Opinion
MDL No: 2785Case No. 17-md-2785-DDC-TJJ
02-27-2020
IN RE: EpiPen (Epinephrine Injection, USP) Marketing, Sales Practices and Antitrust Litigation (This Document Applies to Consumer Class Cases)
MEMORANDUM AND ORDER
Class plaintiffs filed a motion for class certification (Doc. 1353), seeking to represent five classes of end-payors in the United States who paid or reimbursed others for some or all of the purchase price of branded or authorized generic EpiPens. Plaintiffs allege that the Mylan defendants (composed of Mylan N.V., Mylan Specialty, L.P., Mylan Pharmaceuticals, Inc., and Mylan's CEO Heather Bresch) as distributors of the EpiPen, and the Pfizer defendants (composed of Pfizer, Inc., King Pharmaceuticals, Inc., and Meridian Medical Technologies, Inc.) as manufacturers of the EpiPen, have maintained a monopoly over the epinephrine auto-injector ("EAI") market and its profitable revenues by devising an illegal scheme to monopolize the market for EAI devices. For the reasons explained below, the court grants the motion in part and denies it in part.
The Consolidated Class Action Complaint ("Class Complaint") (Doc. 60) asserts the collective claims of five consumer class cases transferred by the Judicial Panel on Multidistrict Litigation. The Class Complaint asserts federal and state antitrust claims, federal RICO Act violations, various state consumer protection law violations, and unjust enrichment claims. The Class Complaint asserts its claims against the two groups of defendants who either sell or manufacture the EpiPen, the Mylan defendants and the Pfizer defendants. When referring to both sets of defendants, this Order calls them, collectively, "defendants."
The JMPL also transferred a case filed by Sanofi-Aventis U.S. LLC ("Sanofi") as part of this Multidistrict Litigation. It proceeds on a separate track, and this certification motion does not affect the claims or defenses in Sanofi's case.
TABLE OF CONTENTS
I. Factual Background................................................................................5
II. Procedural History.................................................................................10
III. Legal Standard.....................................................................................13
IV. Discussion...........................................................................................15
A. Proposed Class Definitions.............................................................15
B. Proposed Class Exclusions.............................................................17
C. Class Certification Under Rule 23...................................................20
1. Ascertainability.................................................................20
2. Standing..........................................................................26
D. Rule 23(a) Requirement...............................................................27
1. Numerosity......................................................................27
2. Commonality....................................................................28
3. Typicality........................................................................30
a. Local 282................................................................31
b. Named Plaintiffs.......................................................36
4. Adequacy........................................................................40
E. Rule 23(b)(3) Requirements...........................................................44
1. Superiority.......................................................................44
2. Predominance...................................................................45
a. Individual Damages Issues..........................................47
b. Uninjured Class Members..........................................53
i. Collateral Source Doctrine.......................................54
ii. The Presence of Uninjured Class Members...................58
(a) Legal Standard....................................59
(b) Prof. Rosenthal's Probability Analysis.......67
(c) Number of Uninjured Class Members........72
(d) Constitutional Considerations.................76
c. RICO......................................................................79
i. RICO Causation......................................................82
(a) Switch to 2-Pak..................................................84
(b) Delay of Generic Competition Through "Pay for Delay" Settlements..................................90
(c) Exclusive Dealing Contracts with PBMs..................95
ii. RICO Damages....................................................103
d. Antitrust.................................................................106
i. Market Power......................................................107
ii. Exclusive Dealing Arrangements..............................108
iii. Generic Delay......................................................111
e. Consumer Protection................................................115
f. Unjust Enrichment...................................................121
F. Rule 23(b)(2) Injunction Class.......................................................123
V. Class Definitions.................................................................................126
VI. Appointment of Class Counsel...............................................................127
VII. Notice..............................................................................................128
To determine whether plaintiffs have met their burden of affirmatively demonstrating compliance with Rule 23's requirements, the court "must accept the substantive allegations of the complaint as true." Shook v. El Paso Cty., 386 F.3d 963, 968 (10th Cir. 2004) (citing J.B. ex rel. Hart v. Valdez, 186 F.3d 1280, 1290 n.7 (10th Cir. 1999)).
The EpiPen is a portable EAI device used to administer epinephrine to treat an anaphylactic reaction to an allergen. In 2007, Mylan acquired the right to market and distribute the EpiPen. Pfizer supplies Mylan with 100% of its EpiPen supply through two wholly-owned subsidiaries—King Pharmaceuticals, Inc., and Meridian Medical Technologies, Inc.—who manufacture the epinephrine and hold the EpiPen patents. Since at least 2009, Mylan's market share of EAI devices in the United States EAI market has remained above 80%. And, since 2009, Mylan's market share consistently has exceeded 90%, and, in 2012, its share was almost 100%. During the same time—and while the cost of the EpiPen's dose of epinephrine has remained about $1—Mylan has raised the EpiPen's price by more than 600%. In 2007, Mylan priced the EpiPen at $100. By 2016, Mylan was charging more than $600. In 2015, Mylan announced that the EpiPen had reached $1 billion in annual sales for the second consecutive year—up from $200 million in 2007.
Mylan's Dealings with Pharmacy Benefit Managers (PBMs)
Pharmacy Benefit Managers ("PBMs") are third-party administrators of prescription drug programs for commercial health plans, self-insured employer plans, Medicare Part D plans, the Federal Employees Health Benefits Program, and state government employee plans. More specifically, PBMs administer a health coverage provider's prescription benefit program by developing the coverage provider's formulary (the list of prescription benefits included in the coverage at various pricing "tiers"), processing claims, creating a network of retail pharmacies who provide discounts in exchange for access to a provider's plan participants, and negotiating with drug manufacturers. A significant majority of patients with prescription drug insurance coverage receive their benefits through a third-party payor whose drug formulary is determined by a PBM. From 2013 to 2015, commercial third-party payors accounted for about 71% of the EAI drug device market in the United States. So, for a competitor to enter and compete vigorously in the EAI drug device market, it is imperative that the competitor access these third-party payors' drug formularies.
Between 2013 and 2015, Mylan significantly increased the EpiPen's price. Plaintiffs allege Mylan then used the additional profit margins to offer PBMs significantly higher rebates and percentage discounts if the PBM would provide exclusive or preferred placement for EpiPen on the PBM's drug formulary. In 2013, when Sanofi launched a rival EAI device—Auvi-Q—Mylan began taking steps to block Auvi-Q from drug formularies, plaintiffs contend. It did so by offering large rebates—30% or higher—to PBMs who controlled formularies for third-party payors and expressly conditioning those rebates on the PBMs: (a) granting the EpiPen exclusive position on the formulary; and (b) removing or severely restricting access to Auvi-Q. Adrenaclick, another rival EAI device, achieved a market share that ranged from 2% in 2013 to 8% in 2016. In 2014, CVS Caremark added Adrenaclick to its Formulary Drug Removals List, effectively eliminating a consumer or other end-payor's opportunity to purchase Adrenaclick as an alternative to the EpiPen. Plaintiffs contend Mylan knew that neither Adrenaclick nor Auvi-Q could raise prices to inflate their margins sufficiently to offer rebates or discounts similar to those Mylan was offering to PBMs.
Defendants' Pricing Scheme
Plaintiffs allege defendants implemented a pricing scheme that defrauded U.S. consumers into paying an inflated price for the EpiPen—one that climbed by more than 500%, in nine years. Plaintiffs assert that defendants' exclusionary pricing scheme deprives patients of a fair price for EAI devices—the price that would result from normal market forces. They describe the pricing scheme this way: Mylan, instead of lowering its prices to gain market share, bargains for market share by providing ever-larger rebates and other kickbacks to PBMs, conditioned on exclusive relationships with those PBMs. Mylan can place itself on the drug formularies by using its monopoly power to charge consumers higher prices for its product. It then can share these revenues with the PBMs (the ones who create the formularies) through substantially enhanced rebates conditioned on excluding insurance coverage for rival products. This conduct, in turn, inflates the prices that consumers pay for the EpiPen so that Mylan can preserve its net realized price and sales volumes. Plaintiffs contend that the net effect of this scheme harms both consumers and competitors alike.
Mylan primarily is a generics pharmaceutical company that makes low margins on drug sales. But the EpiPen, a specialty branded drug, represented a unique and highly profitable revenue stream for Mylan. Plaintiffs assert that Mylan CEO Heather Bresch and other executives recognized this opportunity and decided to exploit the EpiPen to generate billions of dollars in revenue. One of the ways defendants allegedly implemented their pricing scheme was to change the way Mylan sold the EpiPen.
On August 24, 2011, Mylan announced that it no longer would sell individual EpiPens in the United States. Instead, Mylan began selling EpiPens in just one kind of packaging—the EpiPen 2-Pak. This configuration forced U.S. consumers to purchase EpiPens in pairs. Plaintiffs refer to this change as the "hard switch," and allege that it forced consumers and third-party payors to overpay for the EpiPen 2-Pak because the switch effectively doubled the EpiPen's price. While Mylan asserts a medical necessity supported the switch, plaintiffs contend nothing changed in 2011 that required Mylan to sell the EpiPen in a 2-Pak. Moreover, Mylan continued selling EpiPens individually in every other country where Mylan markets the product.
The lone exception is France, where Mylan also sells the EpiPen in just one packaging format: the 2-Pak.
Beginning in 2011, Mylan began raising the EpiPen's price while forcing customers to buy EpiPens in the 2-Pak. In October 2011, Mylan increased the price of an EpiPen 2-Pak to $181. And by May 2016, the price of an EpiPen 2-Pak had jumped to $608.
According to plaintiffs, Mylan developed and implemented a pricing scheme to monopolize the EAI market. This scheme, plaintiffs allege, used a variety of tactics. They included spreading false and misleading information and omitting material information. Plaintiffs contend defendants and the PBMs distracted consumers and regulators from the reality that Mylan was raising the price of the EpiPen from $100 to $600 by launching a campaign of false and misleading statements and actions. These statements and actions include: (a) distributing misinformation about Auvi-Q intended to undermine the FDA's conclusion that Auvi-Q had demonstrated bioequivalence to the epinephrine in the EpiPen; (b) paying formularies to exclude Auvi-Q from coverage, while telling physicians about the exclusion and suggesting the decision was based on clinical recommendations, rather than large, conditional rebate offers; and (c) issuing a fraudulent and misleading press release about the "hard switch" to the 2-Pak that included erroneous medical claims. Plaintiffs also contend that Mylan CEO Heather Bresch testified untruthfully in a Congressional hearing on September 21, 2016. Specifically, they assert, Ms. Bresch misrepresented that: (a) Mylan's profit on the EpiPen as $50 per device when its profits were reportedly 60% higher than that amount; (b) the investment Mylan had made in EpiPen as $1 billion (when Mylan, in fact, had acquired the EpiPen in 2007 without incurring any research and development expenses); (c) Mylan had reduced U.S. healthcare costs by about $180 billion; and (d) 85% of EpiPen patients pay less than $100 for a 2-Pak and a majority pay less than $50.
Patent Infringement Lawsuits
The Pfizer defendants own the patents protecting the EpiPen and serve as contract supplier of the product. The Mylan defendants own the trademarked brand names and control the worldwide marketing and sale of the product. Together, plaintiffs assert, defendants have a unified interest in protecting the EpiPen's alleged monopoly. To that end, defendants allegedly added patents to the already-patented EpiPen to stop generic competitors. For instance, on September 14, 2010, Meridian (a Pfizer subsidiary) secured U.S. Patent No. 7,794,432, just three weeks before Intelliject submitted a New Drug Application ("NDA") to the FDA for Auvi-Q.
Defendants also filed patent infringement lawsuits against two generic EpiPen rivals—Teva and Intelliject—and then entered into settlements with these impending generic manufacturers which, plaintiffs assert, were anticompetitive. In December 2008, Teva filed an Abbreviated New Drug Application ("ANDA") seeking FDA approval to market a generic EpiPen. In August 2009, Pfizer's subsidiaries, King and Meridian, sued Teva for infringing U.S. Patent No. 7,449,012 ("the Teva patent litigation"). In November 2010, King and Meridian filed a First Amended Complaint that included a claim for infringing the newly secured '432 patent. Later, King and Meridian dropped their claims based on the '012 patent, leaving only the infringement claims for the '432 patent. On April 27, 2012, the parties settled the Teva patent litigation. According to plaintiffs, defendants and Teva entered into an unlawful settlement agreement requiring Teva to delay launching its generic EAI for three years—until June 22, 2015—in exchange for defendants providing significant consideration, incentives, and benefits to Teva. Plaintiffs allege the settlement included a substantial "reverse payment" from defendants to Teva to convince Teva to delay bringing its competing product to market. Plaintiffs offer expert testimony from Prof. Andrew K. Torrance about (a) Teva's likelihood of success in the Teva patent litigation, and (b) the costs and duration of that litigation had the parties not settled it.
The Class Complaint also describes patent litigation with Sandoz, but plaintiffs do not include any allegations about that litigation in their certification motion.
Defendants also initiated another patent lawsuit against Intelliject, the inventor of Auvi-Q, when Intelliject sought FDA approval for its competing EAI. On January 11, 2011, King filed suit against Intelliject seeking to block FDA approval of Intelliject's NDA for the product that later became known as Auvi-Q. On August 1, 2011, the FDA announced that it was giving Intelliject's EAI "tentative final approval," pending resolution of the patent litigation King had filed. Doc. 1500 at 33. Six months later, the parties announced that they had settled the Intelliject litigation, with Mylan making the announcement jointly for itself and Pfizer. The settlement included an agreement by Intelliject and Sanofi not to enter the EAI market until November 15, 2012, in exchange for valuable consideration. Plaintiffs allege that defendants, through the Intelliject patent litigation, forestalled competition that Intelliject's competing device would have presented for nearly two years.
II. Procedural History
On August 20, 2018, the court granted in part and denied in part defendants' motions to dismiss the Class Complaint (Doc. 896). The court directed the parties to conduct coordinated discovery for both the consumer class cases and the Sanofi case, followed by dispositive motions in the Sanofi case and a motion for class certification in the consumer class cases. Class plaintiffs now have moved for class certification (Doc. 1353). In response, the Mylan defendants filed an opposition to class plaintiffs' motion (Doc. 1636, Doc. 1834 (sealed)), as did the Pfizer defendants (Doc. 1841). Class plaintiffs filed a reply memorandum supporting their motion for class certification (Doc. 1837, Doc. 1839 (sealed)). On June 11 and 12, 2019, the court conducted a hearing on the class certification motion and took it under advisement.
In their brief, the Pfizer defendants join the arguments in the Mylan defendants' brief, but they also separately argue two other issues. So, arguments made by the Mylan defendants are also attributable to the Pfizer defendants. For ease, the court refers to arguments asserted by the Mylan defendants as ones asserted by Mylan. And it refers to arguments asserted separately by the Pfizer defendants as ones asserted by Pfizer. Also, this Order generally identifies which party made an argument or asserted a position. If neither defendant has addressed an issue that applies to both Mylan and Pfizer, the court notes that defendants have not addressed the issue.
After briefing on plaintiff's Motion for Class Certification had closed, defendants filed a Motion for Leave to File a Sur-reply. Doc. 1574. Defendants assert that plaintiffs improperly presented two new arguments in their Reply to the Motion for Class Certification. Specifically, defendants assert that the Reply raised new arguments about Prof. Rosenthal's probability analysis and the collateral source rule. Defendants thus ask the court for leave to file a Sur-reply to address these two arguments. Defendants have attached their purposed Sur-reply to their motion (Doc. 1574-1).
Plaintiffs respond that defendants' Motion for Leave to File a Sur-reply is baseless. Plaintiffs contend they properly offered their arguments about the collateral source rule and Prof. Rosenthal's probability analysis to respond to arguments defendants had raised in their response brief opposing class certification. Hedging their bets, plaintiffs say they don't oppose defendants' Motion for Leave to File a Sur-reply, so long as the court allows them to file a short response to defendants' arguments in the Sur-reply. Plaintiffs have submitted their proposed response to defendants' Sur-reply as an attachment to their Response (Doc. 1579-1). And, plaintiffs contend, they should have the last word on their Motion for Class Certification because they are the moving party.
The court's local rules limit briefing on motions to the motion, a memorandum in support, a response, and a reply. D. Kan. Rule 7.1(a) & (c). "Surreplies are typically not allowed." Taylor v. Sebelius, 350 F. Supp. 2d 888, 900 (D. Kan. 2004), aff'd on other grounds, 189 F. App'x 752 (10th Cir. 2006). Instead, sur-replies are permitted only with leave of court under "rare circumstances." Humphries v. Williams Nat. Gas Co., No. 96-4196-SAC, 1998 WL 982903, at *1 (D. Kan. Sept. 23, 1998) (citations omitted). For example, when a moving party raises new material for the first time in a reply, the district court has discretion to grant leave to file a sur-reply to afford the opposing party an opportunity to respond to the new material. Green v. New Mexico, 420 F.3d 1189, 1196 (10th Cir. 2005); Doebele v. Sprint/United Mgmt. Co., 342 F.3d 1117, 1139 n.13 (10th Cir. 2003). The rules limiting sur-replies "are not only fair and reasonable, but they assist the court in defining when briefed matters are finally submitted and in minimizing the battles over which side should have the last word." Humphries, 1998 WL 982903, at *1.
Here, the court concludes that plaintiffs' Reply didn't improperly assert new arguments. As plaintiffs explain, they presented arguments about the collateral source rule and Prof. Rosenthal's probability analysis in direct response to defendants' arguments opposing class certification. Specifically, defendants' Opposition to class certification argued that the court should not certify plaintiffs' putative classes because the proposed definitions include uninjured class members—for example, consumers who used a co-pay to purchase their EpiPens. In their Reply, plaintiffs responded that the collateral source rule bars this argument because that rule requires the court to exclude evidence of payments from a third-party (such as an insurer) that would reduce a tortfeasor's liability. Doc. 1837 at 14. Also, plaintiffs cited a probability analysis that Prof. Rosenthal presented in her rebuttal report to show that the number of uninjured class plaintiffs in each class is de minimis. Id. at 19-22. It was proper for plaintiffs to assert these arguments in their Reply because they responded directly to assertions in defendants' Opposition claiming that the proposed classes include uninjured class members.
Yet, it is equally true that these two arguments are "new" in the sense they don't appear in the class certification briefing until the Reply. And, the court finds the parties' additional submissions helpful to the court's consideration of the difficult issues raised in the class certification briefing.
While this court limits sur-replies and such to "rare circumstances," Humphries, 1998 WL 982903, at *1, there's nothing ordinary about this case or this certification motion. The rare circumstances presented in this case warrant granting leave for a sur-reply. Exercising its discretion, the court grants defendants' Motion for Leave to File a Sur-reply (Doc. 1574). Also, the court grants plaintiffs' request to file a response to defendants' Sur-reply. The court will consider the arguments presented by defendants' Sur-reply (Doc. 1574-1) and plaintiffs' Response (Doc. 1579-1) in its analysis, below.
III. Legal Standard
"The class action is 'an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.'" Comcast Corp. v. Behrend, 569 U.S. 27, 33 (2013) (quoting Califano v. Yamasaki, 442 U.S. 682, 700-01 (1979)). The court has considerable discretion when deciding whether to certify a class action. Tabor v. Hilti, Inc., 703 F.3d 1206, 1227 (10th Cir. 2013) (because class certification involves "intensely practical considerations," decision rests within trial court's discretion (citation and internal quotation marks omitted)). But when exercising this discretion, district courts must conduct a "rigorous analysis" and decide whether the putative class satisfies the requirements of Federal Rule of Civil Procedure 23. Comcast, 569 U.S. at 33 (citation and internal quotation marks omitted); Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350-51 (2011).
The elements of the class certification standard are (1) numerosity, (2) commonality, (3) typicality, and (4) adequate representation, plus one of the requirements established by Rule 23(b)(1), (b)(2), or (b)(3). See Fed. R. Civ. P. 23(a)-(b). Here, plaintiffs seek certification under Rules 23(b)(2) and (3). Rule 23(b)(2) applies when "final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." Rule 23(b)(3)-based motions require plaintiffs to demonstrate that "questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy."
Rule 23 "'does not set forth a mere pleading standard.'" Comcast, 569 U.S. at 33 (quoting Dukes, 564 U.S. at 350). As the party requesting class certification, plaintiffs bear the burden of "'affirmatively demonstrat[ing]'" compliance with the rule's requirements. Id. (quoting Dukes, 564 U.S. at 350). Plaintiffs "must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc." Dukes, 564 U.S. at 350. When deciding whether plaintiffs have met their burden, the court "must accept the substantive allegations of the complaint as true," but it cannot "'blindly rely on conclusory allegations which parrot Rule 23.'" Shook v. El Paso Cty., 386 F.3d 963, 968 (10th Cir. 2004) (quoting J.B. ex rel. Hart v. Valdez, 186 F.3d 1280, 1290 n.7 (10th Cir. 1999)). The court "is not limited to the pleadings but may 'probe behind the pleadings' and examine the facts and evidence in the case." Tabor, 703 F.3d at 1227-28 (quoting Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 160 (1982)); see also Ellis v. Costco Wholesale Corp., 657 F.3d 970, 982 (9th Cir. 2011) (rigorous analysis requires evaluations about persuasiveness of evidence). "[A]ctual, not presumed, conformance with Rule 23(a)" is required. Falcon, 457 U.S. at 160.
The court's "rigorous analysis" "[f]requently . . . will entail some overlap with the merits of the plaintiff's underlying claim." Dukes, 564 U.S. at 351. But the careful examination and rigorous analysis requirements do not authorize mini-trials to determinine whether the class, if certified, actually could prevail on the merits of their claims. See Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 465-66 (2013); Dukes, 564 U.S. at 351 n.6. "Rule 23 grants courts no license to engage in free-ranging merits inquiries at the certification stage." Amgen, 568 U.S. at 466. Rather, "[m]erits questions may be considered to the extent—but only to the extent—that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied." Id.
IV. Discussion
A. Proposed Class Definitions
"Defining the class is of critical importance because it identifies the persons (1) entitled to relief, (2) bound by a final judgment, and (3) entitled under Rule 23(c)(2) to the 'best notice practicable' in a Rule 23(b)(3) action." Sibley v. Sprint Nextel Corp., 254 F.R.D. 662, 670 (D. Kan. 2008) (quoting Manual for Complex Litigation § 21.222 (4th ed. 2004)). Thus, "[t]he [class] definition must be precise, objective and presently ascertainable." Id.
Class plaintiffs ask the court to certify five classes. They are:
1. Nationwide Injunctive Relief Class ("Injunctive Class"). All persons and entities in the United States who paid or provided reimbursement for some or all of the purchase price of Branded or authorized generic EpiPens for the
purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries, at any time from August 24, 2011, until the effects of defendants' unlawful conduct cease.
2. Nationwide RICO Damages Class ("RICO Class"). The proposed nationwide RICO Class is coterminous with the Injunctive Class.
3. State Antitrust Damages Class ("State Antitrust Class"). All persons and entities in the Antitrust States who paid or provided reimbursement for some or all of the purchase price of Branded EpiPens at any time from January 28, 2013, until the effects of defendants' unlawful conduct cease, for the purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries.
4. Consumer Protection Damages Class ("CP Class"). All persons and entities in the Consumer Protection States who paid or provided reimbursement for some or all of the purchase price of Branded EpiPens at any time from August 24, 2011, until the effects of defendants' unlawful conduct cease, for
the purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries.
5. Unjust Enrichment Class ("UE Class"). All persons and entities in the Unjust Enrichment States who paid or provided reimbursement for some or all of the purchase price of Branded or authorized generic EpiPens for the purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries, at any time from August 24, 2011, until the effects of defendants' unlawful conduct cease.
Class plaintiffs' original definition for classes 1, 2, and 5 contained the phrase "Branded or AB-rated" EpiPens instead of "Branded or authorized generic EpiPens." An AB-rated device is one the U.S. Food and Drug Administration (FDA) rates as therapeutically equivalent to the brand device. U.S. FOOD & DRUG ADMIN., APPROVED DRUG PRODUCTS WITH THERAPEUTIC EQUIVALENCE EVALUATIONS xvi (2019), https://www.fda.gov/media/71474/download. An authorized generic is an approved brand-name drug that is marketed without the brand name on its label, either by the brand name drug company or another company with the brand company's permission. FDA List of Authorized Generic Drugs, U.S. Food & Drug Admin., https://www.fda.gov/drugs/abbreviated-new-drug-application-anda/fda-list-authorized-generic-drugs (last updated Dec. 19, 2019). Plaintiffs' first version of the proposed definitions seemed to include a mistake because the only AB-rated product available during the defined period is one manufactured by Teva Pharmaceuticals. Plaintiffs report they never intended purchasers of that product to qualify as members of any putative class. And, in their Reply, plaintiffs explain that "reference in the class definitions to 'AB-rated generic EpiPens' should be changed to 'authorized generic EpiPens.'" Doc. 1837 at 13 n.6. Consistent with plaintiffs' request, the court makes this change to the class definitions.
During the hearing on class certification, class plaintiffs' counsel suggested that "until the effects of Defendants' unlawful conduct cease" should refer to the date when class notice is given. Doc. 1789 at 133 (Tr. of Mot. Hr'g Class Certification - Phase II 133:17-18).
Plaintiffs define the "Antitrust States" as: Alabama, Arizona, California, the District of Columbia, Florida, Hawaii, Illinois, Iowa, Kansas, Maine, Michigan, Minnesota, Mississippi, Nebraska, Nevada, New Hampshire, New Mexico, New York, North Carolina, North Dakota, Oregon, Rhode Island, South Dakota, Tennessee, Utah, Vermont, West Virginia, and Wisconsin. As discussed below, even if class certification is warranted under Rule 23, plaintiffs cannot assert class action claims under some of these state laws because no named plaintiff resides in the particular state.
Plaintiffs define the "Consumer Protection States" as: Alaska, California, Connecticut, the District of Columbia, Florida, Hawaii, Illinois, Maine, Maryland, Massachusetts, Missouri, Nebraska, Nevada, New Hampshire, New Mexico, North Carolina, Oklahoma, Rhode Island, Vermont, Washington, and West Virginia. As discussed below, even if class certification is warranted under Rule 23, plaintiffs cannot assert class action claims under some of these state laws because no named plaintiff resides in the particular state.
Class plaintiffs assert all 50 states are "Unjust Enrichment States." If one of the other remedies eventually provides a remedy at law, class plaintiffs intend to exclude Arizona, Delaware, Louisiana, and North Dakota from this class.
B. Proposed Class Exclusions
Plaintiffs' motion lists the following groups as ones who are excluded from each of the classes:
(a) Defendants and their officers, directors, management, employees, subsidiaries, and affiliates;
(b) Government entities, other than government-funded employee benefit plans;
(c) Fully insured health plans (i.e., plans that purchased insurance that covered 100% of the plan's reimbursement obligations to its members);
(d) "Single flat co-pay" consumers who purchased EpiPens or generic EpiPens only via a fixed dollar co-payment that is the same for all covered devices, whether branded or generic (e.g., $20 for all branded and generic devices);
(e) Consumers who purchased or received EpiPens or authorized generic equivalents only through a Medicaid program;
(f) All persons or entities who purchased branded or generic EpiPens directly from defendants; and
(g) The judges in this case and members of their immediate families.
Also, at oral argument on the certification motion, plaintiffs conceded that the classes should exclude another group of entities. Specifically, the court raised concern about including "third-party payors who in effect run their own PBMs." Doc. 1789 at 27 (Tr. of Mot. Hr'g Class Certification - Phase II 27:13-16). Plaintiffs conceded that these "separate entities [who] fill the PBM role . . . should still be excluded from the class[es]." Id. at 27-28 (Tr. of Mot. Hr'g Class Certification - Phase II 27:17-28:4). To the extent plaintiffs seek to limit this exclusion to the PBMs themselves, the court broadens the exclusion to include third-party payors who own or otherwise function as a PBM or control an entity who functions as a PBM. This exclusion is needed to avoid the potential conflict that otherwise could arise by including entities in the classes who allegedly sustained harm from defendants' conduct but who also, allegedly, benefited from defendants' exclusive dealing arrangements.
The following exchange between the court and plaintiffs' counsel occurred at oral argument:
THE COURT: You lost me. Those—you're talking about now CVS and—what was the other one—Humana? They should be excluded?Doc. 1789 at 28 (Tr. of Mot. Hr'g Class Certification - Phase II 28:11-18).
MR. BURNS: Their PBMs should be.
THE COURT: They should be but third-party payors should not?
MR. BURNS: Right. Third-party payors should not.
In their reply brief, plaintiffs offer another exclusion to the class definitions, if necessary. Doc. 1837 at 23. Plaintiffs assert that this exclusion would address the issue of including potential "uninjured class members" in the classes. The proposed exclusion excludes from each class definition individual consumers whose only purchases of an EpiPen occurred before March 13, 2014 (the "Generic Start Date"). To avoid the potential that a class would include a great number of members who sustained no injury, the court adds this exclusion to the class definitions. The court more fully explains the reasons for this additional exclusion to the class definitions in Part IV.E.2.b.i., below.
March 13, 2014 is the date class plaintiffs refer to as the "Generic Start Date," arrived at by their experts as the date when a generic device would have entered the EAI market but for defendants' allegedly anticompetitive behavior. Plaintiffs concede that subsequent discovery may shift this date forward in time, i.e. later than March 2014.
The case authorities permit the court to refine the class definitions. "A court is not bound by the class definition proposed in the complaint and should not dismiss the action simply because the complaint seeks to define the class too broadly." Robidoux v. Celani, 987 F.2d 931, 937 (2d Cir. 1993); see also In re Monumental Life Ins. Co., 365 F.3d 408, 414 (5th Cir. 2004) ("[H]olding plaintiffs to the plain language of their definition would ignore the ongoing refinement and give-and-take inherent in class action litigation, particularly in the formation of a workable class definition."); Sibley v. Sprint Nextel Corp., 254 F.R.D. 662, 671 (D. Kan. 2008) ("If the class definition should require tailoring as the litigation progresses, the Court and parties are authorized to do so." (citing Fed. R. Civ. P. 23(c)(1), (d))).
Plaintiffs also ask the court to select the named plaintiffs as class representatives. For the damages classes, Mylan opposes this request to an extent. It argues certain named plaintiffs are uninjured and are, therefore, inadequate class representatives. The court will address this issue in its analysis of Rule 23(b)(3)'s requirement of typicality, below. See infra Part IV.D.3.b.
Finally, plaintiffs ask the court to appoint Warren T. Burns, Paul J. Geller, Lynn Lincoln Sarko, and Rex Sharp as co-class counsel for class plaintiffs. Mylan does not question the experience of plaintiffs' counsel. But asserts, albeit in conclusory fashion, that it is inappropriate to have the same counsel represent plaintiffs with different theories of injury and competing damages claims. The court will address the request to appoint class counsel at the conclusion of this Order, in Part VI.
C. Class Certification under Rule 23
Plaintiffs seeking to certify a class under Rule 23 must first satisfy Rule 23(a)'s prerequisites of numerosity, commonality, typicality, and adequacy of representation, along with the elements for one of the three types of class actions described in Rule 23(b). As referenced above, class plaintiffs here seek to certify four classes under Rule 23(b)(3) and one class for injunctive relief under Rule 23(b)(2). Mylan argues, that in addition to these explicit requirements, Rule 23 also requires ascertainable classes. In other words, Mylan contends, class plaintiffs must demonstrate the definition of each putative class is sufficiently definite to allow the court to ascertain class membership. Mylan also argues that plaintiffs have ignored the ascertainability requirement. And, their failure to address ascertainability, Mylan contends, requires the court to reject plaintiffs' motion altogether. The court thus begins its analysis of the certification motion by considering whether plaintiffs must satisfy an ascertainability requirement as a prerequisite to certification.
1. Ascertainability
It is unsettled whether ascertainability is a separate and distinct requirement for class certification under Rule 23(b)(3). Some courts, including the Third Circuit, have held that "[t]he predominance and ascertainability inquiries are distinct[.]" Byrd v. Aaron's Inc., 784 F.3d 154, 164 (3d Cir. 2015); see also Brecher v. Republic of Argentina, 806 F.3d 22, 24 (2d Cir. 2015) (recognizing an implied requirement of ascertainability that is "distinct from predominance"). The Third Circuit has explained how the two inquires differ: "'[T]he ascertainability requirement focuses on whether individuals fitting the class definition may be identified without resort to mini-trials, whereas the predominance requirement focuses on whether essential elements of the class's claims can be proven at trial with common, as opposed to individualized, evidence.'" Id. (quoting Grandalski v. Quest Diagnostics Inc., 767 F.3d 175, 184 (3d Cir. 2014)). Thus, courts in the Third Circuit have concluded that "[a]scertainability should not be conflated with other Rule 23 requirements, such as the predominance requirement of Rule 23(b)(3)." In re Wellbutrin XL Antitrust Litig., 308 F.R.D. 134, 147 (E.D. Pa. 2015); see also In re Thalomid and Revlimid Antitrust Litig., No. 14-6997, 2018 WL 6573118, at *18 (D.N.J. Oct. 30, 2018) ("Ascertainability should not be conflated with the predominance requirement.").
The Third Circuit requires a class movant to demonstrate ascertainability by: (1) showing the class is "defined with reference to objective criteria," and (2) providing "some assurance that there can be a reliable and administratively feasible mechanism for determining whether putative class members fall within the class definition." Byrd, 784 F.3d at 164-65 (citation and internal quotation marks omitted); see also Brecher, 806 F.3d at 24 (clarifying that the Second Circuit's test for ascertainability asks "whether the class is sufficiently definite so that it is administratively feasible for the court to determine whether a particular individual is a member" (citation and internal quotation marks omitted)).
In contrast, other Circuits conclude that Rule 23(b)(3) includes an implied requirement of ascertainability, but they have rejected the Third Circuit's "administrative feasibility" standard for determining ascertainability. These Circuits apply a less stringent test. See Mullins v. Direct Dig., LLC, 795 F.3d 654, 657, 662-72 (7th Cir. 2015) (recognizing "an implicit requirement under Rule 23" to establish ascertainability but rejecting the "heightened ascertainability requirement" imposed by the Third Circuit); see also Sandusky Wellness Ctr., LLC v. Medtox Sci., Inc., 821 F.3d 992, 996 (8th Cir. 2016) (discussing criticism of the Third Circuit's standard and refusing to install "a separate, preliminary [ascertainability] requirement"); Rikos v. Procter & Gamble Co., 799 F.3d 497, 525 (6th Cir. 2015) ("We see no reason to follow [the Third Circuit's ascertainability standard], particularly given the strong criticism it has attracted from other courts.").
Here, Mylan begins its ascertainability argument by contending the class definitions for the injunction, RICO, and unjust enrichment classes are defective because they include consumers and entities who purchased or reimbursed Teva's generic EAI. Recognizing this error, plaintiffs have suggested the substitution of "AB-rated" with "authorized generic" in those class definitions, as discussed supra in footnote 6. The court is satisfied that the substituted language cures Mylan's threshold issue.
Returning to the broader ascertainability debate, the court has found no Tenth Circuit case that specifically addresses whether ascertainability is a separate requirement under Rule 23(b)(3). But the most recent class action case from our Circuit includes a footnote acknowledging that the defendant there had asserted Rule 23 includes an implied requirement of ascertainability. See Naylor Farms, Inc. v. Chaparral Energy, LLC, 923 F.3d 779, 788 n.9 (10th Cir. 2019) ("[Defendant] asserts that Rule 23 includes an implied requirement of ascertainability and that [plaintiff] cannot satisfy this implied requirement here." (citations and internal quotation marks omitted)). The Circuit did not address this argument though, because defendant could not show it had raised the issue to the district court. Id. (declining to address the ascertainability argument because defendant "fails to 'cite the precise references in the record where' this ascertainability argument 'was raised and ruled on' below" and "also fails to make a case for plain error on appeal" (quoting 10th Cir. Rule 28.1(A))).
The court recognizes that the Tenth Circuit has held that an ascertainability requirement does not apply to certification under Rule 23(b)(2). See Shook v. El Paso Cty., 386 F.3d 963, (10th Cir. 2004) (noting the Rule 23(b)(2) is "well suited for cases where the composition of a class is not readily ascertainable (emphasis added)). But the court has found no case discussing whether an ascertainability requirement applies to Rule 23(b)(3) cases.
Quoting four words from In re Cox Enterprises, Inc. Set-top Cable Television Box Antitrust Litig., 790 F.3d 1112, 1118 (10th Cir. 2015), Mylan suggests they demonstrate that the Tenth Circuit requires ascertainability. Doc. 1636 at 63. But the phrase "establishing an ascertainable class" appears in a sentence where the court is criticizing plaintiffs for belatedly moving to compel arbitration until after the class was certified. Whether the class should have been certified, and whether it was sufficiently ascertainable, was not at issue in that case. So, the court doesn't view Cox Enterprises the same way Mylan does.
In contrast, this court explicitly considered whether ascertainability is an independent requisite for class certification. See In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2016 WL 5371856, at *2-3 (D. Kan. Sept. 26, 2016). The defendants in that case argued that the putative class proposed by plaintiffs was not sufficiently ascertainable. Id. at *2. Finding no answer in Tenth Circuit decisions, Judge Lungstrum summarized the holdings from other circuits, most of which have imposed a requirement that the proposed class be readily identifiable. Id. (citing Sandusky Wellness Ctr., 821 F.3d at 995 (collecting cases)). Also, Judge Lungstrum recognized that "[a] few circuit courts have applied a stricter standard of ascertainability, by which the plaintiff must show not only that the class is defined by reference to objective criteria, but also that class members may be determined in an economical and 'administratively feasible manner,' such that 'class members can be identified without extensive and individualized fact-finding or mini-trials.'" Id. (first quoting Carrera v. Bayer Corp., 727 F.3d 300, 307 (3d Cir. 2013); then citing Brecher v. Republic of Argentina, 806 F.3d 22, 24 (2d Cir. 2015); and then citing Karhu v. Vital Pharms., Inc., 621 F. App'x 945, 947 (11th Cir. 2015)). But, Judge Lungstrum also noted, "[o]ther courts have criticized and rejected such a heightened standard for ascertainability." Id. (first citing Mullins v. Direct Dig., LLC, 795 F.3d 654, 659-72 (7th Cir. 2015); then citing Sandusky Wellness Ctr., LLC v. Medtox Sci., Inc., 821 F.3d 992, 996 (8th Cir. 2016); and then citing Rikos v. Procter & Gamble Co., 799 F.3d 497, 525 (6th Cir. 2015)).
In the end, Judge Lungstrum was persuaded by the Seventh Circuit's "thorough and well-reasoned analysis" in Mullins. Id. at *3. Mullins had rejected a strict standard requiring an administratively feasible mechanism for identifying class members. Id. at *2 (citing Mullins, 795 F.3d at 662). Specifically, Judge Lungstrum was persuaded that a stricter view of ascertainability furthers no interest not already protected by Rule 23's explicit requirements, nor is it necessary to prevent unfairness to absent class members. Id. (citing Mullins, 795 F.3d at 662-72). Moreover, Judge Lungstrum recognized that the Seventh Circuit had concluded that the risk of fraudulent claims is low, and applying a lower standard does not deny the defendant due process. Id. (citing Mullins, 795 F.3d at 665-72).
After coming to this conclusion, Judge Lungstrum applied the Seventh Circuit's "'weak' version of ascertainability." Id. at *2-3. This test requires that "the class definition must not be too vague, the class must not be defined by subjective criteria, and the class must not be defined in terms of success on the merits." Id. at *2 (citing Mullins, 795 F.3d at 659-60). And, he concluded, the proposed class definitions in Syngenta satisfied the ascertainability requirement because they were "sufficiently definite and objective." Id. at *3-4.
The court concurs with Judge Lungstrum's conclusions. It thus predicts that the Tenth Circuit, if confronted with this question, would decline to recognize ascertainability as a separate, unstated requirement of Rule 23 requiring certification movants to satisfy the more stringent ascertainability standard adopted by the Third Circuit. Instead, the court predicts that our Circuit would follow the Seventh Circuit's approach, applying the less restrictive ascertainability test on certification.
But cf. Cline v. Sunoco, Inc. (R&M), ___ F.R.D. ___, 2019 WL 4879187, at *8-9 (E.D. Okla. Oct. 3, 2019) (recognizing Tenth Circuit does not consider ascertainability a separate Rule 23 factor but concluding that class definition must make class membership determination administratively feasible (citing Carrera, 727 F.3d at 306-07)).
Applying that standard to the class definitions here, the court concludes that they satisfy the Mullins criteria. Specifically, the class definitions are based on objective criteria—i.e., each one defines a class member as a person or entity who paid part of the purchase price for an EpiPen during a specific date range. Also, the definitions aren't impermissibly vague, and they are not defined in terms of success on the merits.
Relying on the opinions of their expert, defendants assert that Prof. James Hughes concluded that none of the data found in the record allows plaintiffs to identify who meets the class definitions and exclusions. See Doc. 1636-4 at 30-40 (Hughes Expert Report ¶¶ 63-84). Thus, defendants argue, plaintiffs can't satisfy the ascertainability standard. But Prof. Hughes's opinions don't identify any ascertainability issues with third-party payors. Doc. 1837-12 at 9 (Hughes Dep. 113:7-114:1). And, as Prof. Hughes testified, to ascertain whether a person or entity qualifies as a class member, he must "demonstrate that [he] paid or provided reimbursement for some or all of the purchase price" of an EpiPen. Id. at 8 (Hughes Dep. 111:16-23). This criteria is sufficiently precise and objective to satisfy the Seventh Circuit's version of the ascertainability standard and, by extension, the standard the court predicts our Circuit would adopt. Thus, the court finds that the proposed class definitions are sufficiently ascertainable.
The court recognizes that Prof. Hughes's Report focuses on the data produced in the case and concludes that "[p]laintiffs have not provided a reliable methodology for identifying who is included or excluded from the class." Doc. 1636-4 at 30-31 (Hughes Expert Report ¶ 63). As discussed above, the court rejects defendants' assertion that plaintiffs must satisfy this heightened standard of ascertainability at certification. But, even if the court applied this heightened standard, plaintiffs assert that they can identify such class members using pharmacy and medical records. Indeed, plaintiffs' expert, Prof. Meredith Rosenthal, has identified various data that, she says, will allow plaintiffs to ascertain who qualifies as a class member. This information includes data from PBMs, pharmacy records, and individuals' own records documenting EpiPen purchases. Doc. 1711-1 at 15-16, 28-29 (Rosenthal Reply Report ¶¶ 20-21, 46-49). Other courts have noted that, in prescription medication cases, this kind of information makes it "fairly easy to identify potential class members through pharmacy and medical records." Krueger v. Wyeth, Inc., No. 03CV2496 JAH (AJB), 2011 WL 8984448, at *2-3 (S.D. Cal. July 13, 2011); see also In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 214 F.R.D. 614, 619 (W.D. Wash. 2003) (contrasting cases "involving . . . prescription medication verifiable by medical and pharmacy records" from the current case involving over-the-counter medication where "the process of simply identifying who rightfully belongs within the proposed class would entail a host of mini-trials"). Thus, even under a heightened ascertainability standard, plaintiffs have identified a viable method for identifying members of the classes.
In reaching this conclusion, the court recognizes that defendants have challenged Prof. Rosenthal's ascertainability opinions with a Daubert motion. By separate Order, the court denies defendants' motion, concluding that Prof. Rosenthal's opinions are sufficiently reliable for the court to consider on class certification. As discussed in that Order, defendants question whether the data identified by Prof. Rosenthal is capable of identifying class members or even if it still exists. The court recognizes that merits discovery may establish that these records can't discharge the task Prof. Rosenthal assigns to them (specifically, identifying consumers who paid nothing for EpiPens, those who used a flat co-pay, or individuals who paid cash for their EpiPens). But, for now, and on the current record, Prof. Rosenthal supplies a reasonably reliable and plausible method for identifying putative class members sufficient to respond to Prof. Hughes's criticisms.
2. Standing
For the most part, defendants concede that plaintiffs have standing. As our court has recognized, "as long as at least one plaintiff has standing with respect to each claim alleged . . . , it will satisfy the injury requirement of standing." Roco, Inc. v. EOG Res., Inc., No. 14-1065-JAR-KMH, 2014 WL 5430251, at *3 (D. Kan. Oct. 24, 2014). But Mylan does challenge plaintiffs' standing in one important respect. It contends plaintiffs lack standing to assert state law claims for states where no named plaintiff resides.
The "reside there" requirement is likely imprecise. The court can imagine a named plaintiff who may have purchased EpiPens in a nearby state where the plaintiff doesn't reside. Such a named plaintiff could have standing to bring a state law claim under the laws of the place of purchase even though the plaintiff doesn't reside in that state. But plaintiffs never argue this nuance and nothing suggests the court needs to consider it here.
Plaintiffs respond, contending they have standing to bring 33 state-law claims because at least one named plaintiff lives in those 33 states. For the other states, plaintiffs cite Roco, and argue they are entitled to amend to add additional plaintiffs to cure standing issues for specific states where, currently, no named plaintiff reside. See id. at *4, 6 (granting leave to amend to add additional plaintiff to fix standing issues). Mylan has the better of this issue. The court declines to certify a class claim for states lacking a named plaintiff to represent the classes. And, the court revises the class definitions to make appropriate exclusions for states where no named plaintiff resides.
Defendants also assert that plaintiffs have a standing problem because many named plaintiffs are uninjured under the legal theories at issue in the certification motion. And, defendants argue, the court cannot certify state law claims for states where the only named representative is an uninjured plaintiff. The court addresses this argument in the typicality section below. See infra Part IV.D.3.b. As that section explains, the current record doesn't establish a standing problem for the named plaintiffs.
In sum, the court finds that plaintiffs have standing to assert their class action claims for states where at least one named plaintiff resides in the state. The court now turns to the Rule 23 requirements.
D. Rule 23(a) Requirements
The universal prerequisites for class certification are (1) numerosity, (2) commonality, (3) typicality, and (4) adequate representation. Fed. R. Civ. P. 23(a). They apply to every certification request made under Rule 23. See Fed. R. Civ. P. 23(b) ("A class action may be maintained if Rule 23(a) is satisfied and if" one of the requirements in Rule 23(b)(1), (b)(2), or (b)(3) is met (emphasis added)). The next four subsections consider these four requirements.
1. Numerosity
To be certified, a proposed class must be so "numerous that joinder of all members is impracticable." Fed. R. Civ. P. 23(a)(1). Defendants do not contest this element, and for good reason. Plaintiffs assert there are hundreds of thousands of EpiPen users and millions of EpiPen devices are sold every year in the United States. Those numbers easily satisfy the numerosity requirement.
See 1 William B. Rubenstein, Newberg on Class Actions § 3:12 (5th ed. 2013) (suggesting that a class of 40 or more members should presumptively satisfy numerosity).
2. Commonality
Next, Rule 23 requires a court to find that "there are questions of law or fact common to the class." Fed. R. Civ. P. 23(a)(2). "Commonality requires the plaintiff to demonstrate that the class members 'have suffered the same injury.'" Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 349-50 (2011) (quoting Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 157 (1982)).
As the Supreme Court has explained, this "same injury" requirement "does not mean merely that they have all suffered a violation of the same provision of law." Id. at 350. Instead, the claims of the putative class plaintiffs "must depend upon a common contention" that is "of such a nature that it is capable of classwide resolution." Id. This requirement "means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke." Id. Although "[e]ven a single [common] question will do," id. at 359 (citation and internal quotation marks omitted), the focus is not on common questions but on "the capacity of a class-wide proceeding to generate common answers apt to drive the resolution of the litigation." Id. at 350 (citation and internal quotation marks omitted).
Defendants do not address commonality as a separate requirement. Instead, Mylan argues that plaintiffs have not met their burden to prove that common questions predominate, nor have they demonstrated commonality for each factual theory of harm plaintiffs assert. Mylan also argues that each of the Rule 23(b)(3) classes fails the commonality and predominance requirements. In other words, Mylan does not make any commonality arguments that are independent of the predominance analysis.
The court find that plaintiffs have satisfied the commonality requirement for each of the five classes they ask the court to certify. They have identified a number of questions with common answers that "would 'resolve an issue that is central to the validity of each one of the claims in one stroke.'" Menocal v. GEO Grp., Inc., 882 F.3d 905, 916 (10th Cir. 2018) (quoting Dukes, 564 U.S. at 350). The questions include: (1) whether defendants used anticompetitive pay-for-delay settlements that delayed generic entry into the EAI market, (2) whether defendants entered unlawful exclusive dealing agreements with PBMs in an effort to protect their monopoly and drive up the price of the EpiPen, and (3) whether defendants unlawfully forced consumers to purchase the EpiPen exclusively in a 2-Pak based on a purported medical necessity when defendants' true motive for removing single EpiPens from the market was to increase profit. Each of these questions involves common factual and legal issues about the existence, scope, and legality of defendants' conduct. See, e.g., Beltran v. InterExchange, Inc., No. 14-CV-03074-CMA-CBS, 2018 WL 1948687, at *4 (D. Colo. Feb. 2, 2018) ("Numerous courts have determined that the commonality prerequisite of Rule 23(a) was satisfied in antitrust and RICO class actions" because they involve common factual and legal questions about defendants' conduct). To resolve these questions, plaintiffs will rely on common evidence to generate common answers. Id. (concluding that plaintiffs satisfied the commonality requirement where "numerous class-wide questions of law and fact may be resolved with common answers drawn from common proof"). Also, as our Circuit has recognized, "any one of these questions alone would satisfy the commonality requirement . . . ." Menocal, 882 F.3d at 916-17; see also id. at 914 ("A finding of commonality requires only a single question of law or fact common to the entire class." (citation and internal quotation marks omitted)).
In sum, each of the putative classes satisfies the commonality requirement of Rule 23(a)(2).
3. Typicality
The typicality element requires a finding that the claims of representative plaintiffs are typical of the claims of the proposed class. Fed. R. Civ. P. 23(a)(3). But, the named plaintiffs' interests and claims "need not be identical" to those of the putative class members. DG ex rel. Stricklin v. Devaughn, 594 F.3d 1188, 1198 (10th Cir. 2010). As long as the claims of the named plaintiff and class members "are based on the same legal or remedial theory, differing fact situations of the class members do not defeat typicality." Id. at 1198-99; see also Menocal v. GEO Grp., Inc., 882 F.3d 905, 914 (10th Cir. 2018) (same).
The typicality requirement asks:
whether a sufficient relationship exists between the injury to the named plaintiff and the conduct affecting the class, so that the court may properly attribute a collective nature to the challenged conduct. In other words, when such a relationship is shown, a plaintiff's injury arises from or is directly related to a wrong to a class, and that wrong includes the wrong to the plaintiff. Thus, a plaintiff's claim is typical if it arises from the same event or practice or course of conduct that gives rise to the claims of other class members, and if his or her claims are based on the same legal theory.In re Pharm. Indus. Average Wholesale Price Litig., 230 F.R.D. 61, 78 (D. Mass. 2005) (quoting In re Am. Med. Sys., Inc., 75 F.3d 1069, 1082 (6th Cir. 1996) (quoting 1 Herbert B. Newberg & Alba Conte, Newberg on Class Actions § 3.13 (3d ed. 1992))).
The typicality and commonality requirements are similar because "'[b]oth serve as guideposts for determining . . . whether the named plaintiff's claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected . . . .'" Dukes, 564 U.S. at 349 n.5 (quoting Falcon, 457 U.S. at 157 n.13). The typicality "requirement has been liberally construed by courts[,]" and "in the antitrust context, typicality 'will be established by plaintiffs and all class members alleging the same antitrust violations by defendants.'" In re Vitamins Antitrust Litig., 209 F.R.D. 251, 260 (D.D.C. 2002) (quoting In re Playmobil Antitrust Litig., 35 F. Supp. 2d 231, 241 (E.D.N.Y. 1998)). "If the claims of the named plaintiffs and putative class members involve the same conduct by the defendant, typicality is established regardless of factual differences." Newton v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 183-84 (3d Cir. 2001).
Plaintiffs assert every class member here relies on the same legal theory—they all paid artificially elevated prices for or were oversold EpiPens—and the claims of the named plaintiffs are typical of the claims of the proposed classes. Mylan disagrees, arguing (1) named plaintiff Local 282 Welfare Trust Fund ("Local 282") is not a typical payor, and (2) many of the named plaintiffs are not typical because they are uninjured. The court examines Mylan's two arguments, below.
Mylan makes these same arguments to assert that the named plaintiffs are inadequate class representatives. Thus, Mylan also contends that plaintiffs can't satisfy Rule 23(a)'s fourth requirement—adequacy. The court addresses the adequacy argument below, infra note 24.
a. Local 282
Local 282 is an employee health and welfare benefit plan, and it is a named plaintiff in this lawsuit. Mylan argues Local 282 is atypical and its interests actually conflict with other third-party payors because it does not bear the price risk of increased drug costs. According to Mylan, Local 282's costs are fully reimbursed by the employers who fund it. In other words, if costs increase, Local 282 simply goes back to the union and negotiates higher contribution rates for union contributions to the Welfare Trust Fund.
To support its argument, Mylan cites the deposition testimony of a Local 282 representative. The representative was asked what it would mean for an insured member of Local 282's health plan if Local 282 prevails in this lawsuit. The representative answered by explaining that members of the plan are not insured. Instead, he testified, plan members are owners of self-insured trust funds. If the plan were to receive reimbursement funds by prevailing in this action, it would:
stem[] the tide of potentially having to go back—for the union to have to go back and negotiate higher contribution rates to the welfare trust fund. If those rates do not have to go up, then the members in a roundabout way can then receive higher wages. . . . They will not be reimbursed with a check per se after settlement if one does occur.Doc. 1636-34 at 4-5 (Bulding Dep. 59:21-60:7).
Citing the testimony of the same union representative, plaintiffs dispute the proposition that Local 282 bears no price risk. Asked who pays the remaining cost of an EAI for a member who pays a co-pay for the device, the witness answered that the Local 282 Welfare Trust Fund pays the cost. Doc. 1839-10 at 4-5 (Bulding Dep. 197:10-198:18). No insurance company pays any portion of the remaining cost, nor does the Welfare Trust Fund seek reimbursement for any part of that cost from any other person or entity. Id.
Plaintiffs also note that Dr. Johnson's report concedes that self-insured plans—such as Local 282's plan—pay for claims using their own funds, including premiums collected from employees, and assume all insurance risk. Dr. Johnson acknowledges that these self-insured plans are responsible for a portion (or even all) of the cost of covered drugs. Doc. 1636-2 at 39-40 (Johnson Expert Report ¶ 41).
John Johnson, Ph.D. is an antitrust economist and statistician, retained by defendants, to evaluate the damage models prepared by plaintiffs' experts, Professors Rosenthal and Elhauge. Defendants did not ask Dr. Johnson to prepare an affirmative methodology for assessing impact or damages, and he did not offer such an assessment. See Doc. 1724 at 91-92 (Tr. of Mot. Hr'g Class Certification - Phase I 91:25-92:5 ("Q: Were you asked to offer any affirmative assessment of a classwide method for assessing impact or damages in this case? A: No. My assignment was not to put forward an affirmative methodology, it was to assess the work of Professor Rosenthal and Professor Elhauge.").
The court rejects Mylan's assertion that Local 282 bears no price risk for increased drug costs. The record simply won't support that conclusion. Even if a union whose members participate in Local 282 negotiated higher contributions from employers sufficient to offset increased costs caused by EpiPen overcharges, that is a forward-looking exercise. Local 282 still has sustained injury when the benefit plan was overcharged (if it was). See In re Solodyn (Minocycline Hydrochloride) Antitrust Litig., No. 14-md-02503, 2017 WL 4621777, at *18 (D. Mass. Oct. 16, 2017) (explaining that even when institutional payors can offset overcharges by increasing premiums after the fact, the institutional payors "experience[ ] antitrust impact at the time of the initial overcharge"). Mylan's argument does not persuade the court that Local 282's injuries are atypical of the classes it seeks to represent.
Mylan next asserts that Local 282, because it is a small not-for-profit trust who does not own its own PBM or negotiate directly with Mylan, does not have the same bargaining power as large payors with millions of members. As a consequence, Mylan argues, the court shouldn't let Local 282 represent the interests of other payors who have not sought to participate in this case and who have different interests. Mylan cites third-party payors such as Anthem, Express Scripts, Inc., Caremark, Change Healthcare, and Humana as companies with a significantly greater number of members and larger revenues. Describing the differences between Local 282 and large third-party payors as a fundamental and uncontroverted conflict, Mylan cites deposition testimony from Humana, Anthem, Cigna Presbyterian, Magellan, MedImpact, Aetna, and United showing each one favors rebates because they produce lower costs. Doc. 1636 at 87-88. It's no surprise those third-party payors would favor higher rebates and the profits they capture from the current pricing regime. But, Mylan offers no evidence tending to establish a "fundamental and uncontroverted conflict" between larger third-party payors and Local 282, id. at 87, nor does it make a convincing argument that the number of members in larger third-party payors affects Local 282's typicality. Our court's precedent is more forgiving than Mylan suggests.
For instance, Mylan quotes Nieberding v. Barrette Outdoor Living, Inc., 302 F.R.D. 600, 610 (D. Kan. 2014) as recognizing that typicality requires all class members to be "at risk of being subjected to the same harmful practices." But this characterization of Nieberding, while accurate, stops too soon. The court's observation about typicality continued beyond the words quoted by Mylan: "The interests and claims of the named plaintiff and class members need not be identical to satisfy typicality; so long as the claims of the named plaintiff and class members are based on the same legal or remedial theory, differing fact situations of the class members do not defeat typicality." Id. (citing Devaughn, 594 F.3d at 1198-99). Under this standard, it is possible for Local 282 to serve as a class representative of third-party payors even if some of them were harmed by defendants' alleged efforts to prevent competitors from entering the market but also benefited from rebates Mylan paid them. Also, the court already has addressed Mylan's concerns. As discussed above, and during oral argument, see Doc. 1789 at 27-28 (Tr. of Mot. Hr'g Class Certification - Phase II 27:13-28:18), the court will exclude from the class definitions third-party payors who own or otherwise function as a PBM (or control any entity that functions as a PBM). Mylan's argument that Local 282 cannot represent the interests of payors who benefit from owning or controlling their own PBM is unpersuasive.
Finally, Mylan contends Local 282 is an atypical representative because it lacks familiarity with how rebating works. Mylan cites testimony by a representative of Local 282 but it does not support the conclusion Mylan attributes to it. See Doc. 1636 at 85 n.120, 88 n.126. The witness was never asked to explain how rebating works. Instead, the submitted transcript pages show that Local 282 used the services of The Segal Group to negotiate with PBMs on its behalf, and that it joined other benefit trust funds as a member of the National Alliance of Labor Health Care Organizations to negotiate collectively with PBMs for a better deal. Doc. 1636-34 at 7, 8 (Bulding Dep. 98:5-14, 103:4-18). Moreover, Mylan never explains why it matters whether Local 282 negotiates its own rebates or, instead, hires The Segal Group to negotiate on its behalf. Indeed, it appears common for third-party payors to employ services of a consultant like The Segal Group. See In re Pharm. Indus. Average Wholesale Price Litig., 230 F.R.D. 61, 72 (D. Mass. 2005) ("A TPP considering the use of a PBM will typically send out a request for proposals, describing its goals. . . . Throughout the process, TPPs are commonly advised by benefits consulting companies like Segal Company . . . In virtually all instances, self-insured employers and union benefit funds retain consultants to represent them in negotiations with PBMs . . . .").
The court is not persuaded that Local 282 is atypical, nor is the court convinced that it should reject Local 282 as a class representative because its interests conflict with the classes it is nominated to represent. Defendants have failed to support their claim that any particular number of third-party payors are potentially adverse, nor have they demonstrated that any third-party payors (other than those with their own PBMs) are indeed adverse. The court has excluded from the class definitions any third-party payors who have separate entities filling the PBM role. In short, the injury claimed by Local 282 "arises from or is directly related to the wrong" claimed by the classes. In re Pharm. Indus. Average Wholesale Price Litig., 230 F.R.D. at 78 (citations and internal quotation marks omitted). Plaintiffs have discharged their burden to demonstrate that Local 282's claims typify those that the classes seek to assert.
Although not highlighted in Mylan's brief, Dr. Hughes's Expert Report notes: "Of the 50,216 group health plans of all types in the US in 2009, only 1,801, or less than 4 percent, are multiemployer, collectively bargained plans like Local 282." Doc. 1636-4 at 24 n.58 (Hughes Expert Report ¶ 47 n.58) (citing U.S. DEPT. OF LABOR, EMP. BENEFITS SEC. ADMIN., GROUP HEALTH PLANS REPORT: ABSTRACT OF 2009 FORM 550 ANNUAL REPORTS REFLECTING STATISTICAL YEAR FILINGS 6, tbl.A2 (2012), https://www.dol.gov/sites/dolgov/files/EBSA/researchers/statistics/retirement-bulletins/annual-report-on-self-insured-group-health-plans-2012-appendix-a.pdf.). While multiemployer, collectively-bargained plans may represent a relatively small proportion of group plans, the raw number of plans similar to Local 282 is nonetheless a significant number of similar plans.
b. Named Plaintiffs
Next, Mylan argues that most of the other named plaintiffs are atypical of the classes they seek to represent because they are uninjured. By Mylan's count, nine named plaintiffs cannot establish injury under any of plaintiffs' legal theories; more than 75% are uninjured under the 2-Pak theory based on their reported purchases; at least 50% are uninjured under the Auvi-Q foreclosure theory based on their reported purchases (or lack of them) during the relevant period; and at least 50% are uninjured under the generic delay theory based on their preferences. Mylan also asserts no named plaintiff can represent the RICO class because every named plaintiff either acknowledged that they did not rely on an alleged misrepresentation or their testimony requires additional inquiry to determine whether they relied on a misrepresentation.
Mylan's RICO argument is unpersuasive. The Supreme Court has held—explicitly—that first-party reliance is not required to establish a RICO claim premised on the predicate act of mail fraud. Bridge v. Phoenix Bond & Indem. Co., 553 U.S. 639, 649 (2008). The Supreme Court recognized that RICO provides a cause of action for "'[a]ny person injured in his business or property by reason of a violation of'" the statute. Id. (quoting 18 U.S.C. § 1964(c)). And, the Court found it "difficult to derive a first-party reliance requirement" from the words of the statute. Id. Thus, the Supreme Court refused to "read a first-party reliance requirement into a statute that by its terms suggests none." Id. at 650.
Applied here, Bridge means the named plaintiffs need not establish first-party reliance to assert their fraud-based RICO claim. Instead, the named plaintiffs just need to establish an injury to business or property "by reason of" defendants' RICO violations. Plaintiffs assert that the named plaintiffs satisfy this requirement for their RICO claim—and all other claims—because, they allege, all named plaintiffs have sustained injury from defendants' conduct. Specifically, plaintiffs assert that each named plaintiff sustained injury under at least one of their theories—the generic delay theory. And, they argue, this showing suffices to establish that each named plaintiff sustained an injury from defendants' conduct.
But Mylan argues that many named plaintiffs aren't injured under this theory. According to Mylan, seven named plaintiffs preferred the branded EpiPen or purchased a branded EpiPen after the launch of the EpiPen AG (an authorized generic of the EpiPen, made by Pfizer and sold by Mylan); at least six named plaintiffs paid less for their EpiPens than plaintiff's expert, Prof. Rosenthal, calculates they would have paid in the but-for world; and at least one named plaintiff concedes that it is his own fault that he never researched whether alternative generic EAIs were available for him to purchase. Plaintiffs respond with three reasons supporting their argument that all named plaintiffs sustained injury under the generic delay theory.
First, plaintiffs argue that a consumer's preference for or purchase of a branded EpiPen instead of the EpiPen AG is not an appropriate proxy to demonstrate the absence of injury. Plaintiffs cite Prof. Rosenthal's opinions and argue that each named plaintiff sustained injury from the delayed generic entry because the delay affected consumer switching patterns. Prof. Rosenthal calculates plaintiffs' alleged damages using "various but-for entry scenarios," using an assumed generic entry date of March 13, 2014. Doc. 1711-1 at 5, 36-37 (Rosenthal Reply Report ¶¶ 4.e, 60). And, in the but-for world, Prof. Rosenthal calculates a generic penetration rate, recognizing several factors that would drive consumers to switch from a branded EpiPen to a generic. See id. at 19-20 (Rosenthal Reply Report ¶ 28 (explaining that consumers are likely to switch to a generic because the EpiPen is more expensive and because it is a product that consumers don't use regularly, thus they aren't likely to develop brand affinity)). As Prof. Rosenthal explained, her opinions take into account consumer switching patterns in a but-for world, which she explained, would have begun once the authorized generic becomes available. See Doc. 1711-5 at 7-8 (Rosenthal Dep. 145:15-146:23). And, she didn't use market share data from the EpiPen AG's entry to form her opinions because the entry of an authorized generic sold by Mylan wasn't a competitive situation and not the same as an independent generic device entering the market. Id. at 9-10 (Rosenthal Dep. 156:2-160:18). Based on Prof. Rosenthal's opinions, plaintiffs have made a plausible showing that each named plaintiff—even if he or she purchased the branded EpiPen in the real world after the launch of the EpiPen AG—sustained injury from the delayed generic entry.
Defendants have sought to exclude Prof. Rosenthal's opinions about injured class members by Daubert motion. By separate Order, the court denies defendants' motion, concluding the opinions are sufficiently reliable for the court to consider at class certification. The court concludes in this Order that Prof. Rosenthal's opinions support plaintiffs' assertion that each named class member sustained injury. The court recognizes that defendants' expert has criticized Prof. Rosenthal's assumptions and attacked her calculated generic penetration rate as flawed. But, the court finds that Prof. Rosenthal has shown she used a reasonable method to reach her opinions, and her opinions plausibly establish that each named class member sustained an injury.
Also, plaintiffs' expert, Prof. Elhauge, opines that brand loyalists sustained harm under the PBM foreclosure theory because "[t]he foreclosure of brand-brand competition between EpiPen and Auvi-Q would predictably have an additional adverse price effect not only on customers who would have continued to buy branded EpiPens but paid a lower price in the but-for world for it, but also on customers who would have bought a generic at a discount from what would have been a lower branded EpiPen price." Doc. 1711-2 at 89-90 (Elhauge Reply Report ¶ 123). So, Prof. Elhauge opines, "during the overlap period [i.e., months when there is an overlap between the foreclosure of Auvi-Q and the exclusion of the generic], all customers would suffer one harm or the other, whether or not they were brand loyalists." Id.
Second, plaintiffs refute defendants' argument that some named plaintiffs aren't injured because they paid less for their EpiPens than Prof. Rosenthal calculates they would have paid in the but-for world. Plaintiffs argue that comparing an insured's actual co-pay amount to a but-for co-pay amount improperly equates real-world transactions with hypothetical averages used as input for a damages calculation. As Prof. Rosenthal explains, her hypothetical average is just that—an average of aggregate damages. Doc. 1711-1 at 21 (Rosenthal Reply Report ¶ 32). Prof. Rosenthal asserts that when the time comes to determine individual damages for a named plaintiff, "the relevant question will be what the Class member's generic copayments would have been at the same point in time as the brand purchase they made." Id. So, Prof. Rosenthal credibly explains, her calculations don't suggest that any named plaintiffs escaped injury because of what they paid for their EpiPens. Id. The court finds this explanation sufficient to establish that each named plaintiff has a colorable claim of injury under the generic delay theory.
Finally, plaintiffs contend, because generic delay involves a but-for world, whether a consumer actually knew about EpiPen alternatives is a meaningless point. As Prof. Rosenthal opines, in the but-for world, the rates at which consumers would switch from the branded EpiPen to a generic differ from what actually happened in the real world where, plaintiffs allege, defendants prevented a generic product from entering the market at an earlier date. Id. at 19-20 (Rosenthal Reply Report ¶ 28); Doc. 1711-5 at 7-8, 9-10 (Rosenthal Dep. 145:15-146:23, 156:2-160:18).
In sum, the court concludes plaintiffs have presented plausible evidence sufficient to establish that each named plaintiff has sustained an injury under at least one of plaintiffs' theories—i.e., the generic delay theory. Prof. Rosenthal's opinions sufficiently support plaintiffs' assertions that each named plaintiff sustained injury sufficient to make their legal theories typical of the other class members they seek to represent. See Menocal v. GEO Grp., Inc., 882 F.3d 905, 914 (10th Cir. 2018) (holding that plaintiffs satisfied the typicality requirement where "the claims of all the class members—including the representatives—share the same theory" and "present no circumstances that would give rise to a different theory of liability"). The court concludes plaintiffs have satisfied Rule 23's typicality requirement.
Because each named plaintiff has sustained injury under one of plaintiffs' theories, the named plaintiffs are typical representatives of the classes they seek to represent. To the extent named plaintiffs were injured by plaintiffs' other two theories (i.e., the 2-Pak and PBM foreclosure theories), those plaintiffs are typical of class members who also allegedly sustained injury under the other two theories. The court recognizes that if plaintiffs' generic delay theory fails on its merits, the court may have to revisit the typicality requirement to determine whether the named plaintiffs are typical of the classes they seek to represent.
4. Adequacy
A certified class must have representative parties who will fairly and adequately protect the interests of the class. Fed. R. Civ. P. 23(a)(4). The adequacy requirement asks whether the named plaintiffs and counsel (a) "have any conflicts of interest with other class members" and (b) will prosecute the action vigorously on behalf of the class. In re Motor Fuel Temperature Sales Practices Litig., 292 F.R.D. 652, 671 (D. Kan. 2013) (citing E. Tex. Motor Freight Sys., Inc. v. Rodriguez, 431 U.S. 395, 403 (1977) (further citations omitted)). "Minor conflicts among class members do not defeat certification[.]" Id. Only a "fundamental conflict" about the specific issues in controversy will prevent a named plaintiff from representing the interests of the class adequately. Id. A fundamental conflict exists where some class members claim an injury resulting from conduct that benefited other class members. Id. Defeating "[the] adequacy requirement of Rule 23" requires a conflict that is "more than merely speculative or hypothetical." Gunnells v. Healthplan Servs., Inc., 348 F.3d 417, 430 (4th Cir. 2003) (citation and internal quotation marks omitted).
Mylan submits that significant intra-class conflicts exist, making the named plaintiffs inadequate class representatives and thus precluding certification. Mylan described the first such conflict by arguing that Local 282 is not typical of other payors who benefit from Mylan's rebating practices that plaintiffs challenge. The court already has addressed this challenge, see supra Part IV.D.3.a., and explained why it is unpersuasive.
As discussed above, supra note 18, Mylan also asserts that Local 282 is an inadequate class representative because it differs from other payors in the classes. Also, Mylan asserts that many named plaintiffs are inadequate class representatives because they are uninjured. The court rejected Mylan's arguments in the preceding section, finding that these arguments don't establish that the named plaintiffs are atypical of the classes they seek to represent. See supra Part IV.D.3. For the same reasons, the court concludes that defendants' arguments don't prevent the named plaintiffs from fairly and adequately representing the classes.
Mylan also asserts that PBMs likely are class members, and that conflicts exist between plaintiffs and PBMs because plaintiffs allege the PBMs are RICO co-conspirators. As plaintiffs point out, however, no PBM can qualify as a class member under the class definitions. The proposed definition requires that the entity must have (a) paid or provided reimbursement for an EpiPen; (b) for the purpose of consumption, not resale; and (c) that purchase was intended for consumption by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries. Also, the court has added an exclusion to the class definitions, excluding all third-party payors who own, operate, or control their own PBM. See supra Part IV.B. So, nothing suggests a PBM will qualify under the class definitions. Mylan certainly hasn't identified any PBMs who would qualify.
Mylan also contends a conflict exists between consumers and payors within the proposed classes because plaintiffs have not tried to apportion among themselves the alleged overcharge for each EpiPen purchase. Delving into what it perceives as plaintiffs' strategy, Mylan describes how plaintiffs could have avoided an alleged conflict by proposing a class consisting solely of consumers, thereby avoiding prejudice to the legal arguments and damage recovery for both consumers and payors. Plaintiffs respond, saying the allocation between consumers and third-party payors will be governed by factors such as the proportion each of them paid toward the cost of the EpiPen in a given transaction. At this point, however, the issue is merely speculative. Mylan has not shown that the composition of any class will create internal competition for damages shares. Moreover, binding authority from the Circuit recognizes that defendants can claim "no interest in the method of distributing the aggregate damages award among the class members." In re Urethane Antitrust Litig., 768 F.3d 1245, 1269 (10th Cir. 2014).
Also, other courts have considered objections similar to Mylan's argument here and have rejected the challenges. For example, in In re Lidoderm Antitrust Litigation, No. 14-md-2521-WHO, 2017 WL 679367 (N.D. Cal. Feb. 21, 2017), defendants asserted that conflicts between consumers and third-party payors precluded certification. Id. at *26. Specifically, defendants in that case asserted (a) the amount of overcharge damages would need to be assigned between third-party payors and consumers, "putting the parties into conflict over who gets what recovery," and (b) some third-party payors "would prefer different legal theories about what would have happened in the but-for world, creating conflicts." Id. The court rejected both challenges. First, the court found defendants had not shown the alleged conflict "would permeate the aggregate damages calculation." Id. Instead, the court reasoned, the conflict would arise when damages are allocated. Id. And, at that point, claims mechanisms may be used to resolve any conflict "between, for example, an end payor consumer and her health insurance plan over how their overcharge damages should be split." Id. Thus, the court concluded the purported conflict "[did] not create a type of conflict that precludes certification." Id. Second, the court found the mere possibility of differing opinions over damage theories might arise among third-party payors did not create a conflict preventing certification. Id. The court noted that, if such a situation arose, then third-party payors who wanted to pursue different damage theories simply could opt out of the class. Id.
Likewise, in In re Solodyn (Minocycline Hydrochloride) Antitrust Litigation, No. 14-md-02503, 2017 WL 4621777 (D. Mass. Oct. 16, 2017), defendants challenged the end payors' showing of adequacy of representation because third-party insurers and consumers are fundamentally different groups. Id. at *12. The court rejected this argument, finding those parties shared a common goal. Each wanted to show they all were injured in the same way by overcharges, and through the same illegal conduct by defendants. Id. This, the court ruled, created an "alignment of incentives" sufficient to overcome the conflict challenge. Id. Also, defendants argued that class members would compete with one another for damage shares. Id. at *13. Citing Lidoderm, the court concluded it could address this issue, if necessary, when damages were allocated. Id. at *13 (citing In re Lidoderm Antitrust Litig., 2017 WL 679367, at *26-27).
Likewise, here, defendants don't persuade the court that the alleged conflict between consumers and third-party payors precludes class certification. To the contrary, the current record shows that consumers and third-party payors share an "alignment of incentives" because they both seek to prove that defendants' illegal conduct caused them to sustain injuries in the form of overpaying for the EpiPen. In re Solodyn (Minocycline Hydrochloride) Antitrust Litig., 2017 WL 4621777, at *12. Defendants haven't shown that the alleged conflict in apportioning damages among consumers and third-party payors "would permeate the aggregate damages calculation." In re Lidoderm Antitrust Litig., 2017 WL 679367, at *26 (emphasis added). As other courts have recognized, the court can address this issue, if a conflict actually emerges, when damages are allocated. In re Solodyn (Minocycline Hydrochloride) Antitrust Litig., 2017 WL 4621777, at *13 ("Defendants have not shown that alleged conflicts among putative class members would 'permeate the aggregate damages calculation' and not merely arise at the time damages are allocated, which can be addressed then." (quoting In re Lidoderm Antitrust Litig., 2017 WL 679367, at *26)). For these reasons, the court doesn't find an intra-conflict exists that precludes certification.
Instead, the court concludes that plaintiffs have established that the named plaintiffs can fairly and adequately represent the interests of the classes. As discussed above, the named plaintiffs' claims are typical of the classes they seek to represent because they allege they sustained the same injuries as a result of defendants' illegal conduct. And, based on the history of the case to date, the named plaintiffs have prosecuted the action vigorously on behalf of the classes through discovery and the class certification proceedings. See Syngenta, 2016 WL 5371856, at *5 (finding the adequacy requirement was met where "[t]he class representatives' claims are typical" and defendant hasn't "suggested any reason why the class representatives lack sufficient incentive to prosecute fully the claims on the behalf of the classes"). The court thus concludes that plaintiffs satisfy Rule 23's adequacy requirement.
In sum, the court finds plaintiffs satisfy each prerequisite for class certification under Rule 23(a). Accordingly, the analysis now turns to the elements of Rule 23(b)(3) in the context of the proposed RICO, state antitrust, consumer protection, unjust enrichment, and injunctive classes.
E. Rule 23(b)(3) Requirements
1. Superiority
To certify a class under Rule 23(b)(3), a class action proceeding must be "superior to other available methods for fairly and efficiently adjudicating the controversy." Fed. R. Civ. P. 23(b)(3). "It is enough that class treatment is superior because it will 'achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.'" CGCHolding Co. v. Broad & Cassel, 773 F.3d 1076, 1096 (10th Cir. 2014) (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 615 (1997)).
Mylan contests superiority. It asserts that differently situated plaintiffs with divergent claims, many of whom are uninjured, will require extended, individualized examination and cross-examination. So, Mylan contends, plaintiffs have not shown a class action is the most efficient way to litigate the claims at issue.
Plaintiffs assert the individual claims are negative value claims: they would cost far more to litigate than they are worth. They also assert that most of the class members are geographically dispersed. So, plaintiffs say, the claims of individual class members will be doomed if individual members are forced to take on defendants alone.
The scale of this case, the desirability for uniformity of decision, and the continued economies of time, effort, and expense all support a finding that class treatment is the superior method to decide this dispute fairly and efficiently. Defendants identify no better alternative method, and a class action will protect defendants from multiple or inconsistent verdicts. The court thus finds that the Rule 23(b)(3) superiority requirement is satisfied here.
2. Predominance
Rule 23(b)(3) also requires a plaintiff to "show that common questions subject to generalized, classwide proof predominate over individual questions." CGC Holding Co. v. Broad & Cassel, 773 F.3d 1076, 1087 (10th Cir. 2014). This requirement does not mean a plaintiff must show "that all of the elements of the claim entail questions of fact and law that are common to the class" or "that the answers to those common questions [are] dispositive" of the claim. Id. (citation omitted). Instead, the predominance inquiry "'asks whether the common, aggregation-enabling issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.'" Id. (quoting 2 William B. Rubenstein, Newberg on Class Actions § 4:40 (5th ed. 2012)). To decide whether a plaintiff can satisfy the predominance requirement, a court must first "characterize the issues in the case as common or not, and then weigh which issues predominate." Id. Critically, so long as at least one common issue predominates, a plaintiff can satisfy Rule 23(b)(3)—even if individual issues (such as damages) remain and require the court to try them separately. Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036, 1045 (2016).
The determination "'whether "questions of law or fact common to class members predominate" begins, of course, with the elements of the underlying cause of action.'" CGC Holding Co., 773 F.3d at 1088 (quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011) (quoting Fed. R. Civ. P. 23(b)(3))). Here, plaintiffs argue, common issues predominate each of their asserted legal claims because they intend to prove the elements of their RICO, antitrust, state consumer protection, and unjust enrichment claims using common, classwide evidence. Mylan disagrees. Mylan argues that plaintiffs have not met their burden to show common issues predominate over individual questions. Generally, Mylan asserts three arguments to support its position that individual issues overwhelm the common questions here, thus precluding certification.
First, Mylan argues, the damage calculations for each plaintiff require individualized inquiries that will overwhelm the common questions for all of their legal claims. Second, Mylan asserts, plaintiffs cannot prove through common evidence that all class plaintiffs sustained injury. Finally, Mylan argues, individual questions predominate over the common issues presented by each of plaintiffs' five proposed classes: the RICO class, the state antitrust class, the state consumer protection class, the unjust enrichment class, and the injunctive class. The court addresses each argument, in turn, below.
a. Individual Damages Issues
Mylan argues plaintiffs have not met their burden to show common issues predominate because, according to Mylan, individual questions about each plaintiff's damages calculation overwhelm the common issues. More specifically, Mylan asserts that plaintiffs haven't provided a common method for measuring damages for each class member. Mylan contends that the damage calculation for each plaintiff requires an individual inquiry about the dates and amounts of each EpiPen purchase, the specifics of the individual's health plan and rebate arrangements, and the action that defendants took to cause that particular plaintiff to sustain harm (i.e., a plaintiff claiming generic delay could not recover damages before the Generic Start Date or after generic delay ended).
Plaintiffs respond that the Tenth Circuit has rejected the argument that individual damages issues preclude certification. Indeed, the Tenth Circuit recently held: "'The fact that damages may have to be ascertained on an individual basis is not, standing alone, sufficient to defeat class certification.'" Naylor Farms, Inc. v. Chaparral Energy, LLC, 923 F.3d 779, 798 (10th Cir. 2019) (quoting Menocal v. GEO Group, Inc., 882 F.3d 905, 922 (10th Cir. 2018)). Instead, the Circuit explained, "'material differences in damages determinations'" will destroy predominance only "if those 'individualized issues will overwhelm . . . questions common to the class.'" Id. (quoting Wallace B. Roderick Revocable Living Tr. v. XTO Energy, Inc., 725 F.3d 1213, 1220 (10th Cir. 2013)). In Naylor Farms, the Tenth Circuit concluded that plaintiffs' expert report provided evidence that the expert could determine damages on a classwide basis through use of a specified model. Id. And, the court noted, defendant had failed to explain why the expert's model was inaccurate or unworkable. Id. The Circuit thus affirmed the district court's conclusion that individual questions about damages did not defeat predominance. Id.
This court has concluded that the Supreme Court's Comcast opinion does not require a classwide damages model for certification under Rule 23(b)(3). In Syngenta, Judge Lungstrum rejected defendant's argument that plaintiffs "must provide a damages model that 'establish[es] that damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3).'" 2016 WL 5371856, at *9 (citing Comcast Corp. v. Behrend, 569 U.S. 27, 35 (2013)). Judge Lungstrum explained: "Multiple circuit courts . . . including the Tenth Circuit, have rejected such a reading of Comcast, which did not create a broadly applicable rule as suggested by [defendant]." Id. (first citing XTO, 725 F.3d at 1220; then citing Neale v. Volvo Cars of N. Am., LLC, 794 F.3d 353, 374-75, 75 n.10 (3d Cir. 2015); then citing In re Deepwater Horizon, 739 F.3d 790, 815 (5th Cir. 2014); then citing In re Urethane Antitrust Litig., 768 F.3d 1245, 1258 (10th Cir. 2014)); see also In re Urethane Antitrust Litig., 768 F.3d at 1258 (refusing to apply Comcast because, "unlike the claimants in Comcast, our plaintiffs did not concede that class certification required a method to prove class-wide damages through a common methodology")); XTO, 725 F.3d at 1220 (explaining that, after Comcast, still "there are ways to preserve the class action model in the face of individualized damages").
Even so, Judge Lungstrum then recognized that "plaintiffs here do rely on a class-wide method to show damages", using their experts' models to calculate aggregate damages. Syngenta, 2016 WL 5371856, at *9-10; see also In re Urethane Antitrust Litig., 768 F.3d at 1268-69 (affirming jury's damage award calculated on an aggregate basis and, after trial, was distributed among the class members using an expert's damages model). And, Judge Lungstrum concluded, this method sufficed to establish "that common questions predominate here despite any possible individual questions" about damages. Syngenta, 2016 WL 5371856, at *10.
Applying the Tenth Circuit's holdings in Naylor Farms and Urethane and Judge Lungstrum's reasoning in Syngenta to the facts here, the court concludes that the individual damages questions that defendants have identified don't preclude class certification. Like the plaintiffs in Syngenta and Urethane, plaintiffs here have proposed a common methodology for calculating damages on an aggregate basis through opinions by their experts, Profs. Rosenthal and Elhauge. Specifically, Prof. Rosenthal has proposed a methodology for calculating damages classwide resulting from defendants' alleged anticompetitive conduct in delaying generic entry and forcing customers to purchase the EpiPen exclusively in a 2-Pak. See Doc. 1500-3 at 21-35 (Rosenthal Expert Report ¶¶ 47-82) (describing her methodology for calculating: (1) damages caused by generic delay for plaintiffs' RICO, antitrust, and consumer protection claims; (2) damages caused by the forced purchase of 2-Paks for plaintiffs' RICO and consumer protection claims; (3) damages based on unjust profits realized as a result of defendants' alleged misconduct for plaintiffs' unjust enrichment claims; and (4) an alternative damages theory for defendants' alleged pricing scheme to support plaintiffs' RICO claim); see also Doc. 1497-3 at 21-35 (Rosenthal Expert Report ¶¶ 47-82) (sealed version). And, Prof. Elhauge has proposed a methodology for calculating aggregate damages in the form of alleged EpiPen overcharges arising from defendants' foreclosure of brand competition through use of exclusionary contracts with PBMs. See Doc. 1500-2 at 56-64, 66 (Elhauge Expert Report ¶¶ 110-120, Table 4) (providing a classwide methodology for calculating share and price impacts of PBM coverage restrictions); see also Doc. 1497-2 at 56-64, 66 (Elhauge Expert Report ¶¶ 110-120, Table 4) (sealed version).
Mylan asserts that Profs. Rosenthal and Elhauge's models contradict each other, and they criticize plaintiffs for not providing a method to determine which model applies to any particular plaintiff. Plaintiffs respond that Mylan's argument presents no problem on class certification. Instead, plaintiffs argue, Mylan's argument goes to the issue of damage allocation, and damage allocation will occur at a later stage of the proceeding, based on a jury's findings. Plaintiffs also explain, at that stage, the court can allocate the damage award based on a jury's findings in a way that prevents double recovery. The court agrees that Mylan's argument goes to damage allocation, not to the question whether plaintiffs can prove damages on a classwide basis.
Mylan criticizes Profs. Rosenthal and Elhauge's methodologies for calculating classwide damages. This criticism relies on defendants' expert's report that identifies purported flaws in: (1) Prof. Rosenthal's method for calculating classwide damages resulting from defendants' switch to the 2-Pak and generic delay, see, e.g., Doc. 1636-2 at 47-58, 133-34 (Johnson Expert Report ¶¶ 54-72, 166-67), see also Doc. 1503-4 at 47-58, 133-34 (Johnson Expert Report ¶¶ 54-72, 166-67) (sealed version); and (2) Prof. Elhauge's method for determining classwide damages caused by defendants' alleged use of exclusionary contracts with PBMs, see, e.g., Doc. 1636-2 at 62-71, 74-81 (Johnson Expert Report ¶¶ 78-89, 92-99), see also Doc. 1503-4 at 62-71, 74-81 (Johnson Expert Report ¶¶ 78-89, 92-99) (sealed version).
Also, by separate motions, defendants ask the court to exclude Profs. Rosenthal and Elhauge's opinions under Daubert. By separate Order, the court denies those motions.
But the court can't decide this issue—i.e., whether plaintiffs' expert reports actually prove classwide damages—on a certification motion. See Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 468-69 (2013) (at class certification stage, plaintiffs "need not . . . prove that the predominating question will be answered in their favor"; ultimate failure of proof on essential element dooms claim, but does not mean individualized claims predominate); see also In re Ethylene Propylene Diene Monomer (EPDM) Antitrust Litig., 256 F.R.D. 82, 100 (D. Conn. 2009) ("[T]he issue at class certification is not which expert is the most credible, or the most accurate modeler, but rather have the plaintiffs demonstrated that there is a way to prove a class-wide measure of damages through generalized proof."); Natchitoches Parish Hosp. Serv. Dist. v. Tyco Int'l, Ltd., 247 F.R.D. 253, 270 (D. Mass. 2008) ("To determine predominance, a court need not plunge into the weeds of an expert dispute about potential technical flaws in an expert methodology.").
At class certification, plaintiffs need not prove that a proposed method for determining damages is valid; instead, the proper inquiry at this stage asks whether damages are capable of proof on a classwide basis. See Comcast Corp. v. Behrend, 569 U.S. 27, 34 (2013) (holding that plaintiffs failed to establish predominance because their damages model "falls far short of establishing that damages are capable of measurement on a classwide basis"); see also In re Wellbutrin XL Antitrust Litig., 282 F.R.D. 126, 144 (E.D. Pa. 2011) ("At the class certification stage, the plaintiffs are not required to prove damages by calculating specific damages figures for each member of the class, but rather they must show that a reliable method is available to prove damages on a class-wide basis." (first citing In re Neurontin Antitrust Litig., No. 02-1390, 2011 WL 286118, at *9 (D.N.J. Jan. 25, 2011); then citing Bell Atl. Corp. v. AT&T Corp., 339 F.3d 294, 303 (5th Cir. 2003))). And, if each expert's approach is plausible based on classwide proof and does not rely on or implicate individualized questions that will predominate over common ones, "[i]t is sufficient at [the class certification stage] to note that both [experts'] methods . . . are plausible approaches . . . ." In re Lidoderm Antitrust Litig., No. 14-md-02521-WHO, 2017 WL 679367, at *18 (N.D. Cal. Feb. 21, 2017) (recognizing that, "in the end," deciding which expert opinion accurately estimates the but-for prices of generic and branded drugs "is subject to further merits-based determinations and findings by the trier of fact"). Cf. Sheet Metal Workers Local 441 Health & Welfare Plan v. GlaxoSmithKline, PLC, No. 04-5898, 2010 WL 3855552, at *6 (E.D. Pa. Sept. 30, 2010) (explaining that "resolving expert disputes in order to determine whether a class certification requirement has been met is always a task for the court" and "[w]eighing expert opinions is not only permissible; it may be integral to the rigorous analysis Rule 23 demands" if the expert opinion is necessary to determine whether plaintiff satisfies the certification requirement (citations and internal quotations omitted)).
Here, plaintiffs have come forward with a plausible method for determining classwide damages through the opinions of Profs. Rosenthal and Elhauge. Mylan's attacks against these experts' methodologies don't undermine their plausibility for establishing classwide damages; instead, they go to the weight the trier of fact should give them when it decides whether the methodologies, in fact, establish that the classes—as a whole—have sustained damage. See Syngenta, 2016 WL 5371856, at *10 (concluding that defendants' criticisms about plaintiffs' experts' methodologies failed to show that they "are so unreliable as to preclude certification" but instead, "[m]ore importantly for purposes of certification, these criticisms present common questions that do not undermine a finding of predominance"); see also In re Solodyn (Minocycline Hydrochloride) Antitrust Litig., No. 14-md-02503, 2017 WL 4621777, at *10 (D. Mass. Oct. 16, 2017) (concluding, at the class certification stage, "any disagreement [about expert's method of calculating damages] does not overcome the Plaintiff's showing that damages will be substantially shown by common proof"). Plaintiffs here have shouldered their burden to establish that common issues predominate the damages inquiry by presenting a common method for calculating classwide damages on an aggregate basis. This showing satisfies Rule 23's predominance requirement.
The court recognizes that this case involves no competing expert methodologies for calculating damages. Defendants' expert, Dr. Johnson, explicitly states that his opinions, to date, advance no methodology for calculating damages. Doc. 1724 at 91-92 (Tr. of Mot. Hr'g Class Certification - Phase I 91:25-92:5 ("Q: Were you asked to offer any affirmative assessment of a classwide method for assessing impact or damages in this case? A: No. My assignment was not to put forward an affirmative methodology, it was to assess the work of Professor Rosenthal and Professor Elhauge.")). And, even if the certification decision required the court to weigh opposing expert opinions, which it does not, see In re Lidoderm Antitrust Litig., 2017 WL 679367, at *18, that exercise would not apply to damage calculations.
b. Uninjured Class Members
Next, Mylan argues that plaintiffs cannot shoulder their burden to establish predominance because they cannot show that all putative class members sustained injury. Instead, Mylan argues, plaintiffs' proposed class definitions contain too many uninjured class members for the court to certify a class action under Rule 23. According to Mylan, predominance fails for each of plaintiffs' liability theories—i.e., the 2-Pak theory, the Auvi-Q foreclosure theory, and the generic delay theory—across all proposed classes because there are uninjured class members who cannot be identified without mini-trials. Mylan criticizes the alleged absence of or flawed analyses by plaintiffs' expert witnesses which, it contends, accounts for plaintiffs' failure to carry their burden to show all class members were injured under any of plaintiffs' legal theories. For these reasons, Mylan urges the court to deny certification for lack of predominance.
Plaintiffs respond with two arguments. First, plaintiffs argue that the collateral source doctrine bars Mylan's argument that the proposed class definitions contain class members who are uninjured because they used a co-pay, coinsurance, or deductible to pay for their EpiPens, and thus had their EpiPens paid for by insurance companies. Second, plaintiffs argue that they have established, through common evidence, that the proposed classes have sustained injury classwide sufficient to satisfy the Tenth Circuit's requirements for certification. According to plaintiffs, Mylan demands a standard of proof that the law in our Circuit does not require. And, plaintiffs assert, their experts provide ample support to show an injury sustained across the classes. Also, plaintiffs argue, all class members have sustained injury because defendants' conduct deprived them of a choice in the market, whether that loss of choice involved the absence of a less expensive generic, foreclosed access to competing brands, or a preference for a single purchase EpiPen. The court addresses each argument, separately, below.
i. Collateral Source Doctrine
Plaintiffs assert that the collateral source doctrine bars Mylan's arguments that certain class plaintiffs—ones who used insurance to pay for their EpiPens—are not injured. The Tenth Circuit has explained that the collateral source rule is "[d]erived from the common law [and] posits that '[p]ayments made to or benefits conferred on the injured party from other sources are not credited against the tortfeasor's liability, although they cover all or a part of the harm for which the tortfeasor is liable.'" Friedland v. TIC-The Indus. Co., 566 F.3d 1203, 1205-06 (10th Cir. 2009) (quoting Restatement (Second) of Torts § 920A(2) & cmt. d (1979)). "Th[is] rule thus permits an injured plaintiff to recover more than the damages he has suffered as the result of an injury." Id. at 1206. The Circuit has explained the rationale for applying the rule: "'[P]ublic policy favors giving the plaintiff a double recovery rather than allowing a wrongdoer to enjoy reduced liability simply because the plaintiff received compensation from an independent source.'" Prager v. Campbell Cty. Mem'l Hosp., 731 F.3d 1046, 1059 (10th Cir. 2013) (quoting Green v. Denver & Rio Grande W. R.R. Co., 59 F.3d 1029, 1032 (10th Cir. 1995)). "Thus, courts have applied the rule when the injured plaintiff has been compensated by, for example, . . . the plaintiff's [own] insurance." Friedland, 566 F.3d at 1206 (citation omitted).
Defendants argue that the collateral source rule doesn't apply here. Specifically, they argue that the collateral source rule contradicts the way plaintiffs have litigated the case to date—i.e., choosing to include both consumers and third-party payors as class members in the proposed class definitions. Defendants argue there is no need to include payors in the classes if consumers can assert injuries for the amounts paid by those payors. Also, defendants assert that applying the collateral source doctrine would raise conflicts between class members over who can recover for what payments.
Defendants also emphasize that plaintiffs repeatedly have described their EpiPen expenses as payments made "out of pocket, after insurance." See, e.g., Doc. 60 at 11, 23-24 (Class Compl. ¶¶ 11, 13, 62-64). Plaintiffs' expert has calculated damages separately for consumers and third-party payors. See, e.g., Doc. 1500-3 at 29 (Rosenthal Expert Report ¶ 67) (separating damages for third-party payors, insured consumers, and cash payers). And, importantly, plaintiffs expressly have excluded from the class definitions: "Single flat co-pay" consumers who purchased EpiPens or generic EpiPens only via a fixed dollar co-payment that is the same for all covered devices, whether branded or generic (e.g., $20 for all branded and generic devices). Defendants suggest that plaintiffs excluded this group because they recognize that a consumer who paid a single, flat co-pay didn't sustain any damage. For all these reasons, defendants argue the collateral source doctrine shouldn't apply to a case where plaintiffs have chosen to bring a class action on behalf of consumers and third-party payors collectively.
At oral argument on the certification motion, plaintiffs' counsel suggested that this exclusion isn't necessary if the court applies the collateral source doctrine to plaintiffs' claims. Doc. 1789 at 26 (Tr. of Mot. Hr'g Class Certification - Phase II 26:9-20). Regardless whether the collateral source doctrine applies to plaintiffs' claims, the court refuses to remove this exclusion from the proposed class definition. As defendants asserted at oral argument, plaintiffs included this proposed exclusion in their original motion. Defendants have litigated the certification motion on the understanding that plaintiffs' class definition excluded single, flat co-pay consumers. Id. at 94 (Tr. of Mot. Hr'g Class Certification - Phase II 94:4-95:11). And defendants' expert prepared his report based on his assumption that this exclusion was part of the class definitions. Thus, defendants argue, the court shouldn't permit plaintiffs to remove the exclusion at this late stage.
The court agrees with defendants. The court will not expand the class definitions from the ones plaintiffs previously proposed. See In re Wholesale Grocery Prods. Antitrust Litig., No. 09-MD-2090 ADM/TNL, 2016 WL 4697338, at *5-6 (D. Minn. Sept. 7, 2016) (rejecting plaintiffs' revised class definition that included more putative plaintiffs on certification motion because "[p]laintiffs [had] impermissibly expanded the class definition"); see also Costelo v. Chertoff, 258 F.R.D. 600, 604-05 (C.D. Cal. 2009) ("The Court is bound to class definitions provided in the complaint and, absent an amended complaint, will not consider certification beyond it."). And, the court finds no reason to remove the exclusion based on an assertion made by plaintiffs at oral argument.
The court concludes that it need not decide—as a matter of law—whether the collateral source rule applies to plaintiffs' claims here on this certification motion. As Judge Lungstrum found in Syngenta, the court didn't need to decide on certification the "legal issue" whether the collateral source doctrine barred defendants from offsetting benefits realized by certain putative class members. 2016 WL 5371856, at *9. Instead, the question was whether "the necessary individual inquiries [to determine whether a class member sustained injury] would be so overwhelming as to defeat predominance . . . ." Id. Judge Lungstrum concluded that they weren't. Id. Judge Lungstrum also found it was "worth noting that the availability of such offsets presents a common question." Id.
Similarly, here, the court must decide whether individual questions about the putative class members' injuries overwhelm the common issues. As discussed below, even without considering or applying the collateral source doctrine, plaintiffs have presented sufficient classwide evidence in their expert reports that the amount of potentially uninjured class members in the putative classes isn't so large that it precludes class certification under Tenth Circuit case law. Also, as Judge Lungstrum recognized in Syngenta, the question whether the collateral source doctrine applies to plaintiffs who used insurance to pay for their EpiPen purchases is a legal issue that presents a common question that applies classwide.
Defendants disagree. They argue that applying the collateral source rule raises even more individual questions because the collateral source rule is a creature of state law, at least when it is applied to plaintiffs' state law claims. See Blanke v. Alexander, 152 F.3d 1224, 1231 (10th Cir. 1998) (identifying the collateral source rule as a "substantive state rule[ ] of evidence" (citation and internal quotation marks omitted)). And, defendants contend, many states have abolished or limited the collateral source rule in recent years. Thus, they argue, differences in state laws will require individual inquiries to determine whether and how to apply a particular state's collateral source rule to specific class members.
Plaintiffs respond, arguing that recent modifications to state rules governing the collateral source doctrine apply to cases involving medical malpractice, personal injury, or wrongful death. See Bryce Benjet, A Review of State Law Modifying the Collateral Source Rule: Seeking Greater Fairness in Economic Damages Awards, 76 Def. Counsel. J. 210, 211 (Apr. 2009) (explaining that "collateral source damages are often limited in health care liability cases as part of broader tort-reform legislation"); see also generally id. (describing the differences among the states' collateral source rules). But, plaintiffs contend, defendants haven't identified any state law departures from the collateral source rule under the federal common law that would apply to the claims asserted here. The court agrees that—to the extent the collateral source doctrine applies to plaintiffs' claims here—any variations in the state laws governing the collateral source rule don't raise individual issues that will overwhelm the common questions.
Finally, in both their Reply and their Response to Defendants' Surreply, plaintiffs offer to exclude another group of individuals from the class definitions. Doc. 1837 at 23; Doc. 1579-1 at 7. Plaintiffs suggest that the class definitions could exclude individual consumers who did not purchase an EpiPen after the Generic Start Date. Plaintiffs question whether more than a de minimis number of such consumers exist when most EAI consumers make repeat purchases every year so that they have an EAI device on hand to treat anaphylaxis. In any event, plaintiffs concede that removing these individuals from the class definitions would reduce the number of purportedly uninjured class members. And, plaintiffs explain, their expert has provided damages calculations that segregate these consumers from her aggregate damage calculations. Doc. 1711-1 at 23-24 (Rosenthal Reply Report ¶ 36, Tables 6-7). Defendants never argue that the court should reject plaintiffs' proposed exclusion. After considering plaintiffs' proposal, the court agrees that this exclusion further protects against including uninjured class members in the classes. And, as our court has recognized, "the [c]ourt and parties are authorized" to narrow the class definitions "[i]f the class definition should require tailoring as the litigation progresses[.]" Sibley v. Sprint Nextel Corp., 254 F.R.D. 662, 671 (2008); see also McCurley v. Royal Seas Cruises, Inc., 331 F.R.D. 142, 162 (S.D. Cal. 2019) ("A plaintiff may properly narrow the class for which it seeks class certification even in a reply brief."); Abraham v. WPX Energy Prod., LLC, 322 F.R.D. 592, 611 (D.N.M. 2017) (holding that "a plaintiff is not bound to the class definition in the operative complaint" and concluding that "[s]ome flexibility . . . is needed in crafting a class action where one is warranted"). So, the court will add this exclusion to the class definitions.
Plaintiffs propose this exclusion as an "alternative to applying the collateral source rule." Doc. 1579-1 at 7. Although the court doesn't decide whether to apply the collateral source rule now, it concludes that this exclusion is warranted because it removes from the class definitions individual consumers who could not have sustained injury because their only EpiPen purchases occurred before the Generic Start Date.
ii. The Presence of Uninjured Class Members
Next, Mylan asserts that plaintiffs fail to shoulder their burden to establish predominance because they cannot prove that all putative class members sustained injury. Relying on cases from other Circuits, Mylan argues that the law requires plaintiffs to "show that they can prove, through common evidence, that all class members were in fact injured by" defendants' conduct. In re Rail Freight Fuel Surcharge Antitrust Litig., 725 F.3d 244, 252 (D.C. Cir. 2013); see also In re Rail Freight Fuel Surcharge Antitrust Litig., 934 F.3d 619, 623-24 (D.C. Cir. 2019) [hereinafter "Rail Freight II"] (affirming district court's denial of class certification because plaintiffs had failed to establish classwide injury; instead, the evidence showed that 2,037 putative class members sustained no injury, and that finding established a need for individual inquiries that defeated predominance);In re Asacol Antitrust Litig., 907 F.3d 42, 57-58 (1st Cir. 2018) (reversing district court's grant of certification because Rule 23 does not permit "a trial in which thousands of class members [must] testify" to establish that they sustained injury).
After oral argument, defendants submitted a letter to the court citing as supplemental authority the Rail Freight II opinion and another case from the District of Massachusetts. Doc. 1859 (first citing Rail Freight II, 934 F.3d 619; then citing In re Intuniv Antitrust Litig., No. 1:16-cv-12396-ADB, 2019 WL 3947262 (D. Mass. Aug. 21, 2019)). Plaintiffs responded to defendants' submission, arguing that the supplement authority doesn't apply to this case because it is authority from outside the Tenth Circuit and factually different from the class certification record here. Doc. 1881. The court addresses these arguments more fully in the analysis below.
Mylan's position relies heavily on the First Circuit's opinion in Asacol. The Tenth Circuit has not examined Asacol or otherwise indicated if it would follow the case. The analysis that follows is, therefore, an assessment of Asacol, relevant Tenth Circuit law, and a prediction whether the Tenth Circuit would adopt Asacol.
(a) Legal Standard
In Asacol, union-sponsored benefit plans sued drug manufacturer Warner Chilcott Limited on behalf of a class of similarly situated indirect purchasers, including individual consumers. 907 F.3d at 45. The plaintiffs alleged that Warner had suppressed entry of a generic drug by removing it from the market at the same time a new drug was introduced. Id. The district court had estimated that 10% of class members had not been injured by defendants' alleged anticompetitive conduct. Id. at 46-47. The court based its finding on conclusions drawn by both sides' experts who inferred the 10% figure from the number of brand loyalists who had purchased similar pharmaceutical products. Id. at 47. In other words, the 10% figure measured the percentage of brand-loyal customers would not have switched to the generic alternative, even if a lower-priced generic alternative were available, had it entered the market. Id. Nevertheless, the district court found that the number of uninjured class members—i.e., 10%—was de minimis, and thus did not preclude class certification. Id. The First Circuit accepted the district court's estimate and found it necessary to consider how to treat these uninjured class members in the face of the principle requiring plaintiffs to show proof of injury. Id. at 51. The First Circuit framed the question as one asking whether a class can be certified "even though the injury-in-fact will be an individual issue, the resolution of which will vary among class members." Id.
In its order granting class certification, the Asacol district court had approved plaintiffs' proposal for a process to cull out uninjured class members before judgment. Id. at 52. Under the process approved by the district court, class members would receive notice asking them to submit a claim form, along with data and documentation. Id. A claims administrator then would evaluate each claim under a court-approved formula, individuals could contest the calculations, and the court would review the administrator's report and make any necessary changes. Id.
On appeal, the First Circuit concluded that review by a claims administrator of defendants' already-announced intention to challenge any affidavits would deprive defendants of their Seventh Amendment right to a jury trial and their due process rights. Id. at 53 ("Plaintiffs' proposed claims process provides defendants no meaningful opportunity to contest whether an individual would have, in fact, purchased a generic drug had one been available."). Also, the First Circuit found, plaintiffs hadn't established that the number of affidavits to which defendants could raise a legitimate challenge "is so small that it will be administratively feasible to require those challenged affiants to testify at trial." Id. at 53. The First Circuit concluded the need to identify thousands of uninjured individuals would predominate over common questions and "render an adjudication unmanageable absent evidence such as . . . unrebutted affidavits . . . or some other mechanism that can manageably remove uninjured persons from the class in a manner that protects the parties' rights." Id. at 53-54.
The First Circuit also rejected plaintiffs' fallback position. This alternative argued that plaintiffs could prove classwide impact through testimony by their expert, who calculated that "a generic substitute drug would have achieved approximately ninety percent market penetration in a but-for world, from which, in part, the district court estimated that about ten percent of the class was likely brand loyal and thus uninjured." Id. at 54. The First Circuit held that where the aggregate damage amount is the sum of damages sustained by a number of individuals and determining whether any individual was injured "turns on an assessment of the individual facts concerning that person," the defendant must receive the opportunity to "challenge each class member's proof that the defendant is liable to that class member." Id. at 55. And whether that inquiry precludes class certification because it threatens predominance "turns on whether such challenges are reasonably plausible in a given case and whether the plaintiff cannot demonstrate that allowing for such challenges in a manner that protects the defendant's rights will be manageable and superior to the alternatives." Id.
No Tenth Circuit case has addressed the First Circuit's Asacol opinion. And Mylan's brief appears to concede that the Tenth Circuit has not embraced Asacol. But during oral argument at the class certification hearing, Mylan tried to distinguish DG ex rel. Stricklin v. Devaughn, 594 F.3d 1188 (10th Cir. 2010), and two cases from this District that appear to conflict with Asacol. Mylan urged the court to adopt Asacol. In sum, Mylan's view is that, as counsel argued, Asacol is "a better read of the law." Doc. 1789 at 64 (Tr. of Mot. Hr'g Class Certification - Phase II 64:17-20).
The court agrees with Mylan on some fronts. DeVaughn doesn't provide direct Tenth Circuit authority on this issue. In Devaughn, the Tenth Circuit explained that "Rule 23's certification requirements neither require all class members to suffer harm or threat of immediate harm nor Named Plaintiffs to prove class members have suffered such harm." 594 F.3d at 1198. The Circuit recognized: "'[A] class will often include persons who have not been injured by the defendant's conduct. . . . Such a possibility or indeed inevitability does not preclude class certification.'" Id. (quoting Kohen v. Pac. Inv. Mgmt. Co., 571 F.3d 672, 677 (7th Cir. 2009)). But, as Mylan posits, Devaughn is different because it was an injunction class case under Rule 23(b)(2), and thus the Tenth Circuit never has discussed whether Rule 23 permits certification of a class with uninjured class members in the context of a class seeking money damages under Rule 23(b)(3).
But, the court disagrees with Mylan on other fronts. For example, Mylan asserts that the two cases from this District—Syngenta and Urethane—differ from the question presented here. To the contrary, the court finds these two cases—both decided by Judge Lungstrum—highly persuasive on the question of what standard the Tenth Circuit would apply when deciding whether plaintiffs have made a sufficient showing of classwide injury for certification under Rule 23.
First, in Urethane, Judge Lungstrum denied a motion to decertify the class he previously had certified. In re Urethane Antitrust Litig., No. 14-1616-JWL, 2013 WL 2097346, at *1 (D. Kan. May 15, 2013). Defendant argued that events after certification confirmed some class members sustained no damages, thereby making certification inappropriate. Id. at *2. Plaintiffs conceded the premise of the argument—that not every class member had sustained compensable damage. Id. Plaintiffs argued, instead, that all class members had been affected by a conspiracy to elevate prices, even if some may have mitigated their damages or otherwise did not sustain damages that could be quantified. Id. Judge Lungstrum found that these facts didn't prevent class certification. Id. He noted that defendant "has not cited any authority supporting the argument that the presence of a few 'zero-damages' class members necessarily defeats certification." Id. And, citing Seventh Circuit case law, Judge Lungstrum recognized that "a class will almost inevitably include persons who have not been injured by the defendant's conduct, and that fact (or even inevitability) does not preclude certification." Id. (citing Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 823 (7th Cir. 2012)). Judge Lungstrum wrote that the Seventh Circuit's opinion in Messner also "noted that a class is too broad to permit certification only if it includes a great number of members who could not have been harmed by the defendant's conduct (as opposed to a great number who ultimately are shown to have suffered no harm)." In re Urethane Antitrust Litig., 2013 WL 2097346, at *2 (citing Messner, 669 F.3d at 824).
Judge Lungstrum's opinion in Urethane doesn't quantify what number would amount to a "great number" of members who sustained no harm. But, Judge Lungstrum denied the motion to decertify, finding that "plaintiffs have shown persuasively that only a very small percentage of class members suffered no damages." Id. And, Judge Lungstrum signaled his agreement with the Seventh Circuit's analysis by concluding, "the presence of those few members does not compel decertification." Id.
The Tenth Circuit affirmed Judge Lungstrum's certification order. In re Urethane Antitrust Litig., 768 F.3d 1245 (10th Cir. 2014), cert. denied 137 S. Ct. 291 (2016). Although the Tenth Circuit never addressed explicitly whether Judge Lungstrum correctly applied the Seventh Circuit's test for determining when certification of a class containing uninjured class members is appropriate, the Circuit affirmed Judge Lungstrum's certification decision, noting that the "presence of individualized damages issues" did not preclude certification because "[c]lass-wide proof is not required for all issues" as long as plaintiffs made "a showing that the questions common to the class predominate over individualized questions." Id. at 1255.
In the second case in this District—Syngenta—Judge Lungstrum again relied on the Seventh Circuit's standard for determining whether the presence of uninjured putative class members defeats certification. Syngenta, 2016 WL 5371856, at *4. Judge Lungstrum held that "the possibility or even inevitability that the class will include members not injured by the defendant's conduct does not preclude class certification." Id. (citing Kohen v.. Pac. Inv. Mgmt. Co., 571 F.3d 672, 677 (7th Cir. 2009)). Instead, again applying Messner, Judge Lungstrum recited the Seventh Circuit's test from that case: "[I]f 'a class is defined so broadly as to include a great number of members who for some reason could not have been harmed by the defendant's allegedly unlawful conduct, the class is defined too broadly to permit certification.'" Id. (quoting Messner, 669 F.3d at 824). Judge Lungstrum then considered defendants' arguments that some plaintiff corn producers may not have sustained injury based on a drop in central market corn prices because that price decrease didn't affect local pricing, or because some plaintiffs sold specialty corn, or because some plaintiffs benefitted by selling other crops or feeding livestock. Id. Judge Lungstrum concluded that the arguments were "merits-based defenses" and provided "no basis to conclude that a 'great number' of members could not have been harmed as alleged by plaintiffs." Id. (citing Messner, 669 F.3d at 824); see also Messner, 669 F.3d at 824 (noting the "critical" distinction between a "fatally overbroad" class and one that "consists largely (or entirely, for that matter) of members who are ultimately shown to have suffered no harm" because "that may not mean that the class was improperly certified but only that the class failed to meet its burden of proof on the merits")).
In Syngenta's predominance analysis, Judge Lungstrum noted several commonalities among the putative class members, including that plaintiffs' expert opinions would show both the fact of damage and the amount using classwide proof. Syngenta, 2016 WL 5371856, at *5. Judge Lungstrum rejected defendants' argument that the fact and amount of damages presented individual issues "that are so important [they] defeat predominance." Id. at *6. Defendants based their argument on the parties' competing expert opinions, which they urged Judge Lungstrum to weigh and to declare defendants the winner. Id. But Judge Lungstrum declined to perform that weighing function at certification, concluding that "the persuasiveness of such evidence [was a question] for the jury at trial." Id. at *7. Plaintiffs' burden at certification is to show predominance; they need not convince the court that their evidence will prevail at trial. Id. And, Judge Lungstrum concluded, "a reasonable juror could believe that" all class members sustained injury from a decrease in "local corn prices [that] reflect[ed] changes in the [Chicago Board of Trade] futures price, as opined by plaintiffs' experts." Id. Because the Supreme Court has allowed the use of such representative evidence to show classwide liability in a class action, Judge Lungstrum concluded that certification was warranted. Id. (citing Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036 (2016)). As Judge Lungstrum noted, "[p]laintiffs are required to offer evidence to show predominance," and, he concluded, "they have done so here." Id.
Defendants filed an interlocutory Rule 23(f) Petition for Permission to Appeal Class Certification Order. The Tenth Circuit denied the request, noting that the decision to grant a Rule 23(f) petition is purely discretionary and finding that defendants had not established a reason warranting appellate review such as manifest error or an unresolved legal issue. Order, In re Syngenta AG MIR 162 Corn Litig., No. 16-607 (10th Cir. Dec. 7, 2016), ECF No. 01019731851. The Circuit also described Judge Lungstrum's certification decision as "well-researched and reasoned, and if any rulings are error, those errors can be addressed on appeal, if necessary." Id. at 3.
As Judge Lungstrum's two opinions show, our court has followed Seventh Circuit precedent when analyzing whether the presence of uninjured class members defeats certification. That is, as the Seventh Circuit has held, "a class should not be certified if it is apparent that it contains a great many persons who have suffered no injury at the hands of the defendant . . . ." Kohen, 571 F.3d at 677-78 (citations omitted). Kohen determined that this issue is best averted by focusing on the class definition. Id. at 677. "[I]f the definition is so broad that it sweeps within it persons who could not have been injured by the defendant's conduct, it is too broad." Id. The Seventh Circuit repeated this observation in Messner and distinguished it from a proposed class consisting "largely (or entirely, for that matter) of members who are ultimately shown to have suffered no harm, [as] that may not mean that the class was improperly certified but only that the class failed to meet its burden of proof on the merits." Messner, 669 F.3d at 824 (citing Schleicher v. Wendt, 618 F.3d 679, 686 (7th Cir. 2010)); see also Schleicher, 618 F.3d at 686 ("Rule 23 allows certification of classes that are fated to lose as well as classes that are sure to win."). But neither Messner nor Kohen provides a formula for determining when a class includes too many uninjured members. See Messner, 669 F.3d at 825 ("There is no precise measure for 'a great many.'"). How many is "too many" is "a matter of degree, and will turn on the facts as they appear from case to case." Id. at 825.
On the other hand, the definition cannot narrow the class so much that it becomes an inappropriate fail-safe class, i.e., one where the court has to decide the merits of each prospective individual class members' claims to determine class membership, thereby defining the class in terms of the legal injury. See Messner, 669 F.3d at 825.
In Asacol, the district court found 10% a de minimis number of uninjured class members. 907 F.3d at 47. When reversing the district court's certification decision, the First Circuit did not specify the percentage of uninjured plaintiffs that are "too many" to permit certification. Though later cases cite Asacol to find that 10% is not de minimis, this finding—if indeed Asacol made it—is an implicit conclusion.
Given the certification motion plaintiffs make here, the court finds persuasive the approach used by the Seventh Circuit and Judge Lungstrum. And, the court predicts the Tenth Circuit would follow those cases if presented with the issue. The court also concludes that the Tenth Circuit wouldn't follow Asacol.
Our court also cited Urethane approvingly in another case. See Sibley v. Sprint Nextel Corp., 315 F.R.D. 642, 664 (D. Kan. 2016) ("[G]iven that so many cases and treatises agree that the need for individualized damage determinations does not defeat commonality or predominance, it follows that the fact that some class members suffered no damages also does not defeat commonality or predominance." (citing In re Urethane Antitrust Litig., 2013 WL 2097346, at *2 (other citations omitted))).
Two district courts have issued recent orders that sharply criticize Asacol and its effect on consumers' ability to bring class actions. See In re Loestrin 24 FE Antitrust Litig., 410 F. Supp. 3d 352, 404 (D.R.I. 2019) (denying motion to certify a class of end-payor plaintiffs ("EPP") because Asacol left the court "no room to rewrite what the First Circuit has clearly scripted" but still the court was "troubled that over ninety percent of consumers in the proposed EPP class may have been injured by Defendants' alleged unlawful conduct, but now have no practical recourse under antitrust law"); see also In re Intuniv Antitrust Litig., No. 1:16-cv-12396-ADB, 2019 WL 3947262, at *7 n.8, *8 (D. Mass. Aug. 21, 2019) (denying motion to certify class because "In re Asacol dictates the outcome here" but also recognizing "In re Asacol is likely a death knell for pharmaceutical, antitrust class actions brought by indirect purchasers [because] [g]iven the myriad ways in which consumers could theoretically be uninjured, once a defendant asserts an intent to challenge each claim to have been affected by their conduct, it becomes nearly impossible for indirect purchasers to show that common issues will predominate under In re Asacol," so "[a]bsent class certification, it is likely that most putative class members' claims will go unremedied," and "assuming that Defendants engaged in anticompetitive conduct, although In re Asacol eliminates the possibility that some consumers might obtain a recovery for damages they did not suffer, it also ensures that a much larger number of Intuniv consumers will receive no remedy for harm [they] actually suffered").
Thus, the court applies the standard articled by the Seventh Circuit and applied by Judge Lungstrum, below. And so, the court must answer the following question: Is class certification precluded here because the putative class definitions "'contain[ ] a great many persons who have suffered no injury at the hands of the defendant[?]'" Messner, 669 F.3d at 825 (quoting Kohen, 571 F.3d at 677).
(b) Prof. Rosenthal's Probability Analysis
Plaintiffs rely on a probability analysis conducted by Prof. Rosenthal to assert that the proposed classes do not contain great numbers of class members who didn't sustain harm. To the contrary, plaintiffs contend that the number of uninjured class plaintiffs in each class, if any, is de minimis. First, Prof. Rosenthal estimates that the likelihood that a third-party payor who covered as few as five prescriptions sustained injury from generic delay is 99.999% or better. Doc. 1711-1 at 30-32 (Rosenthal Reply Report ¶¶ 53, 54, Table 8). Second, Prof. Rosenthal opines the likelihood that a third-party payor who reimbursed at least 10 patients for EpiPens was injured by Mylan's switch to the 2-Pak is 99%. Id. at 32-33 (Rosenthal Reply Report ¶¶ 55-56). Third, Prof. Rosenthal estimates that the likelihood that individual consumers sustained injury from generic delay or Mylan's switch to the 2-Pak is about 95%. Id. at 34-37 (Rosenthal Reply Report ¶¶ 58-60). Finally, Prof. Rosenthal opines that all third-party payors and all cash payor consumers sustained injury from brand foreclosure. Id. at 33-34 (Rosenthal Reply Report ¶ 57).
Defendants assert that plaintiffs don't define the classes to include third-party payors who have covered five or more prescriptions. So, defendants contend, Prof. Rosenthal's conclusion shouldn't apply here. But, as plaintiffs argue, Prof. Rosenthal's probability analysis estimates that the likelihood that a third-party payor who covered just two prescriptions sustained injury is between 98.9% and 99.3%. Plaintiffs assert that these figures for just two purchases also demonstrate a very low probability that the putative classes include uninjured class members. Also, as plaintiffs have demonstrated, most third-party payor plans involve thousands of consumers. And, given the prevalence of anaphylaxis, it seems unlikely that a third-party payor would have covered just two, three, four, or even five EpiPen prescriptions during the class period.
Prof. Rosenthal recognizes that "[c]onsumers with copayments that would not have varied for Auvi-Q vs. EpiPen would not have paid this overcharge directly." Doc. 1711-1 at 33-34 (Rosenthal Reply Report ¶ 57). She estimates that such consumers represent between 41% and 71% of consumers, depending on a consumer's plan structure and EpiPen tier placement. Id. Plaintiffs assert that these consumers still sustained injury because the collateral source doctrine bars defendants from using insurance payments as an offset. As discussed above, the court need not decide whether the collateral source doctrine applies on class certification. But still, it presents another issue common to the classes. And, if the collateral source doctrine doesn't apply and excludes evidence of such plaintiffs, the presence of these consumers doesn't present individual questions that overwhelm the common questions because Prof. Rosenthal has provided a method for identifying these consumers. See supra Part IV.C.1 & n.15. Using that method, plaintiffs can remove those uninjured consumers from the classes, if a merits determination finds that these consumers sustained no injury.
Defendants criticize Prof. Rosenthal's probability analysis and urge the court to disregard it as unreliable. By separate Order, the court denies defendants' motion to exclude Prof. Rosenthal's probability analysis under Daubert. Also, the court finds that defendants' criticisms of the probability analysis don't undermine its usefulness to the court's certification decision. The court already has addressed many of these criticisms in the Daubert Order and in this Order, see supra Part IV.D.3.b.
See pages 37-45 of the court's Memorandum and Order ruling the Daubert motions.
For example, defendants assert that Prof. Rosenthal's analysis fails to account properly for "brand loyalists," i.e., consumers who would continue purchasing the EpiPen even when a generic is available. Prof. Rosenthal's analysis uses a 95% estimate to calculate the rate consumers would have switched from the EpiPen to a generic EAI device. But, defendants' expert provides his own opinion about the proper switching rate to show that Prof. Rosenthal's estimate is wrong. Doc. 1503-4 at 145-46 (Johnson Expert Report ¶ 180) (sealed version). Prof. Rosenthal responds that defendants' expert has used the wrong metric: He estimates brand loyalty by using the switching rates from the EpiPen to Mylan's own authorized generic. As Prof. Rosenthal explains, those switching rates aren't an accurate indicator of brand loyalty because they aren't drawn from a competitive market. Instead, Prof. Rosenthal has explained how her analysis more accurately reflects a 95% switching rate in a but-for world where a generic is available in a competitive market.
Also, defendants argue that some consumers aren't injured because they paid less for their EpiPens than the but-for price calculated by Prof. Rosenthal. But, as plaintiffs argue, comparing the actual price a consumer paid with a but-for price used to calculate aggregate damages is like comparing apples to oranges. As discussed above, supra Part IV.D.3.b., Prof. Rosenthal's average but-for price is an average used to calculate aggregate damages. And, that average isn't relevant to determining an individual class member's damages. Instead, individual damages are assessed by determining the but-for generic price that an individual class member would have paid at the point in time when the class member made a branded purchase. Also, to the extent defendants' arguments about but-for prices are merit-based arguments, the court can't decide them now.
In sum, none of defendants' criticisms convince the court to disregard Prof. Rosenthal's probability analysis on class certification. Instead, the court agrees with plaintiffs: Prof. Rosenthal's probability analysis provides a plausible method for determining—across the classes—the number of class members who may not have sustained injury from defendants' purported conduct. Indeed, the First Circuit recently considered a similar analysis conducted by Prof. Rosenthal, albeit in the summary judgment context. See In re Celexa & Lexapro Mktg. & Sales Practices Litig., 915 F.3d 1, 12-13 (1st Cir. 2019). Decided after Asacol, Celexa & Lexapro reversed the district court's summary judgment ruling against a third-party payor plaintiff's RICO claim, finding that plaintiff's causation evidence "does not seem clearly insufficient" to survive summary judgment. Id. at 12. In particular, the court recognized that Prof. Rosenthal "estimated that if [the third-party payor plaintiff] paid for as few as five independent prescriptions, there would be a 98% chance that at least one was the result of off-label marketing." Id. at 13. Because the record showed that plaintiff "likely paid for the Celexa or Lexapro prescriptions of more than five different patients[,]" the court found that "the odds that [plaintiff] was not harmed if the drugs were, indeed, ineffective was likely infinitesimal (assuming the prescriptions were independent of one another)." Id. (emphasis added).
Here, plaintiffs argue that Prof. Rosenthal performed a similar analysis and with more compelling results. And, they urge the court to accept her findings as classwide evidence of an "infinitesimal" number of uninjured class members that doesn't preclude certification. Doc. 1837 at 19 (citing Celexa & Lexapro, 915 F.3d at 13). Although Celexa & Lexapro affirmed the district court's decision to deny class certification on statute of limitations grounds, the First Circuit, in dicta, disagreed with the district court's conclusion that causation and injury issues defeated predominance. Celexa & Lexapro, 915 F.3d at 14. Instead, the First Circuit found, the third-party payor's "clinical and statistical evidence, if believed, could establish causation and injury at least for any TPP who paid for more than a handful of different patients' prescriptions." Id. Indeed, the First Circuit criticized the district court for "gaug[ing] the substantiality or merit of plaintiffs' proof [of causation] in the context of a Rule 23 motion." Id. at 12. The First Circuit recognized that "[c]ertainly there is room for reasonable disagreement on the merits of Dr. Rosenthal['s] assumption[,]" but, the First Circuit emphasized that "[t]he central issue [on certification] is not whether the method of proof would or could prevail" but "whether the method of proof would apply in common to all class members." Id. (citing Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036, 1047 (2016)).
Here, the court also recognizes the possibility for reasonable disagreement with Prof. Rosenthal's probability opinion. But, the court cannot decide the merits of those disagreements at the class certification stage. Instead, the court considers whether Prof Rosenthal has provided a plausible method for proving injury among the classes sufficient for the court to consider her opinion on certification. She has. So, the court accepts Prof. Rosenthal's analysis on this motion.
(c) Number of Uninjured Class Members
Now, the court must decide, based on the data produced by Prof. Rosenthal, if her numbers estimating the likelihood of uninjured class members preclude class certification because they show that the putative class definitions "contain[ ] a great many persons who have suffered no injury at the hands of the defendant." Messner, 669 F.3d at 825. As recited above, plaintiffs assert that Prof. Rosenthal's calculations establish that the number of uninjured class members in each putative class is de minimis. Thus, plaintiffs argue, Prof. Rosenthal's analysis supports class certification.
Briefly, in their Reply, plaintiffs assert that "[a]ll individual class members were injured by the loss of choice in the market . . . ." Doc. 1837 at 23. At oral argument, plaintiffs again asserted this argument and cited a Sixth Circuit case to support it. See Energy Conversion Devices Liquidation Tr. v. Trina Solar Ltd., 833 F.3d 680, 690 (6th Cir. 2016) (noting "caselaw saying that a reduction in consumer choice can show anti-competitive harm" (first citing Blue Shield of Va. v. McCready, 457 U.S. 465, 481-84 (1982); then citing Conwood Co. v. U.S. Tobacco Co., 290 F.3d 768, 789-90 (6th Cir. 2002)). The court recognizes some courts have applied such a rule in antitrust cases. See Ross v. Bank of Am., N.A., 524 F.3d 217, 226 (2d Cir. 2008) (holding that plaintiffs sufficiently alleged antitrust injuries in the form of reductions in "consumer choice" and the "quality of [ ] services offered"); Laumann v. Nat'l Hockey League, 105 F. Supp. 3d 384, 397 (S.D.N.Y. 2015) ("[A]nticompetitive conduct is injurious if it limits consumer options."). But the court doesn't address plaintiffs' argument here for two reasons. One, plaintiffs didn't develop this argument until the hearing on the certification motion. And two, the court finds that plaintiffs have shown plausibly that they can prove classwide injury using classwide evidence and the presence of uninjured class members doesn't defeat predominance.
Mylan disagrees. It asserts that, even accepting Prof. Rosenthal's calculations, the number of uninjured class members in this putative class action simply is too great to permit certification. For support, Mylan relies on cases from outside our Circuit denying class certification because the proposed classes contained too many uninjured class members. See, e.g., Rail Freight II, 934 F.3d at 623, 625 (affirming the district court's denial of class certification because plaintiffs' damages model showed "that 2,037 members of the proposed class—or 12.7 percent—suffered 'only negative overcharges' and thus no injury" and, as a consequence, the "need [for] individualized adjudications of causation and injury" to determine which class members sustained no injury defeated predominance); Asacol, 907 F.3d at 53-54 (reversing district court's certification order where about 10% of the class was uninjured because, the court found, "there are apparently thousands who in fact suffered no injury" and "[t]he need to identify those individuals will predominate and render an adjudication unmanageable absent . . . [a] mechanism that can manageably remove uninjured persons from the class in a manner that protects the parties' rights"); In re Intuniv Antitrust Litig., No. 1:16-cv-12396-ADB, 2019 WL 3947262, at *8 (D. Mass. Aug. 21, 2019) (concluding that where one putative class contained some 25,000 uninjured class members and another class exceeded 10,000 uninjured members and plaintiffs had "failed to put forth a reasonable and workable plan to weed out the . . . uninjured class members in each putative class[,]" plaintiffs couldn't establish that common issues "'[will] predominate over any questions affecting only individual members'" (quoting Fed. R. Civ. P. 23(b)(3))).
At oral argument, Mylan's counsel provided an estimate of the number of uninjured class members using Prof. Rosenthal's analysis: "If you have a class of 2 million plaintiffs, . . . if you even have just 5 percent of the class that is uninjured, you're talking about 100,000 people who will be getting damages under the plaintiffs' models that we know are not [ ]injured and that the plaintiffs have no way to identify without individualized inquiries and many trials." Doc. 1789 at 73 (Tr. of Mot. Hr'g Class Certification - Phase II 73:13-20).
Plaintiffs respond that the court need not follow these cases because they aren't binding precedent in this Circuit. And, plaintiffs assert, they contradict the case law from our court applying the Seventh Circuit's standard. For example, in Urethane, Judge Lungstrum refused to decertify a class where "plaintiffs [had] shown persuasively that only a very small percentage of class members suffered no damages . . . ." In re Urethane Antitrust Litig., 2013 WL 2097346, at *2. Plaintiffs argue, like the percentage at issue in Urethane, the percentage here is very small—less than 5% for individual consumers and less than .001% for third-party payors.
As discussed above, the Seventh Circuit hasn't provided a formula for determining how many uninjured class members is a "great many" that would preclude certification. The Seventh Circuit says "too many" is "a matter of degree, and will turn on the facts as they appear from case to case." Messner, 669 F.3d at 825. In Messner, the defendant had conceded that "only about 2.4 percent of the putative class members" did not sustain injury. Id. at 826. The Seventh Circuit recognized the percentage "may prove, depending on the ultimate size of the class at issue here, to be a significant number of additional plaintiffs," but, it concluded, "a 2.4 percent decrease in the size of the class is certainly not significant enough to justify denial of certification." Id. Also, the Seventh Circuit noted, "the district court is free to revisit this issue at a later time if discovery shows that the number of members who could not have been harmed . . . was more significant than it appears at this time." Id. (citing Kohen, 571 F.3d at 679).
Other courts have reached similar conclusions based on the percentage of uninjured class members. See, e.g., Brown v. Hain Celestial Grp. Inc., No. C 11-03082 LB, 2014 WL 6483216, at *8 (N.D. Cal. Nov. 18, 2014) (granting certification under Rule 23(b)(3) and recognizing that "a small percentage of uninjured persons does not defeat" certification); Mayo v. UBS Real Estate Sec., Inc., No. 08-00568-CV-W-DGK, 2012 WL 4361571, at *3 (W.D. Mo. Sept. 21, 2012) (holding that a putative class definition was not overbroad when "at least 94%" of its members sustained harm from defendants' purported conduct but denying class certification for other reasons). Cf. Koss v. Norwood, 305 F. Supp. 3d 897, 916 (N.D. Ill. 2018) ("While it is possible that some class members will turn out not to have good claims, that possibility does not defeat certification, and nothing in the definition or evidence tells the court how many, and so it is not 'apparent that it contains a great many persons who have suffered no injury at the hands of the defendant.'" (quoting Messner, 669 F.3d at 825)).
Similarly here, the court finds that plaintiffs have shown, at least at the class certification stage, that the percentages of uninjured class members are small enough that they don't preclude class certification. Still, the court recognizes that "depending on the ultimate size of the class at issue here," the factual record eventually may develop more fully and show that the number of uninjured class members is "more significant" than it now appears. Messner, 669 F.3d at 826. As the Seventh Circuit recognizes, the court "is free to revisit this issue" if that, in fact, occurs. Id. For now, the court finds that the current record doesn't defeat the predominance requirement simply because the putative class definitions contain a percentage of uninjured class members.
Also, as plaintiffs argue, even if the existence of uninjured class members presents an individual issue, it doesn't predominate over the common questions that plaintiffs say they will prove using class wide evidence. As the Tenth Circuit has recognized, "[c]lass-wide proof is not required for all issues" on certification. In re Urethane Antitrust Litig., 768 F.3d 1245, 1255 (10th Cir. 2014). "Instead, Rule 23(b)(3) simply requires a showing that the questions common to the class predominate over individualized questions." Id. (citing Amgen v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 468 (2013)). Just this past year, our Circuit held, "so long as at least one common issue predominates, a plaintiff can satisfy Rule 23(b)(3)—even if there remain individual issues, such as damages, that must be tried separately." Naylor Farms, Inc. v. Chaparral Energy, LLC, 923 F.3d 779, 789 (10th Cir. 2019) (citing Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036, 1045 (2016)); see also In re Urethane Antitrust Litig., 768 F.3d at 1254 (affirming certification and noting that "some of the plaintiffs may have successfully avoided damages" but defendant "has not shown that the district court abused its discretion in finding that class-wide issues predominated over individualized issues" such as the existence and impact of an alleged price-fixing conspiracy that "raised common questions that were capable of class-wide proof").
Here, plaintiffs have presented common questions that are capable of resolution through classwide proof through use of expert opinions and aggregate data. Specifically, those questions are whether defendants committed the alleged unlawful conduct and whether that alleged unlawful conduct affected consumers adversely. The court agrees that these common issues predominate. And, any questions about the absence of injury for some class members don't overwhelm the common issues and defeat predominance on class certification.
(d) Constitutional Considerations
Mylan also argues that the court shouldn't certify classes with uninjured class members because it would violate Article III standing and defendants' constitutional right of due process. Tenth Circuit precedent rejects Mylan's standing argument. The Tenth Circuit has held, in the Rule 23(b)(2) context, that "only named plaintiffs in a class action seeking prospective injunctive relief must demonstrate standing by establishing they are suffering a continuing injury or are under an imminent threat of being injured in the future." Colo. Cross Disability Coal. v. Abercrombie & Fitch Co., 765 F.3d 1205, 1214 (10th Cir. 2014) (citation and internal quotation marks omitted). Although a Rule 23(b)(2) case, Colorado Cross Disability Coalition approvingly cited another case "support[ing] the notion that class 'standing' does not require individualized proof from class members" in the Rule 23(b)(3) context. Id. (citing Denney v. Deutsche Bank AG, 443 F.3d 253, 263 (2d Cir. 2006) (other citation omitted)). Denney held that Article III standing does "not require that each member of a class submit evidence of personal standing." 443 F.3d at 263.
Also, Colorado Cross Disability Coalition cited the concurring opinion in Lewis v. Casey, 518 U.S. 343 (1996), where "[t]hree Justices of the Supreme Court favorably quoted this principle from a leading class action treatise." Id. They said:
"'[Unnamed plaintiffs] need not make any individual showing of standing [in order to obtain relief]. . . . Whether or not the named plaintiff who meets individual standing requirements may assert the rights of absent class members is neither a standing issue nor an Article III case or controversy issue but depends rather on meeting the prerequisites of Rule 23 governing class actions.'"Id. (quoting Lewis, 518 U.S. at 395-96 (Souter, J., concurring) (quoting 1 Herbert B. Newberg & Alba Conte, Newberg on Class Actions § 2.07 (3d ed. 1992))). The Tenth Circuit noted that the majority "seemed to agree" with the concurring opinion's conclusion because it "point[ed] out that its holding did 'not rest upon the application of standing rules.'" Id. (quoting Lewis, 518 U.S. at 360 n.7). Thus, the Tenth Circuit framed the certification question as one "not answered by demanding proof of standing from each class member but by application of Rule 23." Id.
Applying the Tenth Circuit's holding in Colorado Cross Disability Coalition and the authorities that it cites favorably, the court finds that the Article III standing requirement does not demand that plaintiffs establish each putative class member has sustained injury. Instead, here, plaintiffs have satisfied Article III standing by showing that each named plaintiff has sustained injury sufficient to confer standing. See supra Part IV.D.3.b. And, as the Tenth Circuit directs, the court has answered the question whether plaintiffs can represent a class that includes uninjured class members by applying Rule 23. See supra Part IV.E.2.b.ii.(c).
The court also rejects Mylan's due process argument. To support this argument, Mylan again invokes Asacol. As discussed above, Asacol held that plaintiffs' proposed claims process—which involved a claims administrator reviewing contested claim forms—failed to protect defendants' Seventh Amendment and due process rights because the process "provides defendants no meaningful opportunity to contest whether an individual would have, in fact, purchased a generic drug had one been available." Asacol, 907 F.3d at 53. The First Circuit criticized plaintiffs for failing to establish that the number of affidavits to which defendants could raise a legitimate challenge "is so small that it will be administratively feasible to require those challenged affiants to testify at trial." Id. And, because plaintiffs had failed to make the "administratively feasible" showing, the First Circuit held that plaintiffs failed the predominance requirement. Id. at 53-54. But Asacol isn't the governing law in our Circuit. And for reasons already explained, the court declines to apply Asacol.
Instead, as discussed above, our court has refused "to apply a standard—one not adopted by the Tenth Circuit—that would preclude certification without a showing that class members may be determined in an administratively feasible manner." Syngenta, 2016 WL 5371856, at *3. In reaching this conclusion, Judge Lungstrum was persuaded by "the thorough and well-reasoned analysis of the Seventh Circuit" in Mullins v. Direct Digital, LLC, 795 F.3d 654 (7th Cir. 2015). Syngenta, 2016 WL 5371856, at *3. Mullins explained that, when a class action involves an aggregate damage calculation, "the identity of particular class members does not implicate the defendant's due process interest at all" because "[t]he addition or subtraction of individual class members affects neither the defendant's liability nor the total amount of damages it owes to the class." Mullins, 795 F.3d at 670 (citations omitted). In reaching this conclusion, Mullins cited the Tenth Circuit's opinion in Urethane as support because, there, the Tenth Circuit had rejected a "Seventh Amendment challenge to allocation of damages award among class members because defendant 'has no interest in the method of distributing the aggregate damages award among the class members[.]'" Id. (quoting In re Urethane Antitrust Litig., 768 F.3d 1245, 1269 (10th Cir. 2014)).
For the same reasons, the court finds that Mylan's due process rights aren't implicated here. Plaintiffs have made a plausible showing that they can prove damages on a classwide basis using an aggregate calculation based on their experts' opinions. And, using the Seventh Circuit's reasoning in Mullins, the court concludes that Mylan's constitutional arguments don't preclude certification here.
c. RICO
Plaintiffs premise their RICO claim on defendants' alleged EpiPen pricing scheme. Plaintiffs describe this scheme as a national, unified scheme to defraud. And, plaintiffs contend, the scheme did not vary among individual plaintiffs because defendants generally deployed it across the entire U.S. market. Thus, plaintiffs contend, the issues presented by the RICO claim are inherently classwide and subject to classwide proof because all of the claim's elements turn on defendants' conduct. Specifically, plaintiffs identify defendants' conduct susceptible to classwide proof as "the RICO enterprise involving the Mylan Defendants, the Pfizer Defendants, and the co-conspirator PBMs; the pattern of racketeering committed by the Mylan and Pfizer Defendants so that they could raise the price of the EpiPen without any fear of competition, consumer backlash, or regulatory oversight from Congress; the elimination of the EpiPen single pack; and causation and damages stemming from this misconduct." Doc. 1500 at 59-60.
When considering whether the Rule 23 predominance requirement is satisfied in a RICO action, the court must "survey the elements of the class's RICO claims to consider (1) which of those elements are susceptible to generalized proof, and (2) whether those that are so susceptible predominate over those that are not." CGC Holding Co., LLC v. Broad & Cassel, 773 F.3d 1076, 1087 (10th Cir. 2014) (citations omitted). So, the analysis "'whether "questions of law or fact common to class members predominate" begins, of course, with the elements of the underlying cause of action.'" Id. at 1088 (quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011) (quoting Fed. R. Civ. P. 23)). Or, more simply, deciding whether a class can rely on classwide proof starts with what the law requires the class to prove.
RICO "establishes a civil cause of action for persons injured as a result of a prohibited racketeering activity." Id. (citing 18 U.S.C. § 1962(c)). A RICO violation has "four elements: '(1) conduct (2) of an enterprise (3) through a pattern (4) of racketeering activity.'" Robbins v. Wilkie, 300 F.3d 1208, 1210 (10th Cir. 2002) (first quoting Sedima, S.P.R.L. v. Imrex Co., Inc., 473 U.S. 479, 496 (1985); then citing BancOklahoma Mortg. Corp. v. Capital Title Co., Inc., 194 F.3d 1089, 1100 (10th Cir.1999)). And, "[t]o prove a RICO violation, a plaintiff must show that the defendant violated the RICO statute, and the plaintiff was injured 'by reason of' that violation." CGC Holding, 773 F.3d at 1088 (first citing 18 U.S.C. § 1962; then quoting § 1964(c)).
Here, plaintiffs' RICO claim alleges an illegal scheme conducted by an "association-in-fact enterprise," consisting of the RICO Defendants (Mylan N.V., Mylan Specialty, LP, Heather Bresch, Pfizer, King Pharmaceuticals, and Meridian) and the PBM Conspirators (defined as at least four PBMs: CVS Health Corporation, Express Scripts Inc., Optum Rx Inc., and Prime Therapeutics LLC ), "whose purpose was to fraudulently mislead and deceive American consumers to purchase the EpiPen at an inflated price, to purchase the EpiPen as a 2-Pak only, and to cause consumers to pay an artificially inflated price for EpiPens" from 2011 to the present. Class Compl. ¶¶ 600, 604.
Mylan launched the EpiPen 2-Pak on August 24, 2011. Plaintiffs use this date as the earliest date when the putative class sustained damages.
Plaintiffs claim they will prove each element of their RICO claim using classwide evidence. Thus, they contend, the common issues of the RICO claim "predominate over any questions affecting only individual members." Fed. R. Civ. P. 23(b)(3). Specifically, plaintiffs assert that they will prove the first and second elements of their RICO claim with common evidence showing that the RICO Defendants formed an enterprise with the PBM Conspirators and, together, they committed various acts as part of an effort to raise the price of EpiPens. Plaintiffs contend that they will prove the third and fourth elements of their RICO claim with common evidence of a pattern of racketeering activity that includes: (1) removing single EpiPens from the market and forcing consumers to buy EpiPens in the 2-Pak; (2) paying kickbacks to PBMs in exchange for exclusive formulary status; (3) adding patents to the EpiPen to prevent generic competition, filing patent infringement lawsuits against generic rivals, and issuing fraudulent press releases suggesting that settlements of patent lawsuits were legitimate; and (4) making false and misleading statements (a) about rival product, Auvi-Q, (b) about the reasons for the switch to the 2-Pak, (c) about coupons and rebates, and (d) during Heather Bresch's testimony under oath to Congress on September 21, 2016.
Plaintiffs also claim they will prove causation and damages on a classwide basis. Specifically, plaintiffs contend that they will offer generalized proof of RICO causation because their RICO claim is based on a single, common scheme, and thus, class members share common relevant circumstantial evidence showing causation. And, plaintiffs argue, they can prove on a classwide basis that defendants' RICO violations caused plaintiffs to sustain common damages in the form of overpayments for EpiPens. In short, plaintiffs predict their RICO evidence will focus on defendants' conduct.
Defendants respond, arguing the court should not certify the proposed RICO class because individual issues of causation and damages will overwhelm the common issues. Thus, defendants argue, plaintiffs' RICO claim cannot satisfy Rule 23(b)(3)'s predominance requirement. The court addresses each of defendants' predominance arguments separately, below.
Defendants don't argue that the other elements of plaintiffs' RICO claim involve individual issues that defeat predominance.
i. RICO Causation
The Tenth Circuit has explained that RICO's "'by reason of'" requirement requires a plaintiff to show "'a RICO predicate offense "not only was a 'but for' cause of his injury, but was the proximate cause as well."'" CGC Holding, 773 F.3d at 1088 (quoting Hemi Grp., LLC v. City of N.Y.C., 559 U.S. 1, 9 (2010) (quoting Holmes v. Sec. Inv'r Prot. Corp., 503 U.S. 258, 268 (1992))). "Sufficiently establishing the element of causation—both actual and proximate—is crucial to proving any violation of RICO." Id. (citing Bridge v. Phoenix Bond & Indem. Co., 553 U.S. 639, 656-60 (2008)).
The Supreme Court has defined RICO proximate causation as "a flexible concept that does not lend itself to 'a black-letter rule that will dictate the result in every case.'" Bridge, 553 U.S. at 654 (quoting Holmes, 503 U.S. at 272 n.20). Proximate cause provides a "'label generically [for] the judicial tools used to limit a person's responsibility for the consequences of that person's own acts . . . .'" Id. (quoting Holmes, 503 U.S. at 268). And, it includes "a particular emphasis on the 'demand for some direct relation between the injury asserted and the injurious conduct alleged.'" Id. (quoting Holmes, 503 U.S. at 268).
Thus, "'[w]hen a court evaluates a RICO claim for proximate causation, the central question it must ask is whether the alleged violation led directly to the plaintiff's injuries.'" SafeStreets All. v. Hickenlooper, 859 F.3d 865, 889 (10th Cir. 2017) (quoting Anza v. Ideal Steel Supply Corp., 547 U.S. 451, 461 (2006)). The Supreme Court has instructed that "'no need [exists] to broaden the universe of actionable harms to permit RICO suits by parties who have been injured only indirectly.'" Id. (quoting Anza, 547 U.S. at 460).
But, in the context of RICO claims premised on predicate acts of mail fraud, the Supreme Court has held that "a plaintiff is not required to plead that he is a victim of the defendant's underlying crime (e.g., that he relied on the fraudulent mailings) to establish a direct injury." Id. (citing Bridge, 553 U.S. at 649-50). "Rather, a plaintiff may establish proximate causation by plausibly pleading that his business or property has been directly injured as a result of the defendants' § 1962 violation." Id. (citing Bridge, 553 U.S. at 649-50).
When a federal court decides whether issues of RICO causation satisfy the Rule 23 predominance inquiry, "the question is whether the link between defendants' actions and the class's injuries can be adduced through common evidence." GCG Holding, 773 F.3d at 1088. Our Circuit has explained that, in RICO actions, "individualized issues of reliance often preclude a finding of predominance." Id. "Since reliance is often a highly idiosyncratic issue that might require unique evidence from individual plaintiffs, it may present an impediment to the economies of time and scale that encourage class actions as an alternative to traditional litigation." Id. But, the Circuit found, this isn't always the case. Id.
In GCG Holding, the Circuit explained that certification of a RICO claim is proper when "'causation can be established through an inference of reliance where the behavior of plaintiffs and the members of the class cannot be explained in any way other than reliance upon the defendant's conduct.'" Id. at 1089-90 (quoting In re Countrywide Fin. Corp. Mortg. Mktg. & Sales Practices Litig., 277 F.R.D. 586, 603 (S.D. Cal. 2011)). In such a situation, "where circumstantial evidence of reliance can be found through generalized, classwide proof, then common questions will predominate and class treatment is valuable in order to take advantage of the efficiencies essential to class actions." Id.; see also Menocal v. GEO Grp., Inc., 882 F.3d 905, 918 (10th Cir. 2018) (holding that the "causation element is susceptible to generalized proof and thus cannot defeat class certification under Rule 23(b)(3)'s predominance requirement"); see also id. at 20 (explaining that when "class allegations are based on a single, common scheme, class members share the relevant circumstantial evidence in common, thus making class-wide proof possible").
Defendants argue that plaintiffs have not shouldered their burden to show how they can prove RICO causation on a classwide basis. More specifically, defendants assert, plaintiffs cannot show a "causal chain between any specific [RICO predicate] act and the amount paid by each plaintiff" for an EpiPen, much less that they can demonstrate the necessary causal chain through classwide proof. Doc. 1636 at 69. Plaintiffs respond, explaining how they will use classwide evidence to prove three components of the EpiPen pricing scheme using: (1) the "hard switch" to the 2-Pak and the elimination of single EpiPens from the market; (2) the delay of generic competition to the market; and (3) the elimination of branded competition by entering into exclusive dealing contracts with PBMs. The court addresses these three components, below.
(a) Switch to 2-Pak
Plaintiffs contend they will prove that defendants engaged in a pattern of racketeering activity that included removing single EpiPens from the market and forcing consumers to buy EpiPens two at a time in the 2-Pak. And, plaintiffs argue, they will establish RICO causation using generalized proof that defendants' actions caused plaintiffs to pay more for EpiPens because defendants' actions forced them to buy two EpiPens.
To support this argument, plaintiffs rely on the Reports of their expert, Prof. Meredith Rosenthal. Prof. Rosenthal opines that Mylan's switch to the 2-Pak "led to an immediate and permanent increase in the number of [EpiPens] per prescription of approximately 0.5." Doc. 1497-3 at 24 (Rosenthal Expert Report ¶ 54). Prof. Rosenthal opines that this increase "represents the amount by which Class members were required to increase their consumption of EpiPens when they were constrained by the 2-Pak." Id.; see also Doc. 1711-1 at 13-14 (Rosenthal Reply Report ¶ 17) (explaining that her study supports a "causal inference" that "[t]he withdrawal of the single injector EpiPen forced the Class to purchase more pens than it would have [purchased] otherwise," and observing that "Mylan's own analysis makes the same causal inference: An EpiPen market review from 2012 cites a 0.23 pen per prescription increase from the end of 2011 compared to the 2012 average, noting 'Much of this growth was driven by elimination of Single Pack.'").
Defendants have challenged the reliability of Prof. Rosenthal's event study in a Daubert motion. The court's Order deciding that motion explains that defendants' challenges to Prof. Rosenthal's opinions go to the weight of her opinions but they don't preclude admissibility. The court thus considers her event study when deciding plaintiffs' Motion for Class Certification. And, the court finds, her opinions support class certification at this stage of the litigation.
At this stage, plaintiffs don't need to establish that Prof. Rosenthal's opinions are indisputably correct. Instead, they merely must demonstrate that her opinions are capable of showing RICO causation on a classwide basis. The court finds they do. The persuasiveness of Prof. Rosenthal's opinions for proving classwide liability is a matter for the factfinder to consider at a later stage in the case. See In re Syngenta AG MIR 162 Litig., No. 14-md-2591-JWL, 2016 WL 5371856, at *6 (D. Kan. Sept. 26, 2016) (explaining that "'persuasiveness [of an expert's opinion showing classwide liability] is, in general, a matter for the jury'" (quoting Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036, 1049 (2016))).
After the switch to the 2-Pak, Prof. Rosenthal calculates that the ratio of EpiPens sold per prescription was more than 2.0. Doc. 1497-3 at 24 (Rosenthal Expert Report ¶ 54). By comparison, in Canada, where consumers still can buy EpiPens in a single pack, the average number of EAIs dispensed per prescription is about 1.5 or 1.48. Doc. 1499-5 at 4 (Rocheleau Dep. 176:3-25).
This evidence, plaintiffs assert, shows defendants' conduct involved "group coercion." See Menocal v. GEO Grp., Inc., 882 F.3d 905, 922 n.13 (10th Cir. 2018) (recognizing that alleged group coercion can support an inference of reliance). Plaintiffs describe defendants' coercion as having forced all consumers—as a group—to purchase EpiPens in sets of two. And, plaintiffs assert, our Circuit has held that allegations of "group coercion rather than individual arm's length transacting . . . not only allows for a class-wide inference of causation . . . but arguably supports an even stronger inference." Id. (affirming certification of a class of immigration detainees who alleged a private detention facility's sanitation policy violated the Trafficking Victim Protection Act's ("TVPA") prohibition against forced labor, and specifically holding that "a reasonable factfinder [could] conclude by a preponderance of the evidence that each TVPA class member would not have performed his or her assigned cleaning duties without being subject to the Sanitation Policy," and thus "a factfinder could—but need to—accept a class-wide inference of causation," id. at 922).
Plaintiffs also assert that Mylan's internal documents contradict its justification for switching to the 2-Pak. Mylan asserts that it made the switch based on medical guidance recommending that patients carry more than one EpiPen with them at all times. But, plaintiffs argue, Mylan's internal documents confirm that profit was the real reason for imposing the switch to the 2-Pak. See, e.g., Doc. 1499-8 at 4 (Mylan slide deck explaining two reasons for eliminating the single EpiPen: (1) "Double the revenue per RX of 2 pack vs single" and (2) "Strong potential generic defense"); Doc. 1499-22 at 10 (Mylan slide deck projecting the "Impact of Eliminating Single Dose" included "Increase revenue by $81 million and gross profit by $41 million annually"). Plaintiffs argue that this evidence reveals Mylan's true motive for switching to the 2-Pak, and this evidence applies across the board, showing defendants' conduct injured the putative class plaintiffs by requiring them to pay more for EpiPens.
In response, defendants argue that RICO causation is not susceptible to classwide proof because not all plaintiffs would have purchased a single EpiPen instead of the 2-Pak—had the single EpiPen remained available for purchase. Indeed, defendants cite deposition testimony by some named plaintiffs who testified that they preferred to purchase a 2-Pak even if they had the choice to buy one EpiPen at a time, in a single dose. Doc. 1835-2 at 2-10. Thus, defendants argue, the alleged RICO violations did not coerce these plaintiffs into paying more for the EpiPen. And, defendants contend, individualized issues like this one—i.e., identifying which class member would have purchased a 2-Pak even if a single dose was available—outweigh common issues embedded in the causation element. The court disagrees.
Defendants also argue that a RICO plaintiff must show that "commission of the predicate acts" caused plaintiff's injuries. Sedima, S.P.R.L. v. Imrex Co., Inc., 473 U.S. 479, 496 (1985). Here, defendants contend plaintiffs can't show that any specific RICO predicate act caused plaintiffs' injuries as opposed to Mylan's decision to stop selling the EpiPen in single doses. Doc. 1636 at 72. This argument misses the mark. The predicate acts plaintiffs rely on to support their RICO claims are the false and misleading statements that defendants made about the medical justification for the switch to the 2-Pak. See, e.g., Doc. 60 at 172 (Class Compl. ¶ 655.b.). Plaintiffs assert that they will show that defendants committed these predicate acts as part of a pattern of racketeering activity that included removing single doses of the EpiPen from the market and forcing consumers to buy the EpiPen in a 2-Pak. And, plaintiffs assert, they will show that defendants' conduct caused classwide injury because it forced consumers to buy more EpiPens. These allegations suffice to establish that plaintiffs will rely on classwide evidence to support their RICO theory that defendants' alleged commission of predicate acts caused plaintiffs' injuries.
As discussed above—when addressing the issue of uninjured class members—this court has recognized that "a class will almost inevitably include persons who have not been injured by the defendant's conduct, and that fact (or even inevitability) does not preclude certification." In re Urethane Antitrust Litig., No. 04-1616-JWL, 2013 WL 2097346, at *2 (D. Kan. May 15, 2013) (citing Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 823 (7th Cir. 2012)). Instead, "a class is too broad to permit certification only if it includes a great number of members who could not have been harmed by the defendant's conduct (as opposed to a great number who ultimately are shown to have suffered no harm)." Id. (citing Messner, 669 F.3d at 824).
Here, defendants rely on evidence that, ultimately, may show certain putative class members sustained no harm because they would have purchased the 2-Pak, even if a single dose product was available for purchase. But, at this stage of the litigation, the court is unpersuaded that the putative class includes "great numbers" of class members who could not have been harmed by defendants' switch to the 2-Pak. See supra Part IV.E.2.b. Plaintiffs have presented sufficient evidence that defendants' switch to the 2-Pak caused injury to the putative class on a classwide basis in the form of overpayment for the EpiPen. And thus, plaintiffs have shown that common issues of RICO causation predominate. Cf. In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2016 WL 5371856, at *7 (D. Kan. Sept. 26, 2016) (rejecting defendant's argument that "the fact and amount of injury present individual issues because the [Chicago Board of Trade ("CBOT")] price is not sufficiently tied to local corn prices," and explaining that the court would not "weigh the class-wide evidence concerning the relationship between CBOT and local prices" because "the persuasiveness of such evidence is for the jury at trial," and concluding that "a reasonable juror could believe that local corn prices do reflect changes in the CBOT futures price, as opined by plaintiffs' experts.").
In reaching this conclusion, the court recognizes that plaintiffs' 2-Pak theory is but one of the factual theories plaintiffs rely on to support their RICO claim. And, after ruling the class certification motion, the court still retains jurisdiction to refine and limit the scope of the class RICO claims. At this stage of the proceeding, defendants haven't convinced the court that the putative class includes "great numbers" of class members uninjured by the switch to the 2-Pak. But, the court is mindful that defendants have presented evidence in the form of deposition testimony showing that the 2-Pak switch may not have injured some named plaintiffs. If the evidence develops to show that "great numbers" of putative class plaintiffs were unharmed by the 2-Pak switch, the court retains the flexibility to modify its class certification decision and preclude plaintiffs from using the 2-Pak theory to support the RICO class claims. See, e.g., Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 823 (7th Cir. 2012) (recognizing that "depending on the ultimate size of the class at issue here," the estimated percentage of uninjured class members "may prove . . . to be a significant number of additional plaintiffs" but noting, "the district court is free to revisit this issue at a later time if discovery shows that the number of members . . . was more significant than it appears at this time"). The court can revise its certification decision or even limit plaintiffs' use of certain factual theories to support their RICO class claim. See Fed. R. Civ. P. 23(d)(1)(A) ("In conducting an action under this rule, the court may issue orders that . . . determine the course of proceedings or prescribe measures to prevent undue repetition or complication in presenting evidence or argument."). Cf. Shook v. Bd. of Cty. Comm'rs, 543 F.3d 597, 606-07 (10th Cir. 2008) (recognizing that a district court may create subclasses sua sponte as a way to segregate plaintiffs with different factual and individual circumstances used to support their claim); Marisol A. v. Giuliani, 126 F.3d 372, 379 (2d Cir. 1997) (recognizing that certifying subclass serves several purposes including an ability "to conduct the trial in a more orderly manner, by tying the order of proof to particular claims raised by the individual subclasses" and noting that "[o]ne possible method of developing proper subclasses would divide the present class based on the commonality of the [class members'] particular circumstances [or] the type of harm the [class members] allegedly have suffered . . . .")
Should defendants convince the court that "great numbers" of putative class members never could have been harmed by the switch to the 2-Pak because they would have purchased two EpiPens even if defendants had continued to make the single dose available in the market, the court has authority and discretion to modify its ruling.
(b) Delay of Generic Competition Through "Pay For Delay" Settlements
Plaintiffs also assert that their generic delay theory uses classwide proof to establish RICO causation. Plaintiffs allege that one component of the EpiPen pricing scheme was defendants' use of reverse payment settlements to resolve EpiPen patent litigation. Plaintiffs describe the reverse payment settlements as anticompetitive "pay for delay" settlements that effectively delayed generic competition from entering the EAI market for more than three years. Doc. 60 at 172 (Compl. ¶ 655.c.). To support their RICO claim, plaintiffs allege that defendants committed predicate RICO acts by making false statements in press releases about the settlements, describing them as a legitimate means to resolve patent litigation and concealing the true, anti-competitive reasons for the agreements. Id.; see also id. at 176 (Compl. ¶ 659.e.). And, plaintiffs allege, defendants' RICO violations succeeded in delaying generic competition which, in turn, protected Mylan's monopoly in the EAI market and allowed defendants to continue to raise the EpiPen's price.
Through expert testimony by Profs. Einer Elhauge and Andrew Torrance, plaintiffs offer evidence that the patent settlements were not legitimate settlements. Instead, the two professors opine the settlements really were unlawful reverse payment settlements contrived to delay generic competition. Plaintiffs assert that their evidence supporting this theory is common to all class members. The court agrees with them. The evidence focuses entirely on defendants' conduct and motivations. It thus applies to each putative class plaintiffs' claims and either fails or succeeds "in one stroke." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011).
Also, through Prof. Rosenthal's testimony, plaintiffs offer evidence that defendants' misconduct delaying generic competition caused the RICO class—as a whole—to overpay for the EpiPen. Specifically, Prof. Rosenthal calculated overcharges from delayed generic entry on a classwide basis using classwide data and information (calculated separately for TPPs, insured consumers, and cash payors). Doc. 1497-3 at 30, 77-82 (Rosenthal Expert Report ¶ 68, Attach. C.7.a-C.7.e).
Defendants separately filed Daubert motions challenging the reliability of Profs. Torrance, Elhauge, and Rosenthal's opinions. The court's Order denying those motions explains that defendants' challenges go to the weight of expert opinions but don't preclude the court from "admitting them," i.e., considering them when deciding the class certification motion. Thus, on the class certification motion, the court considers these expert opinions. And, the court concludes that they support class certification because they plausibly offer a classwide method for proving RICO liability.
Defendants' challenges to the persuasiveness of these experts' opinions is a different matter. The factfinder can decide how persuasive these opinions are when deciding the merits of the case—i.e., whether the opinions prove that defendants entered unlawful reverse payment settlements that violated RICO. But for current purposes, the opinions present sufficient evidence of classwide causation to support plaintiffs' class certification motion.
Defendants respond that plaintiffs can't rely on their generic delay theory to establish RICO causation on a classwide basis because they can't show that all plaintiffs knew about, much less relied on, defendants' misrepresentations about the legitimacy of the patent litigation settlements. Doc. 1636 at 72-73. Thus, defendants contend, plaintiffs can't connect the specific RICO predicate acts to any overpayment by any particular plaintiff. Id. Plaintiffs disagree. They contend that defendants' argument improperly assumes plaintiffs must show evidence of first-party reliance to prevail on a RICO claim. Plaintiffs have the better end of this dispute.
The Supreme Court has held—explicitly—that first-party reliance is not required to establish a RICO claim that relies on the predicate act of mail fraud. Bridge v. Phoenix Bond &Indem. Co., 553 U.S. 639, 649 (2008). Instead, RICO provides a cause of action for "[a]ny person injured in his business or property by reason of a violation of" the statute. 18 U.S.C. § 1964(c). The Supreme Court reasoned that this statutory language "suggest[s] a breadth of coverage not easily reconciled with an implicit requirement that the plaintiff show reliance in addition to injury in his business or property." Bridge, 553 U.S. at 649. So, plaintiffs argue, RICO doesn't require them to show classwide evidence of first-party reliance on defendants' misrepresentations. Instead, plaintiffs assert, they just need to show classwide evidence of injury to business of property "by reason of" defendants' RICO violations.
Bridge illustrates this principle. In that case, the Supreme Court held that a plaintiff could allege a RICO injury "'by reason of' a pattern of mail fraud even if he has not relied on any misrepresentations." Id. at 649. The Court found that plaintiffs' allegations "clearly" established that they "were injured by [defendants'] scheme." Id. Plaintiffs had alleged: "As a result of [defendants'] fraud, [plaintiffs] lost valuable liens they otherwise would have been awarded." Id. This was "true even though [plaintiffs] did not rely on [defendants'] false attestations of compliance with the county's rules." Id.
Likewise here, plaintiffs offer classwide evidence that one component of defendants' EpiPen pricing scheme—reverse payment settlements that delayed generic competition—injured plaintiffs by forcing them to overpay for the EpiPen. Like Bridge, plaintiffs have asserted an actionable RICO claim even if they didn't rely on defendants' alleged misrepresentations about the settlements' legitimacy.
But, defendants assert, Bridge still requires plaintiffs to support their RICO claim with evidence that "someone relied on the defendant's misrepresentations." Id. at 658. Whether plaintiffs can marshal evidence to make that showing is a merits-based determination that the court cannot make now. Instead, what matters now is whether plaintiffs can prove this requirement of their RICO claim with common evidence.
Plaintiffs assert that they can make this showing on class certification through an classwide inference of reliance. That is, according to plaintiffs, each plaintiff's payment for an EpiPen gives rise to an inference that he relied on defendants' fraudulent statements—or more specifically, defendants' fraudulent omissions—about the EpiPen pricing scheme. Indeed, the Tenth Circuit has recognized that "issues of reliance can be disposed of on a classwide basis without individualized attention at trial." CGC Holding Co., LLC v. Broad & Cassel, 773 F.3d 1076, 1089 (10th Cir. 2014) (citations omitted). "For example, where circumstantial evidence of reliance can be found through generalized, classwide proof, then common questions will predominate and class treatment is valuable in order to take advantage of the efficiencies essential to class actions." Id.
The court recognizes that "deceitful concealment of material facts may constitute actual fraud" sufficient to support a wire fraud claim. United States v. Cochran, 109 F.3d 660, 665 (10th Cir. 1997). "[A] misleading omission is actionable as fraud . . . if it is intended to induce a false belief and resulting action to the advantage of the misleader and the disadvantage of the misled." Id. (citation, internal quotation marks, and internal alternations omitted).
Also, in "certain circumstances . . . it is beneficial to permit a commonsense inference of reliance applicable to the entire class to answer a predominating question as required by Rule 23." Id. (citations omitted). And, "[i]n the RICO context, class certification is proper when 'causation can be established through an inference of reliance where the behavior of plaintiffs and the members of the class cannot be explained in any way other than reliance upon the defendant's conduct.'" Id. at 1089-90 (quoting In re Countrywide Fin. Corp. Mortg. Mktg. & Sales Practices Litig., 277 F.R.D. 586, 603 (S.D. Cal. 2011)).
In CGC Holding, the Circuit held that "evidence of payment for [a] loan commitment—more specifically, the inference that arises from it—is sufficient to present a predominating question related to class member reliance that can resolve a central issue of this litigation in one swoop." Id. at 1091. "More specifically the fact that a class member paid [a] nonrefundable up-front fee in exchange for the loan commitment constitutes circumstantial proof of reliance on the misrepresentations and omissions regarding [one of the defendant's] past and the defendant entities' ability or intent to actually fund the promised loan." Id. at 1091-92.
Similarly, here, plaintiffs argue that the court can infer that the putative plaintiffs relied on defendants' misrepresentations or omissions about the EpiPen pricing scheme when purchasing their EpiPens. Defendants disagree. They assert that the Tenth Circuit has cautioned that an inference of RICO causation "would not be appropriate in most RICO class actions." Id. at 1091 n.9. But, in that same footnote, the Tenth Circuit said: "[T]he inference of reliance here is limited to transactional situations—almost always financial transactions—where it is sensible to assume that rational economic actors would not make a payment unless they assumed that they were receiving some form of the promised benefit in return." Id. That's the scenario we have in this case. It is "sensible to assume" that the putative plaintiffs "would not [have made] a payment for" their EpiPens had they known—as plaintiffs allege—that they were paying an inflated price for that product due to defendants' misconduct in delaying generic competition. Id. But, defendants argue, "the record is replete with examples of named plaintiffs who bought EpiPen products for rational reasons that had nothing to do with Mylan's alleged misconduct." Doc. 1636 at 71 n.79. These are merits-based defenses. Defendants assert they can rebut the classwide presumption of reliance with evidence that some plaintiffs would have purchased their EpiPens even if they had known about defendants' alleged misconduct. But, here, the question on class certification is whether plaintiffs can prove reliance using classwide evidence. The court determines that they can—by using a classwide inference, as approved by our Circuit in CGC Holding. And this showing suffices to establish that common issues of RICO causation predominate under plaintiffs' generic delay theory.
(c) Exclusive Dealing Contracts with PBMs
Last, plaintiffs assert they will rely on common evidence to prove a third component of the EpiPen pricing scheme—i.e., defendants' use of exclusive dealing contracts with PBMs. Plaintiffs assert that defendants' exclusionary contracts blocked branded competition from the market and allowed Mylan to raise the EpiPen's price which, in turn, caused plaintiffs to sustain injury in the form of overpaying for the EpiPen. To support this argument, plaintiffs rely on Prof. Elhauge's Expert Report.
For the same reasons discussed above, supra note 46, the court finds Prof. Elhauge's opinions sufficient to establish that plaintiffs will offer classwide evidence of causation to support their RICO claims.
Prof. Elhauge opines that Mylan's exclusive dealing contracts successfully foreclosed rival product Auvi-Q from accessing a percentage of the EAI market. Doc. 1497-2 at 52-53 (Elhauge Expert Report ¶ 101). Using classwide methodology, Prof. Elhauge has calculated the effects of this foreclosure on EpiPen sales and its price. Id. at 56-66 (Elhauge Expert Report ¶¶ 110-20). He then used those calculations to quantify overcharges incurred by EpiPen consumers. Id. at 64-66 (Elhauge Expert Report ¶ 120, Table 3-4). As Prof. Elhauge explains, his method of calculating the overcharges caused by defendants' exclusive dealing contracts with PBMs "is common to the class, based on classwide evidence, and produces a conclusion about the foreclosure share that is common to the class." Id. at 52-53 (Elhauge Expert Report ¶ 101); see also id. at 54-55 (Elhauge Expert Report ¶ 104) (describing his "method of calculating" the foreclosing effects on Auvi-Q sales and EpiPen's market price inflation as "common to the class, based on classwide evidence, and [one that] produces conclusions about the foreclosing effect and market price inflation that are common to the class").
Defendants respond that too many individual issues exist for the court to conclude that plaintiffs can satisfy the predominance requirement. Defendants argue that the drug supply chain is complex, and it involves too many participants who ultimately affect drug prices. As an example, defendants direct the court to Local 282's experience. Local 282, a named plaintiff and a putative class representative, hired The Segal Group to negotiate drug prices with Local 282's PBM, which, in turn, negotiated with Mylan. Defendants argue that Local 282's experience shows there is an immutable impediment to class certification. And, defendants contend, Local 282's experience is similar to other putative class members' purchasing experiences. According to defendants, Local 282's experience shows there simply are too many intervening and individualized issues in the causal chain between the alleged predicate acts and each plaintiff's ultimate purchase of an EpiPen at a price negotiated by other parties. Thus, defendants argue, common causation issues don't predominate over individuals ones.
Plaintiffs urge the court to find their method of proving causation is sound under the approach approved by the First Circuit in three RICO cases collectively known as In re Neurontin Marketing & Sales Practices Litigation. Drug manufacturer Pfizer, the defendant in each case, asserted that intervening causes—in particular, actions of prescribing doctors—broke the chain of RICO causation between Pfizer's alleged misrepresentations when marketing the drug Neurontin and plaintiffs' injury. The First Circuit rejected Pfizer's argument for several reasons.
The three cases are Kaiser Foundation Health Plan v. Pfizer, Inc., 712 F.3d 21 (1st Cir. 2013), Aetna, Inc. v. Pfizer, Inc., 712 F.3d 51 (1st Cir. 2013), and Harden Manufacturing Co. v. Pfizer, Inc., 712 F.3d 60 (1st Cir. 2013). Harden was a class action of third-party payors; the other two cases were individual cases pursued by third-party payors.
First, the First Circuit recognized that the Supreme Court had held in Bridge that first-party reliance on a defendant's misrepresentations is not required. Kaiser, 712 F.3d at 36 (citing Bridge, 553 U.S. at 642); Aetna, 712 F.3d at 58 (citing Bridge, 553 U.S. at 642); Harden, 712 F.3d at 66-67 (citing Bridge, 553 U.S. at 642). Thus, the First Circuit concluded, Bridge foreclosed Pfizer's argument that any alleged misrepresentations it had made were made to the prescribing doctors—and not to the doctors' patients or others—and broke the causal link for third-party payors. Kaiser, 712 F.3d at 36-37; see also Harden, 712 F.3d at 67; Aetna, 712 F.3d at 58.
Second, the First Circuit found that plaintiffs could prove a "'direct relationship between the injury asserted and the injurious conduct alleged.'" Harden, 712 F.3d at 67 (quoting Holmes, 503 U.S. at 268). The First Circuit noted that plaintiffs in all three cases were the "primary and intended" victims of Pfizer's scheme to defraud. Kaiser, 712 F.3d at 37; see also Harden, 712 F.3d at 67; Aetna, 712 F.3d at 58. Thus, plaintiffs' injuries were "a 'foreseeable and natural consequence' of Pfizer's scheme—a scheme that was designed to fraudulently inflate the number of Neurontin prescriptions for which TPPs paid." Kaiser, 712 F.3d at 37 (quoting Bridge, 553 U.S. at 658); see also Harden, 712 F.3d at 67; Aetna, 712 F.3d at 58.
The First Circuit's conclusion in the Neurontin cases applies with equal force here. Plaintiffs allege that the RICO defendants implemented an EpiPen pricing scheme designed to foreclose competition in the EAI market, protect the EpiPen monopoly, and allow defendants to raise the EpiPen's price. Here, the putative plaintiffs—consumers who paid for the EpiPen—were the primary and intended victims of the alleged pricing scheme. Thus, plaintiffs' alleged injuries—overpayments for the EpiPen—were a foreseeable consequence of defendants' alleged scheme. Defendants' arguments about the complexity of the drug supply chain and the involvement of other actors in that chain "does not add such attenuation to the casual chain as to eliminate proximate cause" but, instead, these arguments "present[ ] a question of proof, to be resolved at trial, regarding the total number of prescriptions (if any) that were attributable to [defendants'] actions." Harden, 712 F.3d at 67; see also Kaiser, 712 F.3d at 39; Aetna, 712 F.3d at 59. And, to prove their allegations, plaintiffs will rely on common evidence about defendants' conduct in the EpiPen pricing scheme, the intended targets of that scheme, and how that scheme injured the plaintiff class as a whole.
Also, the Neurontin cases held that aggregate statistical evidence presented by Professor Meredith Rosenthal—the same expert plaintiffs rely on here—and other circumstantial evidence provided strong evidence of but-for causation. Harden, 712 F.3d at 68; see also Kaiser, 712 F.3d at 45-46; Aetna, 712 F.3d at 57-58. This meant, the First Circuit held, that plaintiffs could show Pfizer's fraud had caused plaintiffs to pay more for Neurontin than they otherwise would have paid. Harden, 712 F.3d at 68; see also Kaiser, 712 F.3d at 45-46; Aetna, 712 F.3d at 57-58. The Circuit recognized that some prescribing doctors had testified that they decided to prescribe Neurontin without relying on Pfizer's alleged misrepresentations. Harden, 712 F.3d at 68; see also Kaiser, 712 F.3d at 45; Aetna, 712 F.3d at 57-58. But, Neurontin also held, such evidence did not preclude plaintiffs from proving but-for causation. Harden, 712 F.3d at 68-69; see also Kaiser, 712 F.3d at 45; Aetna, 712 F.3d at 57-58. Instead, this case explained: "Ultimately, it is a jury's task to weigh the individual testimony presented by Pfizer against the aggregate and circumstantial evidence presented by . . . plaintiffs." Harden, 712 F.3d at 69; see also Kaiser, 712 F.3d at 45; Aetna, 712 F.3d at 57-58. The same is true here. The court finds that a factfinder can weigh any individual issues of causation against plaintiffs' projected common evidence to attempt to prove classwide RICO causation.
Defendants discredit plaintiffs' reliance on the Neurontin decisions, noting that the cases didn't decide class certification. Defendants are right. Defendants also correctly note that other Circuits have refused to apply Neurontin in the class certification context. See Sergeants Benevolent Ass'n Health & Welfare Fund v. Sanofi-Aventis U.S. LLP, 806 F.3d 71, 95-97 (2d Cir. 2015) (affirming district court's decision denying class certification and holding Prof. Rosenthal's regression analysis didn't support plaintiffs' classwide RICO causation theory because plaintiffs hadn't offered "anything beyond mere correlation that might support a reasonable inference that a [drug company's] alleged withholding of safety information played a legally sufficient causal role in the number of . . . prescriptions written"); see also Sidney Hillman Health Ctr. of Rochester v. Abbott Labs., 873 F.3d 574, 578 (7th Cir. 2017) (affirming dismissal of RICO class action claims after concluding that "improper representations made to physicians do not support a RICO claim by Payors" because they are "several levels removed in the causal sequence"). But see Painters & Allied Trades Dist. Council 82 Health Care Fund v. Takeda Pharms. Co. Ltd., 943 F.3d 1243, 1257-58 (9th Cir. 2019) (holding that plaintiffs sufficiently alleged proximate causation because their RICO claims sought "to hold Defendants liable for the consequences of their own acts and omissions toward Plaintiffs: the money spent by Plaintiffs to purchase" a pharmaceutical product and rejecting defendants' argument that "prescribing physicians' and pharmacy benefit managers' decisions constitute an intervening cause to sever the chain of proximate cause" because such a result "would . . . insulate [drug manufacturers] from liability for their fraudulent marketing schemes, as they could continuously hide behind prescribing physicians and pharmacy benefit managers" which is "not the purpose the requirement of proximate cause is intended to serve").
Indeed, the Ninth Circuit has recognized that an "inter-circuit split" exists over the question "whether patients and TPPs suing pharmaceutical companies for concealing an allegedly known safety risk about a drug can satisfy RICO's proximate cause requirement." Painters & Allied Trades, 943 F.3d at 1253. The Ninth Circuit elected to join the First and Third Circuits and concluded that "prescribing physicians do not constitute an intervening cause to cut off the chain of proximate cause" needed to support a RICO claim. Id. at 1257; see also id. (holding that the First and Third Circuits "have it right because their reasoning is more consistent with the Supreme Court's direct relation requirement"). The Ninth Circuit recognized that "prescribing physicians serve as intermediaries between Defendants' fraudulent omission . . . and Plaintiffs' payments for [the drug]," but that "prescribing physicians do not constitute an intervening cause to cut off the chain of proximate cause" because their involvement in defendants' alleged scheme was foreseeable. Id. (concluding that "it was perfectly foreseeable that physicians who prescribed [the drug] would play a causative role in Defendants' alleged fraudulent scheme to increase [the drug's] revenue"). Also, the Ninth Circuit found that defendants "have always known that 'physicians would not be the ones paying for the drugs they prescribed,'" but instead, defendants "are well aware that TPPs and individual patients pay for the drugs." Id. (quoting Neurontin, 712 F.3d at 38-39). And, it recognized that "[d]efendants' alleged fraudulent marketing scheme, which was intended to increase [the drug's] sales, 'only became successful once [they] received payments for the additional [drug] prescriptions [they] induced'—the very injury for which Plaintiffs seek recovery." Id. (quoting Neurontin, 712 F.3d at 39). In sum, the Ninth Circuit held, plaintiffs had satisfied "the Supreme Court's requirement that the proximate cause inquiry focus on the direct relation between the alleged violation and alleged injury." Id. (citing Hemi Grp., 559 U.S. at 12).
While the facts alleged here don't present the exact same issue decided by Painters & Allied Trades or Neurontin, the court still finds the reasoning of these Circuit decisions persuasive. And, the court predicts, the Tenth Circuit would follow their lead and apply their principles to this case. Thus, the court concludes that the presence of intermediary players in the drug supply chain don't present individual causation issues that overwhelm common questions of RICO causation. Instead, the court holds, putative class plaintiffs have shown how they could prove RICO causation from classwide proof that could persuade the factfinder that the purpose of defendants' EpiPen pricing scheme was to raise the price of the EpiPen. In turn, plaintiffs also have shown how classwide proof could provide a basis to conclude that defendants' alleged scheme caused plaintiffs—the ones ultimately paying for the EpiPen—to pay more for the product than they otherwise would have paid without defendants' unlawful acts. This is the very injury that plaintiffs seek to redress.
Plaintiffs have discharged their burden to demonstrate that they can prove RICO causation with classwide evidence. Here, Prof. Elhauge's analysis of the effect of defendants' exclusionary contracts on the price of the EpiPen plausibly demonstrates classwide causation. As his Report explains, Prof. Elhauge used a regression model to "estimate the total impact of the restrictive PBM formulary positions on the quantity of [rival product Auvi-Q] pens Sanofi was able to sell each month that it was on the market." Doc. 1497-2 at 58 (Elhauge Expert Report ¶ 113). Next, Prof. Elhauge estimated Mylan's price elasticity when Auvi-Q was on the market. Id. at 62 (Elhauge Expert Report ¶ 119). Then, using his estimates about the impact of PBM restrictions on quantity and Mylan's price elasticity, Prof. Elhauge calculated the price of the EpiPen caused by PBM foreclosure, multiplied that foreclosure by the quantity of Mylan's actual sales, and determined the total amount of overcharge caused by PBM foreclosure. Id. at 64-66 (Elhauge Expert Report ¶ 120, Table 3-4). Plaintiffs here have come forward with far more than "mere correlation evidence." Cf. Sergeants Benevolent, 806 F.3d at 96 (holding that expert's analysis provided only "mere correlation evidence" that didn't support RICO causation). The court thus finds that Prof. Elhauge's expert opinion suffices to establish how plaintiffs will rely on classwide evidence to show RICO causation. And, the court concludes that the common causation issues predominate over individual ones on plaintiffs' RICO claim premised on the theory that defendants' use of exclusive dealing contracts with PBMs caused plaintiffs to pay more for the EpiPen.
The court recognizes that, in Sergeants Benevolent, the Second Circuit held that plaintiffs failed their burden to show they could prove RICO causation with classwide evidence because their expert (Prof. Rosenthal) provided "mere correlation evidence" that didn't suffice to prove classwide causation. 806 F.3d at 96. But, the Second Circuit contrasted Prof. Rosenthal's analysis there with her analysis in Neurontin that involved an aggregate regression analysis. Id. The Second Circuit recognized: "Neurontin does indicate that where individual physicians' reliance on a pharmaceutical company's misrepresentations forms a necessary link in the causal chain between those misrepresentations and the plaintiffs' injury, such reliance can be proved to a jury with sufficiently powerful aggregate evidence, as opposed to individualized inquiries as to each prescribing physician's actual decisionmaking." Id. at 97. Nevertheless, the Second Circuit concluded, it "need not (and do[es] not intend to) express any view here on whether or when an aggregate regression analysis similar to the one deployed in Neurontin might be sufficient to prove causation on a class-wide basis in other pharmaceutical-marketing cases alleging a pattern of mail fraud actionable under RICO" because plaintiffs' causation evidence there was too "simplistic" to support classwide evidence of causation. Id. In contrast, as discussed above, plaintiffs here have offered classwide evidence of causation through Prof. Elhauge's regression model. And, as the court concludes above, this showing suffices to establish that common issues predominate the RICO causation analysis.
For all these reasons, the court finds that plaintiffs have established that common issues of RICO causation predominate over individual issues sufficient to satisfy the Rule 23(b)(3) requirement.
ii. RICO Damages
Next, when it comes to RICO damages, defendants argue that plaintiffs cannot satisfy Rule 23's predominance requirement. Plaintiffs respond, arguing that the Expert Reports for Profs. Elhauge and Rosenthal sufficiently demonstrate how plaintiffs can prove their RICO damages on a classwide basis. Defendants dispute this proposition, contending plaintiffs' RICO damage calculations run afoul of Comcast Corp. v. Behrend, 569 U.S. 27 (2013).
In Comcast, the Supreme Court determined that at the class certification stage, "any model supporting a 'plaintiff's damages case must be consistent with its liability case.'" 569 U.S. at 35 (quoting ABA Section of Antitrust Law, Proving Antitrust Damages: Legal and Economic Issues 57, 62 (2d ed. 2010)). Thus, in that antitrust case, Comcast held that plaintiffs must "tie each theory of antitrust impact to an exact calculation of damages." Id. at 37. The Comcast plaintiffs had alleged four different theories of antitrust impact, but the district court "accepted [just one of those theories] of antitrust impact as capable of classwide proof and rejected the rest." Id. at 31. Nevertheless, the district court certified a class based on an expert's regression model that "assumed the validity of all four theories of antitrust impact initially advanced" by plaintiffs. Id. at 36. On defendant's Rule 23(f) petition for interlocutory appeal, the Third Circuit affirmed the certification order. Behrend v. Comcast Corp., 655 F.3d 182 (3d Cir. 2011). The Supreme Court reversed, holding the district court had erred when it accepted plaintiffs' regression model because it "falls far short of establishing that damages are capable of measurement on a classwide basis." Comcast, 569 U.S. at 35. And, "[w]ithout presenting another methodology," plaintiffs could not prove "Rule 23(b)(3) predominance." Id. Instead, the Court reasoned, "[q]uestions of individual damage calculations will inevitability overwhelm questions common to the class." Id.
Here, defendants argue Prof. Rosenthal's alternative RICO damages model contravenes Comcast's requirement that class plaintiffs must connect each theory of liability to an exact calculation of damages. They assert that Prof. Rosenthal's calculation would require every class member to have sustained injury under every allegation in the Complaint for every transaction. The court rejects defendants' argument for two reasons.
First, defendants' argument ignores Prof. Rosenthal's actual estimate of RICO damages. In her analysis, she provided separate RICO damage estimates for plaintiffs' generic delay and single pack withdrawal theories and disaggregated each estimate among the affected group, e.g., TPPs, insured consumers, and cash payors. See Doc. 1500-3 at 29, 31 (Rosenthal Expert Report ¶¶ 67, 72). Simply put, defendants' characterization of Prof. Rosenthal's work contradicts the content of the certification record.
Also, Prof. Elhauge provided a separate damage calculation for plaintiffs' other theory—exclusive dealing through defendants' use of exclusionary PBM contract. See Doc. 1500-2 at 56-66 (Elhauge Expert Report ¶¶ 110-120); see also Doc. 1497-2 at 56-66 (Elhauge Expert Report ¶¶ 110-120) (sealed version).
Second, as Prof. Rosenthal has explained, her alternative RICO damages calculation merely represents how plaintiffs' allegations—collectively—show that defendants' conduct permitted Mylan to raise prices on the EpiPen. It presents a summary of aggregate damages caused by defendants' alleged RICO violations, calculated by comparing the actual price per pen to the but-for price that, in her opinion, would have existed absent the EpiPen pricing scheme. Unlike the problematic damages evidence in Comcast, Prof. Rosenthal's alternative theory doesn't include a theory of liability that the court has excluded. Thus, Prof. Rosenthal's theory is consistent with Comcast because her work honors the Supreme Court's rule that "a model purporting to serve as evidence of damages in [a] class action must measure only those damages attributable to that theory." Id. at 35. Plaintiffs here have identified several theories that purportedly injured the putative class members and thus produced damages. Prof. Rosenthal's alternative RICO damages calculation properly combines plaintiffs' theories of liability. And, at the same time, her other analyses disaggregate plaintiffs' damages based on plaintiffs' legal theories.
As the Seventh Circuit has recognized, "it was not the existence of multiple theories in [Comcast] that precluded class certification; it was the plaintiffs' failure to base all the damages they sought on the antitrust impact—the injury—of which the plaintiffs were complaining." Butler v. Sears, Roebuck & Co., 727 F.3d 796, 800 (7th Cir. 2013). The Comcast scenario is not present here. Plaintiffs have presented a damages model that is consistent with their liability theories. It does not contravene Comcast. And, it doesn't matter that they also have presented an alternative model.
Last, defendants assert that plaintiffs can't show classwide proof of damages because "some plaintiffs purchased EpiPen devices prior to events alleged in the Complaint." Doc. 1636 at 73. But, as plaintiffs explain, their class definitions dispense with this argument. The class definitions only include persons and entities who purchased the EpiPen after August 24, 2011—the date when Mylan launched the 2-Pak. Doc. 1353 at 2. Thus, the class definitions only include persons injured during the time defendants allegedly were executing the EpiPen pricing scheme. The court thus rejects defendants' argument that individual issues overwhelm the common issues of RICO damages.
For all these reasons, the court concludes that plaintiffs have demonstrated that common issues predominate the elements of their RICO claim sufficient to satisfy the class certification requirement of Rule 23(b)(3).
d. Antitrust
Plaintiffs ask the court to certify as a class action their state antitrust claims under the laws of 27 states and the District of Columbia. Plaintiffs assert that their theories of antitrust liability involve common questions because they focus squarely on defendants' allegedly unlawful conduct: (a) exclusive dealing arrangements where Mylan gave discounts to certain PBMs who drove and kept EpiPen prices above competitive levels; and (b) reverse payment patent litigation settlements where, plaintiffs allege, the EpiPen patent holder provided consideration to the alleged generic competitor/infringer in exchange for the generic competitor agreeing to delay entry of its generic product into the EAI market, thereby producing inflated EpiPen prices. Defendants argue that plaintiffs cannot prove either antitrust theory on a classwide basis because they can't use common evidence to prove the required element of market power. Also, defendants argue that plaintiffs cannot prove other elements of either one of their antitrust theories using classwide evidence. Instead, defendants contend, individual issues predominate the analysis, thus precluding certification. The court addresses defendants' arguments separately, below.
The 27 states are: Alabama, Arizona, California, Florida, Hawaii, Illinois, Iowa, Kansas, Maine, Michigan, Minnesota, Mississippi, Nebraska, Nevada, New Hampshire, New Mexico, New York, North Carolina, North Dakota, Oregon, Rhode Island, South Dakota, Tennessee, Utah, Vermont, West Virginia, and Wisconsin. And, as noted above, plaintiffs also ask the court to certify a class under the laws of the District of Columbia.
But the court already has dismissed the state law antitrust claims brought under the laws of Arizona, the District of Columbia, Iowa, New Mexico, North and South Dakota, Oregon, Rhode Island, Vermont, West Virginia, and Wisconsin because the Class Complaint includes no named plaintiff residing in any of these states. See Doc. 896 at 127; Doc. 1292 at 3. So, the court cannot certify a class action asserting antitrust violations under the laws of these states. That leaves 17 states subject to this portion of the certification motion.
Plaintiffs' opening brief acknowledges they must use common evidence to show that defendants inflicted antitrust impact, or injury, on the class. Doc. 1500 at 73-74. The antitrust injury requirement "'ensures that a plaintiff can recover only if the loss stems from a competitionreducing aspect or effect of the defendant's behavior.'" Elliott Indus. Ltd. P'ship v. BP Am. Prod. Co., 407 F.3d 1091, 1124-25 (10th Cir. 2005) (quoting Atl. Richfield Co. v. USA Petroleum Co., 495 U.S. 328, 344 (1990)). At the certification stage, plaintiffs do not have to prove impact; instead, they need "only to demonstrate that the element of antitrust impact is capable of proof at trial through evidence that is common to the class rather than individual to its members." Messner v. Northshore Univ. Healthsystem, 669 F.3d 802, 818 (7th Cir. 2012) (citation and internal quotation marks omitted). As discussed above, plaintiffs have provided a common methodology for calculating aggregate damages stemming from defendants' alleged antitrust violations through Profs. Elhauge and Rosenthal's expert opinions, and these calculations apply classwide for each of plaintiffs' antitrust theories. See supra Part IV.E.2.b. The court finds that this showing suffices to establish that plaintiffs are capable of presenting common evidence of antitrust impact.
i. Market Power
Defendants assert that plaintiffs can't show market power in the EAI market with common evidence. Plaintiffs respond, relying on Prof. Elhauge's report and his opinions about market power. He opines that market power can be established through common proof, including proof sufficient to establish the relevant product and geographic markets, and Mylan's dominant position in that market. According to Prof. Elhauge, Mylan held—at least—an 80% market share throughout the class period. Doc. 1500-2 at 11, 15 (Elhauge Expert Report fig.1, ¶ 22); see also Doc. 1497-2 at 11, 15 (Elhauge Expert Report fig.1, ¶ 22) (sealed version). And, according to Prof. Elhauge's calculations, Mylan's market power was even greater than 80% for much of the relevant period. Id.
Mylan asserts, in response, that Prof. Elhauge's opinions don't establish marker power for at least two reasons.First, Mylan asserts that Prof. Elhauge's calculations used incorrect data. According to Mylan's expert, Dr. Johnson, plaintiffs' entire model is flawed and Prof. Elhauge's conclusions about market power are unreliable. Second, Mylan argues, when PBMs shift to the lowest current formulary bid, it quickly changes the market share calculation. Mylan contends that its market share and ability to change prices vary dramatically across PBMs and formularies. Mylan asserts that Prof. Elhauge's failure to account for these variances undermines his conclusions. These showings, Mylan asserts, demonstrate that plaintiffs can't establish market power with evidence common to the entire class.
Defendants haven't asserted a Daubert challenge to Prof. Elhauge's opinions about market definition and market power. But defendants assert that they are reserving their right to challenge these opinions at a later stage. Doc. 1886 at 9 n.1.
Mylan's arguments are not persuasive. As already discussed, on class certification, the court need not determine that Prof. Elhauge's calculations are right. Plaintiffs only need to demonstrate that his analysis applies to the claims for all class members and the issue does not affect class members differently. Plaintiffs have shouldered that responsibility. The court thus finds plaintiffs have made a plausible showing that they can establish market power in a relevant market with common evidence.
ii. Exclusive Dealing Arrangements
Next, defendants deny that plaintiffs can show "exclusive dealing" through common evidence. They argue that the alleged exclusionary contracts that plaintiffs rely on to support their antitrust claims are with different PBMs across different formularies at different times. Consequently, defendants assert, determining whether each consumer or payor's formula constitutes an "exclusive agreement" requires individualized inquiry.
In response, plaintiffs assert that Prof. Elhauge's PBM foreclosure analysis is common to all class members because Mylan's complete foreclosure across the EAI market increased prices for all payors, not just those foreclosed. Prof. Elhauge explained that Mylan, through rebates it offered to PBMs—conditioned on exclusionary terms requiring Auvi-Q to be placed in a less favorable formulary position than EpiPen—gave PBMs the highest rebates and significantly affected the share of EAI doses that EpiPen and Auvi-Q sold within each PBM. Stated differently, Prof. Elhauge opined that Mylan's conditional rebating strategy managed to restrict a substantial portion of the market. According to his calculations, based on Mylan, IQVIA Xponent, and Sanofi's data, Mylan's exclusionary PBM contracts, once fully implemented, consistently foreclosed more than 30% of all formulary lives—and, they typically foreclosed 40%-60% of them. See Doc. 1500-2 at 52-53 (Elhauge Expert Report ¶¶ 99-101) (providing a classwide methodology for calculating share and price impacts of PBM coverage restrictions); see also Doc. 1497-2 at 52-53 (Elhauge Expert Report ¶¶ 99-101) (sealed version). Prof. Elhauge then analyzed the price impact of defendants' exclusionary conduct to estimate prices but for defendants' alleged misconduct. See Doc. 1500-2 at 56-64, 66 (Elhauge Expert Report ¶¶ 110-120, Table 4) (providing a classwide methodology for calculating share and price impacts of PBM coverage restrictions); see also Doc. 1497-2 at 56-64, 66 (Elhauge Expert Report ¶¶ 110-120, Table 4) (sealed version).
The IQVIA Xponent data consists of pharmacy records that plaintiffs produced. Doc. 1711-1 at 29 (Rosenthal Reply Report ¶ 48).
To Mylan's argument that the numerous and varying agreements between Mylan and PBMs preclude predominance, plaintiffs rely on Prof. Rosenthal's findings. She opines that the prices paid by third-party payors and cash payors directly correlated to the EpiPen's WAC price (a 99.9% and 99.8% correlation, respectively). Doc. 1711-1 at 8-9 (Rosenthal Reply Report ¶ 7, n.11, fig.2). Combining this correlation evidence with Prof. Elhauge's analysis, plaintiffs contend, shows that Mylan's actions raised baseline prices for all payors, even those on non-foreclosed formularies and despite the presence of prices negotiated from that baseline. Such common evidence supports a finding of predominance. See In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2016 WL 5371856, at *6-7 (D. Kan. Sept. 26, 2016) (holding plaintiffs satisfied the predominance requirement by showing that they would prove, through expert opinion, that defendants' actions caused a "general market price decrease" that applied across the board to each class member's corn sales); see also In re Urethane Antitrust Litig., 768 F.3d 1245, 1254 (10th Cir. 2014) (holding district court did not abuse its discretion by determining that the antitrust impact of defendants' alleged conduct "involved a common question that would override other individualized issues" because defendants' alleged price-fixing "creat[ed] an inference of class-wide impact even when prices are individually negotiated[,]" especially when the evidence showed that defendants' "conspiracy artificially inflated the baseline for price negotiations"); see also In re Lidoderm Antitrust Litigation, No. 14-md-2521-WHO, 2017 WL 679367, at *20-21 (N.D. Cal. Feb. 21, 2017) (rejecting defendants' argument that "the terms of the rebates [for third-party payors ("TPPs")] vary across the board and require individualized review depending on the size of the TPP, the type of TPP . . . , and the TPP plan" because plaintiffs had provided a "common method of proof by [their] expert, who determined the existence of overcharges on all purchases after determining a but-for price and after taking into account rebates[,]" and thus differences in TPP rebates didn't defeat predominance requirement).
As Prof. Rosenthal explained, the WAC price is the wholesale acquisition cost. Doc. 1724 at 15 (Tr. of Mot. Hr'g Class Certification - Phase I 15:12-22). It is the list price set by the manufacturer. Id. Once a manufacturer has determined a drug's WAC price, that price is typically used in transactions between the manufacturer and wholesaler. Id. It is essentially the benchmark the wholesaler uses for many of its sales to pharmacies. Id.
The court has found Profs. Elhauge and Rosenthal's methodologies sufficiently sound to withstand defendants' Daubert challenges. On certification, Mylan criticizes Prof. Elhauge's opinion as flawed because, defendants contend, he improperly defined "foreclosure." But, for purposes of the certification motion, the court need not determine whether Prof. Elhauge's opinion is correct. Instead, the court must decide whether plaintiffs have presented a plausible method for proving their antitrust claim using classwide evidence. See In re Lidoderm Antitrust Litigation, 2017 WL 679367, at *23 ("A rigorous review of [the experts'] opinions and their reasoning, as required under recent Supreme Court precedent, establishes that the concepts and designs of their models are solid. What facts and assumptions are appropriate to include in those models (and which model is preferred) are not issues [the court] can or should resolve on this [certification] motion."). The court concludes that plaintiffs have satisfied this burden with their experts' opinions. They explain how aggregate foreclosure and price increases applied classwide, even if a particular class member wasn't subject to a foreclosed PBM formulary. The court thus finds that plaintiffs have satisfied the predominance requirement, as used to support their exclusive dealing claim. Some individual questions likely remain, but the common questions dominate the analysis.
iii. Generic Delay
Defendants sharply contest plaintiffs' theory that defendants used a reverse payment settlement to delay entry of a generic EAI competitor. Defendants argue that plaintiffs can't show predominance of common issues for this theory. Plaintiffs respond, arguing that their expert opinions provide common evidence to support their generic delay claim, thus satisfying the predominance requirement.
Specifically, plaintiffs rely on Prof. Elhauge to (a) explain how reverse payment settlements can divide markets over time by delaying generic entry beyond what one otherwise would expect in a way that creates anticompetitive profits split among the settling parties; and (b) describe a methodology to calculate the economically rational entry date for the parties to have accepted in a no-payment settlement (had they been prevented from reaching a reverse payment settlement). Doc. 1500-2 at 17, 33-38 (Elhauge Expert Report ¶¶ 26, 56-66 ); see also Doc. 1497-2 at 17, 33-38 (Elhauge Expert Report ¶¶ ¶¶ 26, 56-66) (sealed version). Prof. Elhauge, in turn, employs Prof. Torrance's analysis of the EpiPen patents and related litigation to derive a generic entry date. Doc. 1500-2 at 39, 42 (Elhauge Expert Report ¶¶ 70, 75 ).
Using Prof. Elhauge's calculated generic entry date, Prof. Rosenthal analyzes how the reverse payment settlement produced supra-competitive prices for the EpiPen, thus forcing consumers to pay inflated EpiPen prices. Doc. 1500-3 at 22-23 (Rosenthal Expert Report ¶¶ 48-53); see also Doc. 1497-3 at 22-23 (Rosenthal Expert Report ¶¶ 48-53) (sealed version). Prof. Rosenthal has calculated the classwide overcharges paid by EpiPen consumers, and she also has provided an aggregate damage calculation for the antitrust class based on defendants' alleged use of an unlawful reverse payment settlement. Doc. 1500-3 at 27-30 (Rosenthal Expert Report ¶¶ 58-68); see also Doc. 1497-3 at 27-30 (Rosenthal Expert Report ¶¶ 58-68) (sealed version).
Defendants take issue with each of these expert opinions. Defendants try to undermine both Profs. Elhauge's and Torrance's opinions by citing the opinion of their own expert, Mr. Folsom, to demonstrate that Teva's chances for success at trial were much lower than Prof. Torrance opines. Also, Mylan criticizes Prof. Rosenthal's opinion about the number of class members who would have switched to a generic product had one been available in the market (i.e., the inverse of the proportion of "brand loyalists"), and her calculation of the average price consumers would have paid in the but-for world where Teva's generic entered the market.
Separately, Pfizer attacks Prof. Elhauge's analysis on two grounds. First, Pfizer contends that Prof. Elhauge's analysis fails to consider (a) Pfizer's economic incentives for settling the patent litigation (and also incorrectly calculates Teva's economic incentives), and (b) the merits of the patent claims in the litigation. Plaintiffs respond, citing Prof. Elhauge's Reply Report. He opines that Pfizer and its subsidiaries—as the sole manufacturer and supplier of the EpiPen—"would benefit from any entry delay caused by the reverse payment, since that would give them more profits under their agreement to supply Mylan." Doc. 1711-2 at 18 (Elhauge Reply Report ¶ 34). Thus, Prof. Elhauge concludes, "Pfizer and its subsidiaries, regardless of their settlement or litigation payoffs, would have found it economically rational to agree to any reverse-payment settlement that was expected to delay Teva's entry." Id. And, Prof. Elhauge correctly notes, "any disputes about the extent to which Mylan drove bargaining about the delay in settlement entry date obtained for the reverse payment would be resolvable based on classwide evidence and any resolution of that issue would be common to the class." Id.
Plaintiffs don't respond specifically to Pfizer's arguments about the merits of the patent claims. Defendants criticize Prof. Elhauge's reliance on Prof. Torrance's opinion that Teva had more than a 70% chance of prevailing in the patent litigation. Defendants levied the same attacks against Prof. Torrance's opinion in the Daubert motion. As the court discussed in its Order denying that motion, defendants' arguments go to the weight of Prof. Torrance's opinions but don't preclude their admissibility. And, here, on certification, the court need not weigh whether Prof. Torrance accurately opines about Teva's chances of success in the patent litigation. Instead, the court must determine whether Prof. Elhauge's opinions, relying on Prof. Torrance's opinions, provide a plausible method for proving plaintiffs' generic delay theory on a classwide basis. The court finds that they do.
Second, Pfizer contends that plaintiffs cannot hold it liable under the generic delay theory because the record contains no evidence that Pfizer induced any generic delay for the EpiPen. Plaintiffs premise their generic delay theory on an allegation that the "reverse payment" that purportedly induced Teva to delay its generic EpiPen was an agreement by Mylan to delay entry of a Nuvigil generic—the product that was the subject of a separate lawsuit between Mylan and a Teva affiliate. Pfizer argues that plaintiffs have adduced no evidence connecting Pfizer to the Nuvigil litigation when it settled the EpiPen patent litigation with Teva. These arguments appear to address the merits of plaintiffs' antitrust claim against Pfizer—i.e., whether plaintiffs can marshal evidence showing that Pfizer participated in the alleged unlawful reverse payment settlement. The court need not—indeed, cannot decide the merits of that issue on certification.
In any event, as plaintiffs argue, they allege an antitrust conspiracy in this case. Members of a conspiracy are liable for acts committed by their co-conspirators in furtherance of the conspiracy. Pinkerton v. United States, 328 U.S. 640, 646-48 (1946); see also Halberstam v. Welch, 705 F.2d 472, 487 (D.C. Cir. 1983) (holding "the law on civil conspiracy" imposes liability on co-conspirators for acts that are "a reasonably foreseeable consequence of the scheme"). So, if plaintiffs can prove that defendants engaged in a conspiracy, the law will hold Pfizer liable for the acts of Mylan as its co-conspirator. Also, as one leading treatise has noted, proof of a conspiracy "is a common question that is thought to predominate over the other issues in the case and has the effect of satisfying the first prerequisite in Rule 23(b)(3)." 7AA Charles Alan Wright, et al., Federal Practice & Procedure § 1781 (3d ed. 2005). The court rejects Pfizer's argument that its purported lack of participation in a reverse payment settlement precludes certification.
The court already has addressed many of defendants' other criticisms about plaintiffs' experts' methodologies and the accuracy of their calculations. By separate Order, the court has found these experts have substantiated their conclusions sufficient to survive Daubert challenges at this stage. Whether these expert opinions ultimately will persuade a factfinder that defendants violated the antitrust laws by delaying generic entry is not something the court can decide at certification. Nor do plaintiffs need to prove that these experts' opinions and calculations are accurate. Instead, to shoulder their certification burden, plaintiffs must show that common issues subject to classwide evidence predominate over individual questions. The court finds that plaintiffs have satisfied that burden. Plaintiffs' experts have presented a plausible method for proving their antitrust claims under a generic delay theory with evidence that applies on a classwide basis. This suffices to establish predominance to support plaintiffs' generic delay theory.
For the above reasons, the court finds that that plaintiffs have shouldered their burden to establish that common issues predominate their antitrust claim sufficient to satisfy the class certification requirement of Rule 23(b)(3).
e. Consumer Protection
Plaintiffs ask the court to certify as a class action their state consumer protection law claims under the laws of 20 states and the District of Columbia that prohibit anticompetitive conduct or unfair practices. Defendants oppose certification of any state consumer protection class action. They contend that variations among the states' laws preclude certification because states use different tests to determine what is unfair conduct violating consumer protection statutes. The court agrees with defendants. The differences among state consumer protection statutes are significant, they present individual issues, and they overwhelm the common questions. As a consequence, plaintiffs fail to establish predominance, and thus, these individual issues defeat class certification of the state consumer protection class under Rule 23(b)(3).
The 20 states are: Alaska, California, Connecticut, Florida, Hawaii, Illinois, Maine, Maryland, Massachusetts, Missouri, Nebraska, Nevada, New Hampshire, New Mexico, North Carolina, Oklahoma, Rhode Island, Vermont, Washington, and West Virginia. Doc. 1500 at 86, 88. And, as noted, plaintiffs also seek certification under the laws of the District of Columbia.
The court notes that the Consolidated Class Action Complaint doesn't plead a violation of the state consumer protection laws in Alaska, the District of Columbia, or Rhode Island. See Doc. 60 at 182-394. It only alleges violation of those three jurisdictions' antitrust laws. See id. at 145-47, 149-50, 152-54, 157-59. But the court has dismissed those three state law antitrust claims because the Class Complaint includes no named plaintiff residing in any of the three. See Doc. 896 at 127 (dismissing "the class plaintiffs' state law antitrust claims asserted against both sets of defendants under the laws of Alaska, . . . the District of Columbia, [and] Rhode Island . . . because the Class Complaint includes no named plaintiff who resides in any of these states"). Also, the court has dismissed the state consumer protection law claims brought under the laws of New Mexico, Vermont, Washington, and West Virginia because the Class Complaint includes no named plaintiff residing in those states. Doc. 896 at 127, Doc. 1292 at 3.
Defendants also make three other arguments. They contend that the state consumer protection statutes at issue here award different types of damages, use varying rules for eligibility to sue, and impose differing statutes of limitations. And, defendants argue, these differences in state laws present individual issues that overwhelm the common questions. Because the court agrees with defendants' argument that differences in how the state laws define unfair conduct prevent certification, the court does not need to reach defendants' other arguments against certification.
As plaintiffs concede, states apply different tests when determining what qualifies as unfair conduct prohibited by a particular consumer protection statute. Doc. 1500 at 89; Doc. 1837 at 39-40. But, plaintiffs dismiss these differences, arguing they are just "[m]inor variations" in the laws. Doc. 1837 at 40. And, plaintiffs assert, the court can manage the differences by submitting special verdict forms or using other "mechanisms." Id. The court disagrees.
Plaintiffs don't identify the other mechanisms the court could employ. But, they cite In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2016 WL 5371856 (D. Kan. Sept. 26, 2016, where our court certified eight, separate statewide class actions. Id. at *12-15 (certifying separate statewide classes consisting of corn producers in Arkansas, Illinois, Iowa, Kansas, Missouri, Nebraska, Ohio, and South Dakota). But this comparison isn't an apt one. Here, plaintiffs don't ask the court to certify separate statewide classes. Instead, they ask the court to certify one "Consumer Protection Damages Class" consisting of all persons and entities in the "Consumer Protection States" who purchased or paid for EpiPens. Doc. 1353 at 3.
The state laws at issue here use a variety of tests to define the form of conduct that will qualify as anticompetitive or unfair conduct prohibited under the consumer statutes. For example, some states consider the following factors: (1) whether the practice offends public policy, the common law, or otherwise; (2) whether it is immoral, unethical, oppressive, or unscrupulous; or (3) whether it causes substantial injury to consumers. Other states examine whether the act or practice: (1) causes or is likely to cause substantial injury to consumers; (2) which is not reasonably avoidable by the consumer; and (3) not outweighed by countervailing benefits to consumers or to competition. Nebraska, New Hampshire, and New Mexico apply their own standards to define unfair conduct under their state consumer protection statutes. Nevada's statute only prohibits deceptive conduct, not unfair conduct. And, in California, there is internal dissonance about the test courts should use under California's act. Courts there have applied two different tests to determine what is an unfair practice under the California Unfair Competition Law.
These states include: Connecticut (Cenatiempo v. Bank of Am., N.A., 219 A.3d 767, 790 (Conn. 2019) (discussing requirements of Connecticut Unfair Trade Practices Act)); Florida (PNR, Inc. v. Beacon Prop. Mgmt., Inc., 842 So.2d 773, 777 (Fla. 2003) (discussing requirements of the Florida Deceptive and Unfair Trade Practices Act)); Hawaii (Hungate v. Law Office of David B. Rosen, 391 P.3d 1, 18 (Haw. 2017) (discussing requirements of Haw. Rev. Stat. § 480-2)); Illinois (Robinson v. Toyota Motor Credit Corp., 775 N.E.2d 951, 961 (Ill. 2002) (discussing requirements of the Illinois Consumer Fraud and Deceptive Business Practices Act)); Massachusetts (Purity Supreme, Inc. v. Attorney Gen., 407 N.E.2d 297, 307 (Mass. 1980) (discussing requirements of Mass. Gen. Laws ch. 93A, § 2)); Missouri (Chochorowski v. Home Depot U.S.A., 404 S.W.3d 220, 226 (Mo. 2013) (en banc) (discussing requirements of the Missouri Merchandising Practices Act)); North Carolina (Bumpers v. Cmty. Bank of N. Va., 747 S.E.2d 220, 228 (N.C. 2013) (discussing requirements of North Carolina's Unfair and Deceptive Practices Act)); Oklahoma (Okla. Stat. tit. 15, § 752(14) (Oklahoma Consumer Protection Act)); and Vermont (Christie v. Dalmig, Inc., 396 A.2d 1385, 1388 (Vt. 1979) (discussing requirements of Vermont Consumer Fraud Act)).
These states include: Maine (Searles v. Fleetwood Homes of Pa., Inc., 878 A.2d 509, 519 n.10 (Me. 2005) (discussing requirements of the Maine Unfair Trade Practices Act)); and Maryland (Sager v. Hous. Comm'n of Anne Arundel Cty., 957 F. Supp. 2d 627, 642 (D. Md. 2013) (discussing the requirements of the Maryland Consumer Protection Act and citing Legg v. Castruccio, 642 A.2d 906, 916 (Md. Ct. Spec. App. 1994)). Also, West Virginia has not adopted this test explicitly, but the West Virginia Consumer Credit and Protection Act instructs courts construing that act to "be guided by the policies of the Federal Trade Commission and interpretations given by the Federal Trade Commission and the federal courts to Section 5(a)(1) of the Federal Trade Commission Act . . . ." W.Va. Code § 46A-6-101(1). Section 45(a)(1) of Title 15 to the United States Code declares unlawful "[u]nfair methods of competition in or affecting commerce" and "unfair or deceptive acts or practices in or affecting commerce." Section 45(n) prohibits the FTC from declaring an act unlawful on the ground that it is "unfair unless the act or practice causes or is likely to cause substantial injury to consumers which is not reasonably avoidable by consumers themselves and not outweighed by countervailing benefits to consumers or to competition." The court predicts that West Virginia would apply this test when determining whether an act is an unfair act prohibited by the West Virginia Consumer Credit and Protection Act.
Nebraska requires the plaintiff asserting a Nebraska Consumer Protection Act claim to "prove that the practice either '(1) fell within some common-law, statutory, or other established concept of unfairness or (2) was immoral, unethical, oppressive, or unscrupulous.'" State ex rel. Stenberg v. Consumer's Choice Foods, Inc., 755 N.W. 2d 583, 591 (Neb. 2008) (quoting Raad v. Wal-Mart Stores, Inc., 13 F. Supp. 2d 1003, 1014 (D. Neb. 1998) and noting that the Nebraska Supreme Court has not defined unfair practices under the state consumer protection act but recognizing that the Nebraska federal court defined the term in Raad). New Hampshire applies "the rascality test" to determine whether an act is an unfair or deceptive act prohibited by the New Hampshire Consumer Protection Act. Axenics, Inc. v. Turner Constr. Co., 164 N.H. 659, 675-76 (N.H. 2013). The rascality test requires that "the objectionable conduct . . . attain a level of rascality that would raise an eyebrow of someone inured to the rough and tumble of the world of commerce." Id. (citation and internal quotation marks omitted). New Mexico's Unfair Trade Practices Act specifically defines the conduct prohibited by the statute. See N.M. Stat. Ann. § 57-12-2(E) (defining an "unconscionable trade practice" as "an act or practice . . . that to a person's detriment: (1) takes advantage of the lack of knowledge, ability, experience or capacity of a person to a grossly unfair degree; or (2) results in a gross disparity between the value received by a person and the price paid"). And New Mexico courts apply the statute's definition when determining whether conduct violates the statute. See, e.g., State ex rel. King v. B & B Inv. Grp., Inc., 329 P.3d 658, 665 (N.M. 2014) (applying N.M. Stat. Ann. § 57-12-2(E)'s definition of "unconscionable trade practice").
The Nevada Deceptive Trade Practices Act allows "any person who is a victim of consumer fraud" to bring an action under the statute. Nev. Rev. Stat. § 41.600(1). "Consumer fraud" includes "deceptive trade practices" as defined in §§ 598.0915 to 597.0925. Nev. Rev. Stat. § 41.600(2)(e). The Nevada federal district court has interpreted the statute's plain language to require a plaintiff to "show a defendant engaged in a consumer fraud of which the plaintiff was a victim." Picus v. Wal-Mart Stores, Inc., 256 F.R.D. 651, 657 (D. Nev. 2009).
The Ninth Circuit has recognized that the California Unfair Competition Law ("UCL") "does not define the term unfair" and "the proper definition of unfair conduct against consumers is currently in flux among California courts." Davis v. HSBC Bank Nev., N.A., 691 F.3d 1152, 1169 (9th Cir. 2012) (citation and internal quotation marks omitted). The Ninth Circuit has cited two tests that the California courts have used to determine whether alleged conduct violates the UCL. Id. at 1169-70.
First, under the South Bay test, "'unfair' conduct occurs when that practice 'offends an established public policy or when the practice is immoral, unethical, oppressive, unscrupulous or substantially injurious to consumers.'" Id. at 1169 (quoting S. Bay Chevrolet v. Gen. Motors Acceptance Corp., 85 Cal. Rptr. 2d 301, 316 (Cal. Ct. App. 1999)). Second, under the Cel-Tech test, a finding of "unfairness" under the UCL "'[must] be tethered to some legislatively declared policy or proof of some actual or threatened impact on competition.'" Id. at 1700 (quoting Cel-Tech Commc'ns, Inc. v. L.A. Cellular Tel. Co., 973 P.2d 527, 544 (Cal. 1999)). In Davis, the Ninth Circuit did not decide which of the two tests defines unfair conduct under the UCL because, it determined, plaintiff had failed "to state a claim under either definition." Id. The Ninth Circuit and California federal courts have followed Davis's guidance and have applied both tests when determining whether alleged conduct qualifies as "unfair" conduct prohibited by the UCL. See, e.g., Hodsdon v. Mars, Inc., 891 F.3d 857, 866-67 (9th Cir. 2018); Rhodeman v. Ocwen Loan Servicing, LLC, No. EDCV 18-2363 JGB(KKx), 2019 WL 5955368, at *17 (C.D. Cal. Nov. 12, 2019); In re Apple Inc. Device Performance Litig., 347 F. Supp. 3d 434, 463-64 (N.D. Cal. 2018).
The court recognizes that differences in state laws don't always preclude class certification under Rule 23(b)(3)'s predominance requirement. Courts have found class certification warranted when "the idiosyncratic differences between state consumer protection laws are not sufficiently substantive to predominate over the shared claims." Hanlon v. Chrysler Corp., 150 F.3d 1011, 1022-23 (9th Cir. 1998), overruled on other grounds by Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011); see also Waste Mgmt. Holdings, Inc. v. Mowbray, 208 F.3d 288, 296 (1st Cir. 2000) ("As long as a sufficient constellation of common issues binds class members together, variations in the sources and application of [state law] will not automatically foreclose class certification under Rule 23(b)(3).").
But "varying state laws may defeat predominance" in other situations. Senne v. Kan. City Royals Baseball Corp., 934 F.3d 918, 928 (9th Cir. 2019); see also Castano v. Am. Tobacco Co., 84 F.3d 734, 741 (5th Cir. 1996) ("In a multi-state class action, variations in state law may swamp any common issues and defeat predominance."). Indeed, the Supreme Court has observed that "[d]ifferences in state law [can] compound . . . the disparities" among the factual questions, thus "undermining class cohesion." Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 624 (1997).
Courts have declined to certify class actions when the legal differences among varying state laws dominate the common questions. See, e.g., Pilgrim v. Universal Health Card, LLC, 660 F.3d 943, 944 (6th Cir. 2011) (holding that "[t]he consumer-protection laws of many States . . . govern these claims and factual variations among the claims abound, making a class action in this setting neither efficient nor workable nor above all consistent with the requirements of Rule 23"); Zinser v. Accufix Research Inst., Inc., 253 F.3d 1180, 1189-90 (9th Cir. 2001) (affirming denial of class certification of a products liability action because "variances in state laws overwhelm common issues of fact"); In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293, 1300-02 (7th Cir. 1995) (decertifying class action and recognizing that negligence laws may vary among states "only in nuance" but "nuance can be important" and a review of differing state pattern instructions and judicial viewpoints about the meaning of negligence show that "the states of the United States sing negligence with a different pitch," thus the district court erred by finding that it could give a single instruction on negligence—one that "merg[ed] the negligence standards of the 50 states and the District of Columbia"); Commander Props. Corp. v. Beech Aircraft Corp., 164 F.R.D. 529, 541 (D. Kan. 1995) (O'Connor, J.) (denying class certification when "the significant number of individual inquiries required by the various state law claims make them ill-suited for class action treatment"); Aks v. Bennett, 150 F.R.D. 187, 192 (D. Kan. 1993) (Lungstrum, J.) (denying class certification because "common questions of law and fact [did] not predominate" when the claims were "subject to various state trust laws" and thus required analysis on a "state by state basis"). Cf. Doll v. Chi. Title Ins. Co., 246 F.R.D 683, 689 (D. Kan. 2007) (Lungstrum, J.) (refusing to certify a class action that involved "varying limitations laws for tort claims in the 18 jurisdictions in which the class members resided" because "[t]he complexity of these analyses of differing states' laws makes the class action proposed by plaintiffs unmanageable and weighs against a finding that a class action would be superior").
This case presents the same situation. The disparity of standards the states use to define unfair conduct under various consumer protection statutes present individual issues, and they overwhelm the common issues. As defendants argue, a factfinder would have to apply each state's law to determine whether the conduct violates a specific state consumer protection statute. And, the factfinder properly could reach different conclusions about the legality of the same conduct based on the varying state law tests it must apply. It is difficult to imagine instructions that could fairly guide a jury to make classwide findings on varying state consumer protection laws and their many standards for defining prohibited conduct. Defendants have persuaded the court that this complexity would confuse the jury or prejudice defendants. The court finds that the variations among state consumer protection laws preclude predominance and thus make it inappropriate to certify the state consumer protection claims as a class action under Rule 23(b)(3).
f. Unjust Enrichment
Unjust enrichment claims generally require plaintiffs to prove three things: (a) defendant received a benefit; (b) at plaintiff's expense; and (c) retention of the benefit would be unjust without compensation. See, e.g., Haz-Mat Response v. Certified Waste Serv. Ltd., 910 P.2d 839, 847 (Kan. 1996) ("The basic elements of a claim based on a theory of unjust enrichment are threefold: (1) a benefit conferred upon the defendant by the plaintiff; (2) an appreciation or knowledge of the benefit by the defendant; and (3) the acceptance or retention by the defendant of the benefit under such circumstances as to make it inequitable for the defendant to retain the benefit without payment of its value." (citation and internal quotation mark omitted)).
Plaintiffs contend defendants received increased profits at plaintiffs' expense, which they acquired unjustly and as a result of their own allegedly unlawful conduct. Plaintiffs assert their unjust enrichment claims satisfy the predominance requirement because they premise their claims on many of the same alleged facts and will prove them using much of the same evidence as their other claims—i.e., evidence focused on defendants' conduct as applied to the entire class. While plaintiffs acknowledge some variations among the states' unjust enrichment laws, they dismiss those differences as minor or idiosyncratic and assert they do not significantly alter the central issue or manner or proof. Mylan disagrees, arguing that the elements of an unjust enrichment claim are not uniform across all the states and the District of Columbia, which could produce different outcomes for different plaintiffs in different states. These differences, Mylan argues, mandate individualized inquiry and preclude certification of an unjust enrichment class.
Plaintiffs urge the court to follow Menocal v. GEO Group, Inc., 882 F.3d 905 (10th Cir. 2018), where the Tenth Circuit affirmed class certification of an unjust enrichment class. But the class claim there was brought under the law of a single state (Colorado), and defendant only challenged the predominance element, as it applied to the unjustness element and damages. Id. at 925-27. The Tenth Circuit concluded that even though unjustness requires a "fact-intensive inquiry," it presented a common question in that case because the theory depended on shared facts focusing on defendant's pay policy that applied uniformly to all class members. Id. at 925. On damages, the Tenth Circuit acknowledged damages would have to be ascertained on an individual basis, but agreed with the district court's finding that they should be easily calculable using a simple formula. Id. at 27.
This case is significantly different than Menocal, where plaintiffs asserted an unjust enrichment claim under the laws of just one state—Colorado. Here, plaintiffs want to certify a nationwide class to assert unjust enrichment claims. Our court has noted that differences among state law definitions of unjust enrichment and its availability as a remedy make federal courts, in general, reluctant to certify a nationwide class on this theory. Thompson v. Jiffy Lube Int'l, Inc., 250 F.R.D. 607, 626 (D. Kan. 2008) ("Because of such variations [among state laws], federal courts have generally refused to certify a nationwide class based upon a theory of unjust enrichment.").
For the same reasons, the court finds here that the individualized inquiries required to determine claims from plaintiffs in 50 states plus the District of Columbia—such as whether a plaintiff conferred a benefit directly on defendants, or whether the type of alleged misconduct satisfies a particular state's law, or assessing compliance with different statutes of limitations—would present significant manageability issues. Also, as plaintiffs concede, certain states preclude plaintiffs from maintaining an unjust enrichment claim if they have another available remedy at law. These questions demand an individualized inquiry that will swamp the common questions. Because these variations exist and pose troubling case management obstacles, the court follows the rule recognized in Thompson. Exercising its discretion, the court denies plaintiffs' request to certify a nationwide class based on an unjust enrichment theory.
See, e.g., Danny Lynn Elec. & Plumbing, LLC v. Veolia ES Solid Waste Se., Inc., No. 2:09cv192-MHT, 2011 WL 2893629, at *6 (M.D. Ala. July 19, 2011) (dismissing Alabama unjust enrichment claim because "the individual defendants enjoyed no direct benefit") (emphasis added)); Mandarin Trading Ltd. v. Wildenstein, 944 N.E.2d 1104, 1111 (N.Y. 2011) (holding that New York unjust enrichment claim failed because "the connection between the parties is too attenuated" and "the pleadings failed to indicate a relationship between the parties that could have caused reliance or inducement").
See, e.g., Bland v. Abbott Labs., Inc., No. 3:11-CV-430-H, 2012 WL 32577, at *2 (W.D. Ky. Jan. 6, 2012) (noting that a Kentucky unjust enrichment claim requires proof of defendants' bad faith and dismissing unjust enrichment claim for failing to allege bad faith); Bimbo Bakeries USA, Inc. v. Pinckney Molded Plastics, Inc., No. 4:06-CV-180-A, 2007 WL 836874, at *6 (N.D. Tex. Mar. 20, 2007) (dismissing Texas unjust enrichment claim because a viable claim requires "some type of unlawful behavior" and there was no "evidence of fraud, duress, or undue advantage" to support the claim).
An unjust enrichment claim in Kansas has a three year statute of limitations, Kan. Stat. Ann. § 60-512, and begins to run "when the enrichment becomes unjust[,]" Estate of Draper v. Bank of Am., N.A., 205 P.3d 698, 715 (Kan. 2009) (internal citation and quotation marks omitted). In contrast, an unjust enrichment claim in Michigan "accrues at the time the wrong upon which the claim is based was done regardless of the time when damage results," Mich. Comp. Laws § 600.5827, and the statute of limitations is six years, Williams v. Chase Bank, No. 15-10565, 2015 WL 4600067, at *4 (E.D. Mich. July 29, 2015) (citing Mich. Comp. Laws § 600.5813)).
See, e.g., KLE Constr., LLC v. Twalker Dev., LLC, 887 N.W.2d 536, 538 (N.D. 2016) (including as one of the elements of a North Dakota unjust enrichment claim "the absence of a remedy provided by law"); Freeman v. Sorchych, 245 P.3d 927, 936 (Ariz. Ct. App. 2011) (listing one of the elements of an Arizona unjust enrichment claims as "the absence of a remedy provided by law"); Cantor Fitzgerald, L.P. v. Cantor, 724 A.2d 571, 585 (Del. Ch. 1998) (reciting the elements of a Delaware unjust enrichment claim as including "the absence of a remedy provided by law").
F. Rule 23(b)(2) Injunction Class
Plaintiffs also seek to certify an injunctive relief class under Rule 23(b)(2). They make this request in addition to their requests that the court certify several Rule 23(b)(3) plaintiff classes to seek monetary damages. See, e.g., Doc. 1353 at 2-3. Plaintiffs explain that the putative Rule 23(b)(2) plaintiff class seeks injunctive relief sufficient to restore consumer choice to the market and the natural competitive equilibrium that would exist but-for Mylan's decision to sell the EpiPen exclusively in the 2-Pak. Plaintiffs allege that selling the EpiPen exclusively in a 2-Pak has caused consumers to pay more to purchase EpiPens than they would have paid if they could have purchased EpiPens in a single dose. And, plaintiffs assert, they will continue to sustain injury as long as defendants continue to sell the EpiPen exclusively in the 2-Pak under the false pretext of medical necessity. In simple terms, plaintiffs ask the court to enjoin Mylan from selling the Epi-Pen exclusively in the 2-Pak.
Rule 23(b)(2) allows a court to certify a class action if the requirements of Rule 23(a) are satisfied and "if the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." Fed. R. Civ. P. 23(b)(2). The Tenth Circuit has explained that Rule 23(b)(2) "demands at the class certification stage plaintiffs describe in reasonably particular detail the injunctive relief they seek 'such that the district court can at least "conceive of an injunction that would satisfy [Rule 65(d)'s] requirements," as well as the requirements of Rule 23(b)(2).'" D.G. ex rel. Stricklin v. Devaughn, 594 F.3d 1188, 1200 (10th Cir. 2010) (quoting Shook v. Bd. of Cty. Comm'rs, 543 F.3d 597, 605 (10th Cir. 2008) (quoting Monreal v. Potter, 367 F.3d 1224, 1236 (10th Cir. 2004))).
The Supreme Court has explained that "Rule 23(b)(2) applies only when a single injunction or declaratory judgment would provide relief to each member of the class." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 360 (2011). But Rule 23(b)(2) "does not authorize class certification when each individual class member would be entitled to a different injunction or declaratory judgment against the defendant," or "when each class member would be entitled to an individualized award of monetary damages." Id. at 360-61.
After Dukes, a "sort of hybrid certification" has "gain[ed] ground in class actions suits." Ebert v. Gen. Mills, Inc., 823 F.3d 472, 480 (8th Cir. 2016) (citing 2 William B. Rubenstein, Newberg on Class Actions § 4:38 (5th ed. 2019)); see also West v. Cal. Servs. Bureau, Inc., 323 F.R.D. 295, 307 (N.D. Cal. 2017) ("District courts may certify both a 23(b)(2) class for the portion of the case concerning injunctive and declaratory relief and a 23(b)(3) class for the portion of the case requesting monetary damages." (citing Newberg on Class Actions § 4:38)). Plaintiffs don't refer to this aspect of their certification motion as a request for a "hybrid" or "dual" class action under both Rule 23(b)(2) and 23(b)(3). But the court assumes that's what plaintiffs intended to do because they ask the court to certify several different classes under either Rule 23(b)(2) or 23(b)(3).
Plaintiffs' Rule 23(b)(2) certification request here doesn't satisfy this requirement. Indeed, defendants oppose plaintiffs' request for certification of a Rule 23(b)(2) class for this very reason. They argue that plaintiffs never identify the specific declaratory or injunctive relief they seek or the legal authority under which they seek it.
Plaintiffs respond to defendants' argument by explaining that "the precise form of this injunctive relief will take shape as the Court rules on the Parties' forthcoming dispositive motions and at trial." Doc. 1837 at 54. As one example, plaintiffs suggest that the court could order defendants to add a label to the EpiPen that recites no medical necessity exists to sell the device in a 2-Pak. Id. at 54 n.228. This vague, partially formulated request for injunctive relief doesn't meet the standard for certification under Rule 23(b)(2).
As our Circuit has explained, "Federal Rule of Civil Procedure 65(d) requires that injunctions be 'specific in terms' and 'describe in reasonable detail, and not by reference to the complaint or other document, the act or acts sought to be restrained.'" Monreal, 367 F.3d at 1236 (quoting Fed. R. Civ. P. 65(d)). Here, plaintiffs' Rule 23(b)(2) certification request fails to describe the injunctive relief they seek with sufficient detail and, as a consequence, the court cannot determine whether classwide injunctive relief is appropriate under Rule 23(b)(2). See id. at 1236 (affirming district court's refusal to certify at Rule 23(b)(2) class because plaintiffs didn't specify the type of injunctive relief they sought and the Circuit couldn't "conceive of an injunction that would satisfy [Rule 65(d)'s] requirements"); see also In re YRC Worldwide, Inc. ERISA Litig., No. 09-2593-JWL, 2011 WL 1303367, at *14 (D. Kan. Apr. 6, 2011) (denying certification under Rule 23(b)(2) when plaintiffs didn't "identify any particular injunctive relief with sufficient specificity to enable the court to 'see how it might satisfy Rule 65(d)'s constraints and thus conform with Rule 23(b)(2)'s requirement'" (quoting Shook, 543 F.3d at 605 n.4)).
In short, plaintiffs have failed to shoulder their burden to show that Rule 23(b)(2) class certification is warranted. They haven't provided sufficient information for the court to conclude that "final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." Fed. R. Civ. P. 23(b)(2). The court thus exercises its discretion and denies plaintiffs' request to certify a Rule 23(b)(2) class.
V. Class Definitions
Consistent with its analysis above, the court certifies the two following classes under Rule 23(b)(3):
1. Nationwide RICO Damages Class ("RICO Class"). All persons and entities in the United States who paid or provided reimbursement for some or all of the purchase price of Branded or authorized generic EpiPens for the purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries, at any time between August 24, 2011, and [specific date to be inserted when class notice is finalized].
2. State Antitrust Damages Class ("State Antitrust Class"). All persons and entities in the Antitrust States who paid or provided reimbursement for some or all of the purchase price of Branded EpiPens at any time between January 28, 2013, and [specific date to be inserted when class notice is finalized], for the purpose of consumption, and not resale, by themselves, their family member(s), insureds, plan participants, employees, or beneficiaries.
As described above, supra Part IV.A. & n.7, plaintiffs' original certification motion proposed that the class period would run "until the effects of Defendants' unlawful conduct cease." See, e.g., Doc. 1353 at 2-3. At oral argument, plaintiffs proposed to change this part of the class definitions to define the class period as ending on the date when class notice is given. The court adopts that change in the certified class definitions.
As discussed above, supra note 52, the court cannot certify a class asserting antitrust violations under the laws of states where no plaintiff resides. So, the court redefines the "Antitrust States" from the definition proposed by plaintiffs to remove those states that have no named plaintiff. The court thus defines the "Antitrust States" to include: Alabama, California, Florida, Hawaii, Illinois, Kansas, Maine, Michigan, Minnesota, Mississippi, Nebraska, Nevada, New Hampshire, New York, North Carolina, Tennessee, and Utah.
The court excludes the following groups from each of the classes:
(a) Defendants and their officers, directors, management, employees, subsidiaries, and affiliates;
(b) Government entities, other than government-funded employee benefit plans;
(c) Fully insured health plans (i.e., plans that purchased insurance that covered 100% of the plan's reimbursement obligations to its members);
(d) "Single flat co-pay" consumers who purchased EpiPens or generic EpiPens only via a fixed dollar co-payment that is the same for all covered devices, whether branded or generic (e.g., $20 for all branded and generic devices);
(e) Consumers who purchased or received EpiPens or authorized generic equivalents only through a Medicaid program;
(f) All persons or entities who purchased branded or generic EpiPens directly from defendants;
(g) The judges in this case and members of their immediate families;
(h) All third-party payors who own or otherwise function as a Pharmacy Benefit Manager or control an entity who functions as a Pharmacy Benefit Manager; and
(i) Individual consumers whose only purchases of an EpiPen occurred before March 13, 2014 (the Generic Start Date).
The court appoints the named plaintiffs proposed as representatives in plaintiffs' Consolidated Class Action Complaint as class representatives for these two classes.
VI. Appointment of Class Counsel
Rule 23(g)(1) requires that a court certifying a class must appoint class counsel. When making this appointment, the court "must consider: (i) the work counsel has done in identifying or investigating potential claims in the action; (ii) counsel's experience in handling class actions, other complex litigation, and the types of claims asserted in the action; (iii) counsel's knowledge of the applicable law; and (iv) the resources that counsel will commit to representing the class[.]" Fed. R. Civ. P. 23(g)(1)(A). The court may appoint an applicant as class counsel "only if the applicant is adequate" under Rule 23(g)(1)'s criteria and if the applicant will adhere to Rule 23(g)(4)'s requirement that the applicant "fairly and adequately represent the interests of the class." Fed. R. Civ. P. 23(g)(2).
Plaintiffs ask the court to appoint Warren T. Burns, Paul J. Geller, Lynn Lincoln Sarko, and Rex Sharp as class counsel. The court finds these attorneys have adequately prosecuted the interests of the classes since the court appointed them as co-lead counsel. Aside from Mylan's conclusory statement that it is inappropriate to have the same counsel represent plaintiffs with different theories of injury and competing damage claims, defendants do not challenge their appointment as class counsel. After considering Rule 23(g)'s factors, the court finds that appointing these attorneys as co-lead counsel is warranted here. Also, the court appoints one more lawyer to the group representing plaintiffs as class counsel—Elizabeth C. Pritzker. Ms. Pritzker has served as Chair of plaintiffs' Steering Committee since the court appointed her in that role. The court thus appoints Warren T. Burns, Paul J. Geller, Lynn Lincoln Sarko, Rex Sharp, and Elizabeth C. Pritzker as co-lead counsel for the class plaintiffs.
The court also appointed Eric Hochstradt, counsel for Sanofi, as co-lead counsel.
VII. Notice
When a court certifies a Rule 23(b)(3) class, Rule 23(c)(2)(B) provides "the court must direct to class members the best notice practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort." Because plaintiffs have not proposed a form or manner of notice under this rule, the court defers its direction of notice to the certified classes. The court orders plaintiffs to submit to the court a proposed plan for notice to the class members consistent with Rule 23(c)(2)(B) on or before March 31, 2020. The plan must include a proposed form of notice.
IT IS THEREFORE ORDERED BY THE COURT THAT the plaintiffs' Motion for Class Certification (Doc. 1353) is granted in part and denied in part. The court certifies a nationwide RICO damages class and a state antitrust damages class under Rule 23(b)(3), as set forth in this Order. The court denies the motion in all other respects.
IT IS FURTHER ORDERED THAT defendants' Motion for Leave to File a Sur-reply (Doc. 1574) is granted. The court orders defendants to file their proposed Sur-reply (Doc. 1574-1) as a separate docket entry. Also, the court orders plaintiffs to file their proposed Response to Defendants' Sur-reply (Doc. 1579-1) as a separate docket entry.
IT IS FURTHERED ORDERED THAT plaintiffs must submit a proposed plan for notice to the class members consistent with Rule 23(c)(2)(B) on or before March 31, 2020.
IT IS SO ORDERED.
Dated this 27th day of February, 2020, at Kansas City, Kansas.
s/ Daniel D. Crabtree
Daniel D. Crabtree
United States District Judge