United States District Court, E.D. Pennsylvania
June 30, 2015
IN RE WELLBUTRIN XL ANTITRUST LITIGATION THIS DOCUMENT RELATES TO INDIRECT PURCAHSER ACTION
This indirect purchaser class action involves claims that the defendants SmithKline Beecham Corporation d/b/a GlaxoSmithKline and GlaxoSmithKline plc (collectively, “GSK”) delayed the entry of generic versions of the drug Wellbutrin XL to the American market by entering into illegal agreements with generic drug companies to settle patent infringement lawsuits. In August 2011, the Court certified the Indirect Purchaser Plaintiff Class (“IPC”) under Federal Rule of Civil Procedure 23(b)(3). See In re Wellbutrin XL Antitrust Litig., 282 F.R.D. 126 (E.D. Pa. 2011).
GSK now moves to decertify the IPC based on a quartet of Third Circuit cases dealing with the requirement that Rule 23(b)(3) classes be ascertainable. See generally Byrd v. Aaron’s Inc., 784 F.3d 154 (3d Cir. 2015); Carrera v. Bayer Corp., 727 F.3d 300 (3d Cir. 2013); Hayes v. Wal-Mart Stores, Inc., 725 F.3d 349 (3d Cir. 2013); Marcus v. BMW of North America, LLC, 687 F.3d 583 (3d Cir. 2012). GSK argues that the IPC cannot satisfy this requirement, and that the IPC should therefore be decertified. In connection with GSK’s motion to decertify the IPC, GSK has filed a Daubert motion to exclude the opinions and testimony of one of the IPC’s experts, Dr. Meredith Rosenthal. The IPC also filed a Daubert motion to exclude the opinions and testimony of GSK’s expert, Dr. Bruce Strombom.
To satisfy the ascertainability requirement, a putative class must show that there is a reliable, administratively feasible mechanism that can identify which potential class members fall within the class definition. Byrd, 784 F.3d at 163. In this case, the IPC must show that it can identify (1) which entities paid some or all of the retail purchase price of Wellbutrin XL and later purchased its generic equivalent (“generic XL”), and (2) which individual consumers and entities paid some or all of the retail purchase price of generic XL. Individual consumers who made only flat co-payments for the generic drug are excluded from the class.
There is no dispute among the parties that third party payers (“TPPs”), such as health insurers and health and welfare benefit plans, may have been entities that paid some or all of the retail purchase price of Wellbutrin XL and/or generic XL. The potential class membership of pharmacy benefit managers (“PBMs”), on the other hand, is hotly contested. PBMs, which generally act as middlemen between TPPs and retail pharmacies, sometimes offer price discount guarantees or spread pricing arrangements on pharmaceutical drugs to their TPP customers. GSK argues that these pricing guarantees caused PBMs to pay for Wellbutrin XL or generic XL. Conversely, the IPC argues that such pricing arrangements constitute “off-transaction financial flows” that do not cause PBMs to pay for Wellbutrin XL or generic XL.
The IPC contends that it has a mechanism for identifying which individual consumers and PBMs (if necessary) are members of the class: utilizing pharmaceutical purchase records maintained by PBMs and retail pharmacies. The IPC must show that such records exist, can identify class members, and can be used in a reliable, administratively feasible fashion to satisfy the ascertainability inquiry.
The Court will grant GSK’s Daubert motion because Dr. Rosenthal’s methodology is not reliable. The Court will deny the IPC’s Daubert motion because it finds that Dr. Strombom is qualified to be an expert on this matter and that his methodology is sufficiently reliable. Finally, the Court will grant GSK’s motion to decertify the IPC because the IPC has not carried its burden of showing that the class is ascertainable.
An understanding of the roles of some of the major players in the retail pharmaceutical industry is necessary to analyze some of the ascertainability issues in this case. Many individual consumers obtain prescription drugs from retail pharmacies. If they are covered by a health insurance plan, consumers may share the cost of prescription drugs with their insurer or health plan. These entities are often referred to as “third party payers” (“TPPs”). GSK’s Daubert Mot. Ex. A, ¶¶ 11, 24 (“Rosenthal Decl.”); GSK’s Mot. to Decertify Ex. A, ¶¶ 48-53 (“Strombom Report”).
In many cases, TPPs employ a PBM as a sort of middleman between the TPP and the retail pharmacy. TPPs provide PBMs with information about individual consumers and the details of their insurance coverage. When a consumer goes to the pharmacy to obtain a prescription drug, he or she provides insurance information to the pharmacist. The pharmacy contacts the PBM with that information, and the PBM determines what price, if any, the consumer is responsible for. The PBM also often forwards the TPP’s portion of the retail price of the drug to the pharmacy. This is usually accomplished via an electronic, automated system. St. Phillip Decl. Ex. 6 at 8-9, Mar. 9, 2015.
II. Relevant Procedural History
The IPC claims that that GSK (along with former co-defendants Biovail Corporation, Biovail Laboratories, Inc., and Biovail Laboratories International SRL (collectively, “Biovail”), who have since settled) delayed the entry of generic versions of the drug Wellbutrin XL (“generic XL”) by entering into illegal agreements with generic drug companies to settle patent infringement lawsuits.
In August 2011, the Court certified the IPC as a Rule 23(b)(3) class after extensive briefing and several days of hearings. See Wellbutrin XL, 282 F.R.D. 126. The Court defined the IPC as follows:
(1) All persons or entities who purchased an AB-rated generic bioequivalent of Wellbutrin XL (“generic XL”) at any time during the “Class Period” (hereafter defined) in California, Florida, Nevada, New York, Tennessee and Wisconsin; and
(2) All entities that purchased 150 mg or 300 mg Wellbutrin XL before an AB-rated generic bioequivalent was available for such dosages AND purchased generic XL in the same state after generic XL became available in California, Florida, Nevada, New York, Tennessee and Wisconsin.
For purposes of the Class definition, persons or entities purchased Wellbutrin XL or generic XL if they paid some or all of the retail purchase price.
Excluded from the Class are “flat co-payers” meaning natural persons whose only purchases of generic XL were made pursuant to contracts with third party payers (“TPP”) whereby the amount paid by the natural person for generic XL was the same regardless of the retail purchase price.
The Class Period begins November 14, 2005 and ends on April 29, 2011.
Order, August 12, 2011 (Docket No. 354). The parties did not raise the ascertainability question in the previous certification proceedings, and the Court did not consider it when it certified the class. See generally Wellbutrin XL, 282 F.R.D. 126.
On September 22, 2014, GSK filed the pending motion to decertify the IPC. The IPC filed a motion to strike GSK’s motion shortly thereafter. Following a status conference in chambers with the parties’ counsel, the Court denied the IPC’s motion to strike. During the status conference, the IPC’s counsel indicated that it needed discovery to properly respond to GSK’s motion to decertify. The Court granted the IPC’s request for discovery, and issued a scheduling order allowing for fact and expert discovery on the motion to decertify and setting deadlines for the filing of any opposition briefs, reply briefs, and related Daubert motions. Scheduling Order, November 19, 2014 (Docket No. 517).
During the fact discovery period, the IPC filed a motion to compel against GSK, mostly seeking information about the TPP-PBM relationship. The Court denied this motion, as the information sought would be in the possession of PBMs and TPPs, not drug manufacturers like GSK. Order, December 10, 2014 (Docket No. 521). The Court also noted that the IPC had served subpoenas for documents and depositions to nine different PBMs regarding their purchase records for Wellbutrin XL and financial arrangements with members of the class. Id. Despite serving these subpoenas, the IPC did not introduce any such PBM documents or deposition testimony to the Court in support of its ascertainability arguments. Oral Ar. Tr. 57:6-60:2, May 29, 2015.
On May 29, 2015, the Court held oral argument on the pending motion to decertify and related Daubert motions.
III. The Ascertainability Record
A. Expert Witnesses
Both GSK and the IPC have introduced expert witnesses to support their arguments on the ascertainability issue. GSK relies on the opinions and conclusions of Dr. Bruce A. Strombom, while the IPC relies on Dr. Meredith Rosenthal and Paul DeBree.
All three experts provided reports and were deposed. A summary of the expert reports follows.
1. Dr. Bruce A. Strombom
Dr. Strombom generally opines that the IPC is not ascertainable because identifying class members would require extensive individualized analysis, and identifies three main theories as to why this is the case. First, Dr. Strombom claims that PBMs, depending on their contracts with TPPs, may have borne financial risk for the purchase of Wellbutrin XL or generic XL due to rebate and price guarantees. He states that determining when and whether a PBM bore financial risk stemming from these types of guarantees would depend on individual contracts with TPPs and would therefore require individualized inquiry.
Second, Dr. Strombom opines that TPPs may have passed on their portion of the increased price of Wellbutrin XL and/or generic XL by, for example, charging higher premiums the following year to consumers of health insurance. Third, Dr. Strombom states that in some cases, available records do not allow individual consumer class members to be ascertained.
a. The PBM Theory
Dr. Strombom highlights the presence of both rebate guarantee provisions and price discount guarantee provisions in PBM-TPP contracts. PBMs may negotiate with pharmaceutical manufacturers to obtain rebates that can be used to lower the price paid for prescription drugs. Similarly, PBMs may negotiate with retail pharmacies to obtain price discounts on purchases made by TPP plan members. Some PBM-TPP contracts include provisions in which the PBM guarantees minimum rebate amounts or minimum price discounts. Dr. Strombom opines that if a PBM is unable to obtain rebates or discounts at the levels guaranteed, the PBM would be responsible to make up the difference. This would cause a PBM to bear financial risk for the retail purchase price of Wellbutrin XL and/or generic XL, making it a member of the class. Strombom Report ¶¶ 16-25.
Dr. Strombom states that these arrangements present an ascertainability problem because they differ from PBM contract to PBM contract and would thus necessitate an individualized inquiry. Additionally, rebate payments from PBMs to TPPs are usually paid in a lump sum incorporating rebates for many different drugs. Determining which PBMs bore financial risk due to rebate payments and discount guarantees for Wellbutrin XL would therefore require individualized inquiries. Strombom Report ¶ 25.
b. The Premium Pass-On Theory
Dr. Strombom also opines that TPPs likely would not be harmed by any overcharge on the price of Wellbutrin XL or generic XL because they would pass on the increased price of the drug to their consumers in the form of higher premiums for the next year’s coverage. Dr. Strombom states that
[a]s health care costs increase for these plans and insurers, they must increase the amount of funds allocated to pay claims. This increase in funds is achieved through various types of adjustments, with the exact adjustments for any given TPP being highly individualized. The specific adjustments that are made, however, determine which parties bear the burden of the alleged overcharge. Thus, class membership is only ascertainable through analysis of these individualized adjustments.
Strombom Report ¶ 28. Dr. Strombom describes the various adjustments that insurance companies can make in order to pass on the cost of increased health care to consumers. Strombom Report ¶¶ 28-29.
Dr. Strombom claims that
there are no records readily identifying the class members who bore the alleged overcharges after pass-on is taken into account. Accordingly, class members can only be ascertained through extensive individualized inquiry into the specific adjustments made by TPPs to their healthcare coverage and premiums, as well as the specific parties (employers versus members) who are ultimately responsible for paying the health insurance premiums.
Strombom Report ¶ 46.
To support this theory, Dr. Strombom cites the fact that insurance companies generally increase insurance premiums to reflect the rising costs of health care, and observes that there is a positive correlation between the cost of health care and the cost of health insurance. Strombom Report ¶¶ 30-33.
Dr. Strombom also supports this theory with the fact that insurance companies use processes such as “experience rating” and “retrospective experience rating” – in which the insurance provider looks back at the costs of particular insured customers – in order to take costs into account in setting future premiums. Strombom Report ¶¶ 34-39.
Finally, Dr. Strombom discusses the Medical Loss Ratio (“MLR”) rule that was enacted as part of the Affordable Care Act (“ACA”). This rule requires that health care spending by insurers be within a certain ratio relative to health care premiums. If the ratio falls below a certain level, insurers are required to rebate to policyholders the amount of the premium sufficient to raise the MLR to the minimum level. Strombom Report ¶¶ 41-42.
Dr. Strombom argues that insurers potentially passed on the cost of Wellbutrin XL or generic XL overcharges because the overcharges could have raised the insurers’ MLR above the minimum level, and thus precluded the issuance of a rebate to the insurance customers. Strombom Report ¶¶ 41-46.
c. Ascertainability of Individual Consumers
Dr. Strombom provides three main reasons why individual consumers of generic XL cannot be ascertained without individual inquiry. All three relate to the fact that the class definition excludes individuals who paid only flat co-payments when they purchased generic XL.
First, Dr. Strombom points out that under their health insurance plans, what many consumers paid for generic XL depended on what pharmacy they obtained the drug from. At certain pharmacies, consumers had to pay only a flat co-payment, while at other pharmacies they may have had to pay a percentage of the retail cost of the drug. Without individualized inquiries of where these consumers purchased generic XL, Dr. Strombom opines, the class could not be ascertained. Strombom Report ¶¶ 48-53.
Second, Dr. Strombom cites consumer surveys that indicate that a majority of respondents could not correctly identify whether they paid flat co-payments or not. Dr. Strombom states that these surveys show that ascertaining the class by affidavit would not be a realistic way of identifying class members. Strombom Report ¶ 54.
Finally, Dr. Strombom analyzed data produced by one of the class representatives, Aetna Health of California, and concluded that the individual consumers could not be ascertained using this data. Although the Aetna data show which Aetna clients purchased generic XL during the class period and the price they paid for the drug, the data do not explicitly state whether those consumers paid a flat co-payment. Strombom Report ¶¶ 55-56.
Dr. Strombom made inferences from the data in an attempt to determine whether Aetna clients paid a flat co-payment and were thus excluded from the class. If the amount paid by the consumer matched the retail cost of the drug, Dr. Strombom characterized the payment as “fully out-of-pocket.” If the consumer paid nothing, Dr. Strombom categorized the purchase “fully covered.” If the consumer paid an integer dollar amount, like $10, Dr. Strombom categorized the payment as “fixed-dollar copay, ” provided that the payment did not also equal a fixed percentage evenly divisible by 5% of the total drug cost (e.g., 20% of the retail price). If the consumer paid a dollar amount equal to a fixed percentage of the total drug cost (e.g., 20%), Dr. Strombom characterized the payment as “coinsurance.” If a payment was consistent with either “fixed-dollar copay” or “coinsurance, ” Dr. Strombom categorized it as “copay or coinsurance.” If the payment fit none of these categories, Dr. Strombom categorized it as “ambiguous.” Strombom Report ¶ 57.
Dr. Strombom used these categories to determine whether a payment qualified a consumer for class membership. Payments in the “fixed-dollar copay” and “fully covered” categories did not qualify consumers to be in the class. Payments classified as “fully out-of-pocket” and “coinsurance” qualified consumers to be in the class. Dr. Strombom opined that payments classified as “copay or coinsurance” and “ambiguous” did not have clear implications for class membership, and listed them as “uncertain.” Strombom Report ¶ 58.
Dr. Strombom then analyzed the Aetna data, and grouped the 15, 167 Aetna clients who made at least one purchase of generic XL based on the types of payments they made. He found that 13, 417 consumers made only non-qualifying payments and should be excluded from the class (“Category A, ” 88.5%); 1, 075 consumers made qualifying payments and should be included in the class (“Category B, ” 7.1%); 336 consumers made only one qualifying payment and also made at least one non-qualifying or uncertain payment which made their status uncertain (“Category C, ” 2.2%); and 339 consumers made no qualifying payments and had at least one uncertain payment (“Category D, ” 2.2%). Strombom Report ¶ 62.
Dr. Strombom conducted a similar analysis of data from OptumHealth, which is publicly available. Of the 23, 463 consumers who made at least one generic purchase, 14, 626 (62.3%) belonged in Category A; 6, 356 (27.1%) belonged in Category B; 993 (4.2%) belonged in Category C; and 1, 488 (6.3%) belonged in Category D. Strombom Report ¶ 64.
2. Dr. Meredith Rosenthal
Dr. Rosenthal provides several opinions related to the ascertainability inquiry. First, she opines that class membership is a function of who pays for the drug at the time of the transaction. Given that, she argues, rebate and price discount guarantees do not make PBMs potential class members. Second, she opines that there are sufficient records in the pharmaceutical industry to identify class members. Dr. Rosenthal also criticizes Dr. Strombom’s analysis of the Aetna and OptumHealth data. Finally, she opines that consumers who could not be identified by pharmaceutical records could be identified by other means.
Dr. Rosenthal states that “Class membership is a function of who pays for the drug at the time of the transaction, ” and offers three reasons for this limitation. GSK’s Daubert Mot. Ex. A, ¶ 4 (“Rosenthal Decl.”). First, Dr. Rosenthal states that she was “instructed by counsel to work from the theory” that class membership is limited to those who paid for the drug at the time of the transaction. Rosenthal Decl. ¶ 4. Second, Dr. Rosenthal states that this limitation is “entirely consistent with the plain meaning of the word ‘paid’ and requires no new theory.” Rosenthal Decl. ¶ 4. Finally, Dr. Rosenthal based her limitation of class membership on the Court’s choice of law analysis, which concluded that “the law of the place of purchase” would govern the transactions. Rosenthal Decl. ¶ 13-14 (citing Wellbutrin XL, 282 F.R.D. at 136).
Dr. Rosenthal opines that rebate and price discount guarantees should not “weigh into the ascertainment of Class members” because they are not computed or allocated at the time of the transaction. Rosenthal Decl. ¶ 25. She argues that Dr. Strombom’s consideration of rebate and price guarantees – what Dr. Rosenthal characterizes as “off-transaction adjustments” – challenges the plain meanings of the words “paid” and “purchase” in the class definition. Rosenthal Decl. ¶ 15. Dr. Rosenthal extends this criticism to Dr. Strombom’s premium pass-on theory, saying it “brings in a highly speculative analysis of long-run economic incidence and incorrectly render absurd the notion of ‘paid.’” Rosenthal Decl. ¶ 20.
Dr. Rosenthal also criticizes Dr. Strombom’s analysis of the Aetna and OptumHealth data. She argues that the consumers in his Category C should be included in the class without any further individualized inquiry because they made at least one qualifying purchase. She also states that consumers in Dr. Strombom’s Category D should be excluded from the class without the need for individualized inquiry because no qualifying payments could be demonstrated. Rosenthal Decl. ¶¶ 18-19, 23.
Finally, Dr. Rosenthal opines that consumers who paid cash for generic XL (that is, those without some form of health insurance) can be ascertained using pharmacy reports, receipts, and prescription bottles. Rosenthal Decl. ¶ 28.
3. Paul DeBree
In his declaration, Mr. DeBree offers two general opinions. First, Mr. DeBree states that those who paid for Wellbutrin XL and generic XL can be easily identified. Second, Mr. DeBree concludes that PBMs did not pay a portion of the retail purchase price of Wellbutrin XL and generic XL, and are therefore not potential members of the IPC.
Mr. DeBree opines that there are readily accessible records that can be used to identify every entity and consumer that paid some or all of the purchase price of Wellbutrin XL and generic XL. Mr. DeBree states that pharmacies record consumers’ prescription and identifying information, and that PBMs necessarily maintain consumer identifying and insurance plan information to facilitate the claim adjudication process with pharmacies. Mr. DeBree believes that such information would be accessible and sufficient to identify all potential class members. During his deposition, however, Mr. DeBree stated that PBMs would not willingly give out whatever pharmaceutical information they maintain, undercutting the statement in his report that such information would be accessible. St. Phillip Decl. Ex. 1, ¶¶ 21, 24-25, Mar. 9, 2015 (“DeBree Decl.”); DeBree Dep. 287:24-288:6.
Mr. DeBree opines that PBMs do not pay a part of the retail purchase price of Wellbutrin XL and/or generic XL because rebate and discount guarantees are calculated in the aggregate over time and concern only the business relationship between the PBM and the TPP. PBMs do not pay a portion of the retail cost of these drugs because such guarantees are directed only towards the PBM’s contractual obligations to TPPs, and create no obligations to consumers to purchase drugs. During his deposition, however, Mr. DeBree conceded that PBMs may bill their TPP clients for pharmaceutical purchases at a lower rate than the rate PBMs paid to retail pharmacies for the purchase of pharmaceutical drugs. DeBree Dep. 79:15-80:6. DeBree Decl. ¶¶ 34, 41-42.
Mr. DeBree argues that PBMs are not potential class members because they are not insurers with an obligation to consumers to pay pharmacies for drug purchases. According to Mr. DeBree, although PBMs do pay pharmacies directly, they only make such payments on behalf of TPPs and use only TPP funds to do so. According to Mr. DeBree, if PBMs did pay a portion of the retail price of pharmaceuticals, they would be operating illegally as unregulated insurance companies. DeBree Decl. ¶¶ 31, 33, 36-38.
B. The IPC’s Other Submissions
The IPC has submitted several other pieces of evidence in its attempt to show that there is a reliable, administratively feasible mechanism for identifying class members. In addition to the Aetna purchase data analyzed by Dr. Strombom, the IPC has submitted pharmaceutical purchase records from two health and welfare plans. Defs.’ Mem. in Opp. to the IPC’s Mot. for Class Certification Exs. 9 & 13, Jan. 28, 2011. The IPC has also submitted the proposed trial plan introduced during the initial class certification proceedings, as well as several documents from the Biovail settlement and related claims process. St. Phillip Decl. Exs. 10-11, Mar. 9, 2015. Finally, the IPC has submitted extensive forms from the National Council for Prescription Drug Programs, including data dictionaries, standards information, and reference guides. St. Phillip Decl. Exs. 16-24, Mar. 9, 2015.
C. GSK’s Other Submissions
GSK has also submitted evidence beyond its expert witness in support of its motion to decertify the IPC. GSK submitted declarations from employees of four different PBMs. These employees have personal knowledge about their PBM employers’ financial relationships with TPP customers. All four PBM employees state, in one form or another, that their PBM employers may absorb a portion of the aggregate cost of their clients’ prescription drug purchases if price discount guarantees are not met. GSK has also submitted a contract between a PBM, Express Scripts, Inc., and a TPP, Plumbers & Pipefitters L.U. No. 572 Health & Welfare Fund. GSK’s Reply in Supp. of its Mot. to Decertify Exs. A-D, I-J.
IV. Daubert Motions
A. Legal Standard
The party offering an expert witness must establish by a preponderance of the evidence the qualifications of the expert and the expert opinion’s compliance with Federal Rule of Evidence 702. See Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 590-95 (1993). To satisfy the requirements of Rule 702, a proffered expert must be qualified to express an expert opinion, the proffered expert opinion must be reliable; and the proffered expert’s testimony must assist the trier of fact and “fit” the facts of the case. In re TMI Litigation, 193 F.3d 613, 664-65 (3d Cir. 1999).
To be qualified as an expert witness, one must “possess specialized expertise.” Calhoun v. Yamaha Motor Corp., U.S.A., 350 F.3d 316, 321 (3d Cir. 2003). This requirement has been interpreted liberally, and a “broad range of knowledge, skills, and training qualify an expert as such.” Id. At a minimum, a proffered expert witness “must possess skill or knowledge greater than the average layman.” Elcock v. Kmart Corp., 233 F.3d 734, 741 (3d Cir. 2000) (quoting Waldorf v. Shuta, 142 F.3d 601, 625 (3d Cir. 1998)).
To be reliable, a proffered expert opinion must be based on the “methods and procedures of science rather than on objective belief or unsupported speculation; the expert must have good grounds for his or her belief.” Calhoun, 350 F.3d at 321 (quoting In re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 742 (3d Cir. 1994)). Several factors guide the reliability inquiry:
(1) whether a method consists of a testable hypothesis; (2) whether the method has been subject to peer review; (3) the known or potential rate of error; (4) the existence and maintenance of standards controlling the technique’s operation; (5) whether the method is generally accepted; (6) the relationship of the technique to methods which have been established to be reliable; (7) the qualifications of the expert witness testifying based on the methodology; and (8) the non-judicial uses to which the method has been put.
Id. The proffered expert’s methodology does not have to be perfect; a court should only exclude the evidence if a flaw is large enough that the expert lacks good grounds for his or her conclusions. In re TMI Litig., 193 F.3d at 665. However, a court must examine the expert’s conclusions to determine whether they “reliably flow from the facts known to the expert and the methodology used.” Id. at 665-66 (quoting Heller v. Shaw Indus., Inc., 167 F.3d 146, 153 (3d Cir. 1999)). If the gap between the data and the opinion proffered is too large, a court should exclude the opinion. Id. at 666.
Finally, in asking whether an expert’s proposed testimony “fits, ” courts ask whether the expert testimony proffered “is sufficiently tied to the facts of the case that it will aid [a finder of fact] in resolving a factual dispute.” U.S. v. Schiff, 602 F.3d 152, 173 (3d Cir. 2010) (quoting Daubert, 509 U.S. at 591).
B. Dr. Bruce Strombom
The IPC claims that Dr. Strombom is not qualified to offer his PBM theory or data analysis opinions, and that these opinions should therefore be excluded. Additionally, the IPC claims that Dr. Strombom’s analysis of pharmaceutical purchase data is flawed. Dr. Strombom’s opinions and testimony will not be excluded because he is sufficiently qualified to be an expert in this case, and his data analysis, though flawed, is reliable enough for the Daubert inquiry.
1. Dr. Strombom’s Qualifications
Dr. Strombom has a Ph.D. in economics from the University of California, Irvine. His graduate research focused on consumer choice in health markets, and his dissertation examined the factors that influence employees’ choices of alternative health plans. GSK’s Mot. to Decertify Ex. A, ¶ 3 (“Strombom Report”).
Dr. Strombom has professional experience in the healthcare industry. He has evaluated the pricing and claims payment practices of health insurers, has reviewed the merger of hospitals, and has conducted economic analyses of healthcare markets. Additionally, Dr. Strombom has conducted economic analyses evaluating the appropriateness of class treatment in cases involving health insurance and pharmaceuticals. Strombom Report ¶ 4.
Dr. Strombom is qualified to give expert opinions because he has years of experience analyzing data in the healthcare field. His qualifications show that he knows more about the healthcare field in general, and the pharmaceutical field specifically, than the average layman.
The IPC argues that Dr. Strombom’s qualifications are not relevant to this case, in that he has “no specialized knowledge about the pharmacy benefit manager or pharmaceutical industry.” IPC’s Daubert Mot. 16. The IPC argues that Dr.
Strombom has never worked or consulted for a pharmacy, a managed care company, or a PBM. Finally, the IPC claims that Dr. Strombom never analyzed, researched, or opined on pharmaceutical data or pharmaceutical platforms until this litigation. See In re TMI Litig., 193 F.3d at 680 (excluding expert testimony because the proffered expert’s only knowledge on the issue was obtained by reviewing literature in connection with his retention as an expert in the case).
Despite the IPC’s claims, Dr. Strombom has in fact analyzed pharmaceutical data before this litigation. In his deposition, he testified that he has used the same publicly available OptumHealth pharmaceutical data in previous assignments. Strombom Dep. 98:21-99:13.
Additionally, despite the fact that he has never worked directly with a PBM, Dr. Strombom’s economics education and his experience in the healthcare field give him the capability to interpret PBM-TPP contracts and determine what risks, if any, are present due to any rebate or price guarantees contained therein. Dr. Strombom has much more skill than the average layperson in data analysis and much more knowledge than the average person regarding the healthcare industry. He is qualified to serve as an expert for purposes of ascertainability.
2. Dr. Strombom’s Analysis of Pharmaceutical Purchase Data
The IPC challenges Dr. Strombom’s analysis of Aetna and OptumHealth data and his conclusions that such data are not sufficient to identify individual consumer class members. Although flawed, Dr. Strombom’s data analysis and conclusions are reliable enough that his opinions and testimony will not be excluded.
The first problem with Dr. Strombom’s analysis is his characterization of “ambiguous” or “uncertain” payments as they relate to class membership. Dr. Strombom opined that if a payment was not consistent with a flat co-payment, a coinsurance payment, or a fully out-of-pocket payment on its face, individualized inquiries would be required to determine the type of payment and thus eligibility for the class. The type of payment, however, does not need to be determined to identify class members; the payment just must have not been a flat co-payment. As long as the consumer’s payment was not a flat co-payment, the consumer would be a member of the class.
Another flaw in Dr. Strombom’s analysis is that he characterizes Category C, which includes consumers who made only one qualifying payment and also made at least one non-qualifying or uncertain payment, as a category in which individualized inquiry is necessary to determine class membership. Membership in the class does not require that every payment not be a flat co-payment, however; a consumer would be a member of the class even if he or she made only one qualifying purchase.
These flaws in the way Dr. Strombom categorizes payments and consumers are not so serious that Dr. Strombom’s opinions on the ascertainability of individual consumers should be excluded. Dr. Strombom’s underlying methodology – analyzing amounts paid by consumers in an attempt to determine whether a given consumer paid only flat co-payments for generic XL – appears to the Court to be a sound way of determining the class membership of individual consumers. Any flaws in Dr. Strombom’s analysis affect only the number of consumers in a given category; they do not undermine Dr. Strombom’s conclusion that there are cases where the Aetna purchase records he analyzed did not provide enough information to ascertain whether or not an individual consumer was a member of the class. His methodology is therefore reliable enough to satisfy the Daubert inquiry.
The IPC makes several other criticisms of Dr. Strombom’s analysis, none persuasive. First, the IPC states that Dr. Strombom’s analysis of the Aetna data was flawed because he analyzed only the data that Aetna produced in this litigation without determining whether Aetna possesses additional data that could help in the ascertainability inquiry. Dr. Strombom analyzed additional data from OptumHealth, however, and came to similar conclusions. This alleged flaw is not large enough to exclude Dr. Strombom’s opinion on the basis of reliability.
The IPC also claims that Dr. Strombom’s analysis is unreliable because he did not rely on data standards set by the National Council for Prescription Drug Programs (“NCPDP”), a standards setting organization. The NCPDP sets pharmaceutical data standards, but does not actively collect or store any pharmaceutical data. The IPC contends that the standards set by the NCPDP include fields to record the amount a consumer pays in co-payments and coinsurance, which would therefore enable class members to be ascertained. In his rebuttal report, Dr. Strombom analyzes data from GSK’s relationship with PBMs, as well as information from the NCPDP, and concludes that NCPDP standards have not been implemented uniformly across the pharmaceutical industry. Strombom Rebuttal Report ¶ 45. The fact that the standards may include fields that could help the ascertainability inquiry does not mean that those fields were actually utilized in generic XL transactions. This alleged flaw in Dr. Strombom’s methodology does not warrant excluding his testimony.
Dr. Strombom has the necessary qualifications to serve as an expert for purposes of the ascertainability inquiry. Although flawed, his analysis is reliable enough for him to offer his expert opinions. The IPC’s Daubert motion is therefore denied.
C. Dr. Meredith Rosenthal
GSK seeks to exclude Dr. Rosenthal’s opinions and conclusions limiting class membership to those who paid for Wellbutrin XL or generic XL at the time of the transaction. These opinions and conclusions will be excluded because Dr. Rosenthal did not support them with independent economic analyses.
Dr. Rosenthal is highly qualified to give opinions on these matters. She is a Professor of Health Economics and Policy at the Harvard School of Public Health and an Academic Affiliate of Greylock McKinnon Associates, a consulting and litigation support firm. Rosenthal Decl. ¶ 1. GSK does not dispute Dr. Rosenthal’s qualifications, and the Court relied on her opinions when it originally certified the IPC.
In her Rebuttal Declaration, Dr. Rosenthal opines that “Class membership is a function of who pays for the drug at the time of the transaction.” GSK’s Daubert Mot. Ex. A, ¶ 4 (“Rosenthal Decl.”). Under this theory, according to Dr. Rosenthal, price discount guarantees would not cause PBMs to pay a portion of the retail purchase price, and PBMs would therefore not be class members. Rosenthal Decl. ¶¶ 15, 24-26.
Dr. Rosenthal cites three reasons why she limited membership of the class to those that paid at the time of the retail transaction, none of which provide adequate support for the limitation or are based on sound economic reasoning. First, Dr. Rosenthal states that she was “instructed by counsel to work from the theory that Class membership is a function of who pays for the drug at the time of the transaction.” Rosenthal Decl. ¶ 4. An instruction from counsel is not a sound basis on which to draw an economic conclusion.
Second, Dr. Rosenthal states that the instruction of counsel is “entirely consistent with the plain meaning of the word ‘paid’ and requires no new theory.” Rosenthal Decl. ¶ 4. Limiting class membership to those who paid at the time of the transaction, however, is not consistent with the plain meaning of “paid some or all of the retail purchase price, ” especially given the realities of the pharmaceutical industry. The class definition does not limit class membership to those who paid at the time of the transaction; it limits class membership to those who paid a portion of the retail purchase price. Indeed, in many cases, TPPs such as health insurers do not pay their portion of the retail purchase price “at the time of the transaction” – a PBM pays for them. Dr. Rosenthal’s reading of the word paid would effectively exclude many TPPs, who the IPC claims are one of two core groups of class members.
Dr. Rosenthal appeared to concede as much during her deposition. She testified that in using the phrase “at the time of the transaction, ” she did not mean literally at the time of the transaction. Rosenthal Dep. 52:15-53:19, Feb. 12, 2015 (“Rosenthal Ascertainability Dep.”). She stated that “if it’s 24, 72 hours later that the TPP funds flow to the PBM, that’s not what I’m talking about. I’m just talking about the specific transaction.” Rosenthal Ascertainability Dep. 52:24-53:1. She further clarified that a payment had to be “related to the filling of that prescription” for it to be a payment qualifying an entity for class membership. Rosenthal Ascertainability Dep. 53:18-19.
Dr. Rosenthal’s deposition testimony seems to indicate that when she used the phrase “at the time of the transaction” repeatedly throughout her declaration, she really meant to say “related to the transaction.” Dr. Rosenthal never explains why price guarantee provisions are not “related to the transaction;” she merely concludes that these are “off-transaction financial flows.” There is an argument that any payment from a PBM to a TPP due to a price guarantee would be “related to the transaction” – the PBM would be responsible for some of the overcharge stemming from a purchase of Wellbutrin XL.
In making this argument, Dr. Rosenthal is not undertaking an economic analysis. Rather, she is attempting to interpret the meaning of words within a class definition – a task that is the duty of the Court, not an economic expert. The plain meaning of the class definition is not a sound basis upon which Dr. Rosenthal can rely in limiting class membership to those who paid at the time of the transaction.
Finally, Dr. Rosenthal based her limitation on class membership on the Court’s choice of law analysis, which concluded that “the law of the place of purchase” would govern the transactions. Wellbutrin XL, 282 F.R.D. at 136. The choice of law inquiry is markedly different from the economic inquiry necessary to determine who “paid” for a given drug. Any reliance by Dr. Rosenthal on the Court’s choice of law analysis for this purpose is misplaced.
Dr. Rosenthal’s conclusions limiting class membership to those who “paid at the time of the transaction” are excluded because her methodology is severely flawed and her conclusions inconsistent and ambiguous. Rather than base her conclusions on economic analysis, Dr. Rosenthal relies on instructions from counsel and the Court’s choice of law analysis.
V. Motion to Decertify
A. Legal Standard
Rule 23(c)(1) requires district courts “to reassess their class rulings as the case develops.” Barnes v. American Tobacco Co., 161 F.3d 127, 140 (3d Cir. 1998). An order granting class certification “may be altered or amended before final judgment.” Fed.R.Civ.P. 23(c)(1)(C).
The party proposing class-action certification “bears the burden of affirmatively demonstrating by a preponderance of the evidence her compliance with the requirements of Rule 23.” Byrd, 784 F.3d at 163. District courts evaluating the propriety of class certification are “obligated to probe behind the pleadings when necessary and conduct a ‘rigorous analysis’ in order to determine whether Rule 23 certification requirements are satisfied.” Id. The party seeking certification of a Rule 23(b)(3) class must “prove by a preponderance of the evidence that the class is ascertainable.” Id.
To satisfy the ascertainability requirement, a plaintiff must show that: “(1) the class is ‘defined with reference to objective criteria’; and (2) there is a ‘reliable and administratively feasible mechanism for determining whether putative class members fall within this definition.’” Id. (quoting Carrera, 727 F.3d at 355). A plaintiff does not need to be able to identify all class members at the class certification stage; rather, “a plaintiff need only show that ‘class members can be identified.’” Id. (emphasis in original) (quoting Carrera, 727 F.3d at 308 n.2). A district court “should ensure that class members can be identified ‘without extensive and individualized fact-finding or mini-trials.’” Carrera, 727 F.3d at 307 (quoting Marcus, 687 F.3d at 593). A party’s assurance “that it intends or plans to meet the requirements of Rule 23 is insufficient. A plaintiff may not merely propose a method of ascertaining a class without any evidentiary support that the method will be successful.” Id. at 306 (internal quotation marks and citation omitted). Finally, a defendant must be able to “test the reliability of the evidence submitted to prove class membership.” Id. at 307.
Ascertainability should not be conflated with other Rule 23 requirements, such as the predominance requirement of Rule 23(b)(3). The ascertainability and predominance requirements are distinct, as the “‘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.’” Byrd, 784 F.3d at 164 (quoting Grandalski v. Quest Diagnostics Inc., 767 F.3d 175, 184 (3d Cir. 2014)).
GSK’s motion to decertify the IPC is granted because the IPC has not shown by a preponderance of the evidence that the class is ascertainable. Specifically, the IPC has not shown that it can ascertain which PMBs, if any, are members of the class, and which individual consumers of generic XL are members of the class.
1. The Potential Class Membership of PBMs
The IPC argues that PBMs are not potential class members, and that it therefore does not need to show that it can ascertain which PBMs are class members and which are not. PBMs are potential class members because they may have paid a portion of the retail purchase price of Wellbutrin XL or generic XL via so-called “spread pricing arrangements” or “price discount guarantees.” An entity is a member of the IPC if it paid some or all of the retail purchase price of generic XL during the class period.
In some contracts with TPPs, PBMs guarantee a certain price discount for prescription drug purchases. GSK’s Reply in Supp. of its Mot. to Decertify Exs. A-D. For example, OptumRx, a large PBM, may make guarantees to its TPP clients that it will obtain a certain level of discounts for prescription drug purchases. GSK’s Reply in Supp. of its Mot. to Decertify Ex. D. If it cannot obtain those discounts from pharmacies, OptumRx is responsible to the TPP for the difference between the guaranteed discount and the actual discount obtained. Id. In this event, it is possible that a PBM will not be fully reimbursed by the TPP for the prescription drug purchases of a TPP’s members. The PBM would, therefore, have paid a portion of the retail purchase price of generic XL.
The IPC does not dispute that such spread pricing arrangements or price discount guarantees exist, but rather characterizes them as “off-transaction financial flows” that do not take place when the retail transaction occurs. In her deposition, Dr. Rosenthal acknowledged that due to price discount guarantees,
it may be possible that with some contract I’m out of sync and I’m losing money on all my prescriptions, and then that, whatever that gap is, the 1 or 2 percent gap, that’s the percentage of the overcharge that the PBM is bearing in that kind of negative risk.
Rosenthal Ascertainability Dep. 75:14-20. The IPC’s arguments are not persuasive on this point. Neither the IPC nor Dr. Rosenthal explain why only the parties who paid at the time of the transaction should be included in the class, and Dr. Rosenthal conceded in her deposition that such a limitation, taken literally, would prevent most TPPs from being class members. Rosenthal Ascertainability Dep. 52:15-53:19. Indeed, in many cases such a limitation would cause PBMs to be included in the class and TPPs to be excluded, as PBMs are often the entities that pay pharmacies “at the time of the transaction.”
The IPC relies on In re Nexium (Esomeprazole) Antitrust Litigation, 297 F.R.D. 168, 179 (D. Mass. 2013), aff’d 777 F.3d 9 (1st Cir. 2015), to support its contention that PBMs are not a part of the class. The district court in Nexium, however, expressly excluded PBMs from the class at the suggestion of the named plaintiff without any discussion. The First Circuit did not mention the class membership of PBMs on appeal. See generally, In re Nexium Antitrust Litig., 777 F.3d 9 (1st Cir. 2015).
The IPC has also offered to include an express exclusion of PBMs in the class definition in this case. IPC’s Opp. to Mot. to Decertify 14 n.28. Although this suggestion would negate the PBM ascertainability issue, it raises potential predominance concerns. The Supreme Court has implied, but not held, that a putative class must show that “damages are susceptible of measurement across the entire class” to demonstrate Rule 23(b)(3) predominance. See Comcast Corp. v. Behrend, 133 S.Ct. 1426, 1433 (2013); see also 7AA Charles Alan Wright et al., Federal Practice and Procedure § 1778 (3d ed. 2015) (stating that Comcast “calls into question [the] long-established understanding” that damages need not be susceptible to common proof); Alex Parkinson, Comment, Comcast Corp v. Behrend and Chaos on the Ground, 81 U. Chi. L. Rev. 1213, 1213-14 (2014) (highlighting the lack of clarity in the wake of Comcast with regards to the role of damages calculations in the Rule 23(b)(3) predominance inquiry).
The Comcast Court held that at the class certification stage, a plaintiff’s damages model must measure only damages attributable to the plaintiff’s theory of liability. Comcast, 133 S.Ct. at 1433. This case does not present what this Court refers to as a “pure” Comcast problem. The IPC’s damages model measures only damages attributable to its theory of liability – that GSK sought to exclude generic XL from the market by entering into settlement agreements with generic companies in patent litigation. The Comcast Court’s implication, however, that a putative class must show that “damages are susceptible of measurement across the entire class” could cause issues for the IPC if PBMs are excluded from the class.
Professor Rosenthal’s damages methodology measures damages in the aggregate across the entire IPC, and would include any damages suffered by PBMs as a result of spread pricing arrangements or price discount guarantees. Oral Arg. Tr. 71:10-20, April 29, 2011; Rosenthal Ascertainability Dep. 160:12-161:2. Professor Rosenthal’s damages methodology does not allocate damages to individual class members, and could not be used to figure out the overcharge suffered by an individual class member. Oral Arg. Tr. 75:20-77:2, April 29, 2011. Professor Rosenthal acknowledged that her model could not currently deduct any damages suffered by PBMs from the damages calculation. Rosenthal Ascertainability Dep. 160:16-20.
It is unclear whether the IPC has shown that if PBMs are excluded, “damages are susceptible to measurement across the entire class” – the IPC’s current damages model would potentially include damages suffered by non-class members, and may therefore overstate the amount of damages suffered by the IPC. The Court does not have enough information at this point to know whether such an overstatement would be significant or relatively minor; indeed, the Court does not know if such an overstatement even exists. Cf. Comcast, 133 S.Ct. at 1433 (stating that at the class certification stage, damages “[c]alculations need not be exact”).
The Court therefore declines to include an express exclusion of PBMs in the class definition. Such an exclusion would potentially create as many problems for the certification of the IPC as it would solve. Additionally, such an exclusion would do nothing to solve the IPC’s problems in showing that individual consumers can be ascertained, which the Court will discuss below.
Price discount guarantees made by PBMs to TPPs potentially caused PBMs to pay some or all of the retail purchase price of Wellbutrin XL or generic XL. If a PBM did pay a part of the retail purchase price, it would be a member of the IPC.
The IPC has not shown by a preponderance of the evidence that there is a reliable and administratively feasible mechanism for determining which PBMs and individual consumers are members of the class. The IPC’s evidence on ascertainability barely goes further than repeated assurances that showing ascertainability in pharmaceutical cases is not difficult and that there are extensive purchase records in the pharmaceutical industry that could be used to ascertain whether individual consumers and PBMs are members of the class. Oral Ar. Tr. 7:7-8:17, 9:20-10:6, May 29, 2015. The IPC has not introduced evidence showing that those assertions are correct, and the Third Circuit has warned district courts against merely relying on a party’s assurance “that it intends or plans to meet the requirements of Rule 23 . . . .” Carrera, 727 F.3d at 306.
The IPC claims that a reliable, administratively feasible method exists for identifying which PBMs and consumers are in the class: utilizing purchase records. The IPC, however, has put forward scant evidence to support that claim. The IPC’s ascertainability experts, Dr. Meredith Rosenthal and Paul DeBree, both claim that there are records at the retail pharmacy and PBM level that can be used to identify class members. DeBree Decl. ¶¶ 27-28; Rosenthal Decl. ¶¶ 11-12. Additionally, the IPC has introduced purchase data from several health plans. Defs.’ Mem. in Opp. to the IPC’s Mot. for Class Certification Exs. 9 & 13, Jan. 28, 2011. Neither expert, however, examined or analyzed these pharmaceutical records, or the Aetna data analyzed by Dr. Strombom, to show that they could be used to ascertain PBMs and individual consumers. The Court is not persuaded by these experts’ conclusory statements.
Even if it were established that such records exist, the IPC has not introduced any evidence showing that such records are obtainable or can be used in an administratively feasible fashion to ascertain class members. The IPC’s own expert testified that it could be difficult to obtain purchase data from PBMs. DeBree Dep. 286:22-288:16. Indeed, the IPC served subpoenas on several PBMs during the recent discovery period, but did not obtain any records from those PBMs. This heightens the Court’s concern that such pharmaceutical records may not be obtainable for use in the ascertainability inquiry.
Nor has the IPC provided any evidence that if it did obtain PBM and retail pharmacy records, it could utilize those records to ascertain the class in an administratively feasible manner. There are thousands of PBMs and retail pharmacies; the IPC has not produced any evidence showing that it could synthesize records from these disparate entities and use them to ascertain PBMs and individual consumers in a reliable and administratively feasible manner. Additionally, the IPC has not shown that such records could identify which consumers paid flat co-payments and which did not.
The IPC also points to the existence of data standards set by the National Council for Prescription Drug Programs (“NCPDP”) as evidence that pharmaceutical records exist that could be used to ascertain consumer class members. The NCPDP sets pharmaceutical data standards, but does not actively collect or store any pharmaceutical data. The IPC contends that the standards set by the NCPDP include data fields for both co-payments and coinsurance, which would enable the IPC to determine which consumers paid flat co-payments.
It is not clear, however, that the players in the pharmaceutical industry utilize all of the fields set out in NCPDP standards. Various surveys conducted by the NCPDP indicate that a substantial portion of the industry is not fully compliant with NCPDP standards. Strombom Rebuttal Report ¶ 45. The fact that NCPDP standards include data fields that could be used to ascertain individual class members does not automatically lead to the conclusion that such fields were utilized by members of the pharmaceutical industry, or that such records would be available to aid in the ascertainability inquiry.
Finally, the IPC argues that Dr. Strombom’s analysis of the Aetna data shows that such data could be used to ascertain whether almost 98% of Aetna customers are in the class or not. Given the flaws in Dr. Strombom’s analysis identified by the Court, however, the Court is unsure what percentage of Aetna customers Dr. Strombom’s analysis could successfully ascertain. This evidence falls short of convincing the Court that such data could be used to ascertain individual class members.
The Court emphasizes that its decision on the ascertainability of the IPC is made on the record before it. The Court does not hold a view regarding whether indirect purchaser classes in other cases involving pharmaceutical purchases are ascertainable or not. The Court’s holding is limited to the record before it in this case.
The IPC has not carried its burden of “affirmatively demonstrating by a preponderance of the evidence” that there is a reliable, administratively feasible method of ascertaining the class. Byrd, 784 F.3d at 163; see also Carrera, 727 F.3d at 306 (“A plaintiff may not merely propose a method of ascertaining a class without any evidentiary support that the method will be successful.”). The IPC’s evidence in support of ascertainability consists mainly of conclusory statements by its experts that records exist that could be used to ascertain the class and the existence of NCPDP standards. This evidence is not enough to show by a preponderance of the evidence that the class is ascertainable. GSK’s motion to decertify the Indirect Purchaser Plaintiff Class is therefore granted. See Vista Healthplan, Inc. v. Cephalon, Inc., 2015 WL 3623005, at *5-*13 (E.D. Pa. June 10, 2015).
An appropriate order shall issue.