93).

In early 2013, Dr. Jeffrey Brown (PAMF's Associate Director for Managed Care) sent letters to physicians with more than 20 Medicare Advantage patients "asking those with higher than PAMF average HCC scores what they thought helped them in HCC coding and ask[ing] those with lower than PAMF average scores what the barriers to HCC coding were." Dr. Brown compiled the results in a survey distributed in March 2013 to Sutter and PAMF executives, including Julie Cheung (Sutter's RAF Program Manager), Nancy McGinnis (Sutter's RAF Director), Roger Larsen (PAMF's Chief Financial Officer and a Sutter Regional Vice President of Finance), Suzy Cliff (PAMF's Vice President of Revenue Cycle), and Kris Anne Crow (PAMF's Director of Coding and Education). The survey found that for the above-average-coding physicians, auto-flagging of diagnosis codes in the Problem Lists and the HCC/RAF cheat sheets (called the "HCC code tip sheet" in the survey) especially helped increase coding. By contrast, the below-average-coding physicians focused on patient care and treatment rather than on coding as exemplified by this statement: "I do not address longstanding stable or prior conditions when that is not important to the care being delivered at the moment." In the survey, Dr. Brown classified that statement as among the "Barriers to Better Coding."

Id. at 20 (¶ 53) (brackets in complaint).

Id.

Using data mining, Sutter and PAMF "pushed" their physicians through messages in the electronic medical record to find and refresh especially high-paying risk-adjusting diagnosis codes to increase patients' risk scores. Similarly, PAMF physicians received "queries" in the electronic medical record from coders reminding the physicians to ensure that all such diagnosis codes were captured. Numerous physicians disliked this practice and felt "pressured" to add diagnosis codes that they did not believe to be clinically accurate or relevant. Furthermore, PAMF coders met one-on-one with physicians to discuss their diagnosis coding. During these meetings, the auditors at times encouraged the physicians to add addendums to their patient records to add risk-adjusting diagnosis codes. (An addendum to a medical record is a note drafted by a physician or other medical professional that clarifies or amends a previous note made by the same professional, typically within 30 days of the encounter. ) At least two doctors, Dr. Williams and Dr. Wong, were prompted to add addendums to records from a prior year and thought that it was unethical to add addendums to old face-to-face encounters. The coders also laid out a plan to address with PAMF physicians other risk-adjusting diagnosis codes that were viewed as "high potential missed opportunity [to] increase RAF score," including for major depression, cachexia, protein-calorie malnutrition, morbid obesity, and chronic obstructive pulmonary disease.

Id. (¶ 54).

Id.

Id. at 20–21 (¶ 54).

Id. at 21 (¶ 54).

Id. (first brackets in complaint).

In August 2013, physicians in PAMF's network began receiving "daily alert" forms for the MA Plan patients on each physician's schedule that day. The daily alerts identified "what HCC codes have not yet been captured this year for the patient[s]." Those codes included not just previously diagnosed conditions but also conditions that data-mining software, using an algorithm, "suspect[ed]" the patient might have. "[T]he focus of the daily alerts was on pressuring physicians to increase RAF scores rather than on improving coding accuracy or meeting the clinical needs of patients."

Id. (¶ 55).

Id.

In addition to the daily alerts, each physician received a weekly list of MA Plan patients scheduled for appointments that week and a monthly report of MA Plan patients needing to schedule "Medicare Wellness Exams" by year's end. The purpose of the forms was to "aid in your capturing of chronic conditions." In response, physicians raised concerns about this pressure and asked that the messages from RAF coders be "nicer." Multiple physicians complained to Dr. Graham Dresden (a PAMF Physician Champion) about the "harshness of the messages" on the daily alerts and about how "they were offended by the messages and ... felt like the message was either confusing, fraudulent, excessive, etc." Dr. Dresden relayed this complaint to PAMF management.

Id. (¶ 56).

Id.

Id. at 34 (¶ 96) (ellipsis in complaint).

Id.

Also by August 2013, PAMF executives, including Suzy Cliff (PAMF's Vice President of Revenue Cycle) and Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion), received these daily alerts, the weekly lists, and the monthly reports in order to interact with and, if necessary, pressure PAMF physicians to increase diagnosis coding during the MA patient encounters. One approach was that a coder would review a physician's documentation after a patient encounter and identify any overlooked risk-adjusting diagnosis codes. The coder would tell the physician to confirm that the codes should be added to the patient's medical record. For example, in August 2013, Dr. Vahamaki developed "Dr. V's PCP [primary care physician] Audit Letter Template" to send to physicians with the message that "[t]he diagnostic coding team has added this code to your visit as an addendum ... Please email back to confirm that this patient has this diagnosis."

Id. at 21 (¶ 57).

Id.

Id. (ellipsis and first brackets in complaint).

3.1.4 Pre-populating patient medical records with diagnosis codes

Over time, Sutter and PAMF began to pre-populate medical records of physician-patient encounters with risk-adjusting diagnosis codes before physicians saw their patients. Sutter and PAMF pre-populated these codes regardless of the health conditions that the physicians managed, evaluated, assessed, or treated during their actual patient encounters.

Id. at 19 (¶ 52).

Id. at 19–20 (¶ 52).

Physicians expressed concern that risk-adjusting diagnoses appeared in patient medical records before they saw their patients. For example, Dr. Thomas Deetz told a PAMF auditor that "pre-populating diagnoses into his visit encounter is possibly fraud ... Does CMS know about what you all are doing?" Physicians also expressed concern that they did not know how to delete incorrect diagnoses from their patients' documentation. For example, Dr. Lisa Gervin told a PAMF auditor that she did not know how to delete an incorrect diagnosis code that a coder entered after her visit with a patient.

Id. at 19 (¶ 52).

Id. at 34 (¶ 96) (ellipsis in complaint).

Id. at 20 (¶ 52).

Id. at 34 (¶ 96).

3.1.5 Adding diagnosis codes as addendums to patient medical records

Sutter and PAMF used data mining to identify and send messages to physicians to find and refresh high-paying risk-adjusting diagnosis codes. Similarly, coders sent "queries" to remind physicians to ensure that all diagnosis codes were captured. Numerous physicians disliked this practice and felt "pressured" to add diagnosis codes that they did not believe were clinically accurate or relevant. For example, Dr. Joann Falkenburg expressed discomfort several times to Physician Champions Drs. Lin and Vahamaki, including (1) "I got two new HCC [daily] alerts today and have concerns about both of them," (2) "they [coders] suggested [a patient] get diagnosed with COPD [asthma ] based on a diagnosis in UC a year and a half ago ... I don't feel it is legitimate to code this," (3) "with my patient on hospice, there is something that seems unseemly about pursuing a new diagnosis of PVD [pulmonary vascular disease ] when she has weeks to live," (4) "it makes me feel a little fraudulent to be considering it," and (5) "I try to be pretty legitimate about how I diagnose, document and chart and want to avoid any possibility that it looks like I am working someone up just for the financial upside."

Id. at 20 (¶ 54).

Id.

Id. at 33 (¶ 96) (brackets and ellipsis in complaint).

Additionally, PAMF coders met one-on-one with physicians to discuss their diagnosis coding. During these meetings, the auditors at times encouraged the physicians to add addendums to their patient records and add risk-adjusting diagnosis codes.

Id. at 20 (¶ 54).

Id.

* * *

The government alleges that Sutter's and PAMF's RAF Campaign "achieved results." The RAF Campaign set a goal of capturing 80% of risk-adjusting diagnosis codes within its HCC focus areas: chronic obstructive pulmonary disease, diabetes, and vascular disease. For example, the "Diagnostic Coding Champions Meeting Minutes" for the August 13, 2013 meeting (attended by most of PAMF's senior executives) identified "gains" over the past two months in these "3 Key Areas" of coding and compared the performance among PAMF's four divisions. Additionally, Sutter data from early 2014 showed $4.4 million in year-end revenue gains from the RAF Campaign from 2012 to 2013. Furthermore, in March 2015, Sutter reported a "20% overall system wide increase" in the risk scores of MA Plan patients, including increases at all four PAMF divisions of between 15% and 23% resulting in an expected $4.173 million in additional Medicare reimbursements. 3.2 "Red Flags" Regarding Coding

Id. at 21 (¶ 58).

Id.

3.2.1 Audits by MA Organizations

As noted above, Sutter and its affiliates furnish healthcare services to MA Plans managed by three MA Organizations: UnitedHealth, Health Net, and Humana. UnitedHealth, Health Net, and Humana contractually require Sutter and its affiliates, including PAMF, to participate and cooperate in medical-chart reviews and audits.

See supra note 48.

Gov't Compl. – ECF No. 41 at 22 (¶ 59).

In October 2012, UnitedHealth conducted an audit focusing on so-called "outlier" risk-adjusting diagnosis codes that certain medical providers submitted much more frequently than the industry average among other large providers. UnitedHealth identified that Sutter and PAMF were outliers with respect to diagnosis codes mapping to the HCC model's code for heart attack (HCC 82). UnitedHealth reviewed the underlying medical charts and found that 27 of 30 of the patient records (90%) containing heart-attack diagnosis codes were erroneous, invalid, unsupported, or otherwise false. UnitedHealth conducted a subsequent audit and found that six of seven patient records (86%) containing heart-attack diagnosis codes similarly were false.

Id. (¶¶ 59–60).

Id. (¶ 60).

Id.

At a December 2012 "PAMF Coding and Compliance Committee" meeting that included Roger Larsen (PAMF's Chief Financial Officer and a Sutter Regional Vice President of Finance), Dr. Jeffrey Brown (PAMF's Associate Director for Managed Care), and Kris Anne Crow (PAMF's Director of Coding and Education), Mr. Larsen said that deleting unsupported codes in patient medical records based on the October 2012 audit results would have a "negative impact to our reimbursement." The Coding and Compliance Committee decided not to perform any follow-up audits to determine whether other medical records similarly contained false diagnosis codes. At the December 2012 meeting, Ms. Crow said that Sutter's and PAMF's coding department "presently does not have the bandwidth to support such an effort [to perform follow-up audits]." Sutter and PAMF did not assign any coders or auditors to perform follow-up audits.

Id. at 23 (¶ 61).

Id.

Id. (brackets in complaint).

Id.

In or shortly after April 2014, UnitedHealth and Health Net hired a consultant to review dates of service in 2012 and 2013. The review identified over 8,000 false diagnosis codes for MA Plan patients that Sutter and its affiliates needed to delete based on "overcod[ing]" and diagnoses "not supported in documentation." Sutter and PAMF executives learned of these results but did not direct any auditors to take remedial action to identify other false diagnosis codes.

Id. at 28 (¶ 78).

Id. (brackets in complaint).

Id.

UnitedHealth conducted additional audits for heart-attack diagnosis codes for dates of service in 2013, 2014, and 2015. At one PAMF location (Camino), UnitedHealth found that 28 of 30 patient records (93%) containing heart-attack diagnosis codes were erroneous, invalid, unsupported, or otherwise false. At another PAMF location (Mills-Peninsula), UnitedHealth found that three of four patient records (75%) containing heart-attack diagnosis codes were erroneous, invalid, unsupported, or otherwise false. Sutter and PAMF deleted the specific diagnosis codes that UnitedHealth's audits identified but refused to expand auditing of diagnosis codes (either of heart-attack diagnosis codes specifically or other risk-adjusting diagnosis codes more generally) beyond the specific codes that UnitedHealth reviewed.

Id. at 23 (¶ 62).

Id.

3.2.2 Lack of compliance and training programs

The government alleges that:

Sutter and PAMF lacked any effective compliance or training program related to diagnostic coding for its Medicare Part C program. While there was a PAMF coding and compliance committee, as noted above its members focused primarily on Sutter's RAF Campaign and little on audits examining the validity of the coding or other compliance efforts.

Id. (¶ 63).

In March 2013, Kris Anne Crow (PAMF's Director of Coding and Education) told Katie Borgstrom (PAMF's Interim Chief Operating Officer) that one of PAMF's divisions, Mills-Peninsula, "has never been audited and we have no idea what is going on there." Ms. Crow said that PAMF's coding and training group "had no credibility" with physicians and that "[h]istorically, the coding department has had no structure, no policies and really no accountability in terms of education provided and timely feedback" to the physicians in PAMF's network. Ms. Crow discussed these problems with Richard Slavin (PAMF's Chief Executive Officer), who agreed that PAMF must improve in these areas.

Id. at 23–24 (¶ 64).

Id. at 24 (¶ 64).

Id.

On May 6, 2013, Sutter hired relator Kathy Ormsby as PAMF's Risk-Adjustment Project Manager. In this position, Ms. Ormsby served as "the primary liaison between [the] coding, revenue cycle, quality & clinical departments with regards to the Medicare Advantage RAF/HCC coding initiative." Ms. Ormsby initially reported to Ms. Crow and at times to Suzy Cliff (PAMF's Vice President of Revenue Cycle).

Id. (¶ 65).

Id. (brackets in complaint).

Id.

Ms. Ormsby previously had earned a coding certification from the American Academy of Professional Coders. Ms. Ormsby also had experience with HCC codes, risk-adjusting diagnosis codes, diagnosis-coding standards, and Medicare Advantage compliance and training through her previous six-year employment at an MA Organization, where her responsibilities included training physicians on accurate coding, supervising risk-adjustment auditors, and helping to ensure compliance with Medicare rules and regulations relating to the Medicare Advantage program.

Id. (¶ 66).

Id.

Within her first few days on the job, Ms. Ormsby became aware that PAMF lacked a coding compliance or training program relating to Medicare Advantage. As she explained, "I was sent to a cube with nothing in it but an empty desk," "with absolutely no support, tools or guidance." Ms. Crow (Ms. Ormsby's supervisor) acknowledged that PAMF had no coding compliance or training program when Ms. Ormsby started at PAMF.

Id. (¶ 67).

Id.

Ms. Ormsby realized that the same problems existed systemwide at Sutter: there were no Sutter coding compliance manuals or training guides on diagnosis coding for physicians. Ms. Ormsby discussed the situation with Julie Cheung (Sutter's RAF Program Manager), who confirmed that Sutter had no compliance program regarding risk-adjustment diagnosis coding.

Id. (¶ 68).

Id.

3.2.3 Internal reviews and audits by Sutter and PAMF

Concerned about the lack of coding compliance and training, Ms. Ormsby personally conducted, within a few weeks of her hiring, a random diagnosis-coding audit of 42 patient encounters at PAMF occurring in the first two quarters of 2013. (The type of audit that Ms. Ormsby conducted was an "Encounter Audit" that evaluates one physician-patient encounter in a given year. Encounter Audits are useful to establish a baseline for coding accuracy. An Encounter Audit alone does not determine the extent of overpayments from CMS. )

Id. at 26 (¶ 71).

Id.

Ms. Ormsby completed this audit in early June 2013. Ms. Ormsby found that 53 of 62 risk-adjustment diagnosis codes (85%) from these patient encounters were false. Sutter and PAMF had submitted all of these codes for reimbursement, which raised the prospect of overpayments from CMS with respect to these patients.

Id. (¶ 72).

Id.

One month later, in July 2013, UnitedHealth sent Nancy McGinnis (Sutter's RAF Director) a letter "identf[ying] your practice as having submitted one or more HCCs at significantly higher rates than your peers," requesting supporting documentation, and noting UnitedHealth's engagement of a consulting firm to conduct a medical-chart review. Julie Cheung (Sutter's RAF Program Manager) forwarded the letter to Ms. Ormsby, among others. Ms. Ormsby responded to Sutter and PAMF executives that the letter "identifies [PAMF] as having some red flags and I want us to be compliant." Ms. Ormsby began lobbying for auditing support, saying that "[w]e really need to get on the ball with our potential HCC auditor[s]."

Id. (¶ 73).

Id.

Id. (brackets in complaint).

In light of her audit results, in August 2013, Ms. Ormsby created a "Corrective Action Plan." Her Plan called for hiring certified coders to perform audits and developing a compliance and training program to improve coding accuracy. Ms. Ormsby cited the 85% diagnosis-coding failure rate in her June 2013 audit and identified the "root cause" as PAMF's ineffective compliance and training. Her Plan explained that the June 2013 audit "confirmed that proper instruction for documentation requirements had not been communicated clearly to providers" and that PAMF "currently lacks a clearly defined procedure for auditing and provider feedback."

Id. (¶ 74).

Id.

Id. at 26–27 (¶ 74).

Ms. Ormsby's Plan called for two types of audits: (1) "Encounter Audits" and (2) "FOCUS Audits" that examine the error rates of several key HCCs (cancer, stroke, and fractures) that Ms. Ormsby understood from her prior experience are often miscoded and result in higher CMS payments. (Unlike Encounter Audits, FOCUS Audits examine diagnosis codes in PAMF patient medical records covering an entire calendar year and thus could be used in determining overpayments from CMS. )

Id. at 27 (¶ 75).

Id.

Ms. Ormsby gave Ms. Crow (her supervisor) a copy of her Plan. PAMF management above Ms. Crow approved the hiring of five certified coders to work as auditors. But PAMF management viewed audits as a tool to increase diagnosis coding— i.e., to submit more risk-adjusting diagnosis codes to CMS — rather than a tool for compliance.

Id.

In January 2014, Dr. Anita Gupta (a PAMF Physician Champion) identified "thousands" of "old, outdated and incorrect" diagnoses on the Problem Lists that place "[us] at risk of incorrectly coding them in a given year." Ms. Ormsby, Ms. Crow, and other PAMF executives, including Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion) and Drs. Amy Lin, Graham Dresden, Anita Gupta, and Susan Schaefer (PAMF's remaining Physician Champions), all learned of this problem.

Id. (¶ 76) (brackets in complaint).

Id.

By April 2014, Ms. Ormsby identified 185 risk-adjusting diagnosis codes that had been "incorrectly captured by providers and submitted for reimbursement" in the first quarter of 2013. Ms. Ormsby proposed to Ms. Crow and Ms. Cliff that they hire five more full-time coders to augment her team, explaining that with these additional coders, "[d]ocumentation across all of PAMF would be better supported to reach the requirements identified by CMS (Center for Medicare and Medicaid Services) and show a marked increase in compliance." Shortly after Ms. Ormsby made this proposal, Sutter and PAMF executives learned that the audit conducted by UnitedHealth's and Health Net's consultant (described above) had identified over 8,000 false diagnosis codes that Sutter and its affiliates needed to delete based on "overcod[ing]" and diagnoses "not supported in documentation."

Id. (¶ 77).

Id. (brackets in complaint).

See supra notes 141–143.

On June 3, 2014, Ms. Ormsby informed PAMF management that the preliminary results of her 2013 Encounter Audits found 1,082 false risk-adjusting diagnosis codes in 2,226 patient encounters audited. A few weeks later, on June 27, 2014, Ms. Ormsby informed Ms. Crow that the Physician Champions "have been training our providers on aortic arthrosclerosis and morbid obesity incorrectly." Ms. Ormsby suggested an audit to uncover the extent of the problem. PAMF did not authorize any audit.

Gov't Compl. – ECF No. 41 at 28 (¶ 79).

Id. (¶ 80).

Id.

In late July 2014, Ms. Ormsby identified (to PAMF management) false diagnosis codes relating to a specific Medicare Advantage beneficiary (identified as "Patient A") dating back to 2010. In 2010, PAMF had submitted a risk-adjusting diagnosis code for prostate cancer for Patient A. UnitedHealth asked for medical records that supported that diagnosis code. Ms. Ormsby pulled Patient A's medical records and found nothing in them to support the prostate-cancer code for 2010. Ms. Ormsby gave Patient A's medical records to UnitedHealth and brought the issue to the attention of Ms. Crow (her supervisor) and Ms. Cheung (Sutter's RAF Program Manager). Ms. Crow asked Ms. Ormsby to calculate the potential reimbursements to CMS from the false coding related to Patient A, in light of UnitedHealth's understanding that CMS took the position that "if one HCC failed in audit, [CMS] could assume that for every patient in the plan that submitted the same HCC, [CMS] can ask for the payment back." Ms. Ormsby identified 484 codes for prostate cancer submitted for payment in 2010 and estimated the potential reimbursement at $1.936 million, which she said was "probably low." Ms. Cheung reprimanded Ms. Ormsby for giving Patient A's medical records to UnitedHealth and told her to never do that again and instead to send patient medical records to her.

Id. (¶ 81).

Id.

Id. at 29 (¶ 82) (brackets in complaint).

Id.

Id. at 28 (¶ 81).

In July 2014, Ms. Ormsby performed another Encounter Audit like the one she conducted in June 2013. Ms. Ormsby reviewed 20 physician-patient encounters covering March 2013. She found false risk-adjusting diagnosis codes in 18 of 20 encounters (90%).

Id. at 29 (¶ 83).

Id.

Throughout 2014, Ms. Ormsby circulated to Sutter and PAMF management — including Roger Larsen (PAMF's Chief Financial Officer and a Sutter Regional Vice President of Finance), Suzy Cliff (PAMF's Vice President of Revenue Cycle), and Ms. Cheung — a "RAF Dashboard" summarizing various data metrics, including (1) Sutter's and PAMF's Medicare Advantage patient population and how their average risk scores compared to state and national benchmarks, (2) a "prevalence" rate identifying the percentage of MA Plan patients assigned certain lucrative diagnosis codes, and (3) the "HCC Recapture Rate" and "HCC Score Comparison" designed to track the performance of "acuity capture and reporting" by PAMF physicians. The RAF Dashboard detailed hundreds of erroneous, invalid, unsupported, or otherwise false diagnosis codes identified during Ms. Ormsby's Encounter and FOCUS Audits.

Id. at 15 (¶ 40), 30 (¶ 88).

Id. at 30 (¶ 88).

In December 2014, Ms. Ormsby and her audit team memorialized the final tally of the 2013 FOCUS Audit of risk-adjusting diagnosis codes for cancer, stroke, and fracture. For cancer, 164 of 182 patient records (90%) were erroneous, invalid, unsupported, or otherwise false. For stroke, 162 of 169 patient records (96%) were erroneous, invalid, unsupported, or otherwise false. For fracture, 57 of 86 patient records (66%) were erroneous, invalid, unsupported, or otherwise false.

Id. at 29 (¶ 84).

Id.

On December 16, 2014, Ms. Ormsby met with Marcella Alaniz (a PAMF Compliance Analyst) to discuss the results of the 2013 FOCUS Audit. Ms. Ormsby also discussed the results with Jessica Driver-Zuniga (Sutter's lead RAF/HCC coder). Three days later, on December 19, 2014, Ms. Ormsby widely distributed the results of the 2013 FOCUS audit to PAMF senior management, flagging the failure rates of 90% for cancer, 96% for stroke, and 64% for fracture.

Id. (¶ 85).

Id. at 29–30 (¶ 85).

Id. at 30 (¶ 86). While 57 of 86 is approximately 66%, the complaint alleges that Ms. Ormsby flagged for PAMF senior management a failure rate for fracture of 64%. Id.

Also on December 19, 2014, Ms. Ormsby notified PAMF senior management that there were additional false risk-adjusting diagnosis codes that had been submitted to CMS and that required returning overpayments to CMS. Ms. Ormsby wrote, "[w]e have identified 94 encounters that have been submitted to CMS without supporting documentation for HCC conditions billed" from PAMF physicians and that she "expect[ed] this number to increase daily until a resolution can be implemented."

Id.

Id. (second brackets in complaint).

The 2013 FOCUS Audit further showed that over 3,500 physician-patient encounters from 2013 remained unreviewed. Ms. Ormsby identified the 2013 FOCUS Audit as a high-priority compliance issue that encompassed over 7,500 physician-patient encounters in one year of service. Ms. Ormsby's auditing team had the capacity to review only a small percentage of the physician-patient encounters with cancer, stroke, or fracture diagnosis codes for dates of service in 2013. Ms. Ormsby raised (in writing to Ms. Cliff) her concern that no other dates of service had been reviewed.

Id. (¶ 87).

Id. (¶ 86).

Id. (¶ 87).

Id.

In early January 2015, Ms. Ormsby raised with Sutter and PAMF management, including Mses. Cheung and Cliff, an issue concerning "misleading labels" for stroke in Sutter's electronic medical-record system. Ms. Ormsby said (and attached a screenshot showing) that Sutter's labels stated that various types of stroke were considered acute and carried an HCC label as long as the stroke took place within eight weeks of the physician-patient encounter. Ms. Ormsby explained that this diagnosis code should not be captured after a patient is discharged from the hospital in an inpatient setting (much less eight weeks after discharge) and that "[t]he labels are causing providers to capture the incorrect ICD-9 [diagnosis] codes and we are being reimbursed inappropriately." Ms. Ormsby asked to remove the words "8 weeks" from the label. Ms. Cheung responded that "[t]he Compliance Reimbursement Team hasn't yet weighed in." Ms. Cheung stated that these misleading electronic medical-record labels were a systemwide problem at all of Sutter's affiliates, not just PAMF, and that a change in labeling "won't be made for just one organization."

Id. (¶ 89).

Id.

Id. at 30–31 (¶ 89) (first brackets in complaint).

Id. at 31 (¶ 89).

Id. (brackets in complaint).

Id.

On January 21, 2015, Ms. Ormsby met with Ms. Cliff, and on January 22, 2015, Ms. Ormsby sent a follow-up email to PAMF management — including Ms. Cliff, Dr. Veko Vahamaki, (PAMF's Lead RAF Physician Champion), Debbie Troklus (PAMF's Compliance Director), Marcella Alaniz (a PAMF Compliance Analyst), and Dr. Criss Morikawa (a PAMF executive) — "reiterat[ing her] concerns" regarding five key HCC-coding compliance problems:

1. Accuracy rates of cancer, fracture and stroke (2013 dates of service and beyond);

2. Concerns regarding the payments received without supporting documentation (60 day window);

3. Encounters by providers who are no longer at PAMF and have unsupported HCC submissions;

4. Providers who are not responding to staff messages regarding specificity and clarification for HCC's submitted to CMS;

5. Discontinued use of the auditing billing notes/corrections to rectify unsupported ICD-9 [diagnosis codes].

Id. at 39 (¶ 113).

Ms. Ormsby also notified management that the monthly electronic medical-record "sweeps" of risk-adjusting diagnosis codes from physician-patient encounters would lead to Sutter's and PAMF's submitting many false codes to the MA Organizations and then to CMS. Jessica Driver-Zuniga (Sutter's lead RAF/HCC coder) acknowledged in a February 24, 2015 memo that unsupported diagnosis codes that Sutter and PAMF knew were invalid were being "swept" into their reimbursement system.

Id. at 39–40 (¶ 114).

Id. at 40 (¶ 114).

In March 2015, Ms. Ormsby updated her Encounter and FOCUS Audit results. The updated 2013 FOCUS Audit results showed Ms. Ormsby's and her auditing team's deletion of 1,001 false diagnosis codes that had been submitted to CMS for reimbursement. The 2013 and 2014 Encounter Audit results showed deletion of 777 false diagnosis codes in 2013 and 517 false diagnosis codes in 2014. Ms. Ormsby's new supervisor, Christian Gabriel (hired by Sutter and PAMF in 2015 ) reviewed Ms. Ormsby's data and asked Ms. Ormsby to calculate how much CMS overpaid Sutter and PAMF for these false codes. Ms. Ormsby estimated overpayments at approximately $4.2 million. She explained that this likely was the tip of the iceberg: 3,844 physician-patient encounters remained unaudited for dates of service in 2013 alone, and thousands of later physician-patient encounters were never audited.

Id. at 31 (¶ 90).

Id.

Id. at 25 (¶ 70), 31 (¶ 90), 40 (¶ 117).

Id. at 31 (¶ 90).

Id.

3.3 Knowingly Ignoring Red Flags and Actual Notice of False Claims

The government alleges that:

Sutter and PAMF management knew about the ineffective compliance and training that would inevitably result in substantial false coding. They also knew about the internal and external audits highlighting years of substantial false coding at PAMF. Instead of addressing these problems, Sutter and PAMF management continued to engage in the RAF Campaign and encouraged aggressive diagnosis coding, resulting in the submission of false codes and inflated Medicare reimbursements.

Before Ormsby's arrival at PAMF in May 2013, Sutter and PAMF management recklessly disregarded and were deliberately indifferent to problems of false diagnosis coding, with few attempts made to audit or otherwise identify such problems even in the face of the high failure results of audits and chart reviews by UHG [UnitedHealth], Peak [UnitedHealth's consultant] and Optum [UnitedHealth's affiliate]. Indeed, the RAF Campaign itself, with the goal of increasing lucrative diagnosis coding, highlighted Sutter and PAMF's focus on Part C profits over compliance....

In addition to the frequent complaints from physicians, after Ormsby's arrival Sutter and PAMF management received actual notice from Ormsby and her audit team about rampant false diagnosis coding and ineffective compliance and training. Initially, Sutter and PAMF management ignored her and continued the RAF Campaign unabated. However, as Ormsby and her auditing team deleted false diagnosis codes that mapped to HCCs and negatively impacted the reimbursement from CMS, Sutter and PAMF management took steps to impede her efforts and stop her ability to delete false codes.

Id. at 32 (¶¶ 91–92), 34 (¶ 97).

For example, on September 19, 2014, Ms. Ormsby attended a PAMF executive meeting with Roger Larsen (PAMF's Chief Financial Officer and a Sutter Regional Vice President of Finance), Suzy Cliff (PAMF's Vice President of Revenue Cycle), Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion), and many others. Ms. Ormsby delivered a PowerPoint presentation summarizing the results, to date, of the Encounter and FOCUS Audits. The results identified 6,082 false risk-adjusting diagnosis codes in 12,220 physician-patient encounters. Ms. Ormsby made compliance and training recommendations at the meeting, including proposing "Diagnosis Champions" to work alongside the Physician Champions and to train physicians in PAMF's network about proper coding and other compliance issues. Several days after this meeting, Mr. Larsen wrote an email to Dr. Vahamaki, Ms. Cliff, and Dr. Michael Conroy (PAMF's Chief Medical Officer) discussing increasing reimbursements from CMS. Among other things, Mr. Larsen wrote, "We are now over a year into the HCC improvement effort [i.e., the formal RAF Campaign] and I see that we are making some limited progress but are behind where we could be and will likely not achieve more than a modest improvement if we continue as is. Given our existing efforts and the general mindset of the physicians[,] I predict we will achieve at most a 10% improvement" and that "I think we may need to ask the Board to reconsider implementing a physician compensation incentive along with a refocus on the other key parts of the plan to effectively change the PCP [primary-care physician] culture necessary for HCC success." Dr. Vahamaki forwarded Mr. Larsen's email to Julie Cheung (Sutter's RAF Program Manager) and Nancy McGinnis (Sutter's RAF Director) and agreed with Mr. Larsen's "strategic" focus on profits. Dr. Vahamaki identified one barrier as Ms. Ormsby's and the coding department's focus on compliance.

Id. at 35 (¶ 103).

Id.

Id. at 35–36 (¶ 103).

Id. at 36 (¶ 104).

Id. (first brackets in complaint).

Id. at 36–37 (¶ 105).

Id. at 37 (¶ 105).

On October 15, 2014, Ms. Cliff wrote an email to Ms. Ormsby titled "HCC Coding Corrections" instructing Ms. Ormsby: "[p]er our conversation this morning, please remind your team to stop performing any charge corrections on accounts until we can map out the downstream [e]ffects."

Id. (¶ 106) (brackets in complaint).

On November 12, 2014, Dr. Criss Morikawa (a PAMF executive) distributed to PAMF management and Ms. Ormsby's auditing team an email describing PAMF's new policy prohibiting "submitting charge corrections to payors (esp. Medicare) more than 30 days after date of service." Dr. Edward Yu (PAMF's Medical Director) asked, "What happens if the incorrect diagnosis code puts us at risk of [M]edicare fines for inaccurate coding?" Dr. Morikawa responded that due to the "close to a million charge transactions" that PAMF submits every month, "it is not scalable to hold and review every encounter — even on say all [M]edicare."

Id. (¶ 107).

Id. (brackets in complaint).

On November 26, 2014, Ms. Ormsby attended a meeting with Marcella Alaniz and Jessica Lin (PAMF Compliance Analysts), among others. Ms. Alaniz said that Ms. Ormsby lacked authority to delete false diagnosis codes from physician-patient encounters in the electronic medical records that were submitted to MA Organizations and, in turn, to CMS. Ms. Ormsby tried to reverse the directive to stop her auditing team from deleting false diagnosis codes. On December 1, 2014, Ms. Ormsby warned PAMF management — including Ms. Alaniz, Ms. Cliff, Dr. Morikawa, and Debbie Troklus (PAMF's Compliance Director) — that "I don't think this recommendation is a compliant solution." The next day, on December 2, 2014, Ms. Ormsby warned PAMF management about the importance to Medicare of proper diagnosis coding and Medicare's upcoming focus on coding compliance. On December 10, 2014, Ms. Ormsby wrote an email to Dr. Conroy, Dr. Vahamaki, and Ms. Cliff stating, "I am very concerned about the large number of non-compliant chronic HCC conditions that have been submitted to the health plans for reimbursement." Ms. Ormsby emphasized that MA Organizations would be auditing diagnosis coding. Ms. Ormsby also discussed her concerns about Sutter's and PAMF's requiring physicians (as opposed to coders) to delete unsupported codes: she explained that, based on her experience, most physicians will not delete unsupported codes due to time pressure and inattention (and that some physicians had left PAMF and thus could not delete unsupported codes in any event). Ms. Cliff subsequently chastised Ms. Ormsby for escalating the issue to Dr. Conroy.

Id . at 38 (¶ 108).

Id. More specifically, Ms. Ormsby had authority to make changes on the "billing side" of the electronic medical record (which does not affect claims to or payments from CMS) but was told that she did not have authority to delete diagnosis codes from the "encounter side" of the electronic record (the "side" that is submitted to CMS for payment). Id.

Id. (¶ 109).

Id.

Id. (¶ 110).

Id.

Sutter and PAMF continued the RAF Campaign without implementing any of Ms. Ormsby's recommendations. In December 2014, Dr. Susan Schaefer (Sutter's Regional Physician Champion for Diagnostic Coding) allegedly increased coding pressure on physicians because "as we approach end of year [we] are trying to maximize capture of HCC's." In another widely distributed email in December 2014 discussing a coding issue, Dr. Schaefer wrote, "I am not sending this to Kathy [Ormsby] as we know what happens."

Id. at 39 (¶ 111).

Id. (brackets in complaint).

Id. (¶ 112).

As noted above, in February 2015, Sutter and PAMF hired a new supervisor, Christian Gabriel, to oversee Ms. Ormsby. Initially, Ms. Ormsby emphasized to Mr. Gabriel the importance of coding-compliance issues. Ms. Ormsby presented him with a recent CMS presentation highlighting the importance of appropriate HCC coding and documentation and identifying the Medicare rules that related to coding and documentation. Ms. Ormsby also gave Mr. Gabriel a self-assessment (relating to PAMF's MA Plan program) describing the need for more auditors and compliance.

See supra note 229.

Gov't Compl. – ECF No. 41 at 40 (¶ 117).

Id.

When Ms. Ormsby tried to raise compliance issues, Mr. Gabriel told her to discuss any "differences in private" with him rather than via emails that included the auditing team. On March 9, 2015, Mr. Gabriel issued a "verbal warning" to Ms. Ormsby based on misgivings that Ms. Ormsby had expressed to him and her auditing team about the RAF Campaign.

Id. (¶ 118).

Id.

Mr. Gabriel told Ms. Ormsby that the audit "team, structure and process is my #1 focus," "[g]iven the lack of progress in improving our RAF/HCC scores." Similarly, at a "Strategy Meeting" in mid-March 2015, Mr. Gabriel stressed that the new focus of Ms. Ormsby's auditing was to "rais[e] the RAF score." A few days later, on March 18, 2015, Mr. Gabriel stressed this new focus in a three-hour meeting (with the auditing team) that Ms. Ormsby could not attend. Mr. Gabriel wrote in a follow-up email to the auditing team that he had "dropped a ‘bomb’ on you in terms of a new initiative ...." Ellie Kamkar (Manager of Coding, Training, and Auditing at PAMF) reported to Ms. Ormsby that Mr. Gabriel had directed the auditing team "to take off the compliance hat and put on the revenue hat" based on directives from senior management. Ms. Kamkar also believed that Mr. Gabriel "was asking her to teach physicians how to up code."

Id. (brackets in complaint).

Id. at 41 (¶ 119).

Id.

Id. (ellipsis in complaint).

Id.

A week later, on March 25, 2015, Mr. Gabriel had a meeting with Ms. Ormsby where he told her that "[w]e need to audit to raise [RAF] score[s]." Mr. Gabriel directed Ms. Ormsby and the auditing team to conduct more data-mining audits that would "support leadership's directives for this year" for the RAF Campaign. On March 26, 2015, Mr. Gabriel wrote an email to Ms. Ormsby and the auditing team where he explained that the goal of the audits was to increase Medicare reimbursement. Mr. Gabriel identified "our overall goals" as:

Id. (¶ 120) (brackets in complaint).

Id.

• Identification of new HCC's

• Decreasing the # of patients without HCC's

• Maintain/improve HCC capture rate for 2015

• Improve RAF/HCC scores through several techniques

• Data-mining for HCC pockets of opportunities

• Focus on providers that have a large volume of HCC eligible patients and target for review

Id. (bullet points in original email).

The audits targeted certain conditions (diabetes with manifestations, thrombocytopenia, pulmonary vascular disease, chronic kidney disease, major depressive disorder, and pathological fractures ) that the government alleges are all "lucrative" risk-adjusting diagnosis codes that drive up the reimbursement amounts that CMS pays out.

Id. at 41–42 (¶ 120).

Mr. Gabriel thereafter instituted additional policies to increase RAF scores. He required that — before physicians actually saw their patients — coders pre-populate patients' medical records with risk-adjusting diagnosis codes that the coders suspected (but did not know) might be applicable to the patient. Sutter's and PAMF's systems swept these pre-populated codes into the electronic medical record and submitted them through MA Organizations to CMS for payment unless the physician affirmatively deleted the codes. (Sutter's and PAMF's policies prevented coders from deleting codes: only physicians could do so. ) The government alleges that by prohibiting coders from deleting false risk-adjusting diagnosis codes and having coders add diagnosis codes that physicians had not reported or verified in the electronic medical record, Sutter and PAMF knowingly pursued policies designed to yield inflated reimbursements through the over-reporting of diagnosis codes. At a March 27, 2015 strategy meeting involving Arvin Magusara (a Sutter Senior Analyst), Mr. Gabriel, and Ms. Ormsby, Julie Cheung (Sutter's RAF Program Manager) stated that increasing MA Plan patient risk scores "had been a concern for several years" among the RAF Steering Committee, which included Dr. Jeffrey Burnich (a Sutter Senior Vice President and Executive Officer), Nancy McGinnis (Sutter's RAF Director), Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion), and Ms. Cheung. Ms. Cheung said that false-coding problems remained and that "CMS is still receiving HCC's that we know are not correct." No follow-up discussions took place at the meeting, and no efforts were taken to correct this problem.

Id. at 42 (¶ 121).

Id.

Id. As discussed above, coders could delete diagnosis codes from the "billing side" of the electronic medical records, but this would not prevent the codes from being swept to CMS for payment. See supra note 247; accord Gov't Compl. – ECF No. 41 at 42 (¶ 121).

Gov't Compl. – ECF No. 41 at 42 (¶ 121).

Id. (¶ 122).

Id.

Several days later, at a March 31, 2015 "Champions Meeting" that included Drs. Vahamaki, Graham Dresden, Anita Gupta, Amy Lin (PAMF's Physician Champions), Mr. Gabriel, and Ms. Ormsby, Mr. Gabriel explained that each PAMF division had met the prior week and (according to the notes of the March 31 meeting) discussed "[h]ow both the physician and the auditor could work together to identify areas to increase the RAF scores for each division." During the meeting, Ms. Ormsby told management, "I want to go on the record saying that I do not agree with any auditor reviewing/auditing in search of reimbursement. I don't believe that this is a compliant practice." Ms. Ormsby also complained about management stopping the Encounter and FOCUS Audits. Mr. Gabriel "interrupted" her and said that (1) "we are not doing any encounter audits this year," (2) PAMF's compliance department (not the coding department) would focus on compliance, and (3) "[u]nfortunately our compliance department does not have the bandwidth to investigate compliance concerns" related to coding.

Id. at 42–43 (¶ 123) (brackets in complaint).

Id. at 43 (¶ 123).

Id.

Id. (brackets in complaint).

On May 7, 2015, Ms. Ormsby left PAMF for another position.

Id. at 7 (¶ 13).

Sutter did not resume using internal auditing to find and delete erroneous, invalid, unsupported, or otherwise false diagnosis codes until the second quarter of 2016. At that time, the Office of Patient Experience began an audit to attempt to establish an accuracy baseline for physician-patient encounters for dates of service in 2015 that resulted in diagnoses of stroke and heart attack. The audit randomly selected a sample of Medicare Advantage beneficiaries from each Sutter affiliate, including PAMF, and reviewed patient encounters by primary-care physicians and specialists. The audit reviewed the medical records for 38 beneficiaries and found that the accuracy rate at PAMF for heart-attack diagnosis codes was 39.29% and the accuracy rate for stroke diagnosis codes was 22.22%. The accuracy rate at PAMF, when combined with the Mills Peninsula Division of PAMF and the Mills Peninsula Medical Group (a provider affiliated with PAMF), for stroke diagnosis codes was 10.87%. A similar review of 206 Medicare Advantage beneficiaries throughout Sutter system-wide showed that the accuracy rate for heart-attack diagnosis codes was 53.7%, and the accuracy rate for stroke diagnosis codes was 22.8%.

Id. at 43 (¶ 124).

Id.

Sutter and PAMF knew that they had to report and return overpayments to CMS. Sutter's written policies stated, "Overpayment Refund, 13-540 ... POLICY [:] Sutter Health and its Affiliates will report and refund overpayments from state and federal health care programs within 60 days of identification, or the due date for any applicable reconciliation" and required that "[a]s appropriate, Sutter Health and its affiliates will take remedial steps to prevent identified overpayments from recurring." Sutter's policy defined "Overpayment " to include "incorrect code or modifier assignment resulting in a higher level of reimbursement, insufficient or lack of documentation to support billed services[,] ... lack of medical necessity, ... or any other finding that reflects [that] an overpayment was received as a result of inaccurate or improper coding or reporting of healthcare items or services."

Id. at 15 (¶ 42).

Id. at 15–16 (¶ 42) (emphasis in policy, brackets and ellipsis in complaint).

Id. at 16 (¶ 42) (emphasis in policy, ellipsis in complaint).

3.4 Examples of False Diagnosis Codes and Estimate of Total Overpayments

The government alleges:

During the period from January 2010 through December 31, 2016, Sutter and PAMF, through their unlawful conduct discussed ... above, knowingly caused the submission of thousands of erroneous, invalid, unsupported or otherwise false risk-adjusting diagnosis codes to CMS for tens of thousands of Medicare Advantage beneficiaries at PAMF. The MA beneficiary population at PAMF Mills Peninsula tallied approximately 28,000 over those six years, while PAMF served approximately 74,000 MA beneficiaries during that period. These false claims inflated CMS's reimbursements by tens of millions of dollars.

Sutter and PAMF knew that they were required to submit accurate diagnosis data to the MA Plans and delete erroneous, invalid, unsupported or otherwise false diagnoses. Sutter and PAMF were also on notice from Ormsby and her audit team, as well as from other audits and chart reviews, of thousands of such coding problems. Yet, Sutter and PAMF knowingly disregarded that information and failed to investigate the prevalence of this miscoding or delete these codes. Instead, they knowingly retained the resulting overpayments.

Id. at 45–46 (¶¶ 131–32).

The government cites ten examples of false claims that Sutter and PAMF submitted through MA Organizations to CMS:

1. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient A" for prostate cancer dating back to a physician-patient encounter in 2010. Nothing in Patient A's medical

Id. at 28 (¶ 81).

record supported the prostate-cancer code for that year.

2. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient B" for "malignant neoplasm of thyroid gland" for a date of service in 2012. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient B's Medicare Advantage coverage. Patient B's medical records show that this diagnosis code was false because (1) Patient B's thyroid cancer was treated by thyroidectomy in 2007, (2) the cancer had not recurred, and (3) there was no evidence of treatment, evaluation, or management of thyroid cancer in 2012 in Patient B's medical records.

3. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient C" for "malignant melanoma of skin of scalp and neck" for a date of service in 2012. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient C's Medicare Advantage coverage. Patient C's medical records show that this diagnosis code was false because (1) Patient C had last been treated for malignant skin cancer in 2006 and (2) there was no evidence of treatment, evaluation, or management of skin cancer in 2012 in Patient C's medical records.

4. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient D" for stroke (specifically, "cerebral artery occlusion, unspecified, with cerebral infarction") for a date of service in 2014. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient D's Medicare Advantage coverage. Patient D's medical records show that this diagnosis code was false because (1) Patient D's past medical history reflected a cerebellar infarction, a transient rather than a chronic medical condition that does not map to any HCC code, (2) the cerebellar infarction took place in 1990 without further recurrence, and (3) no stroke or cerebrovascular-accident event was noted in 2014 in Patient D's medical records.

5. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient E" for stroke (specifically, "cerebral artery occlusion, unspecified, with cerebral infarction") for a date of service in 2014. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient E's Medicare Advantage coverage. Patient E's medical records show that this diagnosis code was false because (1) Patient E's diagnostic test of magnetic resonance angiography reflected normal results without signs of a

Id.

Id. at 46 (¶ 133).

Id.

Id. at 46–47 (¶ 133).

Id. at 47 (¶ 133).

Id.

stroke and (2) no stroke or cerebrovascular-accident event was noted in 2014 in Patient E's medical records. A billing note said that the diagnosis code should be corrected to a "history of cerebrovascular accident" but said that the physician's original code is "associated with an order and cannot be removed."

6. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient F" for stroke (specifically, "cerebral artery occlusion, unspecified, with cerebral infarction") for a date of service in 2013. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient F's Medicare Advantage coverage. Patient F's medical records show that this diagnosis code was false because (1) Patient F was admitted in 2011 for a stroke, (2) no stroke had recurred, (3) Patient F had been on an anticoagulant long-term since her stroke in 2011, and (4) there was no evidence of treatment, evaluation, or management of a stroke event or cerebrovascular accident in 2013 in Patient F's medical records.

7. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient G" for "cerebral embolism with cerebral infarction; cerebral artery occlusion, unspecified, with cerebral infarction" for a date of service in 2013. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient G's Medicare Advantage coverage. Patient G's medical records show that this diagnosis code was false because (1) no stroke or cerebrovascular-accident event was noted in 2014 [sic] in Patient G's medical records, (2) Patient G's medical records show a history of a cerebrovascular accident taking place in 2004 for which the patient subsequently underwent rehabilitation, and (3) there was no evidence of treatment, evaluation, or management of a cerebrovascular accident in 2012 [sic] in Patient G's medical records.

8. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient H" for hip/femur fracture, specifically, "traumatic fracture of the mid-cervical section and unspecified part of the neck of the femur" for a date of service in 2014. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient H's Medicare Advantage coverage. Patient H's medical records show that this diagnosis code was false because (1) there was no evidence of treatment, evaluation, or management of an acute hip or femur fracture in 2014 in Patient H's medical records, (2) Patient H was last treated for fracture of the right hip in 2011, and (3) Patient H received a total hip replacement in December 2011.

Id.

Id. at 47–48 (¶ 133).

Id. at 48 (¶ 133).

Id.

9. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient I" for hip/femur fracture, specifically, "traumatic fracture of the pelvis" for a date of service in 2014. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient I's Medicare Advantage coverage. Patient I's medical records show that this diagnosis code was false because (1) there was no evidence of treatment, evaluation, or management of an acute hip or femur fracture in 2014 in Patient I's medical records, (2) Patient I's fracture occurred in 2013, and (3) Patient I's medical records document only a history of pelvic fracture (as opposed to actual fracture) for 2014.

10. Sutter and PAMF submitted a risk-adjusting diagnosis code for "Patient J" for "benign neoplasm of the brain for cerebral meninges" for a date of service in 2014. Sutter's and PAMF's submission of this diagnosis code caused CMS to pay out more money for Patient J's Medicare Advantage coverage. Patient J's medical records show that this diagnosis code was false because (1) Patient J's last brain MRI was in 2012, (2) a 2014 physician-patient encounter noted a history of a "small meningioma" and did not note any recurrence of a brain neoplasm, and (3) there was no evidence of treatment, evaluation, or management of brain cancer in 2014 in Patient J's medical records.

Id.

Id. at 48–49 (¶ 133).

Id. at 49 (¶ 133).

Id.

4. The Relator's Allegations Against Sutter Regarding Its Non-PAMF Affiliates

Ms. Ormsby's First Amended Complaint is similar to the government's complaint regarding PAMF but includes additional allegations about Sutter and its affiliates other than PAMF.

Sutter has two operating units (consolidated from a former five-region structure): a Bay Area operating unit and a Valley operating unit. Sutter's Bay Area operating unit includes one medical-foundation corporation: Sutter Bay Medical Foundation ("Sutter Bay"). Sutter Bay does business as PAMF, Sutter East Bay Medical Foundation, and Sutter Pacific Medical Foundation. Sutter's Valley operating unit includes one medical-foundation corporation: Sutter Valley Medical Foundation ("Sutter Valley"). Sutter Valley does business as Sutter Medical Foundation and Sutter Gould Medical Foundation. Sutter is the sole member of Sutter Bay and Sutter Valley.

Relator FAC – ECF No. 52 at 5 (¶ 11).

Id.

Id. at 5–6 (¶ 11).

Id. at 6 (¶ 11).

Id.

Sutter has two other affiliates, Sutter Connect, LLC, and Sutter Medical Network. Sutter Connect does business as Sutter Physician Services and supports Sutter's various medical foundations with services such as administration, billing, managed-care management, financial-management reporting, and provider relations. Sutter Medical Network is the network of the approximately 5,500 physicians at Sutter's affiliated hospitals and foundations.

Id. (¶¶ 12–13).

Id. (¶ 12).

Id. (¶ 13).

Sutter's medical foundations all use an electronic-medical-record system called EpicCare, and Sutter Physician Services sends out the diagnosis-code records in this system to the MA Organizations, which submit the codes to CMS.

Id. at 16 (¶ 45), 47 (¶ 149); see also id. at 31 (¶ 99) (one individual at Sutter Physician Services performed Medicare Advantage submissions Sutter-wide).

When Ms. Ormsby started working at PAMF in May 2013, she searched on Sutter's intranet for relevant RAF (Risk Adjustment Factor, i.e., risk score) policies and procedures at Sutter's other affiliates. Other than Sutter's Overpayment Policy, Ms. Ormsby found no policies or procedures relevant to a RAF program. Ms. Ormsby asked her peers at Sutter's other affiliates and the Physician Champions, but none could point her to any relevant materials. Thus, Ms. Ormsby alleges, as of May 2013, there was no formalized support at Sutter for the Medicare Advantage program (which had approximately 48,000 patients Sutter-wide).

Id. at 33 (¶ 104).

Id. at 19–20 (¶ 60); see also supra notes 294–296.

Relator FAC – ECF No. 52 at 33 (¶ 104).

Id. at 33 (¶ 104), 35 (¶¶ 109–10).

Id. at 33 (¶ 104), 35–36 (¶ 111).

While she was at PAMF, Ms. Ormsby participated in Sutter's RAF Coder User Group, which was made up of individuals at all Sutter affiliates doing jobs similar to Ms. Ormsby's at PAMF, i.e., supporting Sutter's RAF program. The RAF Coder User Group held monthly calls over WebEx and met in person quarterly. The purpose of the calls and meetings was to keep employees across Sutter's affiliates up to date on Sutter's systemwide effort to increase beneficiary RAFs (i.e., beneficiary risk scores). For example, each meeting included time for a "Round Robin" discussion, where Physician Champions or RAF employees from each affiliate would share what they were doing to increase RAF scores. At one meeting, the Round Robin was subtitled "Proactive Coding Strategies." Before the meeting, Sutter had collected objections from physicians to Sutter's RAF Campaign, such as "I don't see the purpose of doing annual wellness visits. I know it doesn't extend life" and "I know what RAF means — Revenue for Sutter at My Expense!" At the meeting, Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion) coached coders on how to overcome objections from physicians. Following these types of exchanges, Sutter circulated strategies through the RAF portal so that anyone supporting the RAF Campaign had access to the tools that affiliates used to raise RAF scores. From January 2013 to January 2015, Sutter increased its average HCC coding by 21% across all affiliates. From January 2014 to January 2015, Sutter had a systemwide increase of 25% in its RAF scores. When Sutter circulated the preliminary numbers reporting the 2014–2015 increase, Dr. Vahamaki forwarded them to Ms. Ormsby with his observation that Sutter's RAF Campaign to increase RAF scores was producing results.

Id. at 25 (¶ 75).

Id. (¶ 76).

Id. at 26 (¶ 76).

Id. (¶ 78).

Id.

Id. (¶ 79).

Id.

Id. (¶ 78).

Id. at 28 (¶ 85).

Id.

Ms. Ormsby repeatedly urged Sutter management, the Physician Champions, and members of the RAF Coders User Group to understand that Medicare required compliance and that Sutter's focus on raising RAF scores was not proper. Sutter did not implement Ms. Ormsby's suggestions. Julie Cheung (Sutter's RAF Program Manager) repeatedly confirmed that Ms. Ormsby was the only person at Sutter conducting audits to evaluate whether Sutter's RAF Campaign was generating unsupported diagnosis codes that resulted in overpayments from CMS.

Id. at 27 (¶ 83).

Id.

Id. (¶ 84).

As noted above, PAMF management agreed to hire five additional coders to audit risk-adjustment data in PAMF's Medicare Advantage program. Sutter did not authorize additional resources to expand audits Sutter-wide, even though Sutter's other affiliates needed help training their physicians on accurate coding.

Id. at 34 (¶ 107); see also supra note 180.

Relator FAC – ECF No. 52 at 34–35 (¶ 107). For example, in 2015, a coder from Sutter Gould "[a]cknowledged that they need to get out of the office to support the clinicians, but this is a challenge with just two of them to support 100 physicians." Id. at 35 (¶ 107).

In February 2014, following UnitedHealth's request for medical records for Patient A in connection with an audit of diagnosis codes, Ms. Ormsby spoke with Ms. Cheung to tell her that Sutter's coding issues were not limited to PAMF and instead affected Sutter systemwide. Ms. Cheung responded that there was invalid and unsupported diagnosis coding at Sutter systemwide and that PAMF was not "unique." Ms. Ormsby explained to Ms. Cheung what the potential liability could be if audit failures were extrapolated across Sutter's patient population. Ms. Cheung expressed concern about the amount of money (millions of dollars) that Sutter could be made to repay. Ms. Ormsby told Ms. Cheung that she had forwarded to UnitedHealth the medical records of audited patients, as United Health had requested. Several days later, Sutter had a mandatory RAF Coder User Group call with Ms. Cheung and coders from all of Sutter's affiliates, including Ms. Ormsby. Ms. Ormsby and other coders from other Sutter affiliates reported that Sutter had badly failed the audit, particularly in the areas of cancer, stroke, and myocardial infarction (heart attack ). Ms. Ormsby and coders from other Sutter affiliates had forwarded medical records to UnitedHealth, as UnitedHealth had requested. Ms. Cheung directed all RAF Coder User Group members to not, under any circumstances, submit medical records as they had done and, in the future, to instead forward the medical records solely to her.

Id. at 28–29 (¶¶ 88–89); see also supra note 193.

Relator FAC – ECF No. 52 at 29 (¶ 91).

Id.

Id. at 30 (¶ 92).

Id.

Id. (¶ 93).

Id. (¶ 94).

Id.

On March 31, 2014, Ms. Ormsby met with Ms. Cheung, Jessica Driver-Zuniga (Sutter's lead RAF/HCC coder), Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion), Arvin Magusara (a Sutter Senior Analyst), and Michelle Tulier from Optum, a UnitedHealth affiliate. The purpose of the meeting was to strategize about further improving RAF scores. On April 4, 2014, Ms. Tulier sent an email to the meeting's participants, copying Nancy McGinnis (Sutter's RAF Director), and referenced training the physicians whose patients were audited to improve their coding and documentation. Ms. Tulier did not discuss expanding physician training on accurate coding Sutter-wide or expanding the auditing to remove known improper diagnosis codes beyond the specific codes reviewed in the audit.

Id. (¶ 95).

Id.

In July 2014, Ms. Cheung invited Ms. Ormsby to participate in a "Peak Audit" (undertaken at Optum's request), which was a medical-record review by an outside vendor for dates of service from 2013 to July 2014 across all Sutter affiliates. Ms. Ormsby responded that "it might be a better investment to hire our own (additional) auditors" to "improve[ ] documentation and increase ‘compliant’ capture of HCC in the future." Ms. Cheung responded that "one vocal leader believes that it's worthwhile as long as the $ earned exceeds $ spent." Ms. Cheung expressed frustration that Sutter was not taking the necessary steps to prevent coding errors from recurring, writing that "[w]e keep spending money to find the same issues, but we're not preventing it from happening again."

Id. at 31 (¶ 98).

Id. (brackets in complaint).

Id.

The Peak Audit revealed widespread false coding across Sutter's affiliates, requiring Sutter to delete thousands of unsupported diagnosis codes. In December 2014, Ms. Ormsby exchanged emails with Michael Aguilar (a member of Sutter Physicians Services), the person performing Medicare Advantage submissions Sutter-wide. Mr. Aguilar confirmed that, except for the Peak Audit, Sutter had no process in place to submit "deletes" for unsupported diagnosis codes.

Id. (¶ 99).

Id.

Id. at 32 (¶ 99).

Ms. Ormsby was aware of two other discrete auditing "projects" at PAMF and Sutter Gould during her time at PAMF: a project to find unsupported diagnosis codes for 2012 dates of service and a project to find diagnosis codes for physician visits for immunizations or therapeutic injections when a medical assistant (as opposed to a physician) administered an injection. Other than those projects and her own efforts to perform audits at PAMF, Ms. Ormsby is unaware of any other audits that Sutter performed to address the issue of unsupported diagnosis codes being submitted to CMS. Ms. Ormsby alleges that she would have known about any other audits through her participation in the RAF Coder User Group and her interactions with Physician Champions. Instead, at these meetings and in interactions with the RAF Coder User Group and Physician Champions, Ms. Ormsby heard that no Sutter affiliates were conducting audits to remove or prevent unsupported codes and that Sutter management resisted any such auditing.

Id. (¶ 100).

Id. (¶ 101).

Id.

Id. at 32–33 (¶ 101).

As discussed above, Ms. Ormsby began auditing diagnosis codes at PAMF and implementing a "Corrective Action Plan." Ms. Ormsby urged Ms. Cheung, Jessica Driver-Zuniga (Sutter's lead RAF/HCC coder), and her peers at Sutter's other affiliates to conduct similar audits. In the fall of 2013, at a RAF Coder User Group meeting, Ms. Ormsby said that she was conducting encounter audits to establish accuracy baselines for PAMF and would soon start FOCUS audits for cancer, fracture, and stroke. Ms. Ormsby impressed on the group the need to focus on improper coding for cancer, fracture, and stroke because if these diagnosis codes for "active" cancer, fracture, or stroke were left in patient medical records when the conditions were no longer active, CMS would pay Sutter higher amounts for conditions that patients did not have and that Sutter was not treating. The payment for these conditions could be thousands of dollars per patient per year. As of the time that Ms. Ormsby left Sutter in May 2015, however, none of Sutter's other affiliates undertook the baseline-accuracy testing that Ms. Ormsby had initiated at PAMF or conducted any subsequent comparisons or auditing. (By contrast, Sutter conducted audits to identify HCCs that CMS had rejected in order to resubmit them to CMS for payment. )

Id. at 37–38 (¶¶ 116–19); see also supra notes 162–178.

Relator FAC – ECF No. 52 at 39 (¶ 124).

Id.

Id. at 40 (¶ 125).

Id.

Id. at 38 (¶¶ 120–21).

Id. at 43–44 (¶ 138).

At all of its affiliates, Sutter limited its auditors to removing unsupported diagnosis codes only from the "billing" side of electronic medical records, not from the "encounter data" that was submitted to CMS for payment. In February 2015, Ms. Cheung admitted that Sutter knew that unsupported diagnosis codes (caught by the auditors and removed from the "billing" file) nonetheless were being submitted to CMS for payment when the "encounter data" was swept (automatically pushed) to CMS. This caused CMS to pay Sutter based on false diagnosis codes. Ms. Cheung shared Ms. Ormsby's concern that this Sutter practice did not comply with Medicare Advantage program requirements and that this was a Sutter-wide problem. Ms. Cheung told Ms. Ormsby that they needed to "brainstorm" how to fix it because she (Ms. Cheung) did not know how. In a February 24, 2015 "Meeting Preparation Memo" to the RAF Coder User Group, Ms. Driver-Zuniga confirmed that Sutter had a systemwide problem that caused it to submit false diagnosis codes and retain overpayments predicated on those codes:

Id. at 47–48 (¶¶ 148–51).

Id. at 48 (¶ 151).

Id.

Due to limitations with the current preformatted electronic claim form in the Sutter E[lectronic]H[ealth]R[ecord], only 12 diagnosis codes can be submitted per encounter. To overcome this limitation, a monthly "data sweep" was implemented several years back. While the [data] sweep supports the capture and reporting of diagnostic information for RAF reporting, [Sutter Medical Network] has learned of an unintended consequence — the inclusion of HCC diagnosis codes removed from claims, but remaining in the Sutter E[lectronic]H[ealth]R[ecord]. To improve quality control, [Sutter Medical Network] would like for you to brainstorm with your affiliate, the pros and cons of potential solutions that can be used system-wide.

Id. at 48–49 (¶ 152) (emphasis and brackets in complaint). Ms. Ormsby alleges that this monthly "data sweep" (as an end run around the 12-diagnosis claim-form limit) was a red flag because it is implausible that Sutter physicians routinely were treating patients for 12 or more conditions in a standard office visit (that typically was less than 30 minutes). Id. at 48 (¶ 152 n.4).

At a strategy meeting to raise RAF scores a month later, Ms. Cheung confirmed that Sutter still did not have a plan to stop the submission of false diagnosis codes or to return overpayments. Ms. Cheung confirmed that CMS was still receiving HCCs that Sutter knew were false. Ms. Ormsby had suggested allowing coders to delete unsupported codes from the "encounter data" of patient medical records, but Sutter management allegedly shut down those mechanisms in order to increase RAF scores and would not consider them as a "potential solution."

Id. at 49 (¶ 153).

Id.

At all of its affiliates, Sutter employed a strategy of using patients' "problem lists" and a daily alert form to encourage physicians to capture new HCCs and recapture old HCCs that had not been documented for the year. At some point, Sutter modified the process at its affiliates to pre-populate diagnosis codes into patients' encounter data before physicians met with their patients. Christian Gabriel (PAMF's Director of Education and Coding) described this practice as "aggressive," and Dr. Veko Vahamaki (PAMF's Lead RAF Physician Champion) questioned whether pre-populating diagnosis codes into patients' actual encounter data was proper. After Ms. Ormsby left PAMF in May 2015, PAMF joined Sutter's other affiliates in this pre-population practice, which effectively captured HCCs regardless of whether physicians actually diagnosed patients with the medical conditions that had been pre-populated.

Id. at 50–51 (¶ 157).

Id. at 51 (¶ 157).

Id.

On April 19, 2019, Sutter, including all of its affiliates with Medicare Advantage patients other than PAMF, signed a settlement agreement to refund $30 million to CMS to resolve allegations by the government and CMS that it submitted improper payment data that inflated the payments it received. The settlement covered improper billing for medical conditions, including cancer, hip and vertebral fractures, strokes, and myocardial infarction (heart attack ), involving six HCCs between 2010 and 2012 and seven HCCs between 2013 and 2016. The settlement expressly reserved and did not release any liability arising under the FCA.

Id. at 53 (¶ 166).

Id.

Relator Req. for Judicial Notice Ex. B (Settlement Agreement) – ECF No. 81 at 14 (¶ 3.b). The court grants the unopposed request for judicial notice.

THE FALSE CLAIMS ACT

The plaintiffs each bring claims under (1) 31 U.S.C. § 3729(a)(1)(A)–(B) (the "direct-FCA provision") and (2) 31 U.S.C. § 3729(a)(1)(G) (the "reverse-FCA provision").

The government additionally brings common-law claims for payment by mistake and unjust enrichment.

Broadly speaking, (1) the direct-FCA provision imposes liability on a party that fraudulently presents a claim for the government to pay it money, and (2) the reverse-FCA provision imposes liability on a party that fraudulently avoids an obligation for it to pay money to the government. The plaintiffs allege that Sutter and PAMF committed direct-FCA violations by knowingly submitting false diagnosis codes to CMS, which caused CMS to pay Sutter and PAMF more money than it otherwise would have paid. They allege that Sutter and PAMF committed reverse-FCA violations by knowingly failing to delete false diagnosis codes that they had submitted and knowingly failing to return payments from CMS that were predicated on those false codes.

1. Direct-FCA Claims

The direct-FCA provision makes liable anyone who "[ (1) ] knowingly presents, or causes to be presented, a false or fraudulent claim for payment or approval; [or (2) ] knowingly makes, uses, or causes to be made or used, a false record or statement material to a false or fraudulent claim[.]" 31 U.S.C. § 3729(a)(1)(A)–(B). The direct-FCA provision "requires a showing of: ‘(1) a false statement or fraudulent course of conduct, (2) made with scienter, (3) that was material, (4) causing the government to pay out money or forfeit moneys due.’ " Godecke ex rel. United States v. Kinetic Concepts, Inc. , 937 F.3d 1201, 1208 (9th Cir. 2019) (quoting United States ex rel. Campie v. Gilead Scis., Inc. , 862 F.3d 890, 899 (9th Cir. 2017) ).

Regarding the false-statement element, "[t]o state an FCA claim, a [plaintiff] is not required to identify actual examples of submitted false claims; instead, ‘it is sufficient to allege particular details of a scheme to submit false claims paired with reliable indicia that lead to a strong inference that claims were actually submitted.’ " Id. at 1209 (some internal quotation marks omitted) (quoting Ebeid ex rel. United States v. Lungwitz , 616 F.3d 993, 998–99 (9th Cir. 2010) ). "A [plaintiff] is not required to identify representative examples of false claims to support every allegation, although the use of representative examples is one means of meeting the pleading obligation." Id. (citing Ebeid , 616 F.3d at 998 ).

Regarding the scienter element, "[l]iability under the FCA is established only when the defendant ‘knowingly’ presents a false or fraudulent claim for payment." Id. at 1211 (citing 31 U.S.C. § 3729(a)(1)(A) ). " ‘Knowingly’ is defined as having: (1) actual knowledge of the information; (2) deliberate ignorance of the truth or falsity of the information; or (3) reckless disregard of the truth or falsity of the information." Id. (citing 31 U.S.C. § 3729(b)(1)(A) ). "The FCA's ‘knowingly’ requirement ‘requires no proof of specific intent to defraud.’ " Id. (internal brackets omitted) (quoting 31 U.S.C. § 3729(b)(1)(B) ). "Instead of pleading specific intent to defraud, it is sufficient to plead that the defendant knowingly filed false claims, or that the defendant submitted false claims with reckless disregard or deliberate ignorance as to the truth or falsity of its representations." Id. (citing United States v. Bourseau , 531 F.3d 1159, 1167 (9th Cir. 2008) ). "The deliberate ignorance standard can cover ‘the ostrich type situation where an individual has buried his head in the sand and failed to make simple inquiries which would alert him that false claims are being submitted.’ " Id. (quoting Swoben , 848 F.3d at 1174 ). " ‘Congress adopted the concept that individuals and contractors receiving public funds have some duty to make a limited inquiry so as to be reasonably certain they are entitled to the money they seek.’ " Id. (quoting Swoben , 848 F.3d at 1174 ).

Regarding the materiality element, the FCA defines materiality as " ‘having a natural tendency to influence, or be capable of influencing, the payment or receipt of money or property.’ " Id. at 1213 (quoting 31 U.S.C. § 3729(b)(4) ). "Although the requirement is ‘demanding,’ the Supreme Court has held that there is not a bright-line test for determining whether the FCA's materiality requirement has been met." Id. (citing Universal Health Servs., Inc. v. United States ex rel. Escobar , ––– U.S. ––––, 136 S. Ct. 1989, 2003, 195 L.Ed.2d 348 (2016) ). "Instead, the Supreme Court has given a list of relevant, but not necessarily dispositive, factors in determining whether the false claims were material, such as whether the government decided ‘to expressly identify a provision as a condition of payment.’ " Id. (quoting Escobar , 136 S. Ct. at 2003 ). " ‘Likewise, proof of materiality can include, but is not necessarily limited to, evidence that the defendant knows that the government consistently refuses to pay claims in the mine run of cases based on noncompliance with the particular statutory, regulatory, or contractual requirement.’ " Id. (quoting Escobar , 136 S. Ct. at 2003 ). " ‘Conversely, if the government pays a particular claim in full despite its actual knowledge that certain requirements were violated, that is very strong evidence that those requirements are not material.’ " Id. (quoting Escobar , 136 S. Ct. at 2003 ). " ‘Or, if the government regularly pays a particular type of claim in full despite actual knowledge that certain requirements were violated, and has signaled no change in position, that is strong evidence that the requirements are not material.’ " Id. (quoting Escobar , 136 S. Ct. at 2003–04 ). " ‘Materiality, in addition, cannot be found where noncompliance is minor or insubstantial.’ " Id. (quoting Escobar , 136 S. Ct. at 2003 ).

2. Reverse-FCA Claims

The reverse-FCA provision makes liable anyone who "[ (1) ] knowingly makes, uses, or causes to be made or used, a false record or statement material to an obligation to pay or transmit money or property to the Government, or [ (2) ] knowingly conceals or knowingly and improperly avoids or decreases an obligation to pay or transmit money or property to the Government." 31 U.S.C. § 3729(a)(1)(G).

The elements of a violation under the first prong of the reverse-FCA provision are that (1) a record or statement was false, (2) the defendant had knowledge of the falsity, (3) the defendant made or used (or caused to be made or used) the false record or statement, (4) the defendant's purpose was to conceal, avoid, or decrease an obligation to pay the government, and (5) the false record or statement was material. Bourseau , 531 F.3d at 1164–71.

The elements of a violation under the second prong of the reverse-FCA provision are that the defendant (1) concealed or improperly avoided or decreased an obligation to pay the government and (2) did so knowingly. United States ex rel. Customs Fraud Investigations, LLC v. Victaulic Co. , 839 F.3d 242, 255 (3d Cir. 2016) (citing 31 U.S.C. § 3729(a)(1)(G) ); accord United States v. Vandewater Int'l, Inc. , No. 2:17-cv-04393RGK-KS, 2019 WL 6917927, at *4 n.1 (C.D. Cal. Sept. 3, 2019) (citing Victaulic , 839 F.3d at 255 ). There is no requirement under the second prong to show that the defendant used a false record or statement or that a record or statement was material. Victaulic , 839 F.3d at 255 ; Vandewater , 2019 WL 6917927, at *4 n.1.

As first enacted in 1986, the reverse-FCA provision contained only the first prong. False Claims Amendment Act of 1986, Pub. L. No. 99-562, § 2, 100 Stat. 3153, 3153 (1986). "The plain language of this statute require[d] that a defendant make or use a false record or statement in order to conceal, avoid or decrease an obligation to the government." Bourseau , 531 F.3d at 1169. In 2009, Congress amended the reverse-FCA provision to additionally make liable anyone who "knowingly conceals or knowingly and improperly avoids or decreases an obligation to pay or transmit money or property to the government." Fraud Enforcement and Recovery Act of 2009 (FERA), Pub. L. No. 111-21, § 4, 123 Stat. 1617, 1622 (2009). "A false statement is no longer a required element, since the post-FERA FCA specifies that mere knowledge and avoidance of an obligation is sufficient, without the submission of a false record, to give rise to liability." Victaulic , 839 F.3d at 255.

Regarding the obligation-to-pay-the-government element, in 2010, as part of the Patient Protection and Affordable Care Act ("ACA"), Congress promulgated "Enhanced Medicare and Medicaid Program Integrity Provisions," including provisions codified at 42 U.S.C. § 1320a-7k. Patient Protection and Affordable Care Act, Pub. L. No. 111-148, § 6402, 124 Stat. 119, 753–56 (2010). Section 1320a-7k provides that a person (including a medical provider) who has received a Medicare or Medicaid overpayment must report and return the overpayment within 60 days after the overpayment is identified. 42 U.S.C. § 1320a-7k(d)(2)(A). Any overpayment retained for more than 60 days becomes an "obligation" for purposes of the reverse-FCA provision. 42 U.S.C. § 1320a-7k(d)(3).

"The term ‘person’ means a provider of services, supplier, [M]edicaid managed care organization (as defined in section 1396b(m)(1)(A) of this title), Medicare Advantage organization (as defined in section 1395w-28(a)(1) of this title), or PDP sponsor (as defined in section 1395w-151(a)(13) of this title). Such term does not include a beneficiary." 42 U.S.C. § 1320a-7k(d)(4)(C) (internal headings omitted).

More specifically, § 1320a-7k(d)(2) provides that a person must report and return an overpayment "by the later of — (A) the date which is 60 days after the date on which the overpayment was identified; or (B) the date any corresponding cost report is due, if applicable." The "corresponding cost report" condition is inapplicable in the Medicare Advantage program because, in general, CMS makes payments to MA Organizations based on their bids and not based on their actual incurred costs. Contract Year 2015 Policy and Technical Changes to the Medicare Advantage and Medicare Prescription Drug Benefit Programs , 79 Fed. Reg. 29,844, 29,920 (May 23, 2014).

As discussed below, § 1320a-7k does not expressly delineate the scope of "overpayment" and does not define "identified." A central issue in the case is how to define these terms.

Regarding the scienter element and (with respect to the first prong only) the false-statement and materiality elements, the standards for these elements are the same for the reverse-FCA provision and the direct-FCA provision. 31 U.S.C. § 3729(b)(1), (4) ; Bourseau , 531 F.3d at 1164–68, 1170–71.

STANDARD OF REVIEW

A complaint must contain a "short and plain statement of the claim showing that the pleader is entitled to relief" to give the defendant "fair notice" of what the claims are and the grounds upon which they rest. Fed. R. Civ. P. 8(a)(2) ; Bell Atl. Corp. v. Twombly , 550 U.S. 544, 555, 127 S.Ct. 1955, 167 L.Ed.2d 929 (2007). A complaint does not need detailed factual allegations, but "a plaintiff's obligation to provide the ‘grounds’ of his ‘entitlement to relief’ requires more than labels and conclusions, and a formulaic recitation of the elements of a cause of action will not do. Factual allegations must be enough to raise a claim for relief above the speculative level[.]" Twombly , 550 U.S. at 555, 127 S.Ct. 1955 (internal citations omitted).

To survive a motion to dismiss, a complaint must contain sufficient factual allegations, which when accepted as true, " ‘state a claim to relief that is plausible on its face.’ " Ashcroft v. Iqbal , 556 U.S. 662, 678, 129 S.Ct. 1937, 173 L.Ed.2d 868 (2009) (quoting Twombly , 550 U.S. at 570, 127 S.Ct. 1955 ). "A claim has facial plausibility when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged." Id. "The plausibility standard is not akin to a ‘probability requirement,’ but it asks for more than a sheer possibility that a defendant has acted unlawfully." Id. (citing Twombly , 550 U.S. at 557, 127 S.Ct. 1955 ). "Where a complaint pleads facts that are merely consistent with a defendant's liability, it stops short of the line between possibility and plausibility of ‘entitlement to relief.’ " Id. (some internal quotation marks omitted) (quoting Twombly , 550 U.S. at 557, 127 S.Ct. 1955 ).

FCA claims also are subject to the heightened pleading requirements of Federal Rule of Civil Procedure 9(b). Swoben , 848 F.3d at 1180. Under Rule 9(b), "the plaintiff must allege ‘the who, what, when, where, and how of the misconduct charged,’ including what is false or misleading about a statement, and why it is false." Id. (quoting Ebeid , 616 F.3d at 998 ). "Knowledge, however, may be pled generally." Id. (citing United States v. Corinthian Colls. , 655 F.3d 984, 996 (9th Cir. 2011) ).

Under Rule 9(b), " ‘mere conclusory allegations of fraud are insufficient.’ " Id. (citing Wool v. Tandem Computers Inc. , 818 F.2d 1433, 1439 (9th Cir. 1987), overruled as stated in Flood v. Miller , 35 F. App'x 701, 703 n.3 (9th Cir. 2002) ). "Broad allegations that include no particularized supporting detail do not suffice[.]" Id. (citing Bly-Magee v. California , 236 F.3d 1014, 1018 (9th Cir. 2001) ). "[B]ut ‘statements of the time, place and nature of the alleged fraudulent activities are sufficient[.]’ " Id. (quoting Wool , 818 F.2d at 1439 ). "Because this standard ‘does not require absolute particularity or a recital of the evidence,’ a complaint need not allege ‘a precise time frame,’ ‘describe in detail a single specific transaction’ or identify the ‘precise method’ used to carry out the fraud." Id. (quoting Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 1298 (3d ed. 2016) ; Cooper v. Pickett , 137 F.3d 616, 627 (9th Cir. 1997) ).

ANALYSIS

Sutter and PAMF move to dismiss the complaints on the following grounds: (1) a statutory requirement — that CMS ensure "actuarial equivalence" between traditional Medicare and Medicare Advantage, 42 U.S.C. § 1395w-23(a)(1)(C)(i) — means that to state an FCA claim, the plaintiffs must allege that an MA Participant's rate of unsupported diagnosis codes exceeds the rate of unsupported diagnosis codes in the traditional Medicare data CMS uses in its Medicare Advantage risk model (and the plaintiffs have not alleged this); (2) Ms. Ormsby cannot pursue her claims against Sutter's non-PAMF affiliates because the claims exceed the scope of the government's intervention; and (3) the complaints fail to allege facts showing that Sutter and PAMF knowingly failed to return overpayments (for a reverse-FCA claim) or knowingly submitted false claims (for a direct-FCA claim).

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 7–21.

Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 5–10.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 9, 21–28; Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 6, 10–13.

The court denies the motions to dismiss. "Actuarial equivalence" is not a defense to an FCA claim (and nothing in 42 U.S.C. § 1395w-23(a)(1)(C)(i) imposes the pleading standard that Sutter and PAMF assert here), the clear weight of authority establishes that Ms. Ormsby may pursue her claims, and the plaintiffs otherwise sufficiently plead their claims.

1. The "Actuarial Equivalence" Argument Is Not a Valid Defense

Sutter's and PAMF's main argument is that MA Participants are not overpaid merely because they submit unsupported diagnosis codes. Instead, they argue, MA Participants are overpaid only if they have a rate of unsupported codes that is higher than the error rate in the traditional Medicare fee-for-service program. They move to dismiss the FCA claims on the ground that the plaintiffs did not allege any facts about this foundational requirement for false claims.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 14, 17–18.

Id.

Id. at 17–18.

This argument is predicated on 42 U.S.C. § 1395w-23(a)(1)(C)(i), which provides that CMS must pay MA Organizations in a manner that ensures "actuarial equivalence" between payments for healthcare under traditional Medicare and payments for health care under MA Plans:

[CMS] shall adjust [Medicare Advantage payments] for such risk factors as age, disability status, gender, institutional status, and such other factors as [CMS] determines to be appropriate, including adjustment for health status ..., so as to ensure actuarial equivalence. [CMS] may add to, modify, or substitute for such adjustment factors if such changes will improve the determination of actuarial equivalence.

To support their argument that payments to an MA Participant predicated on false diagnosis codes are not "overpayments" unless the MA Participant's diagnosis-code "error rate" exceeds the "error rate" for traditional Medicare, Sutter and PAMF cite a District of Columbia decision (now on appeal to the D.C. Circuit) that addressed a challenge by the MA Organization UnitedHealth (and its affiliates) under the Administrative Procedures Act ("APA") to CMS's Medicare Advantage payment practices. UnitedHealthcare Insurance Co. v. Azar , 330 F. Supp. 3d 173 (D.D.C. 2018) ( UnitedHealthcare II ), recons. denied , No. 16-157 (RMC), 2020 WL 417867 (D.D.C. Jan. 27, 2020) ( UnitedHealthcare III ), appeal docketed , No. 18-5326 (D.C. Cir. filed Nov. 18, 2018). In 2014, CMS issued guidance that said, among other things, that "a risk adjustment diagnosis that has been submitted for payment but is found to be invalid because it does not have supporting medical record documentation would result in an overpayment." Medicare Program; Contract Year 2015 Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Drug Benefit Programs , 79 Fed. Reg. 29,844, 29,921 (May 23, 2014). The UnitedHealthcare court held that this guidance (which it called the "2014 Overpayment Rule" ) violated the statutory mandate of actuarial equivalence. UnitedHealthcare II , 330 F. Supp. 3d at 176. The court found that CMS uses unaudited records of the payments it makes in the traditional Medicare program in the risk model it uses to set payment rates for the Medicare Advantage program. Id. at 176, 179. Because the traditional-Medicare data are unaudited, they necessarily contain errors. Id. at 179. The court held that using traditional Medicare data with errors to set payment rates — and then requiring MA Participants to return payments predicated on erroneous or unsupported diagnosis codes — imposed higher scrutiny on Medicare Advantage than traditional Medicare and thus violated actuarial equivalence. Id. at 176, 186–87.

As discussed below, CMS disputes that its 2014 guidance was a new rule because it has long required (since well before 2014) that diagnosis codes be supported by medical-record documentation. See infra note 424. The court uses the UnitedHealthcare court's term "2014 Overpayment Rule" for convenience, and the use of that term is not an opinion that it was a new rule.

Relying on UnitedHealthcare , Sutter and PAMF assert that requiring MA Participants to return payments predicated on false diagnosis codes leads to the "inevitable" result that CMS will pay less for beneficiaries enrolled in MA Plans than it would pay for those beneficiaries if they were enrolled in traditional Medicare, thereby violating "actuarial equivalence." Sutter and PAMF thus contend that they must return overpayments (predicated on false diagnosis codes) only if "the error rate for a Medicare Advantage contract is greater than the CMS error rate" for traditional fee-for-service Medicare. They contend that the plaintiffs did not plead FCA claims because they did not allege (1) the rate of unsupported diagnosis codes for traditional Medicare providers, (2) Sutter's and PAMF's "overall rate of unsupported diagnosis codes," and (3) a comparison of those rates showing that Sutter's and PAMF's "error rate" (in the form of the unsupported diagnosis codes) exceeded the traditional Medicare "error rate." They also argue that at minimum, this argument shows that their position was reasonable, thereby defeating the plaintiffs' allegations about the scienter needed for FCA claims.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 15 (quoting UnitedHealthcare II , 330 F. Supp. 3d at 176, 187 ).

Id. at 8, 15 (quoting UnitedHealthcare II , 330 F. Supp. 3d at 186 ).

Id. at 21 (emphasis in original).

Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 6.

The "actuarial equivalence" argument is not a ground to dismiss the complaints. First, Sutter and PAMF have not established that their premise — CMS's using unaudited traditional-Medicare data in its Medicare Advantage risk model and requiring MA Participants to return payments predicated on false diagnosis codes inevitably leads to CMS's paying less for Medicare Advantage than for traditional Medicare — is sound. Second, even if it were, it is not a defense to an FCA claim. CMS's possibly underpaying MA Participants does not entitle the MA Participants to correct that supposed wrong by submitting false diagnosis codes and failing to report or return payments predicated on the codes. Additionally, nothing in the "actuarial equivalence" provision imposes the pleading standard Sutter and PAMF seek to impose here. Third, Sutter's and PAMF's position — which, at its core, is that MA Participants do not have to report or return payments predicated on false codes — is not objectively reasonable and does not defeat scienter.

1.1 The Premise — That CMS Is Violating "Actuarial Equivalence" — Is Not Sound

In the next sections, the court examines (1) CMS's historical audit process and its promulgation of the 2014 Overpayment Rule, (2) the UnitedHealthcare court's invalidation of the Rule, and (3) whether Sutter's and PAMF's contention — that the return of payments predicated on false diagnosis codes inevitably violates "actuarial equivalence" — is sound.

1.1.1 CMS's Medicare Advantage audit process and the "2014 Overpayment Rule"

Understanding how the actuarial-equivalence analysis in UnitedHealthcare applies to this case begins with the consideration of CMS's historical audit process for MA Organizations and its promulgation of the 2014 Overpayment Rule.

1.1.1.1 Audits of MA Organizations

Historically, CMS conducted "Risk Adjustment Data Validation" ("RADV") audits of a subset of MA Organizations, comparing their submitted diagnosis codes to underlying patient medical records to verify that the records supported the diagnosis codes. UnitedHealthcare II , 330 F. Supp. 3d at 180 ; accord Silingo , 904 F.3d at 672–73 ("With data for millions of people being submitted each year, CMS is unable to confirm diagnoses before calculating capitation rates. Instead, the agency accepts the diagnoses as submitted, and then audits some of the self-reported data a few years later to ensure that they are adequately supported by medical documentation.") (citing 42 C.F.R. §§ 422.310(e), 422.311 ; 79 Fed. Reg. at 2001 ). CMS required MA Organizations to return any payments predicated on unsupported diagnosis codes. UnitedHealthcare II , 330 F. Supp. 3d at 180.

In 2008, CMS announced that it would apply the RADV error rate from the audited samples to the entire MA Plan for the MA Organization. Id. More specifically, previously MA Organizations returned only payments for the specific unsupported diagnosis codes identified in the audit. The new proposal meant that MA Organizations also would return an extrapolated payment for diagnosis-code submissions that CMS did not audit, essentially on the presumption that the other submissions would include similar unsupported codes. Id.

In response to CMS's request for comments, MA Organizations objected that the proposed methodology would violate the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i) because CMS (1) sets the payment rates for Medicare Advantage based on diagnosis codes from traditional Medicare providers that are not audited or verified in any way and (2) requires MA Organizations to make repayments based on audited results. Id. As a result of the comments, in 2012, CMS announced that it would apply a "Fee-for-Service Adjuster" or "FFS Adjuster" — essentially, according to the UnitedHealthcare court, CMS's own estimate of the error rate in risk factors and diagnosis codes submitted in traditional Medicare — to the results of RADV audits of the MA Organizations (meaning, to its extrapolation of the RADV audit results to the entire MA Plan). Id. The UnitedHealthcare court described the outcome from CMS's application of the FFS Adjuster: "Medicare Advantage providers must return to CMS any audited ‘overpayments’ to the extent that the insurer's errors exceed the estimated error rate in CMS payments under traditional Medicare." Id. at 180–81 (citing Ctrs. For Medicare and Medicaid Servs., Notice of Final Payment Error Calculation Methodology for Part C Medicare Advantage Risk Adjustment Data Validation Contract-Level Audits 1–5 (Feb. 24, 2012), available at https://www.cms.gov/Research-Statistics-Data-and-Systems/Monitoring-Programs/recovery-audit-program-parts-c-and-d/Other-Content-Types/RADV-Docs/RADV-Methodology.pdf (last visited Mar. 16, 2020)). CMS's announcement of the FFS Adjuster puts it differently, saying that CMS would conduct a review to calculate the actual amount of the FFS Adjuster without forecasting the amount or linking it to "error rates." See Notice of Final Payment Error Calculation Methodology 4–5.

CMS's announcement stated that "[t]he FFS adjuster accounts for the fact that the documentation standard in RADV audits to determine a contract's payment error (medical records) is different from the documentation standard used to develop the Part C 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." Notice of Final Payment Error Calculation Methodology 4–5.

1.1.1.2 The "2014 Overpayment Rule"

In 2010, as part of the ACA (as discussed above), Congress promulgated "Enhanced Medicare and Medicaid Program Integrity Provisions" that require MA Participants to return Medicare "overpayments" within 60 days after the overpayments are "identified." 42 U.S.C. § 1320a-7k(d)(2)(A). "The term ‘overpayment’ means any funds that a person receives or retains under [Medicare or Medicaid] to which the person, after applicable reconciliation, is not entitled." 42 U.S.C. § 1320a-7k(d)(4)(B). An overpayment that is not returned within 60 days of when it is identified becomes an "obligation" under the reverse-FCA provision. 42 U.S.C. § 1320a-7k(d)(3).

In January 2014, CMS issued a notice of proposed rulemaking, including a rule to clarify for § 1320a-7k's statutory definition of "overpayment" by defining the terms "funds" and "applicable reconciliation." 79 Fed. Reg. at 1996. The proposed definition of "funds" was "payments an MA organization ... has received that are based on data that these organizations submitted to CMS for payment purposes and for which they have responsibility for the accuracy, completeness, and truthfulness of such data under existing [ 42 C.F.R.] § 422.504(l ) and § 423.505(k) ). For Part C [Medicare Advantage], the data submitted by the MA organization to CMS includes § 422.308(f) (enrollment data) and § 422.310 (risk adjustment data)." Id. The proposed definition of "applicable reconciliation" was that it "occurs on the date that CMS announces as the final deadline for risk adjustment data submission." Id.

CMS also stated that "[i]t is our expectation that MA organizations ... must be continuously diligent regarding the accuracy and completeness of payment-related data they submit to CMS for a payment year, whether during or after that payment year, and whether before or after applicable reconciliation dates." Id. at 1997. It disclaimed that this was a new requirement, explaining that "[t]his expectation is based on existing requirements at [ 42 C.F.R.] § 422.310, § 422.504(l ), § 423.329(b)(3)(ii), and § 423.505(k), and our proposed amendments that clarify and strengthen these requirements." Id. In May 2014, CMS published its final rules. 79 Fed. Reg. 29,844. It said in its notice that it "did not receive any comments on the proposed definitions of the terms ‘funds’ or ‘overpayment’ " and finalized (in 42 C.F.R. § 422.326(a) ) its proposed definitions for those terms. Id. at 29,920, 29,958. Additionally, CMS "reminded all stakeholders that even in the absence of a final regulation on these statutory provisions, MA organizations ... are subject to the statutory requirements found in [ 42 U.S.C. § 1320a-7k ] and could face potential False Claims Act liability, Civil Monetary Penalties (CMP) Law liability, and exclusion from Federal health care programs for failure to report and return an overpayment." Id. at 29,919. It reaffirmed that an "overpayment" occurs when "the MA organization ... has submitted erroneous data to CMS that caused CMS to overpay the organization." Id. at 29,921 ("identification of an overpayment means knowing that the MA organization ... has submitted erroneous data to CMS that caused CMS to overpay the organization"). "For example, a risk adjustment diagnosis that has been submitted for payment but is found to be invalid because it does not have supporting medical record documentation would result in an overpayment." Id. (The UnitedHealthcare court refers to CMS's May 2014 announcement as the "2014 Overpayment Rule." UnitedHealthcare II , 330 F. Supp. 3d at 182. )

CMS did not describe its guidance about "overpayments" — including its guidance that submissions of unsupported diagnosis codes result in overpayments — as a new rule. Instead, it characterized its guidance as in keeping with its longstanding requirements that providers submit accurate diagnosis codes and return overpayments arising from inaccurate diagnosis codes. See 79 Fed. Reg. at 29,921 ; see also id. at 29,919 ("MA organizations ... continue to be obliged to comply with our current procedures for handling inaccurate payments.... MA organizations ... have responsibility for the accuracy, completeness, and truthfulness of data they submit under existing [42 C.F.R.] §§ 422.504(l ) and 423.505(k). For Part C, the data submitted by the MA organization to CMS includes §§ 422.308(f) (enrollment data) and 422.310 (risk adjustment data)."); accord UnitedHealthcare II , 330 F. Supp. 3d at 180 (noting that for years before CMS proposed its 2014 rules, it had "required repayment to CMS of any costs that were based on unsupported diagnosis codes").

CMS addressed a commenter's contention that (given CMS's announcement that it would use an FFS Adjuster in its RADV audits in the context of extrapolating error rates from an RADV audit to the entire MA Plan), an overpayment cannot exist for a particular Medicare Advantage contract unless CMS applied an FFS Adjuster to the entire contract. 79 Fed. Reg. at 29,921. CMS disagreed:

Our RADV methodology does not change our existing contractual requirement that MA organizations must certify (based on best knowledge, information, and belief) the accuracy, completeness, and truthfulness of the risk adjustment data they submit to CMS. Further, this decision does not change the long-standing risk adjustment data requirement that a diagnosis submitted to CMS by an MA organization for payment purposes must be supported by medical record documentation.

Id. at 29,921 –22.

1.1.2 UnitedHealthcare 's invalidation of the 2014 Overpayment Rule

The genesis of the UnitedHealthcare lawsuit was FCA lawsuits against UnitedHealth, including the Swoben lawsuit, charging that UnitedHealth submitted false risk-adjusting diagnosis codes (and false certifications that its codes were accurate, complete, and truthful) to CMS. UnitedHealthcare Ins. Co. v. Price , 255 F. Supp. 3d 208, 209–10 (D.D.C. 2017) ( UnitedHealthcare I ). UnitedHealth and its affiliates then sued CMS in UnitedHealthcare , claiming, in part, that CMS's May 2014 notice implemented a new rule that any payment for an unsupported diagnosis code constituted an "overpayment" and that this violated the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i) and the APA.

The UnitedHealthcare lawsuit was filed when Sylvia Burwell was Secretary of Health and Human Services and originally was captioned UnitedHealthcare Insurance Co. v. Burwell . It was recaptioned as UnitedHealthcare Insurance Co. v. Price and UnitedHealthcare Insurance Co. v. Azar as Tom Price and Alex Azar, respectively, were appointed as Secretary.

Complaint, UnitedHealthcare Ins. Co. v. Burwell , No. 1:16-cv-00157-RMC (D.D.C. Jan. 29, 2016) – ECF No. 1 at 9–13 (¶¶ 9–15), 33 (¶ 71), 39–40 (¶¶ 91–94).

CMS moved to stay the UnitedHealthcare lawsuit in light of the FCA lawsuits. The court denied the motion, holding that "[t]his APA case is limited to whether CMS acted beyond its authority when promulgating its 2014 Overpayment Rule." Id. at 211. It explained:

[A]ny decision in this matter will not answer the most relevant questions in the FCA Cases. Whether a government contractor knowingly engaged in fraud, and whether a government agency appropriately promulgated a rule several years later, are simply too different from one another to warrant a stay, even if such lawsuits may touch upon similar questions of statutory interpretation.

Id.

UnitedHealth moved for summary judgment in the UnitedHealthcare lawsuit. The court granted the motion on the ground that the 2014 Overpayment Rule — that "any diagnostic code that is inadequately documented in a patient's medical chart results in an ‘overpayment,’ " UnitedHealthcare II , 330 F. Supp. 3d at 182 (emphasis in original) (citing 79 Fed. Reg. at 29,921 ) — established a system where actuarial equivalence purportedly cannot be achieved. Id. at 187. It concluded that (1) CMS's using unaudited data from traditional Medicare (which contain errors) in the risk model it uses to set Medicare Advantage payment while (2) requiring MA Participants to return payments predicated on erroneous diagnosis codes, led to CMS's paying less to MA Participants to provide healthcare coverage than it pays for the coverage of comparable beneficiaries in traditional Medicare, thereby violating "actuarial equivalence." Id. at 184–85.

More specifically (and based solely on the administrative record ), the UnitedHealthcare court made the following findings and conclusions.

See UnitedHealthcare I , 255 F. Supp. 3d at 211 ("[A]ny analysis by the Court would be limited to the administrative record behind the 2014 Overpayment Rule and the statute.").

For traditional Medicare Part B, physicians bill for services and identify the reasons for the services with diagnosis codes. Id. at 177, 184. Payment depends only on the services; the diagnosis codes — while required — are irrelevant to payment. Id. at 177, 184. This means that "physicians are essentially indifferent to the diagnosis.... There's no financial incentive to be particularly careful" with diagnosis coding. Id. at 179 (ellipsis in original) (quoting hearing). As a result, "diagnosis reports for Medicare Part B are considered much less reliable[.]" Id. Also, because diagnosis codes do not affect what CMS pays for Part B, CMS purportedly does not audit "traditional Medicare patient records; to the contrary, it accepts their diagnosis codes as given." Id. at 184.

CMS uses the traditional Medicare data (as discussed above) as inputs in its HCC model for calculating risk scores for (and thus payments under) Medicare Advantage. Id. at 178. But because CMS and medical providers are not careful about diagnosis codes in Part B (because they do not matter for payment), the Part B diagnosis codes have errors. Id. (citing UnitedHealth's claim that the error rate "can be as high as 20%."). "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.’ " Id. at 184. The UnitedHealthcare court concluded, "the effect of the 2014 Rule, without some kind of adjustment, is that Medicare Advantage insurers will be paid less to provide the same healthcare coverage to their beneficiaries than CMS itself pays [to physicians under Medicare Part B] for comparable patients." Id. at 184–85.

The problem with the Rule, said the court, was that it had no such adjustment. Id. at 187. According to the court, CMS recognized (in the context of RADV audits) the actuarial need to apply an FFS Adjuster. Id. But "the fly in the ointment" was that CMS did not implement a similar adjuster in the 2014 Overpayment Rule." Id. "The consequence is inevitable: while CMS pays for all diagnostic codes, erroneous or not, submitted to traditional Medicare, it will pay less for Medicare Advantage coverage because essentially no errors would be reimbursed." Id. The court thus invalidated the Rule under the APA because it "establishes a system where ‘actuarial equivalence’ cannot be achieved."

The UnitedHealthcare court also held that the putative 2014 Overpayment Rule should be set aside because it (1) purportedly violated a separate statutory provision, 42 U.S.C. § 1395w-23(b)(4)(D), and (2) was arbitrary and capricious. UnitedHealthcare II , 330 F. Supp. 3d at 187–90.

1.1.3 UnitedHealthcare does not compel the conclusion that there is no actuarial equivalence here

Sutter and PAMF contend that — under the analysis in UnitedHealthcare — requiring MA Participants to return payments predicated on false diagnosis codes also leads to the "inevitable" result that CMS will pay less for beneficiaries enrolled in MA Plans than for beneficiaries enrolled in traditional Medicare, thereby violating the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i). The court does not reach that conclusion for several reasons.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 15 (citing UnitedHealthcare II , 330 F. Supp. 3d at 176, 187 ).

First, the UnitedHealthcare court predicated its decision — that CMS inevitably pays less for beneficiaries in MA plans — on CMS's application of an FFS Adjuster when extrapolating RADV audit results to the entire MA Plan. UnitedHealthcare II , 330 F. Supp. 3d at 180–81 (citing Notice of Final Payment Error Calculation Methodology 1–5). But as discussed above, CMS's 2012 announcement of the FFS Adjuster says only that CMS would conduct a review to calculate the actual amount of the FFS Adjuster. See Notice of Final Payment Error Calculation Methodology 4–5. It does not identify the amount or state that CMS is paying less for Medicare Advantage coverage. Thus, the court cannot conclude from CMS's 2012 announcement that CMS inevitably pays less for beneficiaries in MA plans.

In 2018, CMS completed its FFS Adjuster review and concluded that while there is a significant error rate in traditional Medicare diagnosis codes, the overall impact of those errors on Medicare Advantage payments is less than 1% and results in a 1% overpayment to MA Organizations (i.e., CMS pays more, not less, for Medicare Advantage coverage). Ctrs. for Medicare and Medicaid Servs., Fee for Service Adjuster and Payment Recovery for Contract Level Risk Adjustment Data Validation Audits 5 (Oct. 26, 2018), available at https://www.cms.gov/Research-Statistics-Data-and-Systems/Monitoring-Programs/Medicare-Risk-Adjustment-Data-Validation-Program/Other-Content-Types/RADV-Docs/FFS-Adjuster-Excecutive-Summary.pdf (last visited Mar. 16, 2020). The UnitedHealthcare court found CMS's FFS Adjuster review to be unpersuasive, UnitedHealthcare III , 2020 WL 417867, at *5, but CMS's FFS Adjuster review shows that CMS had not concluded in 2012 that it was paying less for Medicare Advantage coverage, much less that it was inevitably doing so.

Second, Sutter and PAMF advance a hypothetical to demonstrate that CMS inevitably pays less. The hypothetical does not demonstrate this point. The following is a simplified version of the hypothetical. Assume that CMS used traditional Medicare data from 2011 as inputs in its HCC model to determine risk scores in 2014. Assume that in 2011, there were 1,000 beneficiaries in traditional Medicare with a diagnosis code for diabetes (with a resulting $1 million in bills to CMS), and 500 of the 1,000 diagnosis codes were unsupported. Because CMS allegedly does not audit the traditional Medicare data it uses in its HCC model, it would assume that the 1,000 diabetes diagnosis codes all are valid and would calculate (and pay MA Participants) an average cost per code of $1,000. Now assume that in 2014, the beneficiaries enrolled in either traditional Medicare or Medicare Advantage and had the same costs and codes that they had in 2011. For enrollees in traditional Medicare, CMS would pay $1 million. For enrollees in MA plans, CMS requires MA Participants to delete unsupported codes (and return the related payments), which means that after deleting the 500 false diagnosis codes, CMS would pay the MA Participant $500,000. Sutter and PAMF contend that by paying less for Medicare Advantage coverage ($500,000) than traditional Medicare coverage ($1 million), CMS violates "actuarial equivalence." Sutter and PAMF contend that CMS therefore cannot require the MA Participant to return payments for false diagnosis codes (unless the MA Participant's submission rate of false codes exceeds the error rate for traditional Medicare providers).

Cf. Ctrs. for Medicare and Medicaid Servs., Announcement of Calendar Year (CY) 2014 Medicare Advantage and Part D Payment Policies and Final Call Letter 3 (Apr. 1, 2013), available at https://www.cms.gov/medicare/health-plans/medicareadvtgspecratestats/downloads/announcement2014.pdf (last visited Mar. 16, 2020) ("We will rebase the FFS capitation rates for 2014, using historical claims data for 2007 through 2011.").

Cf. Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 13. Sutter and PAMF propose a hypothetical where 1,000 beneficiaries have 3,000 different diagnosis codes (i.e., the average beneficiary has multiple diagnosis codes), and 1,500 codes are unsupported. Id. The court begins with a hypothetical that assumes one diagnosis code per beneficiary and then addresses below the more complicated issues that may arise when beneficiaries have multiple diagnosis codes.

This is not to say that CMS does not audit traditional Medicare data at all. CMS audits traditional Medicare data, just as it audits Medicare Advantage data. What Sutter and PAMF claim is that CMS does not audit traditional Medicare data before using the data as inputs in its HCC model. Id. ("CMS does not audit fee-for-service data before calculating the incremental costs associated with diagnosis codes"). The government agrees that "[t]he claim dataset used to calibrate the HCC model reflects a snapshot in time" and that "[c]ertain claims and diagnoses could have been corrected by the [traditional Medicare] providers that submitted them or rejected by CMS before or after the claims dataset was generated to calibrate the HCC model." Gov't Opp'n – ECF No. 82 at 28.

Of course, CMS would not pay exactly $1 million for traditional-Medicare coverage of these beneficiaries in 2014. Traditional Medicare uses a fee-for-services model, and the beneficiaries likely would require different services in 2014 than in 2011. A beneficiary in perfect health on Monday might suffer an accident on Tuesday, incurring additional medical costs. But the relevant issue here is actuarial predictions, not actual costs, and this hypothetical therefore assumes that actuarial predictions accurately predict actual costs.

But the hypothetical captures only part of the story. If CMS really does not audit the traditional Medicare data that it uses as inputs in its HCC model, then the traditional Medicare data likely contain not only unsupported diagnosis codes but also, unsupported costs. In this hypothetical involving data from traditional Medicare in 2011, Sutter and PAMF posit that half of the diagnosis codes were unsupported, but it might be that half of the costs were unsupported too (based on bills for services that were medically unnecessary or not performed), and CMS did not catch the misbilling because it purportedly does not audit traditional Medicare data. CMS would assume that the $1 million in costs were valid and pay MA Participants $1,000 for a diabetes diagnosis code. Now assume that in 2014, the beneficiaries enrolled in either traditional Medicare or Medicare Advantage and had the same costs and codes that they had in 2011. For enrollees in traditional Medicare (with providers who billed only for legitimate services instead of overbilling, as they did in 2011), CMS would pay $500,000 (which means that it would not pay less for Medicare Advantage beneficiaries than for traditional Medicare beneficiaries).

The Ninth Circuit in Silingo cited a GAO report that found that in 2016, $16.2 billion — nearly 10% — of all Medicare Advantage payments were improper. Silingo , 904 F.3d at 673 (citing Medicare Advantage Program Integrity 1). The same report found that 11% of traditional Medicare payments also were improper, in part, because reviewers could not conclude that the traditional Medicare provider actually provided the billed services, that it provided services for the amounts billed, or that the billed services were medically necessary. Medicare Advantage Program Integrity 3–4 & n.8.

The UnitedHealthcare court observed that traditional Medicare providers do not have an incentive to be careful about their diagnosis coding because coding does not affect what they are paid. UnitedHealthcare II , 330 F. Supp. 3d at 179. Traditional Medicare providers do have an incentive to bill as many services as possible to increase payments to them.

Cf. Gov't Opp'n – ECF No. 82 at 28 ("If both the traditional Medicare provider and the MA provider comply with their obligation to correct their data, they continue to receive comparable payments."). And, of course, if the traditional Medicare providers knowingly overbilled and received $1 million in payment for $500,000 in legitimate services, they would be subject to FCA liability, like Medicare Advantage providers that knowingly overcoded. See, e.g. , United States ex rel. Martinez v. KPC Healthcare Inc. , No. 8:15-cv-01521-JLS-DFM, 2017 WL 10439030, at *6 (C.D. Cal. June 8, 2017) ("[U]nder [42 U.S.C. § 1320a-7k ]'s plain meaning, a Medicare fee-for-service payment recipient has violated 31 U.S.C. § 3729(a)(1)(G) if it fraudulently bills the government and then knowingly fails to return the overpayment within sixty days of the identification of the overpayment.") (citing cases).

The actuarial-equivalence argument is even less persuasive for patients with multiple diagnoses. An example is patients who fracture their tibia (the front bone in the lower leg) or their fibula (the rear bone in the lower leg). Assume that in 2011, there were 1,000 traditional Medicare beneficiaries, 200 with a diagnosis code for a tibia fracture, 200 with a diagnosis code for a fibula fracture, and 600 erroneously coded for both (when they broke only one bone). All patients need an X-ray of their leg at an assumed cost (say, $100). Sutter and PAMF assume that CMS's HCC model divides the total cost ($100 × 1,000 patients = $100,000) by the total number of diagnosis codes (200 + 200 + (600 × 2) = 1600) for a cost of $62.50 per diagnosis code. But CMS's HCC model is a regression model, not a cost averager. UnitedHealthcare II , 330 F. Supp. 3d at 178. Sutter and PAMF provide no support for their assumption that CMS's HCC model divides total costs by total diagnosis codes (as opposed to using a regression model that correctly identifies that diagnosis codes (for a tibia fracture, a fibula fracture, or both) all result in the same $100 X-ray cost). Assuming the regression model and the $100 cost, CMS would pay an MA Participant $100 (not $62.50), and the alleged errors in the double-coded traditional Medicare patients would not change those payment rates. Under this scenario, it is not inevitable that CMS pays less for MA plan beneficiaries than for traditional Medicare beneficiaries.

See Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 13.

"A regression analysis is ‘a common statistical tool designed to isolate the influence of one particular factor — e.g., sex — on a dependent variable — e.g., salary.’ " EEOC v. Gen. Tel. Co. of Nw., Inc. , 885 F.2d 575, 577 n.3 (9th Cir. 1989) (internal brackets and ellipses omitted) (quoting Sobel v. Yeshiva Univ. , 839 F.2d 18, 21–22 (2d Cir. 1988) ). More specifically, a regression analysis is " ‘[t]he use of an algebraic formula to express the influence of one or more independent variables (e.g., racial status) on the average level of a dependent variable (e.g., selection rate).’ " Id. (internal ellipses omitted) (quoting David Baldus and James Cole, Statistical Proof of Discrimination 357 (1980)).

In sum, the arguments do not persuade that CMS inevitably pays less for MA Plan beneficiaries than for traditional Medicare beneficiaries, thereby violating the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i).

1.2 "Actuarial Equivalence" Is Not a Defense to an FCA Claim

In any event, actuarial equivalence is not a defense to an FCA claim: it does not authorize MA Participants to report false diagnosis codes or keep (and not report and return) overpayments predicated on the false codes, and it does not vitiate scienter.

1.2.1 MA Participants cannot retain payments predicated on false diagnosis codes

Even if CMS pays less for beneficiaries in MA plans than it pays for traditional Medicare beneficiaries (thereby violating the "actuarial equivalence" requirement of 42 U.S.C. § 1395w-23(a)(1)(C)(i) ), controlling Ninth Circuit authority establishes that this is not a defense that permits MA Participants such as Sutter and PAMF to submit false claims based on false diagnosis codes and to keep (and not report and return) inflated payments predicated on the false claims.

42 U.S.C. § 1320a-7k requires an MA Participant to report and return "overpayments." "Overpayments" are "any funds that a person receives or retains under [Medicare or Medicaid] to which the person, after applicable reconciliation, is not entitled." 42 U.S.C. § 1320a-7k(d)(2)(A), (d)(4)(B). Necessarily, then, an MA Participant must be entitled to a payment to keep it. Ninth Circuit precedent establishes the following. A properly documented medical record must support each diagnosis code. Swoben , 848 F.3d at 1168 (citations omitted). An MA Participant should withdraw unsupported diagnosis codes. Id. at 1177 n.8 ; see id. at 1174 (rejecting the argument that MA Participants are not obliged to withdraw unsupported diagnosis codes). Payments based on unsupported diagnosis codes are "improperly inflated." Silingo , 904 F.3d at 673. The UnitedHealthcare court similarly noted that for more than a decade, CMS has "required repayment to CMS of any costs that were based on unsupported diagnosis codes." UnitedHealthcare II , 330 F. Supp. 3d at 180. These cases compel the conclusion that MA Participants are not entitled to payments predicated on unsupported diagnosis codes. And if MA Participants are not entitled to payments predicated on unsupported diagnosis codes, then the payments are "overpayments" subject to the reporting-and-return requirements of § 1320a-7k.

Sutter and PAMF do not challenge the "applicable reconciliation" provision and do not contend that payments to them are not overpayments because "applicable reconciliation" had not taken place.

MA Participants dispute whether CMS should require them to return payments based on extrapolating RADV-audit results across non-audited diagnosis codes, but there is no dispute that CMS has long required MA Participants to return all payments based on the unsupported diagnosis codes identified during audits. UnitedHealthcare II , 330 F. Supp. 3d at 180.

A contrary conclusion would permit MA Participants to submit or not report false diagnosis codes to adjust for a payment rate that is less than the rate for traditional Medicare beneficiaries. That would defy the aphorism that " ‘[m]en must turn square corners when they deal with the government.’ " PAMC, Ltd. v. Sebelius , 747 F.3d 1214, 1217 (9th Cir. 2014) (quoting Rock Island, A. & L.R. Co. v. United States , 254 U.S. 141, 143, 41 S.Ct. 55, 65 L.Ed. 188 (1920) ). And "[i]f the [defendants] did indeed knowingly submit false claims in order to receive payment ... not covered under the [existing] rule, the invalidity of the rule will be no defense." Cedars-Sinai Med. Ctr. v. Shalala , 125 F.3d 765, 769 (9th Cir. 1997). In Cedars-Sinai , hospitals submitted claims to Medicare and Medicaid for reimbursement of non-FDA-approved medical devices. Id. at 767. A Health Care Financing Administration ("HCFA") policy barred coverage. Id. In a separate qui tam FCA lawsuit, a relator alleged that the hospitals knowingly submitted the false claims. Id. The hospitals challenged the HCFA policy under the APA in a separate lawsuit. Id. In part, the district court held that the issues in the qui tam case and the APA case were different. Id. at 769. On appeal, the Ninth Circuit affirmed, holding that the invalidity of the policy was not a defense to the submission of false claims: " ‘One who elects fraud as a means of self-help may not escape the consequences by urging that his conduct be excused because the [rule] which he sought to evade is invalid.’ " Id. (internal brackets omitted) (quoting Bryson v. United States , 396 U.S. 64, 68, 90 S.Ct. 355, 24 L.Ed.2d 264 (1969) ).

This precedent establishes that even if CMS's payment methodology does not ensure "actuarial equivalence" between payments for healthcare under traditional Medicare and payments for health care under MA Plans, and "were declared invalid, ... that will be no defense to the [ ] claims under the False Claims Act." Id. (citing United States v. Weiss , 914 F.2d 1514, 1522–23 (2d Cir. 1990) ); accord UnitedHealthcare I , 255 F. Supp. 3d at 211 ("[w]hether a government contractor knowingly engaged in fraud [an FCA claim], and whether a government agency appropriately promulgated a rule years later [an APA claim], are simply too different from one another" to warrant staying one action in lieu of the other). Also, the defendants' proposed pleadings standard — requiring the government to allege facts establishing that the defendants' "error rate" of unsupported codes is higher than the "error rate" in the traditional Medicare fee-for-service program — is not obviously accomplishable at the pleadings stage and would make it unworkable for the government to pursue reverse-FCA claims against MA Participants. This is not what Congress intended. Cf. Kane ex rel. United States v. Healthfirst, Inc. , 120 F. Supp. 3d 370, 390 (S.D.N.Y. 2015) (rejecting the defendants' narrow interpretation of 42 U.S.C. § 1320a-7k because it placed an unworkable burden on the plaintiffs at the pleadings stage).

MA Participants know in advance what they will be paid: each year, CMS announces well in advance the annual Medicare Advantage capitation rate and the risk and other factors to be used in adjusting such rate. 42 U.S.C. § 1395w-23(b)(1)(B) ; see, e.g. , Announcement for Calendar Year (CY) 2014 Medicare Advantage and Part D Payment Policies . If MA Participants think that they should be paid more or that the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i) requires CMS to pay them more, they can raise that objection with CMS.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 14, 17–18.

Here, the defendants propose no workable methodology to define or compare an MA Participant's "error rate" against the traditional-Medicare error rate. The error rate is not simply total unsupported diagnosis codes divided by total diagnosis codes, in part because some errors are serious and some are not. Put another way, the severity of the code and the magnitude of the error matter. For example, miscoding the wrong broken bone (e.g., coding a patient who broke his tibia as having broken his fibula ) is different than falsely coding a cancer diagnosis (e.g., coding a healthy patient as having lung cancer ).

In addition, because it is impracticable to review every single diagnosis code that an MA Participant submits (much less every diagnosis code in CMS's traditional Medicare data), any calculation of "error rates" necessarily must involve sampling and extrapolation — which in turn necessitates some mechanism to select appropriately representative samples and then give the samples their appropriate weight. Sutter and PAMF offer no mechanism for doing so.

Sutter's and PAMF's proposed pleading standard would mean that the government could not sue an MA Participant — even if it had proof that all of the MA Participant's codes for, say, pneumonia were false, and the MA Participant refused to repay — unless it (1) took an appropriate sample of all of the codes the MA Participant submitted across its contract with CMS, including codes unrelated to the fraudulent-pneumonia-coding scheme, (2) obtained the underlying medical records, (3) compared the codes and the records to identify and weight the extent and seriousness of the errors, (4) extrapolated the errors over the MA Participant's contract with CMS, (5) took an appropriate sample of all of its traditional Medicare data, (6) obtained the underlying medical records, which might require it to seek medical records from hundreds of separate third-party traditional Medicare providers, (7) compared the codes and the records to identify and weight the extent and seriousness of the errors, (8) extrapolated the errors over the MA Participant's contract with CMS, and (9) compared the resulting error rates. Even after the pleadings stage, the case would get swallowed by disputes about sample sizes, extrapolation, and error-rate calculations, thereby preventing the government from litigating the underlying fraudulent-pneumonia-code scheme and frustrating Congress's intent for robust FCA enforcement to prevent healthcare fraud. Cf. Kane , 120 F. Supp. 3d at 390.

Cf. Gov't Compl. – ECF No. 41 at 33 (¶ 96) (PAMF physician complaining to PAMF Physician Champion and others in PAMF management that "changing a diagnosis from acute bronchitis to pneumonia is not a simple or unimportant change" and "it is so obviously unethical").

Cf. Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 21 (arguing that the government must "make allegations about the overall rate of unsupported diagnosis codes in Defendants' data" to plead an FCA claim) (emphasis in original).

As the Supreme Court has said, "Congress ‘does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions — it does not, one might say, hide elephants in mouseholes.’ " Epic Sys. Corp. v. Lewis , ––– U.S. ––––, 138 S. Ct. 1612, 1626–27, 200 L.Ed.2d 889 (2018) (quoting Whitman v. Am. Trucking Ass'ns, Inc. , 531 U.S. 457, 468, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001) ). Nothing suggests that the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i) — part of a statute that does not mention the FCA — imposes a pleading or proof standard for FCA cases. "It's more than a little doubtful that Congress would have tucked into the mousehole of [ 42 U.S.C. § 1395w-23(a)(1)(C)(i)'s actuarial-equivalence provision] an elephant that tramples the work done by [the FCA]." Cf. id. at 1627. That conclusion is bolstered by Congress's subsequent 2010 enactment, as part of the ACA, of "Enhanced Medicare and Medicaid Program Integrity Provisions" that expand, not contract, the scope of FCA liability. Cf. Kane , 120 F. Supp. 3d at 391 ("Each time Congress has weighed in on the purpose and power of the FCA, it has endorsed a reading of that statute as a robust, remedial measure aimed at combatting fraud against the federal government as firmly as possible.").

In Epic Systems , the Supreme Court upheld under the Federal Arbitration Act ("FAA") an arbitration clause in an employment contract that waived the right to collective litigation against the employer and held that Section 7 of the National Labor Relations Act did not alter the FAA. Epic Sys. , 138 S. Ct. at 1626–27.

Swoben also supports this result. In Swoben , the MA Organizations allegedly submitted certifications regarding unsupported diagnosis codes. 848 F.3d at 1166. CMS audits put them on notice that their submissions likely included a significant number of erroneous codes. Id. at 1167. Under the relevant regulations, MA Organizations may (but are not required to) conduct retrospective reviews of their enrollees' medical records to ensure the accuracy of the diagnosis codes. Id. at 1169. According to the relator, the MA Organizations conducted one-sided reviews to capture under-reporting errors but not over-reporting errors, resulting in inflated capitation payments, and rendering their periodic certifications false, in violation of the FCA. Id. at 1170. The MA Organizations argued (in part) that they reasonably believed that they were not required to take affirmative steps to find unsupported diagnosis codes because (1) 42 C.F.R. § 422.310(d) provided only that MA Organizations must submit "data that conform to CMS'[s] requirements for data equivalent to [traditional] Medicare fee-for-service data," and (2) CMS does not verify diagnosis codes from traditional Medicare providers, and the "data equivalent" requirement in the regulation meant that MA Organizations did not have to verify their diagnosis codes either. Id. at 1179. The Ninth Circuit rejected the argument: "because nothing in § 422.310(d) ['s ‘data equivalent’ requirement] speaks to a Medicare Advantage organization's obligations to ensure the accuracy of risk adjustment data, it does not modify a Medicare Advantage organization's obligations" to submit accurate diagnosis codes under the relevant regulations. Id. (citing 42 C.F.R. §§ 422.503(b)(4)(vi) & 422.504(l ) ). Like the "data equivalent" regulation in Swoben , nothing in 42 U.S.C. § 1395w-23(a)(1)(C)(i)'s "actuarial equivalence" provision speaks to an MA Participant's obligation to ensure the accuracy of its diagnosis codes, and thus nothing in the provision modifies an MA Participant's obligation to ensure that it is submitting accurate diagnosis codes or allows it to keep (and not report and return) inflated payments predicated on false diagnosis codes.

Sutter and PAMF distinguish Swoben on the ground that it did not address the "actuarial equivalence" provision in 42 U.S.C. § 1395w-23(a)(1)(C)(i). Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 8–9 & n.3 (citing United States ex rel. Poehling v. UnitedHealth Grp., Inc. , No. CV 16-8697-MWF (SSx), 2019 WL 2353125 (C.D. Cal. Mar. 28, 2019) (Poehling II )). But the reasoning for the holding in Swoben (a regulation requiring data equivalence reported does not alter the obligation to report data accurately) applies here: the requirement of "actuarial equivalence" between Medicare Advantage coverage and traditional Medicare coverage does not alter the accurate-reporting obligation either. The failure to discuss "actuarial equivalence" does not distinguish Swoben .
Sutter and PAMF also contend that Swoben does not control because the direct-FCA claims there were based on false certifications of accuracy, completeness, and truthfulness under 42 C.F.R. § 422.504(l ), and the direct-FCA claims here are based on false diagnosis codes, not certifications. Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 8–9. This argument does not change the outcome. The Swoben court did not limit its holding — that MA Participants must submit accurate diagnosis codes and withdraw unsupported codes — to false certifications and compels the conclusion that MA Participants cannot bill CMS for false diagnosis codes (and keep and not report the resulting overpayments). Swoben , 848 F.3d at 1168, 1774, 1777 n.8 ; accord Silingo , 904 F.3d at 673 (submitting unsupported codes results in "improperly inflated" payments).

1.2.2 The "actuarial equivalence" requirement does not vitiate scienter

Sutter and PAMF — quoting Safeco Insurance Co. of America v. Burr , 551 U.S. 47, 70 n.20, 127 S.Ct. 2201, 167 L.Ed.2d 1045 (2007) — also contend that even if one rejects their actuarial-equivalence argument, their interpretation of that requirement vitiates scienter because " ‘Congress could not have intended’ to subject a defendant to liability under a knowing or reckless standard where the defendant ‘followed an interpretation that could reasonably have found support in the courts, whatever their subjective intent might have been.’ " This argument fails.

Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 6, 14–16.

In Safeco , the Supreme Court said, "[w]here ... the statutory text and relevant court and agency guidance allow for more than one reasonable interpretation, it would defy history and current thinking to treat a defendant who merely adopts one such interpretation as a knowing or reckless violator. Congress could not have intended such a result for those who followed an interpretation that could reasonably have found support in the courts, whatever their subjective intent may have been." Safeco , 551 U.S. at 70 n.20, 127 S.Ct. 2201. By contrast, "even if a regulated party adopts a ‘reasonable’ interpretation of an ‘ambiguous’ statute, it can nonetheless be deemed liable for knowingly making a false statement if it ‘had been warned away from that interpretation’ by authoritative agency guidance." Ipsen Biopharmaceuticals, Inc. v. Azar , 943 F.3d 953, 957–58 (D.C. Cir. 2019) (quoting United States ex rel. Purcell v. MWI Corp. , 807 F.3d 281, 288 (D.C. Cir. 2015) ); accord Swoben , 848 F.3d at 1178 (rejecting MA Organizations' argument that their interpretation of the regulation in question was objectively reasonable in light of "CMS'[s] clear statements in the [regulation's] preamble," which "resolved any ambiguity" against the MA Organizations).

CMS regulations have long warned MA Participants that they must certify that the data they submit is accurate, complete, and truthful, 42 C.F.R. § 422.504(l )(3), that they will be required to submit a sample of medical records for data validation, 42 C.F.R. § 422.310(e), and that there may be penalties for the submission of false data, id. Additionally, for more than a decade, CMS has "required repayment to CMS of any costs that were based on unsupported diagnosis codes." UnitedHealthcare II , 330 F. Supp. 3d at 180. No court — including the UnitedHealthcare court — has held that an MA Participant can keep and not report or return payments based on unsupported codes. Cf. id. at 189 ("UnitedHealth does not contend that Medicare Advantage insurers should be permitted knowingly or recklessly to bill CMS for erroneous diagnosis codes.").

The UnitedHealthcare decision issued in 2018, i.e., after the core time period of the defendants' alleged misconduct. See Gov't Compl. – ECF No. 41 at 6 (¶ 10), 45 (¶ 131). Before 2018, agency guidance warned medical providers that they could not keep (and must report or return) overpayments based on false diagnosis codes. 42 C.F.R. §§ 422.310(e), 422.504(l )(3) ; see also, e.g. , 79 Fed. Reg. at 29,921 ("For example, a risk adjustment diagnosis that has been submitted for payment but is found to be invalid because it does not have supporting medical record documentation would result in an overpayment."). Sutter's written policies and procedures acknowledge this: they state that Sutter would report and return any payment based on (among other things) an "incorrect code or modifier assignment resulting in a higher level of reimbursement" or "insufficient or lack of documentation to support billed services." See supra notes 294–296. Safeco addresses circumstances where the statutes, regulations, and court decisions allow for more than one reasonable interpretation. Safeco , 551 U.S. at 70 n.20, 127 S.Ct. 2201. That is not the case here. The 2018 UnitedHealthcare decision does not vitiate the defendants' scienter at the time of their alleged misconduct. See Halo Elecs., Inc. v. Pulse Elecs., Inc. , ––– U.S. ––––, 136 S. Ct. 1923, 1933, 195 L.Ed.2d 278 (2016) ("Nothing in Safeco suggests that [courts] should look to facts that the defendant neither knew or had reason to know at the time he acted."); accord Cedars-Sinai , 125 F.3d at 769 ("Even if the Hospitals succeed in having the rule [later] declared invalid, however, that will be no defense to the Relator's claims under the False Claims Act.") (citing Weiss , 914 F.2d at 1522–23 ). In any event, as discussed above, the UnitedHealthcare court did not hold that MA Participants could keep (and not report or return) overpayments based on false diagnosis codes. Cf. UnitedHealthcare II , 330 F. Supp. 3d at 180, 189. To the contrary, it acknowledged that CMS requires repayment of any costs predicated on unsupported diagnosis codes. Id. at 180.

In sum, the actual-equivalence argument does not vitiate scienter.

* * *

Sutter and PAMF complain that "the government ... attempt[s] to impose a heightened standard on Medicare Advantage providers that it has never imposed in traditional Medicare." It does not. Traditional Medicare providers and Medicare Advantage providers both must submit accurate data to CMS. The government brings FCA claims against both if they submit false claims that result in overpayments or fail to return overpayments. Traditional Medicare providers — who submit claims based on medical services — are liable if they (with the requisite scienter) submit false bills for medical services (or fail to return overpayments). Medicare Advantage providers — who submit claims based on diagnosis codes — also are liable if they submit (with the requisite scienter) false or unsupported codes (or fail to return overpayments). The court denies the motion to dismiss on the ground that "actuarial equivalence" is not a defense to an FCA claim.

Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 5.

Sutter and PAMF argue if the court were to dismiss the FCA claims, the government would not lack resources: "[i]t is free to use administrative and other mechanisms to pursue reimbursement when it believes that it has overpaid a participant in the Medicare Advantage program." Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 24. But that approach shifts the burden for ascertaining the accuracy of diagnosis codes to CMS (presumably through its audits) and away from MA Participants. That is not what the statutory scheme requires. Instead, MA Participants must certify that their diagnosis codes and data are true and accurate, and MA Participants must report and return any overpayments. See supra notes 17–25, 41–47 and accompanying text. CMS audits MA participants to confirm the accuracy of the diagnosis coding. See supra note 44 and accompanying text. That audit process does not excuse MA Participants' obligation to report diagnosis codes and overpayments accurately in the first instance.

This holding means that going forward (including through discovery, motions, or trial), this case is not about "actuarial equivalence" calculations or the overall error rates in CMS's traditional Medicare data. Instead, and unlike the APA dispute in UnitedHealthcare , the case is about whether Sutter and PAMF submitted false diagnosis codes to CMS or failed to report and return overpayments predicated on the false codes. Cf. Cedars-Sinai , 125 F.3d at 769 (invalidity of applicable rule "will be no defense to ... claims under the False Claims Act"); UnitedHealthcare I , 255 F. Supp. 3d at 211 ("any decision in this [APA] matter will not answer the most relevant questions in the FCA Cases[, namely, w]hether a government contractor knowingly engaged in fraud").

2. The Relator Can Pursue Her Sutter-Wide Claims

The government's complaint challenges PAMF's unsupported diagnosis codes, but it did not allege FCA or other violations at Sutter affiliates other than PAMF. Ms. Ormsby filed a First Amended Complaint after the government intervened in the action, adding additional allegations about Sutter affiliates other than PAMF (as described above in the Statement). The government previously settled with Sutter and its non-PAMF affiliates for $30 million in a non-FCA settlement. It does not object to Ms. Ormsby's expansion of the FCA claims to non-PAMF Sutter affiliates. Sutter moves to dismiss Ms. Ormsby's complaint on the ground that she cannot maintain a broader FCA action than the government's FCA action, and that her lawsuit thus is limited to PAMF (and cannot extend to Sutter's non-PAMF affiliates).

Gov't Compl. – ECF No. 41; Hr'g Tr. – ECF No. 97 at 18 ("[THE GOVERNMENT:] If you look at the claims in the respective [latest] complaints and we're dealing with, as you already noted, PAMF and Sutter as to PAMF[.]").

Relator FAC – ECF No. 52.

Id. at 3–4 (¶ 7); see Relator Req. for Judicial Notice Ex. B (Settlement Agreement) – ECF No. 81 at 10–22.

Hr'g Tr. – ECF No. 97 at 18 ("[THE GOVERNMENT:] .... [W]e think, just like Judge Donato did in [United States ex rel. Jahr v. Tetra Tech EC, Inc. , No. 3:13-cv-03835-JD (N.D. Cal. filed Aug. 19, 2013)], that we don't have a problem with [the relator] pursuing those claims.").

Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 5–7. Ms. Ormsby's initial complaint focused on PAMF, but she alleged too that there was a system-wide failure at Sutter. Relator Compl. – ECF No. 1 at 4–5 (¶¶ 7–8) (system-wide failure), 15–32 (¶¶ 51–115) (PAMF).

Practically, the government's intervention means that it has the primary responsibility for prosecuting the action, see 31 U.S.C. § 3730(c)(1), and the litigation in intervened cases generally focuses initially on the government's case (here, PAMF). That said, the government's limit of its case to PAMF does not bar Ms. Ormsby's claims regarding Sutter's non-PAMF affiliates. The court thus denies Sutter's motion to dismiss Ms. Ormsby's claims.

Under the FCA, a relator may bring a civil FCA lawsuit on behalf of the government, and thereafter, the government may elect to intervene and proceed with the action. 31 U.S.C. § 3730(b)(1)–(2). If the government intervenes, it "shall — (A) proceed with the action, in which case the action shall be conducted by the government; or (B) notify the court that it declines to take over the action, in which case the person bringing the action shall have the right to conduct the action." 31 U.S.C. § 3730(b)(4).

When the government intervenes in a case, it becomes a party to the lawsuit as a whole, not merely a party to particular claims:

[ 31 U.S.C. § 3730 ] states that "[t]he Government may elect to intervene and proceed with the action within 60 days after it receives both the complaint and the material evidence and information." It does not state that the government may intervene in part of the action or as to certain counts or certain claims for relief. Neither does it state that the government declines party status in those claims it chooses not to prosecute or settle.... [T]he government becomes a "party" to the suit as a whole when it intervenes. It does not become a "party" to a particular claim or number of claims."

United States ex rel. Bennett v. Biotronik, Inc. , 876 F.3d 1011, 1020–21 (9th Cir. 2017).

After the government intervenes, it has primary responsibility for prosecuting the action "and shall not be bound by an act of the person bringing the action," meaning, the relator. 31 U.S.C. § 3730(c)(1). The relator has the right to continue as a party in the action. Id. But § 3730(c) limits the relator's participation and essentially gives the government the authority to run the plaintiffs' show. Id. In summary form, the limitations to the relator's participation are as follows: (1) the government may dismiss the case over the relator's objections (on the government's motion and after the court provides the relator the opportunity for a hearing on the motion); (2) the government may settle the action over the relator's objections (after the court determines, after a hearing, that the settlement is fair, adequate, and reasonable under all the circumstances); (3) the court — upon the government's showing that the relator's participation causes undue delay or is repetitious, irrelevant, and harassing — may impose limits on the relator's participation (such as limiting the number of witnesses, limiting the length of testimony, limiting cross-examination, and otherwise limit participation); and (4) the court — upon a defendant's showing that the relator's conduct is harassing or would cause the defendant undue burden or expense — may limit the relator's participation. 31 U.S.C. § 3730(c)(2).

Section 3730 demonstrates that the government is a party to the case and has primary responsibility for prosecuting it. But it does not address whether a relator may pursue claims when the government does not. That said, nothing in § 3730 suggests that the government's intervention bars the relator from pursuing claims beyond those in the government's complaint, and the weight of authority suggests that she can.

Section 3730 does not require the government — on intervention — to pursue all of the relator's claims. Bennett , 876 F.3d at 1020 (after the government intervenes in an FCA case, the case may include "claims [the government] chooses not to prosecute or settle"). As then-Judge Samuel Alito wrote for the Third Circuit:

When a qui tam action is filed, the government may "proceed with the action ,"

§§ 3730(b)(2) and (4) (emphasis added) or "decline to take over the action ," § 3730(b)(4)(B) (emphasis added), but the government often decides to take over only certain claims in a multi-claim action, and we are aware of no decision holding that this is improper.

United States ex rel. Merena v. SmithKline Beecham Corp. , 205 F.3d 97, 102 (3d Cir. 2000) (emphasis in original). Indeed, the government need not intervene at all, and the relator can prosecute the case. 31 U.S.C. § 3730(b)(4)(B). The statute and case law thus contemplate that the government can pursue some or all of the relator's claims, and the relator can pursue claims when the government does not.

In cases where the government intervenes in part, necessarily there are other claims that the relator has. Courts regularly allow relators to pursue their separate claims after the government's intervention. See, e.g. , United States ex rel. Ketroser v. Mayo Found. , 729 F.3d 825, 826 (8th Cir. 2013) (the relator filed an FCA case, the government intervened in part and settled the intervened claim, and the relator then filed an amended complaint asserting additional claims against the defendant); United States ex rel. Fallon v. Accudyne Corp. , 97 F.3d 937, 938 (7th Cir. 1996) ("The Attorney General took over the prosecution of Count I, see 31 U.S.C. § 3730(b)(4)(A), but left Count II in the hands of the relators."); Fed. Recovery Servs., Inc. v. United States , 72 F.3d 447, 449 n.1 (5th Cir. 1995) ("The United States elected to intervene in that portion of the suit against [some] defendants but declined to intervene against [another defendant]. Although the United States' intervention vested it with control of the litigation against [the first group of defendants], [relator] retained the authority to proceed against [the other defendant] on its own.").

To the extent that Sutter argues that the government's intervention bars Ms. Ormsby's claims outright, courts also reject arguments to dismiss the complaints as a matter of course. See, e.g. , United States ex rel. Dresser v. Qualium Corp. , No. 5:12-cv-01745-BLF, 2016 WL 3880763, at *10 (N.D. Cal. July 18, 2016) ("Defendants argue that if [relator] has any claims that overlap with the United States' claims, those claims should be dismissed because [relator] has no cause of action under the FCA once the United States intervenes .... Defendants' view of the FCA is contrary to the text of the statute, which gives relators the right to continue as a party to an FCA action even when the United States chooses to intervene."); United States ex rel. Shemesh v. CA, Inc. , 89 F. Supp. 3d 36, 55–56 (D.D.C. 2015) ("[D]ismissal [of the relator's complaint] is not automatically triggered by the government's intervention. ‘[T]here is no presumption under the statute against allowing both complaints to proceed.’ ") (quoting United States ex rel. Landis v. Tailwind Sports Corp. , 51 F. Supp. 3d 9, 28 (D.D.C. 2014) ).

Courts also allow relators to amend their complaints after the government intervenes. See, e.g. , Dresser , 2016 WL 3880763, at *1 ("After the United States intervened in part and filed an Intervenor Complaint, [the relator] amended her complaint."); Shemesh , 89 F. Supp. 3d at 43 (noting that the government intervened in March 2014 and the relator amended his complaint in April 2014).

Sutter acknowledges the many cases that allow "relators to litigate individual claims for relief under the False Claims Act after the government intervenes." But it contends that the court should follow the only case — United States ex rel. Brooks v. Stevens-Henager College, Inc. , 359 F. Supp. 3d 1088 (D. Utah) — where a district court held that relators have no right to maintain the non-intervened portion of an FCA case after the government intervenes. Id. at 1120 ("relators may not maintain the non-intervened portion of an action") (capitalization removed). Sutter asserts that Brooks rightly identified the unique procedural context of a relator's amendment after the government's intervention and appropriately held that a relator's "limited right to continue as a party to the action ... does not allow the relator to amend his or her complaint to add defendants and claims to the Government's action" because "[t]hose rights necessarily belong to the party with the primary responsibility for conducting the action — in this case, the Government." Id. at 1116. "The Government's complaint in intervention superseded the relators' amended complaint, and any pleading subsequently filed by the relator lacked legal effect." Id.

Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 8.

Id.

The Brooks court reached its decision in part because the FCA does not say that a relator can prosecute the non-intervened part of the complaint, and "Congress'[s] silence as to a relator's right to prosecute the non-intervened claims leads to the conclusion that no such right exists." Id. at 1120. It examined the structure and legislative history of the FCA and found that the statute's use of the word "action" meant "civil action" and not "cause of action." Id. at 1118, 1121–26.

Other courts reach the contrary conclusion that the word "action" in the FCA means "cause of action" or claim. For example, in Merena , then-Judge Alito wrote that "the draftsmanship of the qui tam statute has its quirks, and one of those quirks is that the statute is based on the model of a single-claim complaint.... [T]he qui tam statute is phrased as if every qui tam complaint contained only one claim." Merena , 205 F.3d at 101–02 (citing United States ex rel. Mistick v. Hous. Auth. , 186 F.3d 376, 387 (3d Cir. 1999) ). He then examined the qui tam statute's use of the word "action" and concluded that the "each claim in a multi-claim [FCA] complaint must be treated as if it stood alone." Id. at 102 ; accord United States ex rel. Schumann v. AstraZeneca Pharm. L.P. , 769 F.3d 837, 846 (3d Cir. 2014) ("FCA's reference to ‘action’ may reasonably be read to mean ‘claim’ because the statute envisions a single-claim complaint") (citing Merena , 205 F.3d at 101–02 ); United States ex rel. Rauch v. Oaktree Medical Centre, P.C. , No. 6:15-cv-01589-DCC, 2020 WL 1065955, at *6–9 (D.S.C. Mar. 5, 2020) (declining to follow Brooks , applying Merena , noting that the Brooks court (and not the defendants) sua sponte raised the issue, determining that the word "action" meant "cause of action," and holding that the FCA allowed relators to pursue non-intervened claims) (citing Merena , 205 F.3d at 101–02 ). Given the clear weight of authority that allows a relator to pursue non-intervened claims, the court follows that approach (and not Brooks ) as persuasive. The government can pursue some or all of a relator's claims, and a relator can pursue claims that government does not.

The Brooks court held that the word "action" 31 U.S.C. § 3730 cannot mean "claim" or "cause of action" because 31 U.S.C. § 3730(e)(4)(A) refers to "an action or claim under this section," thereby distinguishing the term "action" from a "claim." Brooks , 359 F. Supp. 3d at 1118 (emphasis in original). At the time that 31 U.S.C. § 3730 was amended, in 1986, to first allow relators to continue as parties after the government intervened, it did not contain any reference to "an action or claim under this section." See Pub. L. No. 99-562, § 2, 100 Stat. at 3154–57. Instead, it appeared to use "action" and "claim" somewhat interchangeably. See id. at 3156 (awarding relator "at least 15 percent but not more than 25 percent of the proceeds of the action or settlement of the claim"). The reference to "an action or claim under this section" in what is now 31 U.S.C. § 3730(e)(4)(A) was not added until 24 years later with the passage of the ACA. See Pub. L. No. 111-148, 124 Stat. at 901.

A contrary decision not only is inconsistent with the many decisions that allow relators to pursue non-intervened claims but also is inconsistent with the FCA. The FCA allows the government and private parties to pursue civil actions for false claims. 31 U.S.C. § 3730(a) – (b). It allows the government to pursue some or all claims. Bennett , 876 F.3d at 1020. A relator can pursue claims if the government does not intervene. 31 U.S.C. § 3730(c)(3). The government's failure to intervene on any or all claims does not mean that the claims lack merit (or that the government thinks that they do). The decision to intervene can turn on "any number of reasons. For example, a decision not to intervene may ‘not necessarily be an admission by the United States that it has suffered no injury in fact, but rather the result of a cost-benefit analysis.’ " United States ex rel. Williams v. Bell Helicopter Textron Inc. , 417 F.3d 450, 455 (5th Cir. 2005) (internal brackets omitted) (quoting United States ex rel. Berge v. Bd. of Trs. , 104 F.3d 1453, 1458 (4th Cir. 1997) ). Also, for resource reasons, the government cannot pursue every meritorious claim. If relators cannot pursue non-intervened claims, then there is no recourse at all. Surely the FCA does not require that result.

3. The Plaintiffs Sufficiently Plead Their Claims

Sutter and PAMF move to dismiss the complaints on the grounds that (1) for the reverse-FCA claims, the plaintiffs failed to allege (with the particularity that Rule 9(b) requires) that they identified any overpayments or knowingly avoided repaying them and (2) for the direct-FCA claims, the plaintiffs failed to plead that they knowingly submitted materially false claims or statements. The court denies the motions to dismiss because the plaintiffs sufficiently pleaded these elements of the claims.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 9, 21–28; Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 6, 10–13.

3.1 Reverse-FCA Claims

The elements of the reverse-FCA claims are that Sutter and PAMF (1) concealed or improperly avoided or decreased an obligation to pay the government and (2) did so knowingly. 31 U.S.C. § 3729(a)(1)(G).

Because the plaintiffs sufficiently plead a claim under the second prong of the reverse-FCA provision, the court need not address whether they plead a claim under the first prong, i.e., whether Sutter and PAMF made or used a material false statement to conceal or avoid an obligation to pay the government. Victaulic , 839 F.3d at 255 ; see also Gov't Opp'n – ECF No. 82 at 19–21 (addressing only the sufficiency of the complaint under the second prong of the reverse-FCA statute).

3.1.1 Concealing or avoiding an obligation to pay the government

As discussed above, 42 U.S.C. § 1320a-7k requires MA Participants to return Medicare overpayments within 60 days after the overpayments are "identified." If they do not, the overpayments become "obligations" under the reverse-FCA provision. 42 U.S.C. § 1320a-7k(d)(2)(A). Payments based on false diagnosis codes are overpayments. Sutter and PAMF move to dismiss on the ground that the plaintiffs have not alleged that they "identified" any overpayments and failed to return them within 60 days. Because the plaintiffs sufficiently allege that internal and external reviews of diagnosis codes put the defendants on notice of the potential overpayments, the court denies the motion to dismiss on this ground.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 22–24.

Section 1320a-7k uses the word "identified" but does not define it further. Another court — in the similar context of an employee's identifying potential overpayments and telling his employer — conducted a comprehensive analysis of the statute and held that an overpayment is "identified" when the "provider is put on notice of a potential overpayment, rather than the moment when an overpayment is conclusively ascertained." Kane , 120 F. Supp. 3d at 387–88 (analyzing the text, legislative history, and legislative purpose of § 1320a-7k and the FCA). Kane involved a software glitch that resulted in the submission of improper claims to Medicaid. Id. at 375. The New York State Comptroller alerted the defendants about the glitch and identified specific wrongful claims, and an employee then put them on notice of a set of claims likely to contain numerous overpayments. Id. at 388. At this point, the defendants were on notice, the claims were identified, and the defendants had a duty to report and return wrongly collected money. Id. at 388, 390.

The employee's list of claims was overinclusive and underinclusive: (1) roughly half of the claims on the list were not overpaid, and (2) the list omitted certain overpayments. Kane , 120 F. Supp. 3d at 377 & n.7. The court held that the list "identified" overpayments for the employer. Id. at 390.

The Kane court acknowledged that its holding — that an overpayment is "identified" when a party is put on notice of a potential overpayment — could be viewed as "unforgiving." Kane , 120 F. Supp. 3d at 389. Under such a rule, "an overpayment would technically qualify as an ‘obligation’ even where a provider receives an email [putting it on notice], struggles to conduct an internal audit, and reports its efforts to the government within the sixty-day window, but has yet to isolate and return all overpayments sixty-one days after being put on notice of potential overpayments." Id. The Kane court noted that the potential harshness of the rule was tempered by the fact that the FCA separately imposes a scienter requirement above and beyond § 1320a-7k's "identified" requirement. As the Kane court explained, "while [‘identified’ overpayments] might qualify as ‘obligations,’ the mere existence of an ‘obligation’ does not establish a violation of the FCA. Rather, in the reverse false claims context, it is only when an obligation is knowingly concealed or knowingly and improperly avoided or decreased that a provider has violated the FCA." Id. (emphasis in original). In the situation described above, where a medical provider is put on notice of a potential overpayment and makes a good-faith effort to investigate and to report and return any overpayment it finds, "the provider would not have acted with the reckless disregard, deliberate ignorance, or actual knowledge of an overpayment required to support an FCA claim." Id. at 389–90.

Other courts have followed this approach. See United States ex rel. Strauser v. Stephen L. LaFrance Holdings, Inc. , No. 18-CV-673-GKF-FHM, 2019 WL 1086363, at *16 (N.D. Okla. Mar. 7, 2019) (allegation that defendant "had notice that the price-match program was unlawful and failed to exercise reasonable diligence in investigating, reporting, and returning overpayments" pleads a claim under § 1320a-7k and the reverse-FCA provision); Graves v. Plaza Med. Ctrs., Corp. , 276 F. Supp. 3d 1335, 1347 (S.D. Fla. 2017) (" ‘[T]he sixty day clock begins ticking when the provider is put on notice of a potential overpayment, rather than the moment when an overpayment is conclusively ascertained, which is compatible with the legislative history of the FCA and the FERA [Fraud Enforcement Recovery Act].’ ") (quoting Kane , 120 F. Supp. 3d at 388 ); United States ex rel. Ortiz v. Mount Sinai Hosp. , No. 13 Civ. 4735(RMB) , 2015 WL 7076092, at *13 (S.D.N.Y. Nov. 9, 2015) (allegation that defendants had received overpayments and "had notice of those overpayments" but failed to return them pleads a claim under § 1320a-7k and the reverse-FCA provision).

Sutter and PAMF argue that the plaintiffs must allege that they "actually identified " unsupported diagnosis codes and failed to return the payments predicated on those codes within 60 days. To the extent that Sutter and PAMF argue that § 1320a-7k requires the plaintiffs to plead that they had actual knowledge of specific unsupported diagnosis codes and payments based on those codes in order for the payments to be "identified" for purposes of triggering their 60-day-return obligation, their argument fails. The defendants in the Kane action made a similar argument, asserting that the government had to allege that they "classified with certainty" overpayments in order for the overpayments to be "identified." Id. at 384. The Kane court rejected this argument, holding that:

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 22 (emphasis in original).

Defendants' interpretation ... would make it all but impossible to enforce the reverse false claims provision of the FCA in the arena of healthcare fraud. In the Government's words, "Permitting a healthcare provider that requests and receives an analysis showing over 900 likely overpayments to escape FCA liability by simply ignoring the analysis altogether and putting its head in the sand would subvert Congress's intent in amending § 3729(a)(1)(G)." Doc. 59 at 19 (citing United States v. Lakeshore Med. Clinic, Ltd. , No. 11 Civ. 00892, 2013 WL 1307013, at *4 (E.D. Wis. Mar. 28, 2013) ). Sure enough, the government's Complaint in this action alleges that Defendants, upon receiving [employee]'s email and analysis, did nothing with the set of claims he pointed out as potentially overpaid and paid back hundreds of claims only after receiving the government's CID. Gov't's Compl. ¶¶ 36, 38. If [employee]'s email did not "identify" overpayments within the meaning of the statute, there will be no recourse for the government when providers behave as [defendants] allegedly behaved here. It would be an absurd result to construe this robust anti-fraud scheme as permitting willful ignorance to delay the formation of an obligation to repay the government money that it is due.

....

Congress expressly created FCA liability for the retention of Medicaid [and Medicare] overpayments in the ACA. By requiring providers to self-report overpayments and imposing a relatively short deadline for repayments, violation of which risks the severe liability of the FCA, Congress intentionally placed the onus on providers, rather than on the Government, to quickly address overpayments and return any wrongly collected money. This reading is in line with the legislative purpose of the FCA, the 1986 FCA amendments, and the FERA, which together reflect Congress's more than 150-year commitment to deterring fraud against the federal government and ensuring that Government losses due to fraud are recouped in a timely fashion. Based on this understanding of legislative purpose, Defendants' proposed reading of the ACA would frustrate Congress's intention to subject willful ignorance of Medicaid [and Medicare] overpayments to the FCA's stringent penalty scheme.

Id. at 390–91 (emphasis in original).

Sutter and PAMF nonetheless argue that the standard is "actually identified" (rather than "put on notice and thus should have identified") and — in support of that argument — cite the first version (introduced by the House of Representatives) of what ultimately became the ACA. That version provided that parties had to return "known," rather than "identified," overpayments within 60 days, and "known" was defined as it is in the FCA (i.e., including deliberate ignorance and recklessness, not just actual knowledge). H.R. 3200, 111th Cong. § 1641 (as introduced by House, July 14, 2009). Congress ultimately enacted the Senate version, which provided that overpayments had to be reported and returned within 60 days of when they were "identified," rather than "known." Pub. L. No. 111-148, § 6402(a), 124 Stat. at 755–56. Sutter and PAMF suggest (without citing any supporting authority) that the change from "known" to "identified" reflects an intent to impose a higher standard than the FCA's "knowing" standard for triggering the 60-day-return obligation. The defendants in the Kane action made the same argument. Kane , 120 F. Supp. 3d at 386. The Kane court rejected it, noting that the legislative history of the ACA was silent as to why Congress made that change and holding that "[t]o define ‘identified’ such that the sixty day clock begins ticking when a provider is put on notice of a potential overpayment, rather than the moment when an overpayment is conclusively ascertained" better comported with the legislative history of the FCA and "Congress['s] inten[t] for FCA liability to attach in circumstances where, as here, there is an established duty to pay money to the government, even if the precise amount due has yet to be determined." Id. at 387–88.

Id. at 24.

Id.

Sutter and PAMF argue that CMS in 2014 promulgated a regulatory definition for "identified" and that the UnitedHealthcare court set aside that definition. In the absence of a controlling agency definition, the court still must construe the statutory term "identified," and the court follows as persuasive Kane , which construed the term without relying on agency interpretation. See id. at 391–93. Following Kane , an overpayment is "identified" (as the word is used in § 1320a-7k ) when the "provider is put on notice of a potential overpayment, rather than the moment when an overpayment is conclusively ascertained." Id. at 387–88.

Id. (citing UnitedHealthcare II , 330 F. Supp. 3d at 191 ).

The plaintiffs alleged sufficient facts to show that Sutter and PAMF had notice that they had submitted false diagnosis codes and had potential overpayments. By 2013 or 2014, internal reviews by Ms. Ormsby and external reviews by UnitedHealth, HealthNet, and their consultant confirmed that many of the risk-adjusting diagnosis codes they submitted were false. Sutter and PAMF recognized that "CMS is still receiving HCC's that we know are not correct." They were aware that they might have to return millions of dollars in overpayments predicated on the false diagnosis codes. Under Kane , the plaintiffs sufficiently pleaded that reviewers "identified" overpayments to Sutter and PAMF and triggered their 60-day-return obligation under § 1320a-7k(d)(2)(A). The plaintiffs also allege that Sutter and PAMF did not report or return the overpayments within 60 days of their being identified, and, thus, that the overpayments are "obligations" under the reverse-FCA provision.

See supra notes 133–147, 162–236, 359–368, 373–379.

See supra notes 225, 281, 392, 398.

See supra notes 137, 196–198, 230–232, 363, 387–388.

See supra note 288; see also Relator FAC – ECF No. 52 at 4–5 (¶ 7), 46–47 (¶ 147) (Sutter and its non-PAMF affiliates returned (pursuant to the settlement with the government) $30 million in overpayments for unsupported diagnosis codes only in 2019, i.e., at least five years after Ms. Ormsby put it on notice, and only after Ms. Ormsby filed this FCA action). As Sutter acknowledges, a plaintiff can base a reverse-FCA claim on an allegation that "the overpayments that Sutter did identify and return ... had been identified and improperly retained for more than 60 days before they were returned." Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 12. The plaintiffs have done so here.

In sum, the plaintiffs have sufficiently pleaded that Sutter and PAMF concealed or avoided obligations to pay the government.

3.1.2 Scienter

Sutter and PAMF also contend that — while Ms. Ormsby may believe that diagnosis codes were unsupported and resulted in overpayments — they do not share that assessment based on their assignment of review authority to physicians, who can evaluate diagnosis codes better. At the pleadings stage, the plaintiffs sufficiently allege that Sutter and PAMF knowingly concealed or avoided their obligations to pay the government (i.e., that they had actual knowledge or were deliberately ignorant or reckless as to whether they had an obligation to repay the government).

Id. at 23.

See supra notes 133–147, 169–286, 356–403.

Sutter and PAMF allegedly knew that about their potential liability for overpayments (totaling millions of dollars) predicated on false diagnosis codes. Sutter and PAMF point to their deletion of codes identified in the internal and external audits. But those audits addressed only limited samples, and the plaintiffs allege that the defendants were deliberately ignorant or reckless by not reviewing other risk-adjusting diagnosis codes. Allegations that a defendant " ‘has buried [its] head in the sand and failed to make simple inquiries which would alert it that false claims are being submitted’ " plead scienter. Godecke , 937 F.3d at 1212 (internal brackets omitted) (quoting Swoben , 848 F.3d at 1174 ); see Bourseau , 531 F.3d at 1168 (allegations that a defendant "has buried [its] head in the sand" plead scienter with respect to reverse-FCA claims just like direct-FCA claims).

See supra notes 137, 196–198, 230–232, 363, 387–388.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 22–23.

The court cannot address Sutter's fact challenges to the diagnosis codes on a motion to dismiss. The plaintiffs have alleged sufficiently that the defendants had notice of the unsupported diagnosis codes. On a motion to dismiss, the court accepts those allegations as true. Godecke , 937 F.3d at 1210 ("a motion to dismiss is too early a stage to render a judgment on the reliability of [relator]'s recollections"). Also, while Sutter and PAMF contend that they appropriately assigned the authority to delete codes to physicians, the plaintiffs allege that Sutter and PAMF knew, were deliberately ignorant, or were reckless about the physicians' failure to delete codes for reasons such as time pressure or inattention, lack of knowledge about how to delete codes, and pressure by Sutter and PAMF to keep their coding levels high. By assigning the authority to delete codes to physicians who they knew would not actually delete codes, Sutter and PAMF knowingly concealed and avoided their obligations to repay the government. In sum, the plaintiffs have sufficiently pleaded Sutter's and PAMF's scienter.

Id.

See supra note 253.

See supra notes 118–119.

See supra notes 76–113, 119–125.

3.2 Direct-FCA Claims

The elements of the direct-FCA claims are that (1) Sutter and PAMF submitted false claims for payment or used false records or statements, (2) they did so knowingly, and (3) the false claims were material and (4) caused the government to pay out money. 31 U.S.C. § 3729(a)(1)(A)–(B) ; Godecke , 937 F.3d at 1208. Sutter and PAMF move to dismiss the direct-FCA claims on the ground that the plaintiffs (1) do not identify any diagnosis code that Sutter and PAMF identified as false when they submitted it and (2) instead rely on inadequate audits and red flags. Because the plaintiffs sufficiently plead false claims, scienter, and materiality (the elements at issue), the argument fails.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 25.

3.2.1 False claims, or false records and statements

Sutter and PAMF do not meaningfully dispute that the government has sufficiently alleged that they submitted false claims predicated on false records or statements, namely, the false diagnosis codes submitted to CMS. In any event, the government has alleged with the requisite particularity that Sutter and PAMF submitted claims predicated on false diagnosis codes.

See id. at 7 ("The government's Complaint-in-Intervention alleges that defendants Sutter Health and Palo Alto Medical Foundation (collectively, ‘Defendants’) submitted diagnosis codes to the Medicare Advantage program that they should have known were not adequately documented by medical charts.").

Sutter also contends that Ms. Ormsby's claims against non-PAMF affiliates fail because she does not identify any unsupported diagnosis code that a non-PAMF affiliate submitted to CMS. But "[t]o state an FCA claim, a [plaintiff] is not required to identify actual examples of submitted false claims; instead, ‘it is sufficient to allege particular details of a scheme to submit false claims paired with reliable indicia that lead to a strong inference that claims were actually submitted.’ " Godecke , 937 F.3d at 1209 (some internal quotation marks omitted) (quoting Ebeid , 616 F.3d at 998–99 ). Ms. Ormsby has alleged particular details of a scheme at Sutter's non-PAMF affiliates to submit false diagnosis codes, including by having its coders pre-populate diagnosis codes into patients' encounter data before physicians met with their patients, in violation of Medicare Advantage requirements that diagnosis codes be based on a "face-to-face" visit between a physician and a patient that is documented in the medical record. Ms. Ormsby also alleged reliable indicia that lead to a strong inference that false diagnosis codes were submitted, including acknowledgements by Sutter executives that non-PAMF affiliates were submitting false codes.

Defs. Mot. to Dismiss Relator FAC – ECF No. 68 at 11.

See supra notes 401–403.

See supra notes 361, 376, 394–399.

Sutter asserts that Rule 9(b) requires Ms. Ormsby to differentiate between Sutter's affiliates and to link these affiliates to the alleged fraudulent scheme. It cites Swoben to support this assertion: " Rule 9(b) does not allow a complaint to merely lump multiple defendants together but requires plaintiffs to differentiate their allegations when suing more than one defendant and inform each defendant separately of the allegations surrounding his alleged participation in the fraud." This argument is inapposite. Ms. Ormsby does not name Sutter's affiliates as individual defendants (and thus does not lump multiple defendants together). Instead, she charges Sutter for misconduct that it conducted through its affiliates. In any event, "[t]here is no flaw in a pleading, however, where collective allegations are used to describe the actions of multiple defendants who are alleged to have engaged in precisely the same conduct." Swoben , 848 F.3d at 1184 ; accord Silingo , 904 F.3d at 677 ("A good claim against one defendant did not become inadequate simply because a co-defendant was alleged to have committed the same wrongful acts.").

Defs. Mot. to Dismiss Relator Compl. – ECF No. 68 at 10.

Id. at 11 (quoting Swoben , 848 F.3d at 1184 ).

Ms. Ormsby alleges that Sutter — through all of its affiliates (including PAMF) — used the same electronic-medical-records system and submitted diagnosis codes the same way, had false codes and high error rates in audits, had no process in place to delete unsupported diagnosis codes, prevented coders from deleting unsupported diagnosis codes they found and instead submitted those codes to CMS, and ultimately began pre-populating its patients' medical records with diagnosis codes. Ms. Ormsby sufficiently pleads her claim against Sutter regarding its non-PAMF affiliates.

See supra note 337.

See supra note 366, 377.

See supra notes 378–379.

See supra notes 391–399.

See supra notes 401–403.

In sum, the plaintiffs have sufficiently pleaded that Sutter and PAMF submitted false claims and used false records or statements.

3.2.2 Scienter

Sutter and PAMF contend that the plaintiffs "do[ ] not identify a single diagnosis code that Defendants had identified at the time it was submitted." They argue that backward-looking audits (like those conducted by Ms. Ormsby and UnitedHealth) "did not give Defendants retroactive knowledge that they were submitting unsupported codes before those codes had even been submitted." This argument fails.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 25 (emphasis in original).

Id. at 26.

The plaintiffs need not allege that Sutter and PAMF had actual knowledge of any specific falsity. Allegations that they were deliberately ignorant or reckless with respect to the truth or falsity of their submissions are sufficient to plead a claim. 31 U.S.C. § 3729(b)(1)(A) ; Godecke , 937 F.3d at 1211. And "where [an] organization turns a blind eye to [diagnostic-code] over-reporting errors, it exhibits reckless disregard and deliberate ignorance toward the truth or falsity of the data submitted to CMS." Swoben , 848 F.3d at 1175–76. Sutter and PAMF allegedly had no compliance program regarding risk-adjustment diagnosis coding. They nonetheless pressured physicians to submit more diagnosis codes and told their internal auditors "to take off the compliance hat and put on the revenue hat," stop removing false diagnosis codes, and focus instead on raising (not lowering) patient risk scores so that CMS would pay out more money to them. They allegedly had non-physician coders pre-populate diagnosis codes into patient medical records (before physicians saw their patients) and add diagnosis codes retroactively to patient medical records, in violation of Medicare Advantage requirements that diagnosis codes must be based on a "face-to-face" visit between a physician and a patient that is documented in the medical record. Collectively, these allegations plead, at the least, that Sutter and PAMF submitted risk-adjusting diagnosis codes with deliberate ignorance or reckless disregard of the truth or falsity of their submissions. Cf. Silingo , 904 F.3d at 680 (allegations that health-assessment reports were not properly signed by physicians in violation of Medicare requirements supported a claim that the defendant showed reckless disregard or deliberate ignorance to potential false claims and thus acted with scienter); Swoben , 848 F.3d at 1175 (where defendants "were on notice that their data included a significant number of erroneously reported diagnosis codes[, w]e do not see how a Medicare Advantage contractor who has engaged in such conduct can in good faith certify that it believes the resulting risk adjustment data reported to CMS are accurate, complete and truthful"). Sutter and PAMF assert that the plaintiffs are "arguing that there must be something improper about Defendants' efforts to increase their Medicare Advantage reimbursements. But ... there is nothing improper — let alone fraudulent — about attempting to do a better job of capturing members' diagnoses." But the plaintiffs do not challenge legitimate Medicare Advantage reimbursements. They challenge false diagnosis codes that, as Sutter and PAMF executives acknowledged, "CMS is still receiving ... [and] we know are not correct."

See supra notes 148–161, 338–342, 356.

See supra notes 54–113, 238–242, 255–257, 264–274, 283, 343–353, 400.

See supra notes 114–125, 275–278, 401–403.

Sutter and PAMF suggest that the government must (1) identify specific individuals at Sutter and PAMF who submitted claims or signed certifications to CMS and (2) allege that those individuals had knowledge that the claims or certificates were false. See Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 26. This is not required. The Ninth Circuit has allowed FCA claims to proceed where a plaintiff alleges only that (1) a defendant organization submitted claims and (2) the defendant organization had deliberate ignorance or reckless disregard (because its management had deliberate ignorance or reckless disregard), without the plaintiff's identifying specific individuals who submitted the claims. Godecke , 937 F.3d at 1211–12 ; see Fourth Amended Complaint at 57–70 (¶¶ 171–200), United States ex rel. Godecke v. Kinetic Concepts, Inc. , No. 2:08-cv-01885-CAS-AGR (C.D. Cal. filed June 26, 2017), ECF No. 257 (making allegations against defendant organization generally and not identifying specific individuals who submitted claims with scienter).
Sutter and PAMF cite United States v. Scan Health Plan , No. CV 09-5013-JFW (JEMx), 2017 WL 4564722 (C.D. Cal. Oct. 5, 2017) (Swoben II ), holding that "[a] complaint may not rely on the notion that a corporation has ‘collective scienter’ separate from the scienter of any actual human.’ This rule — that a complaint asserting claims based on false statements must plead the unlawful state of mind of the speaker — applies to claims alleged under the False Claims Act." Swoben II , 2017 WL 4564722, at *5 (citing United States v. Sci. Applications Int'l Corp. , 626 F.3d 1257, 1274 (D.C. Cir. 2010) and other authorities). The putative prohibition on "collective scienter" (or "collective knowledge") that Swoben II references does not require the government to identify the specific individuals who submitted claims or signed certifications and plead or prove that those individuals had scienter.
As the D.C. Circuit's decision in Science Applications (on which Swoben II relies) explains, "the ‘collective knowledge’ theory allows ‘a plaintiff to prove scienter [of a corporation] by piecing together scraps of "innocent" knowledge held by various corporate officials, even if those officials never had contact with each other or knew what others were doing in connection with a claim seeking government funds.’ " Sci. Applications , 626 F.3d at 1275 (quoting United States ex rel. Harrison v. Westinghouse Savannah River Co. , 352 F.3d 908, 918 n.9 (4th Cir. 2003) ). Take, for example, a scenario where one official wrote up a hypothetical false claim for use in a compliance training exercise, and a second official submitted it to CMS (without knowing the claim was false and without being deliberately ignorant or reckless). The putative prohibition on a collective-scienter theory might prevent a plaintiff from piecing together the knowledge of those two officials — each of whom was "innocent" on his own — to impute "collective scienter" to the organization. But see id. at 1275–76 ("Under the FCA, if a plaintiff can prove that a government contractor's structure prevented it from learning facts that made its claims for payment false, then the plaintiff may establish that the company acted in deliberate ignorance or reckless disregard of the truth of its claims.").
But the putative prohibition on a collective-scienter theory does not require a plaintiff to plead or prove that the individuals who — on behalf of the defendant organization — submitted claims or certifications had scienter themselves. "If [courts] established such a rule, corporations would establish segregated ‘certifying’ offices that did nothing more than execute government contract certifications, thereby immunizing themselves against FCA liability." Harrison , 352 F.3d at 919. It is sufficient to plead that the defendant organization had scienter (whether imputed knowledge from any employee who had scienter or otherwise), regardless of whether the individuals who submitted the claims or certifications were wholly innocent and had no scienter themselves. See, e.g. , Harrison , 352 F.3d at 919–20 ; United States ex rel. Zissa v. Santa Barbara Cty. Alcohol, Drug, & Mental Health Servs. , No. CV 14-6891-DMG (RZx), 2019 WL 3291579, at *5 (C.D. Cal. Mar. 12, 2019) (allegations that the organization's employees had knowledge of falsity pleads that the organization had knowledge, under a "basic agency theory of liability[, which] applies to FCA cases") (citing cases); see also United States ex rel. Ling v. City of Los Angeles , No. CV 11-974 PSG (JCx), 2018 WL 3814498, at *10–11 (C.D. Cal. July 25, 2018) (finding Swoben II unpersuasive because it relied on cases addressing the Private Securities Litigation Reform Act "without noting the differing statutory bases").

Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 6–7 (emphasis in original) (citing United States ex rel. Integra Med Analytics, LLC v. Baylor Scott & White Health , No. 5:17-CV-886-DAE, 2019 WL 3713756 (W.D. Tex. Aug. 5, 2019), appeal docketed , No. 19-50818 (5th Cir. filed Sept. 5, 2019)).

See supra notes 225, 281, 392, 398. Sutter's and PAMF's reliance on Integra Med Analytics is misplaced. In that case, "nothing in Plaintiff's complaint implicates a conclusion that these targeted efforts requested, demanded, or encouraged doctors and staff to diagnose in a way that was not justified by the physicians['] own medical opinions, judgments, and the medical record[.]" Integra Med Analytics , 2019 WL 3713756, at *5. Here, by contrast, Sutter and PAMF allegedly pressured physicians to diagnose and code in a way that the physicians thought was not justified (and thought was potentially fraudulent), had non-physicians add diagnosis codes, and knew that false diagnosis codes were being submitted. See, e.g. , Gov't Compl. – ECF No. 41 at 33–34 (¶ 96) (PAMF physicians complaining about Sutter's and PAMF's pressure: "I don't feel it is legitimate to code this," "it makes me feel a little fraudulent to be considering [upcoding]," "it is so obviously unethical," and "pre-populating diagnoses into [physician's] visit encounter is possibly fraud ... Does CMS know about what you are doing?") (ellipsis in complaint), 42 (¶ 121) (Sutter's RAF Program Manager acknowledging that "CMS is still receiving HCC's that we know are not correct.").

In sum, the plaintiffs have sufficiently pleaded Sutter's and PAMF's scienter.

3.2.3 Materiality and causation

The false claims at issue here are false diagnosis codes. Diagnosis codes are the only factors that CMS uses to determine a beneficiary's health status to calculate Medicare Advantage payments for that beneficiary. When MA Participants submit false risk-adjusting diagnosis codes, CMS pays more money (and, conversely, when they delete risk-adjusting diagnosis codes, CMS pays less money). This establishes that the diagnosis codes are material. United States ex rel. Poehling v. UnitedHealth Grp., Inc. , No. CV 16-08697-MWF (SSx), 2018 WL 1363487, at *9–10 (C.D. Cal. Feb. 12, 2018) (government's allegations that diagnosis codes are "the sole determinant in the calculation of any risk adjustment payment based on a beneficiary's health status" and that CMS adjusts payments upwards or downwards based on addition or deletion of diagnosis codes sufficiently pleads that diagnosis codes are material).

Gov't Opp'n – ECF No. 82 at 16, 18 ("[T]he diagnoses submitted to CMS because of Defendants' fraud are the material false claims, not the certifications.... The unsupported diagnosis codes are the false claims that Defendants caused to be submitted to Medicare.").

See supra note 35–37.

See supra notes 38–40.

Sutter and PAMF challenge materiality with respect to a separate issue: whether the separate certifications that they had to submit to CMS — attesting to the accuracy, completeness, and truthfulness of the data (including diagnosis codes) — are material. The diagnosis codes are material. Also, by alleging that Sutter and PAMF submitted false diagnosis codes that caused CMS to pay them money, the plaintiffs sufficiently plead causation.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 26–28. As noted above, Medicare regulations require, as a condition of receiving payment, that MA Participants such as Sutter and PAMF certify the accuracy, completeness, and truthfulness of data they submit to CMS. 42 C.F.R. § 422.504(l ). Separate and apart from the underlying diagnosis codes, these certifications can also constitute "claims" and provide an additional basis for an FCA claim. Swoben , 848 F.3d at 1173 (" ‘[A] claim under the False Claims Act can be false where a party merely falsely certifies compliance with a statute or regulation as a condition to government payment.’ ") (quoting United States ex rel. Hendow v. Univ. of Phoenix , 461 F.3d 1166, 1171 (9th Cir. 2006) ).

Sutter and PAMF do not meaningfully challenge that the diagnosis codes are material. See Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 27–28.

3.3 Common-Law Claims

Sutter and PAMF move to dismiss the common-law claims on the ground that if they were not overpaid, the government has no claim for unjust enrichment or payment by mistake. Because the court has held that payments predicated on false diagnosis codes are overpayments, the court denies the motion to dismiss the common-law claims.

Defs. Mot. to Dismiss Gov't Compl. – ECF No. 66 at 18 n.4; Defs. Reply in Supp. of Mot. to Dismiss Gov't Compl. – ECF No. 86 at 14.

CONCLUSION

The court denies Sutter's and PAMF's motions to dismiss and grants the motion for judicial notice.

This disposes of ECF Nos. 66, 68, and 81.

United States ex rel. Ormsby v. Sutter Health

Opinion

ORDER DENYING DEFENDANTS' MOTIONS TO DISMISS

TABLE OF CONTENTS

INTRODUCTION

STATEMENT

1. Medicare Advantage Generally

2. The Government's Allegations Regarding Diagnosis Codes

3. The Government's Allegations Against Sutter and PAMF

3.1 Sutter's and PAMF's "RAF Campaign"

3.1.1 Characterizing conditions as "chronic"

3.1.2 Using a "pit crew" to add diagnosis codes to patient medical records

3.1.3 Pressuring physicians to add diagnosis codes to patient medical records

3.1.4 Pre-populating patient medical records with diagnosis codes

3.1.5 Adding diagnosis codes as addendums to patient medical records

3.2.1 Audits by MA Organizations

3.2.2 Lack of compliance and training programs

3.2.3 Internal reviews and audits by Sutter and PAMF

3.3 Knowingly Ignoring Red Flags and Actual Notice of False Claims

3.4 Examples of False Diagnosis Codes and Estimate of Total Overpayments

4. The Relator's Allegations Against Sutter Regarding Its Non-PAMF Affiliates

THE FALSE CLAIMS ACT

1. Direct-FCA Claims

2. Reverse-FCA Claims

STANDARD OF REVIEW

ANALYSIS

1. The "Actuarial Equivalence" Argument Is Not a Valid Defense

1.1 The Premise — That CMS Is Violating "Actuarial Equivalence" — Is Not Sound

1.1.1 CMS's Medicare Advantage audit process and the "2014 Overpayment Rule"

1.1.1.1 Audits of MA Organizations

1.1.1.2 The "2014 Overpayment Rule"

1.1.2 UnitedHealthcare 's invalidation of the 2014 Overpayment Rule

Id.

1.1.3 UnitedHealthcare does not compel the conclusion that there is no actuarial equivalence here

1.2 "Actuarial Equivalence" Is Not a Defense to an FCA Claim

1.2.1 MA Participants cannot retain payments predicated on false diagnosis codes

1.2.2 The "actuarial equivalence" requirement does not vitiate scienter

2. The Relator Can Pursue Her Sutter-Wide Claims

3. The Plaintiffs Sufficiently Plead Their Claims

3.1 Reverse-FCA Claims

3.1.1 Concealing or avoiding an obligation to pay the government

3.1.2 Scienter

3.2 Direct-FCA Claims

3.2.1 False claims, or false records and statements

3.2.2 Scienter

3.2.3 Materiality and causation

3.3 Common-Law Claims

CONCLUSION

IT IS SO ORDERED.

United States ex rel. Ormsby v. Sutter Health

United States ex rel. Ormsby v. Sutter Health

Case Details

CitationsCopy Citation

Citing Cases

Citations