Bond v. Clover Health Invs.

1

TIMOTHY BOND, Lead Plaintiff and JEAN-NICOLAS TREMBLAY, Named Plaintiff, Individually and on behalf of all others similarly situated, v. CLOVER HEALTH INVESTMENTS, CORP. f/k/a SOCIAL CAPITAL HEDOSOPHIA HOLDINGS CORP. III, VIVEK GARIPALLI, ANDREW TOY, JOE WAGNER, and CHAMATH PALIHAPITIYA, Defendants.

No. 3:21-cv-00096

United States District Court, M.D. Tennessee, Nashville Division

February 14, 2023

MEMORANDUM

ALETA A. TRAUGER, United States District Judge.

Firas Jabri and Jean-Nicholas Tremblay have filed a Motion for Class Certification (Doc. No. 101), to which defendants Clover Health Investments, Corp. f/k/a Social Capital Hedosophia Holdings Corp. III (“SCH”), Vivek Garipalli, Andrew Toy, Joe Wagner, and Chamath Palihapitiya have filed a Response (Doc. No. 108), and Jabri and Tremblay have filed a Reply (Doc. No. 110). For the reasons set out herein, the motion will be granted.

I. BACKGROUND

A. Medicare Advantage and the Problem of Risk Adjustment Fraud

In 1997, “Congress created ‘Medicare Part C,' sometimes referred to as Medicare Advantage [‘MA']. Under Part C, beneficiaries may choose to have the government pay their private insurance premiums rather than pay for their hospital care directly.” Azar v. Allina Health Servs., 139 S.Ct. 1804, 1809 (2019). The federal government funds MA plans on a “capitated” basis meaning that the private company operating a plan “receive[s] in advance a monthly lump sum from CMS for every beneficiary that [it] enroll[s], without regard to the services that the beneficiaries will actually receive.” UnitedHealthcare Ins. Co. v. Becerra, 16 F.4th 867, 873 (D.C. Cir. 2021). The capitated monthly rate for each patient, however, is not necessarily the same. The Center for Medicare and Medicaid Services (“CMS”) “uses a model-called the CMS Hierarchical Condition Category, or CMS-HCC, risk-adjustment model”-that assigns each Part C beneficiary a “risk assessment score” based on a formula considering various “demographic characteristics” that are “predictive of differing costs of care.” Id. at 874 (citing 42 U.S.C. § 1395w-23(a)(1)(C)(i)). CMS's formula allows the agency to “determine prospectively, based on Medicare Advantage beneficiaries' actuarially relevant, known demographic and health characteristics, the per-capita payment rate that will fairly compensate th[e] Medicare Advantage insurer” for providing the beneficiary's coverage for that month. Id. at 873.

If capitation rates were based solely on general, verifiable demographic traits such as age and sex, CMS's job in setting the scores would be as simple as consulting its own enrollment data and applying a formula to the numbers it found. In order for the risk assessment formula to be as effective as possible, however, Congress has authorized CMS to consider any “such other factors as the Secretary determines to be appropriate, including adjustment for health status.” 42 U.S.C. § 1395w-23(a)(1)(C)(1). The result is a somewhat more complex model that relies, among other things, on data provided by beneficiaries' healthcare providers, including with regard to “individuals' medical diagnoses.” U.S. ex rel. Anita Silingo v. WellPoint, Inc., 904 F.3d 667, 672 (9th Cir. 2018) (citing Policy and Technical Changes to the Medicare Advantage and the Medicare Prescription Drug Benefit Programs, 74 Fed.Reg. 54,634, 54,673 (Oct. 22, 2009)). While this additional data allows risk assessment scores to be more precise, it also introduces a vulnerability into the system. “With data for millions of people being submitted each year, CMS is unable to confirm diagnoses before calculating capitation rates.” Id. As a result, capitated rates may be skewed upward or downward based on providers' misreporting (or non-reporting) of patient information. In order to mitigate that problem, “Medicare regulations require risk adjustment data to be produced according to certain best practices.” Id. (citing 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)).

While those regulations may have had some success in combating data reporting errors- particularly inadvertent ones-they have not been able to wholly eradicate the practice of intentional overreporting of risk factors or, as it has come to be known, “risk adjustment fraud.” Because risk factors are reported by healthcare providers, but the inflation of a patient's risk score benefits only the insurer, there is no inherent incentive to commit the fraud-at least as long as those parties are kept at an arm's length from each other. However, if a Part C plan operator can find a way to induce, encourage, or trick healthcare providers into over-reporting a beneficiary's risk factors without actually implementing a more expensive course of treatment, the insurer can receive a higher capitated payment for that beneficiary without actually having to pay for more services.

B. Clover and the Clover Assistant

Garipalli founded the original Clover^[1] in 2013 for the purpose of providing Medicare Advantage plans. (Id. ¶ 56.) Historically, the Part C/MA market has been concentrated among a handful of large insurers with patient bases far larger than Clover's. Clover has publicly acknowledged its comparatively weak market position, but the company and its executives characterized the MA field as, in the words of one press release, “ripe for disruption” after having “seen little innovation” for years. (Doc. No. 70 ¶ 61.^[2]) According to that press release, Clover was designed to provide that disruption with its “unique model,” through which it “partner[ed] with primary care physicians using its software platform, the Clover Assistant, to deliver data-driven, personalized insights at the point of care.” (Id. ¶ 63.)

Clover described the Clover Assistant as its “flagship software platform . . . to provide America's seniors with PPO and HMO plans that are the obvious choice for Medicare-eligible consumers.” (Id. ¶ 66.) According to Clover, the Assistant used “machine learning” to analyze “millions” of data points in order to provide “actionable and personalized insights at the point of care.” (Id. ¶ 66.) The plaintiffs, however, claim that the purpose of the Clover Assistant was far simpler: it was “designed to identify opportunities to assign higher Medicare risk adjustments so that Clover [could] obtain larger reimbursements from Medicare.” (Id. ¶ 67.) It did this by, among other things, guiding physicians' offices to record information that, based on the Medicare risk assessment formula, were likely to result in upward revisions of the patient's risk assessment score. Those upward revisions made Clover's deals with Medicare more profitable by allowing Clover to receive higher capitated payments for individual patients. (Id. ¶¶ 57, 6769.)

While there is nothing inherently wrong with encouraging physicians to report information likely to result in higher risk assessment scores, doing so through improper means or without sufficient guardrails to ensure accuracy risks crossing over the line into risk adjustment fraud. The court has previously detailed some of the mechanisms through which the Clover Assistant is alleged to have improperly driven inflated risk assessments-such as by nudging users to diagnose complex conditions and failing to recognize when old diagnoses should be removed. See Bond v. Clover Health Invs., Corp., 587 F.Supp.3d 641, 651 (M.D. Tenn. 2022). Exacerbating that problem was the fact that physicians themselves allegedly did not appear to be very interested in using the Assistant. Rather, Clover had to pay physicians to use the software as a “loss leader,” and many physicians who agreed to do so actually left the task to support staff, who should not have been permitted to diagnose patients and may have been particularly vulnerable to the Assistant's influence. Id. at 654-55.

C. The Clover/SCH Merger and this Case

United States securities laws allow the public trading of so-called “special purpose acquisition companies”-also referred to as “SPACs” or “blank check companies.” A SPAC typically has “no operating history, assets, revenue, or operations” of its own. Daniel S. Riemer, Special Purpose Acquisition Companies: SPAC and Span, or Blank Check Redux?, 85 Wash. U. L. Rev. 931, 933 (2007). Rather, a SPAC exists to become publicly traded itself and then “to buy a private company”-one that actually provides a good or service but is not yet publicly traded- thereby allowing investors to “effectively [take the acquired] company public while avoiding the tradition[al] initial public offering [‘IPO'] process.” Phillips v. Churchill Cap. Corp. IV, No. 1:21-CV-00539-ACA, 2021 WL 4220358, at *1 (N.D. Ala. Sept. 16, 2021). SCH was a SPAC, and, on October 6, 2020, it announced that it would be fulfilling its mission by acquiring Clover.

As the Clover/SCH merger moved forward, the companies and their executives made various public statements about the transaction and/or Clover's business model, either to the press or in filings to the SEC. According to the plaintiffs, many of those statements were false or misleading because they concealed or failed to disclose that:

(i) the Company had committed multiple legal and regulatory violations since January 1, 2018 and was . . . under investigation by the DOJ for violations of the False Claims Act;

(ii) the Company's growth and positive performance stemmed from illegal gifts and/or payments to healthcare practitioners and/or office staff in violation of the Anti-Kickback Statute, the [False Claims Act], and [CMS Guidelines], and unreported related party transactions;

(iii) only a small fraction of the healthcare providers who had contracted with

...