MEMORANDUM AND ORDER

DANIEL D. CRABTREE, District Judge.

This multi-district litigation ("MDL") involves antitrust, civil RICO, consumer protection, and unjust enrichment claims asserted by a putative class of plaintiffs. The putative class members assert these claims against suppliers and manufacturers of the EpiPen ("the Mylan and Pfizer defendants"). The EpiPen is an epinephrine auto-injector ("EAI") that delivers epinephrine to treat severe allergic reactions known as anaphylaxis. The putative class members—i.e., end-payors who purchased the EpiPen—have filed a motion seeking class certification under Fed. R. Civ. P. 23. Doc. 1353. Contemporaneously, the parties have filed motions seeking to exclude certain expert testimony offered either to support or oppose the putative class members' Motion for Class Certification. This Order rules those Daubert motions.

As explained more fully below, the court rules the parties' motions as follows:

The putative class plaintiffs' motions:

The Mylan and Pfizer defendants' motions:

I. Legal Standard

The court has a "gatekeeping obligation" to determine whether expert testimony is admissible. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147 (1999) (citing Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589 (1993)). When performing this gatekeeping role, the court has broad discretion. Kieffer v. Weston Land, Inc., 90 F.3d 1496, 1499 (10th Cir. 1996) (citing Orth v. Emerson Elec. Co., 980 F.2d 632, 637 (10th Cir. 1992)). Courts exercise this discretion under Federal Rule of Evidence 702. It provides:

Fed. R. Evid. 702.

The Tenth Circuit has not yet ruled whether a court must conduct a full Daubert analysis at the class certification stage. One federal court has noted that "[t]he issue of how to evaluate expert testimony at the class-certification stage `ha[s] beguiled the federal courts.'" Campbell v. Nat'l R.R. Passenger Corp., 311 F.Supp.3d 281, 294 (D.D.C. 2018) (quoting 3 William B. Rubenstein, Newberg on Class Actions § 7:24 (5th ed. 2014)). In dictum, the Supreme Court has expressed "doubt" about the notion that "Daubert [does] not apply to expert testimony at the certification stage of class-action proceedings." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 354 (2011). But the Court has not yet ruled the question definitively.

Previously, this court recognized that "the case law contains some dissonance about the proper breadth of Daubert-style challenges at the class certification stage." Doc. 1532 at 4. As one leading treatise has explained, "two approaches . . . have emerged in the case law." 3 William B. Rubenstein, Newberg on Class Actions § 7:24 (5th ed. 2013). Some Circuits have held that a court must perform a "full Daubert analysis" before certifying a class. See Am. Honda Motor Co. v. Allen, 600 F.3d 813, 815-16 (7th Cir. 2010) (per curiam) ("We hold that when an expert's report or testimony is critical to class certification, . . . the district court must perform a full Daubert analysis before certifying the class . . . ."); see also In re Carpenter Co., No. 14-0302, 2014 WL 12809636, at *3 (6th Cir. Sept. 29, 2014) (holding district court did not abuse its discretion by applying Daubert to determine admissibility of expert testimony offered to support class certification); Sher v. Raytheon Co., 419 F. App'x 887, 890-91 (11th Cir. 2011) ("Here the district court refused to conduct a Daubert-like critique of the proffered experts' qualifications. This was error.").1 The Eighth Circuit has adopted a different approach. It directs trial courts to apply a "focused Daubert analysis" instead of a "full and conclusive Daubert inquiry before certification." In re Zurn Pex Plumbing Prods. Liab. Litig., 644 F.3d 604, 610-11 (8th Cir. 2011). The Third Circuit has considered both approaches and recognized that some differences might exist between these two approaches. Nonetheless, the Third Circuit has declined "to examine whether there might be some variation between the Seventh and Eighth Circuit formulations" because "both courts limit the Daubert inquiry to expert testimony offered to prove satisfaction of Rule 23's requirements." In re Blood Reagents Antitrust Litig., 783 F.3d 183, 187, 188 n.8 (3d Cir. 2015) (joining with "sister courts to hold that a plaintiff cannot rely on challenged expert testimony, when critical to class certification, to demonstrate conformity with Rule 23 unless the plaintiff also demonstrates, and the trial court finds, that the expert testimony satisfies the standard set out in Daubert").

Here, the court likewise declines to decide whether the two approaches differ in a material fashion. Instead, because the parties' motions seek to exclude expert opinions used to support or oppose Rule 23's class certification requirements, the court applies the Daubert standard to those proffered opinions to determine whether the court should consider them at the class certification stage.

Outside the class certification phase, our Circuit has directed trial judges to apply a two-part test when determining the admissibility of expert testimony under Daubert and Rule 702. Conroy v. Vilsack, 707 F.3d 1163, 1168 (10th Cir. 2013). First, the court must determine "whether the expert is qualified `by knowledge, skill, experience, training, or education' to render an opinion." United States v. Nacchio, 555 F.3d 1234, 1241 (10th Cir. 2009) (quoting Fed. R. Evid. 702). Second, the court "`must satisfy itself that the proposed expert testimony is both reliable and relevant, in that it will assist the trier of fact, before permitting a jury to assess such testimony.'" Id. (quoting United States v. Rodriguez-Felix, 450 F.3d 1117, 1122 (10th Cir. 2006) (further citations omitted)).

To qualify as an expert witness, the witness must possess "such skill, experience or knowledge in that particular field as to make it appear that his opinion would rest on substantial foundation and would tend to aid the trier of fact in his search for truth." LifeWise Master Funding v. Telebank, 374 F.3d 917, 928 (10th Cir. 2004) (citation and internal quotation marks omitted). To determine whether the expert's testimony is reliable, the court must assess "whether the reasoning or methodology underlying the testimony is scientifically valid and . . . whether that reasoning or methodology properly can be applied to the facts in issue." Daubert, 509 U.S. at 592-93.

In Daubert, the Supreme Court identified four factors that—though not exhaustive—trial courts should consider when determining the reliability of proffered expert testimony under Fed. R. Evid. 702. They are: (1) whether the theory used can be and has been tested; (2) whether it has been subjected to peer review and publication; (3) the known or potential rate of error; and (4) general acceptance in the scientific community. Id. at 593-94. The Supreme Court has emphasized, however, that these four factors are not a "definitive checklist or test," and that a court's gatekeeping inquiry about reliability "must be tied to the facts of a particular case." Kumho Tire, 526 U.S. at 150 (citations and internal quotation marks omitted).

But in some cases, "the relevant reliability concerns may focus upon personal knowledge or experience," rather than the Daubert factors and scientific foundation. Id. For such testimony to satisfy the reliability standard, it "must be `based on actual knowledge, and not mere "subjective belief or unsupported speculation."'" Pioneer Ctrs. Holding Co. Emp. Stock Ownership Plan & Tr. v. Alerus Fin., N.A., 858 F.3d 1324, 1341-42 (10th Cir. 2017) (quoting Mitchell v. Gencorp, Inc., 165 F.3d 778, 780 (10th Cir. 1999) (quoting Daubert, 509 U.S. at 590)). "When expert opinion `is not supported by sufficient facts to validate it in the eyes of the law, or when indisputable record facts contradict or otherwise render the opinion unreasonable, it cannot support a jury's verdict' and will be excluded." Id. at 1342 (quoting Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 242 (1993)).

"The proponent of expert testimony bears the burden of showing that the testimony is admissible." Conroy, 707 F.3d at 1168 (citing Nacchio, 555 F.3d at 1241). "[R]ejection of expert testimony is the exception rather than the rule." Fed. R. Evid. 702 advisory committee's notes to 2000 amendments. While Daubert makes the court the gatekeeper for expert testimony, "[v]igorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof" remain "the traditional and appropriate means of attacking shaky but admissible evidence." Daubert, 509 U.S. at 596 (citation omitted).

The court has discretion to determine how to perform its gatekeeping function under Daubert. Bill Barrett Corp. v. YMC Royalty Co., LP, 918 F.3d 760, 770 (10th Cir. 2019). "The most common method for fulfilling this function is a Daubert hearing, although such a process is not specifically mandated." Goebel v. Denver & Rio Grande W. R.R., 215 F.3d 1083, 1087 (10th Cir. 2000) (citations omitted); see also United States v. Charley, 189 F.3d 1251, 1266 (10th Cir. 1999) ("The trial judge is granted great latitude . . . in deciding whether to hold a formal [Daubert] hearing."). Alternatively, the district court may satisfy its gatekeeping role without a formal Daubert hearing "so long as the court has sufficient evidence to perform `the task of ensuring that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand.'" Goebel, 215 F.3d at 1087 (quoting Daubert, 509 U.S. at 597). Here, exercising its discretion, the court concludes that it need not conduct a separate Daubert hearing to rule the parties' motions to exclude. The court has reviewed the parties' filings and attached exhibits carefully. And the court finds that the parties have provided a sufficient record to render a decision without a hearing.

II. Discussion

A. The Putative Class Plaintiffs' Motions to Exclude

The putative class plaintiffs have filed two motions seeking to exclude expert testimony: (1) Motion to Strike the Partial Testimony of Dr. Michael Blaiss (Doc. 1584); and (2) Motion to Strike the Testimony of Professor James Hughes (Doc. 1852). The court addresses each motion, in turn, below.

1. Motion to Strike the Partial Testimony of Dr. Michael Blaiss

The putative class plaintiffs seek to exclude just one aspect of Dr. Michael Blaiss's testimony. Their motion asks the court to strike Dr. Blaiss's proffered opinions expressed in Section 6.0 of his expert report. Dr. Blaiss is a medical doctor with training in allergy and immunology. Section 6.0 of Dr. Blaiss's expert report opines about the likelihood that existing EpiPen users would have switched to the generic EAI product of Teva Parenteral Medicines Inc. ("Teva") had the Teva generic launched in the market at an earlier date. Dr. Blaiss concludes EpiPen users may not readily have switched to the Teva generic because: (1) EAI products are unique, (2) the Teva product is not identical to the EpiPen EAI, and (3) the Teva generic is not widely available. Doc. 1585-1 at 18-19 (Blaiss Expert Report ¶ 6.0).

Plaintiffs assert that the court should exclude Dr. Blaiss's opinions because his conclusions in Section 6.0 are unsupported by data or analysis. Thus, plaintiffs argue, Dr. Blaiss's opinions are: (1) unreliable, and (2) not relevant because they do not fit with the facts of the case.2 The court rejects both of plaintiffs' arguments.

First, the court finds Dr. Blaiss's opinions sufficiently reliable. Dr. Blaiss explains that he bases his opinions about patients' reluctance to switch from the EpiPen to the Teva generic on more than 30 years of clinical experience treating patients who present a risk of anaphylaxis and prescribing EAI products to those patients. Also, Dr. Blaiss has reviewed the Teva generic product information and determined that it is "not identical to the EpiPen Auto-Injector." Id. at 19 (Blaiss Expert Report ¶ 6.0); see also id. at 12-13 (Blaiss Expert Report ¶ 5.1.6). Thus, Dr. Blaiss concludes, patients are likely to stick with EpiPen products instead of having to learn how to use another product with different instructions. Id. at 19 (Blaiss Expert Report ¶ 6.0).

Plaintiffs characterize Dr. Blaiss's opinions as ones based on anecdotal evidence drawn from his clinical experience. Plaintiffs assert that opinions about patient conversion present economic questions, but, plaintiffs argue, Dr. Blaiss has no training or education in the law, finance, business, marketing, consumer behavior, pharmacy, or economics. Also, plaintiffs criticize Dr. Blaiss's opinions because he never used any economic methodologies to reach his conclusions about the likelihood that patients would switch products. Thus, plaintiffs argue, Dr. Blaiss's opinions are not reliable because he can't ground them in the study of economics. The court disagrees.

Dr. Blaiss has grounded his opinions using his observations and experiences as a medical doctor. Plaintiffs' criticisms about the absence of economic methodologies used to support Dr. Blaiss's opinions go to their weight. But, the court declines to find Dr. Blaiss's opinions inadmissible under Daubert at the class certification stage. See In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2016 WL 5371856, at *10 (D. Kan. Sept. 26, 2016) (holding criticisms about reliability of plaintiffs' experts' opinions "go to the weight of those opinions and not . . . their admissibility" and concluding "the experts' methodologies [weren't] so unreliable as to preclude certification"); see also Bell v. 3M Co., No. 16-cv-02351-RBJ, 2018 WL 4599689, at *3 (D. Colo. Sept. 25, 2018) (concluding "defendant's challenges go to the weight of the evidence, not whether the Court can consider the opinions in its analysis of the class certification issues"); Campbell v. Nat'l R.R. Passenger Corp., 311 F.Supp.3d 281, 304 (D.D.C. 2018) (holding expert's opinions were sufficiently reliable for court to consider at class certification and challenges to the opinions went to "what weight to afford [the] opinions as part of [the court's] class-certification analysis").

Second, Dr. Blaiss's opinions are relevant to the class certification issues. Specifically, plaintiffs allege that defendants engaged in unlawful conduct that delayed Teva's entry of a generic product into the EAI market. Plaintiffs assert that the delay prevented generic competition, protected the EpiPen monopoly, and allowed defendants to continue raising the EpiPen's prices. Plaintiffs rely on their own expert testimony to demonstrate the classwide impact of the delay. Prof. Meredith Rosenthal opines that almost 95% of EpiPen users would have switched to a Teva generic within one year of its launch. And, using that figure, she provides an opinion about the damages sustained classwide from the delayed entry of a generic competitor. Dr. Blaiss's opinions—based on his own observations and experience—are relevant to the generic-delay theory and, specifically, the question whether plaintiffs sustained damages classwide from that delay. Thus, the court rejects plaintiffs' argument that Dr. Blaiss's opinions don't fit within the facts of the case.

For the above reasons, the court denies the putative class plaintiffs' Motion to Strike the Partial Testimony of Dr. Michael Blaiss.

2. Motion to Strike the Testimony of Professor James Hughes

The putative class plaintiffs also move to strike the testimony of Prof. James Hughes. Prof. Hughes is a college professor who specializes in "the fields of health economics, industrial organization, law and economics, and labor economics." Doc. 1852-4 at 3 (Hughes Expert Report ¶ 1). In this MDL, Prof. Hughes generally provides expert testimony on two subjects.

First, Prof. Hughes provides an overview of "the economic and commercial context surrounding the distribution of and payment for prescription drugs in the United States." Id. at 8 (Hughes Expert Report ¶ 12). More specifically, he explains "the flow of products and payments in the prescription drug space and the roles that various actors, including third-party payors, pharmacy benefit managers ("PBMs"), and consumers, play in the pharmaceutical supply and payment chain." Id. at 5 (Hughes Expert Report ¶ 7).

Second, Prof. Hughes opines about the ascertainability of the putative class in this lawsuit. Id. (Hughes Expert Report ¶ 8). He concludes that the putative class plaintiffs cannot ascertain with sufficient certainty who qualifies as a member of the putative class. This is so, Prof. Hughes asserts, because plaintiffs rely on data that is insufficient to: (1) identify and include the individuals who fit within the class definition, and (2) identify and exclude individuals who never actually paid for an EpiPen device. Id. at 8-9 (Hughes Expert Report ¶¶ 13-14).

The putative class plaintiffs argue that the court should strike Prof. Hughes's testimony on both subjects because, they contend, it is neither relevant nor reliable.3

a. Testimony about the Pharmaceutical Industry

The putative class plaintiffs contend that Prof. Hughes's testimony about the pharmaceutical industry is not reliable because it consists of generalized opinions about the industry as a whole—and not a particular drug. Plaintiffs' attack disregards the advisory committee's note to Rule 702. It recognizes that:

Fed. R. Evid. 702 advisory committee's note to 2000 amendments; see also United States v. Brinson, 772 F.3d 1314, 1320 (10th Cir. 2014) (holding that district court did not abuse its discretion by allowing a detective to testify as an expert about the general characteristics of the prostitution trade).

Consistent with the guidance from the advisory committee's note, defendants assert that they offer Prof. Hughes's expert testimony to provide important and relevant background information to help the court understand the highly technical operation of a complex industry. And, they contend, the court can consider Prof. Hughes's opinion about the industry because Rule 702 allows the "venerable practice of using expert testimony to educate the factfinder on general principles." Fed. R. Evid. 702 advisory committee's note to 2000 amendments.

In response, plaintiffs assert that Prof. Hughes's generalized testimony here just doesn't apply to the facts of the EpiPen market. Thus, plaintiffs contend, the court should exclude the testimony because it is unhelpful and irrelevant. For support, plaintiffs cite a portion of Prof. Hughes's deposition transcript where he opines he wouldn't expect to see a positive correlation between the drug price set by the manufacturer (the Wholesale Acquisition Cost, "WAC") and the pharmacy's price for cash paying customers (the Usual and Customary Price, "U&C price") for a given drug. The putative class plaintiffs argue that Prof. Hughes performed no analysis to reach this conclusion. And, they contend, his testimony directly conflicts with the analysis that plaintiffs' expert performed on EpiPen price trends. According to plaintiffs, their expert—Prof. Meredith Rosenthal—analyzed the actual correlation between EpiPen WAC and U&C price and concluded it is "near-perfect." Doc. 1852 at 16-17.

Defendants respond that Prof. Hughes never opines about the relationship between WAC and U&C price in the pharmaceutical industry. Also, he offers no opinions about the correlation between these prices specifically for EpiPen products. Instead, he just offers generalized background information about the industry. The court agrees with defendants. Plaintiffs rely on just this one example—Prof. Hughes's response to a deposition question—to argue that his opinions don't fit the facts of the case. But Prof. Hughes never opines specifically about price correlation. Instead, his expert opinions consist of general background information about the pharmaceutical industry. And Rule 702 allows this kind of generalized expert testimony.

Next, the putative class plaintiffs argue, Prof. Hughes's opinion that one of the members of the putative class—i.e., Local 282 Welfare Trust Fund ("Local 282")—bears no financial responsibility for drugs dispensed to its insured members is unreliable because the evidence in the case contradicts his conclusion. Prof. Hughes opines that it's the employers paying into Local 282—not Local 282 itself—who pay the costs in full for drugs dispensed to Local 282's insured members. Doc. 1852-4 at 24 (Hughes Expert Report ¶ 48). Plaintiffs argue that Local 282's Fed. R. Civ. P. 30(b)(6) witness's testimony establishes that Prof. Hughes is incorrect. This 30(b)(6) witness testified that Local 282 "pays for every prescription drug covered under its plans," and does not "see[k] reimbursement for any part of that cost from any other entity." Doc. 1852-9 at 4 (Bulding Dep. 198:14-22). But the 30(b)(6) witness also testified that when Local 282's costs increase, it "potentially [has] to go back . . . and negotiate higher contribution rates to the welfare trust fund." Doc. 1873-4 at 3 (Bulding Dep. 59:3-25). So, Prof. Hughes opines, it's the employer-sponsors who ultimately bear the financial responsibility for drug costs. But, plaintiffs argue, Prof. Hughes conceded in his deposition that the 30(b)(6) witness's testimony "says nothing about the employer sponsor being obligated to make up a shortfall when it occurs," and he doesn't know whether Local 282's employer sponsors actually are obligated to make up a shortfall if one occurs. Doc. 1852-2 at 104-05 (Hughes Dep. 102:19-103:2).

Based on this competing testimony, plaintiffs argue that the evidence in the case contradicts Prof. Hughes's opinion that Local 282 bears no financial responsibility for drugs dispensed to its insured members. Also, plaintiffs challenge the methodology Prof. Hughes used to reach his conclusions because he supposedly relied on testimony that—plaintiffs contend— doesn't support his conclusions. But, as summarized above, the parties offer competing evidence that either supports or refutes Prof. Hughes's opinion about Local 282's financial responsibility. These arguments go to the weight that the court should give his opinion, not its admissibility.

Also, plaintiffs argue that Prof. Hughes's opinions aren't supported by the evidence because he conceded that asking an employer sponsor to make more contributions to the fund is "similar" to "a commercial insurer who has to go back and increase premiums" when faced with cost increases. Id. at 102-03 (Hughes Dep. 100:20-101:11). And, plaintiffs argue, courts are clear that antitrust defendants may not avoid liability for overcharges simply because the plaintiff was able to pass any additional costs on to someone else. Doc. 1852 at 17 (citing In re Lidoderm Antitrust Litig., No. 14-md-02521-WHO, 2017 WL 679367, at *22 (N.D. Cal. Feb. 21, 2017)); see also In re Nexium Antitrust Litig., 777 F.3d 9, 27 (1st Cir. 2015) ("[D]efendants incorrectly assume that if a class member offsets an overcharge through later savings attributable to the same or related transaction, there is no injury. But antitrust injury occurs the moment the purchaser incurs an overcharge, whether that injury is later offset or not. Here, if a class member is overcharged, there is an injury, even if that class member suffers no damages." (internal citations omitted)).

This argument challenges Prof. Hughes's opinion using a legal argument—that is, plaintiffs contend that Prof. Hughes's opinion does not preclude defendants' liability for antitrust claims under well-settled law. But plaintiffs' legal argument provides no reason to exclude his testimony. This legal argument does not challenge the reliability of Prof. Hughes's reasoning or the methodology used to form his opinions. Thus, the court denies the putative class plaintiffs' motion to exclude Prof. Hughes's opinions on this basis.

Finally, the putative class plaintiffs assert that the court should exclude Prof. Hughes's expert testimony about the pharmaceutical industry because, they contend, he reached his conclusions before he reviewed the relevant literature and without reviewing the entire contents of the sources he relies on to support his opinions. Thus, plaintiffs argue, Prof. Hughes's opinions are the product of an unreliable and unscientific methodology. The court disagrees.

Plaintiffs take issue with part of Prof. Hughes's deposition testimony where he described the information he reviewed to prepare his expert report. Prof. Hughes testified that he "looked at fact depositions from PBM people, from Local 2[8]2" and he wrote part of his report "based largely on [his] 20 years of experience in examining the pharmaceutical market." Doc. 1852-2 at 48 (Hughes Dep. 46:7-21). Then, Prof. Hughes described that "afterwards [he] went back to look for secondary sources that were consistent with what [he] knew about the industry so that [he] could [c]ite an outside source for [his] outside authority for the fact that [he] was presenting in the report." Id. Plaintiffs contend that this testimony shows that Prof. Hughes "cherry-picked" evidence to support his opinions—something an expert may not do. Doc. 1852 at 18 n.79 (citing Fish v. Kobach, 309 F.Supp.3d 1048, 1102 (D. Kan. 2018)).

The court doesn't read Prof. Hughes's testimony the way plaintiffs describe it. Instead, Prof. Hughes testified that he drafted the report based on his 20 years of experience in the field, and then he found secondary sources to support his knowledge. Daubert requires an expert to base his testimony either "upon professional studies or personal experience." Kumho Tire, 526 U.S. at 152. As defendants assert, Prof. Hughes did both here. He relied on his personal knowledge—drawn, he says, from his 20 years of experience studying the pharmaceutical industry—to draft the portions of his report describing that industry. And he verified that knowledge by finding and citing secondary sources that substantiated his assertions. In short, Prof. Hughes didn't "cherry pick evidence in support of his opinion" like the expert did in Fish v. Kobach. 309 F. Supp. 3d at 1102. Instead, Prof. Hughes backed up his knowledge by citing other authorities to support it. The court finds no reason to exclude Prof. Hughes's opinions under Daubert based on the methodology he used to cite secondary sources when preparing his expert report.

Also, the court doesn't agree with plaintiffs' assertions that Prof. Hughes's opinions are unreliable because he performed an incomplete or cursory review of the materials. Plaintiffs contend that Prof. Hughes testified that he relied on deposition testimony to form his opinions. But Prof. Hughes also conceded that he didn't read the full deposition transcript of any of the named plaintiffs or Local 282's 30(b)(6) witness. Doc. 1852-2 at 48, 104 (Hughes Dep. 46:7-16, 102:9-10). Plaintiffs argue that Prof. Hughes's limited review has caused him to mischaracterize deposition testimony, thus making his opinions unreliable.

Defendants respond that Daubert does not require an expert to review every line of every deposition where the case involves hundreds of pages of deposition testimony. Instead, an expert need only review the relevant parts of the record to form a reliable opinion. See, e.g., Milam v. Ranger Ins. Co., No. CIV-04-1749-HE, 2006 WL 5347771, at *3 n.9 (W.D. Okla. June 13, 2006) ("Contrary to defendants' suggestion, it is not necessary that an expert witness `read all the depositions' and turn himself into some sort of `super juror' in order to offer expert testimony on an issue, assuming other prerequisites to admissibility are met."). And, defendants contend, Prof. Hughes has characterized the deposition testimony properly based on its context.

The court finds that the parties' dispute about Prof. Hughes's characterization of the evidence goes to the weight his opinions deserve, but does not undermine their admissibility. Plaintiffs properly have challenged the way Prof. Hughes interpreted the evidence to argue that his opinions don't preclude the court from certifying a class action in this MDL. See Doc. 1837 at 47, 49 (Pls.' Reply Mem. in Further Supp. of Mot. for Class Certification). The court will consider these arguments on class certification, but the court refuses to exclude Prof. Hughes's opinions under Daubert.4

b. Opinions about Ascertainability

Next, the putative class plaintiffs argue that the court should exclude Prof. Hughes's opinions about ascertainability of the putative class because, plaintiffs contend, they are unreliable. Plaintiffs advance three arguments to support this request.

First, plaintiffs assert, Prof. Hughes's ascertainability opinion is an impermissible legal conclusion. Our Circuit recognizes that "testimony on ultimate facts is authorized under [Federal Rule of Evidence] 704," but an expert "may not give an opinion on ultimate issues of law." Specht v. Jensen, 853 F.2d 805, 808 (10th Cir. 1988) (emphasis added). But here, the court concludes that Prof. Hughes is not offering opinions about ultimate issues of law.

As Prof. Hughes explained in both his expert report and his deposition, he provides an expert opinion about whether plaintiffs have proposed a methodology for analyzing data so that one can identify which consumers qualify as members of the class under the class definition. Doc. 1852-4 at 5 (Hughes Expert Report ¶ 8); Doc. 1852-2 at 111 (Hughes Dep. 109:3-14 ("[M]y assignment is very narrow, is that can you tell who is in—is there a methodology to tell who is in the class and who is not in the class.")). Prof. Hughes concludes that "based on the data and other materials [he has] reviewed, the Plaintiffs do not provide data or a methodology for identifying which individuals should be excluded and included in the class." Doc. 1852-4 at 9 (Hughes Expert Report ¶ 16). Prof. Hughes thus bases his opinion on facts he has reviewed in this case. He never opines that plaintiffs fail to satisfy a legal burden to establish ascertainability. To the contrary, Prof. Hughes testified he was not "applying any court's test because it's a legal standard" and "[t]hat would be outside [his] expertise." Doc. 1852-2 at 111 (Hughes Dep. 109:3-14). The court thus concludes that Prof. Hughes isn't offering an impermissible legal opinion.

Second, plaintiffs argue that Prof. Hughes bases his ascertainability opinions on a legal standard that simply doesn't apply in our Circuit. Thus, plaintiffs contend, his opinions aren't relevant to this case. To support this argument, plaintiffs cite Prof. Hughes's deposition transcript. Prof. Hughes testified that ascertainability requires "a methodology upfront that can be used to identify who is in the class and who is not in the class." Doc. 1852-2 at 111-12 (Hughes Dep. 109:14-110:4). And, according to Prof. Hughes, one performs that methodology by "using data that is produced in discovery . . . ." Id. at 117-18 (Hughes Dep. 115:22-116:6). So, as Prof. Hughes testified, he reached his conclusions about ascertainability by reviewing the data that insurers and PBMs had produced in discovery. Doc. 1852-2 at 118 (Hughes Dep. 116:11-22).

Plaintiffs assert that Prof. Hughes's testimony shows that he bases his opinions on the Third Circuit's standard used to determine ascertainability, as defined in Carrera v. Bayer Corp., 727 F.3d 300 (3d Cir. 2013). See id. at 307 (holding the "method of determining whether someone is in the class must be `administratively feasible'" and a plaintiff cannot "satisfy the ascertainability requirement if individualized fact-finding or mini-trials will be required to prove class membership"). Indeed, defendants argue that plaintiffs bear the burden to demonstrate ascertainability before the court can certify a class. Doc. 1873 at 18. But, plaintiffs respond that defendants are citing a legal standard that doesn't apply in our Circuit.

The Tenth Circuit hasn't imposed a requirement that a putative class must demonstrative an "administratively feasible method" for determining whether putative class members fall within the class definition. And this court has refused to apply "the Third Circuit's stricter standard from Carrera that requires an administratively feasible mechanism for identifying class members . . . ." In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591, 2016 WL 5371856, at *2-3 (D. Kan. Sept. 26, 2016) (declining "invitation 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"). For this reason, plaintiffs ask the court to exclude Prof. Hughes's opinions as irrelevant.

Daubert does not require the court to exclude Prof. Hughes's testimony for this reason. As already discussed, Prof. Hughes isn't offering a legal conclusion. Instead, Prof. Hughes bases his ascertainability opinions on data produced in the case that he used to determine whether one can identify who qualifies as a class member (and who doesn't) under the proposed class definition.5 Whether the law requires this kind of methodology to prove ascertainability at the class certification stage is a question of law for the court to decide. And the court can determine whether Prof. Hughes's ascertainability opinions are relevant under the applicable legal standard when it decides the class certification motion.6 The court thus declines to exclude Prof. Hughes's opinions under Daubert because it relies on a legal standard from another Circuit.

Finally, plaintiffs contend that Prof. Hughes did not perform an appropriate analysis before reaching his ascertainability opinions. Thus, plaintiffs argue, Prof. Hughes's opinions are unreliable because he bases them on speculation. Plaintiffs criticize Prof. Hughes for failing to review plaintiffs' expert reports. Doc. 1852-2 at 51 (Hughes Dep. 49:11-15). But, Prof. Hughes explained, he didn't read those reports because plaintiffs' experts had not included any opinions about ascertainability in their reports. Id. (explaining that he reviewed none of plaintiffs' expert reports because "they were not on the topic [he] was opining on"). Plaintiffs don't refute the proposition that their expert reports had not proposed a methodology for identifying class members when Prof. Hughes submitted his expert report. But, they contend, Prof. Hughes's methodology is flawed because he relied on a consultant to provide the information about plaintiffs' expert reports but he never verified independently that the expert reports included no opinions about ascertainability. Plaintiffs' argument here asks too much. The court declines to find Prof. Hughes's opinions unreliable simply because he relied on information from a consultant—information that indeed was accurate—when he formed his opinions.

Also, plaintiffs argue that Prof. Hughes's opinions are flawed because he didn't use the data produced in the case to determine whether a consumer paid with a fully-funded HSA, a Mylan coupon, or a flat co-pay. But, Prof. Hughes testified that he doesn't believe it's possible to determine whether a consumer paid for the EpiPen in one of those three ways based on the data produced in the case. Doc. 1852-2 at 113-14, 145-46, 150 (Hughes Dep. 111:16-112:15, 143:25-144:6, 148:15-20). Prof. Hughes's report identifies the data he reviewed and analyzed when forming his opinions. Doc. 1852-4 at 8, 30-40 (Hughes Expert Report ¶¶ 13, 63-84). And, the court finds, Prof. Hughes's methodology for using that data is sufficiently reliable for the court to consider his ascertainability opinions on class certification. Plaintiffs' criticisms about how Prof. Hughes interpreted the data—and what he concluded the data shows and does not show—go to the weight of his opinions, not their admissibility. The court declines to exclude Prof. Hughes's opinions from the court's consideration of the class certification motion.

For all these reasons, the court denies the putative class plaintiffs' Motion to Strike the Testimony of Prof. James Hughes.

B. Defendants' Motions to Exclude

The Mylan and Pfizer defendants have filed three motions seeking to exclude expert testimony: (1) Motion to Exclude Expert Opinions of Professor Meredith Rosenthal (Doc. 1602); (2) Motion to Exclude Expert Opinions of Professor Einer Elhauge (Doc. 1604); and (3) Motion to Exclude Opinions and Proposed Testimony of Andrew K. Torrance (Doc. 1847). The court addresses each motion, in turn, below.

1. Motion to Exclude Expert Opinions of Professor Meredith Rosenthal

Defendants move the court to exclude the expert opinions expressed by the putative class plaintiffs' expert, Prof. Meredith Rosenthal. Prof. Rosenthal opines about the damages purportedly sustained by the putative class plaintiffs. See generally Doc. 1500-3 (Rosenthal Expert Report). She has identified certain categories of damages that, she contends, the putative class plaintiffs have sustained on a classwide basis. Id. And, she has applied certain methodologies to calculate those purported damages. Id.

Defendants argue that the court should exclude some of Prof. Rosenthal's opinions because, they contend, her opinions are unreliable and thus inadmissible under Daubert.7 Defendants assert four arguments supporting this request. First, defendants contend that Prof. Rosenthal's damages methodology for the 2-Pak theory is an unreliable model for measuring classwide injury. Second, defendants contend that Prof. Rosenthal's ascertainability opinion is flawed and improper rebuttal testimony. Third, defendants argue that Prof. Rosenthal's probability analysis cannot estimate or identify uninjured class members in a reliable fashion. Defendants also assert that her probability analysis constitutes improper rebuttal testimony. Finally, defendants argue that the methodologies Prof. Rosenthal used to determine generic delay and Auvi-Q foreclosure damages rely on Prof. Einer Elhauge's unreliable opinions. So, defendants contend, if the court excludes Prof. Elhauge's opinions under Daubert, it also must exclude Prof. Rosenthal's opinions that rely on Prof. Elhauge's opinions. The court addresses all four arguments, below.

a. Methodology for 2-Pak Damages

First, defendants seek a Daubert ruling excluding Prof. Rosenthal's opinion about the damages allegedly sustained by the putative class members from purchasing the EpiPen in the 2-Pak. In August 2011, Mylan stopped selling EpiPen products in single packages. And instead, it began offering the EpiPen exclusively in the 2-Pak (i.e., two EpiPen devices in one package). Mylan asserts that it switched to the 2-Pak in response to medical guidance recommending that patients always carry two EAI devices in case the patient requires more than one dose of epinephrine to treat an anaphylactic reaction. The putative class plaintiffs allege that Mylan's switch to the 2-Pak was not clinically necessary, but instead, Mylan implemented the change to force consumers to purchase two EpiPens instead of one. And they have retained Prof. Rosenthal to provide a methodology for calculating the classwide damages sustained because Mylan allegedly forced EpiPen customers to buy at least two EpiPens at a time.

Prof. Rosenthal's report explains that she calculated the "overcharge damages" allegedly sustained from Mylan's switch to the 2-Pak using "the difference between expenditures resulting from the average number of pens per prescription before vs. after the withdrawal." Doc. 1500-3 at 24 (Rosenthal Expert Report ¶ 54). Specifically, Prof. Rosenthal compared "the average pen per prescription from the first two quarters of 2011 with the average pen per prescription in the years following that point in time." Id. at 30 (Rosenthal Expert Report ¶ 69). Prof. Rosenthal asserts that—in a but-for world—consumers would have continued to purchase single pens at the same rate they had manifested in the first two quarters of 2011 for the next seven years, but-for Mylan's switch to selling the EpiPen exclusively in the 2-Pak. Id. at 30-31, 33 (Rosenthal Expert Report ¶¶ 69-73, 77-78); see also Doc. 1497-3 at 84-90, 101-104 (Sealed Rosenthal Expert Report Attach. D.1-D.4, Attach. E.3.a-E.3.c). But, Prof. Rosenthal opines that, in the actual world, Mylan forced an increase in EpiPen consumption by requiring consumers to buy EpiPens in the 2-Pak. Id. Thus, Prof. Rosenthal concludes, consumers overpaid for the EpiPen after August 2011 because EpiPen purchasers had to buy two EpiPens instead of just one. Id. And then, Prof. Rosenthal calculates the alleged overcharge damages sustained by EpiPen consumers from the forced increase in their consumption of EpiPens after the switch to the 2-Pak. Id.

In her Reply Report, Prof. Rosenthal explains that her 2-Pak damages methodology "uses an event study approach." Doc. 1711-1 at 12-13 (Rosenthal Reply Report ¶ 15). Prof. Rosenthal explains how event studies are "widely employed in economic analysis." Id. And, she describes how, "[i]n each case, the event study relies on the timing of an event of interest to identify its effects." Id. According to Professor Rosenthal, "[t]he power of the event study is the ability to abstract from complex phenomena that are not changing over the short time window of the study—in this case, factors such as patient and physician preferences, comorbidities, and product attributes." Id.

Defendants argue that Prof. Rosenthal's analysis here isn't an event study at all. Instead, they say, Prof. Rosenthal has used an unreliable methodology that she invented just to support the putative class plaintiffs' class certification motion. Defendants contend that real event studies are used in securities and finance litigation—not consumer class actions like this one—to show how stock prices have reacted to announcements about economic events. Prof. Rosenthal responds, explaining, "[e]vent studies have also been used in health policy and pharmaceutical economics." Doc. 1711-1 at 13 (Rosenthal Reply Report ¶ 15). And, to support this rebuttal, she cites several peer-reviewed articles from economic publications where, she asserts, event studies were used in pharmaceutical economic analysis. Id. (Rosenthal Reply Report ¶ 15 n.18). Defendants respond that each of Prof. Rosenthal's cited articles aren't helpful to her point because each one used event studies to examine the effect on stock prices. Doc. 1884 at 7 n.2.

The court finds that defendants' distinction doesn't preclude the court from considering Prof. Rosenthal's opinion at the class certification stage. The fact that event studies typically are used in securities and finance litigation doesn't undermine the reliability Prof. Rosenthal's methodology. And defendants don't explain why it is improper to use an event study in a consumer class action case like this one. To satisfy Daubert's reliability requirement, the party offering the expert opinion "`need not prove that the expert is undisputably correct or that the expert's theory is "generally accepted" in the scientific community.'" Bitler v. A.O Smith Corp., 400 F.3d 1227, 1233 (10th Cir. 2005) (quoting Mitchell v. Gencorp Inc., 165 F.3d 778, 781 (10th Cir. 1999)). Instead, the expert testimony's proponent "`must show that the method employed by the expert in reaching the conclusion is scientifically sound and that the opinion is based on facts which sufficiently satisfy Rule 702's reliability requirements.'" Id. (quoting Mitchell, 165 F.3d at 781). Here, plaintiffs sufficiently have established that Prof. Rosenthal used a reliable method to reach her conclusions about the 2-Pak. At best, defendants' attacks on the appropriateness of an event study in a consumer class action go to the weight of her opinion, not its admissibility.

Defendants also contend that a valid event study must consider the potential effect of competing factors that may influence the outcome. Defendants argue that Prof. Rosenthal failed to consider such other factors here. Specifically, defendants assert that Prof. Rosenthal's model studies the potential effect of just one factor—the switch to the 2-Pak. But, defendants contend, Prof. Rosenthal failed to consider other factors that could influence consumers' purchases, such as widespread medical guidance issued after 2011 recommending that patients purchase two EAI devices. Defendants argue that the medical guidance could have caused consumers to purchase the 2-Pak after August 2011. And thus, consumers would have based their decisions to purchase in packs of two on a reason entirely independent of Mylan's decision to withdraw the single-pack from the market. But, they assert, Prof. Rosenthal didn't account for other factors like this one. As a result, defendants contend, her study fails to demonstrate that the switch to the 2-Pak was the sole reason why the putative class members bought more EpiPens than they otherwise would have. According to defendants, this omission makes her analysis unreliable. See Bricklayers & Trowel Trades Int'l Pension Fund v. Credit Suisse Sec. (USA) LLC, 752 F.3d 82, 95-97 (1st Cir. 2014) (affirming district court's decision to exclude expert opinion in form of event study that didn't appropriately account for other confounding factors that may have affected company's stock price); see also Sergeants Benevolent Ass'n Health & Welfare Fund v. Sanofi-Aventis U.S. LLP, 806 F.3d 71, 95 (2d Cir. 2015) (affirming district court's denial of a class certification motion and holding Dr. Rosenthal's regression analysis didn't support plaintiffs' RICO causation theory because she "conceded that she had not been asked by Plaintiffs' counsel to perform the kind of regression analysis that might have isolated the relative causal effect of the numerous variables [other than defendant's alleged withholding of safety information] bearing on the decline in [an antibiotic drug's] sales").

But, Prof. Rosenthal explains why she didn't need to consider whether medical guidance recommending patients carry two EAI devices would have influenced the outcome of her study. She opines that the medical guidance "did not change coincident with the removal of the single injector packaging." Doc. 1711-1 at 13 (Rosenthal Reply Report ¶ 16). Instead, she explains, "[t]he effects of clinical recommendations are constant before and after the event [she] stud[ies] and so do not need to be separately accounted for." Id. As support, she cites deposition testimony by the putative class plaintiffs' medical expert, Dr. Jay Portney, who testified that the medical guidance recommending two EAI devices wasn't a landmark document because it only published what allergists already knew. Id. (Rosenthal Reply Report ¶ 16 n.20). Also, Prof. Rosenthal explains, her event study "assess[es] the impact of the withdrawal of the single injector EpiPen using a narrow time window (two quarters pre-withdrawal) to avoid picking up the effects of other influences," and it "strongly support[s] the casual inference" that withdrawing "the single injector EpiPen forced the Class to purchase more pens than it would have [purchased] otherwise." Id. at 13-14 (Rosenthal Reply Report ¶ 17).

Plaintiffs argue that the cases defendants cite to support excluding Prof. Rosenthal's opinion involve a failure to consider competing factors that could have influenced the study's outcome. In contrast, here, plaintiffs contend, Prof. Rosenthal didn't have to consider other factors because any such factors—like the prevalence of medical guidance—remained constant before, during, and after Mylan switched to selling the EpiPen exclusively in the 2-Pak.

Defendants argue that the only support for Prof. Rosenthal's conclusion that the other factors remained constant is her own subjective analysis. Defendants criticize Prof. Rosenthal's reliance on plaintiffs' medical expert's testimony. Instead, defendants argue, Prof. Rosenthal should have performed an empirical analysis to determine whether the medical guidance, in fact, influenced customers' purchasing decisions. These arguments go to the weight of Prof. Rosenthal's opinion, but they don't preclude its admissibility. Certainly, "[v]igorous cross-examination" and "presentation of contrary evidence" are "the traditional and appropriate means of attacking shaky but admissible evidence." Daubert, 509 U.S. at 596 (citation omitted). And defendants appropriately can challenge (and have challenged) Prof. Rosenthal's opinions on class certification in this fashion. See Doc. 1636 at 41-45 (Mylan Defs.' Opp'n to Class Pls.' Mot. for Class Certification).8 But, exercising its discretion, the court concludes that defendants' arguments don't warrant a Daubert order excluding Prof. Rosenthal's opinions about the alleged damages the putative class members sustained from Mylan's switch to the 2-Pak.

b. Ascertainability Opinion

Second, defendants ask the court to exclude Prof. Rosenthal's opinion about ascertainability of the class for two reasons: (1) her ascertainability opinion is unreliable because she can't identify the customers who actually qualify as class members under the class definition, and (2) her ascertainability opinion is improper rebuttal testimony.

The court addresses the second argument first. This argument contends that the putative class plaintiffs bear the burden to prove that their proposed class is ascertainable. Thus, defendants contend, plaintiffs should have offered any expert opinion about ascertainability in their experts' opening reports. Prof. Rosenthal didn't do so. Instead, Prof. Rosenthal first offered her ascertainability opinion in her Reply Report. Thus, defendants argue, the court should exclude this opinion as improper rebuttal testimony.

Plaintiffs respond that Prof. Rosenthal's Reply provides proper rebuttal testimony in response to criticisms levied by expert opinions that defendants presented in their Opposition to the class certification motion. Specifically, defendants' experts opine that large numbers of proposed class members are uninjured because they never paid for their EpiPens. And, defendants argue, plaintiffs cannot ascertain from the data produced during discovery who actually paid for their EpiPens, and thus who qualifies as a member of the proposed class. Defendants' expert, Prof. James Hughes, asserts that he has reviewed the data and documents produced in discovery and he has concluded they are "insufficient to . . . exclude individuals from the class who did not actually pay for an EpiPen device." Doc. 1852-4 at 9 (Hughes Expert Report ¶ 14). Plaintiffs argue that Prof. Rosenthal's Reply Report merely responds to Prof. Hughes's criticisms about her analysis. See Doc. 1711-1 at 3 (Rosenthal Reply Report) (explaining that she is "respond[ing] to the expert report[ ] of . . . Dr. Hughes as [it] pertain[s] to [her] affirmative report on class certification submitted in this matter," and specifically, she "disagree[s] with Dr. Hughes and provide[s] documentation that contradicts his conclusions regarding the so-called `ascertainability' of Class members"). And thus, plaintiffs contend, Prof. Rosenthal's ascertainability opinion is proper rebuttal testimony.

Our Circuit has explained that "[r]ebuttal evidence is evidence which attempts to `disprove or contradict' the evidence to which it is contrasted." Tanberg v. Sholtis, 401 F.3d 1151, 1166 (10th Cir. 2005) (quoting Black's Law Dictionary 579 (7th ed. 1999)). The Federal Rules specifically limit rebuttal expert testimony to evidence that is "intended solely to contradict or rebut evidence on the same subject matter identified by another party" in its expert disclosures. Fed. R. Civ. P. 26(a)(2)(D)(ii) (emphasis added). A court properly may admit rebuttal evidence on a topic when "a party opens the door to [that] topic." Tanberg, 401 F.3d at 1166. But generally, courts will exclude "use of a rebuttal expert to introduce evidence more properly a part of a party's case-in-chief, especially if the alleged rebuttal expert is used to introduce new legal theories." Foster v. USIC Locating Servs., LLC, No. 16-2174-CM, 2018 WL 4003354, at *2 (D. Kan. Aug. 17, 2018) (citations and internal quotation marks omitted); see also Koch v. Koch Indus., 203 F.3d 1202, 1224 (10th Cir. 2000) ("When plaintiffs, however, seek to rebut defense theories which they knew about or reasonably could have anticipated, the district court is within its discretion in disallowing rebuttal testimony."). In short, the decision to admit or exclude rebuttal testimony is within the trial court's discretion. Tanberg, 401 F.3d at 1166.

Here, plaintiffs assert that they couldn't have anticipated Prof. Hughes's criticisms about identifying the individuals who qualify as class members because, they contend, Prof. Hughes relies on an "administrative feasibility" standard for identifying class members and that standard doesn't apply in our Circuit. As with the discussion about the putative class plaintiffs' motion to exclude the opinions of Prof. James Hughes, the parties sharply disagree about the legal requirements for showing ascertainability at class certification. With this motion, seeking to exclude Prof. Rosenthal's opinions, defendants cite a Pennsylvania case where the court defined the "ascertainability requirement" as "show[ing] that there is a reliable, administratively feasible mechanism that can identify which potential class members fall within the class definition." In re Wellbutrin XL Antitrust Litig., 308 F.R.D. 134, 137 (E.D. Pa. 2015) (citing Byrd v. Aaron's Inc., 784 F.3d 154, 163 (3d Cir. 2015)). In contrast, plaintiffs argue the Tenth Circuit has no "administrative feasibility" requirement to prove ascertainability on class certification. And, they argue, our court explicitly has rejected such a requirement. See In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591, 2016 WL 5371856, at *2-3 (D. Kan. Sept. 26, 2016) (declining an "invitation 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"). Instead, plaintiffs assert, this court has adopted an ascertainability standard that 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 v. Direct Digital, LLC, 795 F.3d 654, 659-60 (7th Cir. 2015)).

On this Daubert motion, the court declines to decide what standard to apply when determining whether the putative class presently is ascertainable. That is a question the court will decide in its class certification analysis.9 Here, the court exercises its discretion and refuses to exclude Prof. Rosenthal's opinion as improper rebuttal testimony on this Daubert motion. Prof. Rosenthal's Reply Report explicitly recites that she offers her rebuttal opinions in response to Prof. Hughes's criticisms about her analysis. See Doc. 1711-1 at 3. The Reply Report cites specific portions of Prof. Hughes's report and explains why Prof. Rosenthal disagrees with his analysis of the ascertainability question. See id. at 27-28 (Rosenthal Reply Report ¶¶ 44, 46-49). The court finds that Prof. Rosenthal's ascertainability opinion is proper rebuttal testimony.

Now, the court turns back to defendants' first argument supporting their request that the court exclude Prof. Rosenthal's opinion about the ascertainability of the class. Defendants contend that Prof. Rosenthal doesn't have a reliable methodology for ascertaining three types of EpiPen consumers to determine whether these consumers qualify as members of the class under the class definition. The three types of consumers are: (1) consumers who paid nothing, (2) consumers who had flat co-pays for the EpiPen, and (3) consumers who paid cash for their EpiPen products.10

Defendants assert that the law in our Circuit requires the putative class plaintiffs to define the class using a definition that is "`"precise, objective, and presently ascertainable."'" See In re Syngenta AG MIR 162 Corn Litig., 2016 WL 5371856, at *1 (quoting In re Urethane Antitrust Litig., 237 F.R.D. 440, 444 (D. Kan. 2006) (quoting Manual for Complex Litig. § 221.222, at 270 (4th ed. 2005))); see also Shook v. El Paso Cty., 386 F.3d 963, 972 (10th Cir. 2004) (recognizing that "the lack of identifiability is a factor that may defeat Rule 23(b)(3) class certification"). And, defendants contend, Prof. Rosenthal hasn't used a reliable methodology to ascertain who qualifies (and does not qualify) as a member of the class under the Tenth Circuit's standard for any of the three categories of consumers. The court addresses each argument for each type of consumer, below.

i. Consumers Who Paid Nothing

In the first category of consumers, defendants explain that Mylan issued coupons to consumers who, in many instances, used the coupons to pay nothing for their EpiPens. And, if the consumer paid nothing to purchase an EpiPen, that consumer doesn't qualify as a member of the class under the class definition.

Prof. Rosenthal offers a methodology for identifying this first category of consumers so that they aren't included in the class. Doc. 1711-1 at 21-22 (Rosenthal Reply Report ¶ 33). She asserts, "[t]here are individual identifiers, provider identifiers, and dates in the Mylan coupon data that will permit matching with paid pharmacy claims." Id. To illustrate, Prof. Rosenthal used CVS/Caremark pharmacy claims data to "cross reference with the Mylan coupon data and identify the prescriptions where copay coupons were used." Id. By "[m]atching on exact combinations of fill date, physician identification number, [National Drug Code] and/or Rx number," Prof. Rosenthal was able to "link the copay coupon information with the CVS/Caremark pharmacy claims . . . ." Id. Prof. Rosenthal opines that "[m]any examples like this can be observed in the merged data," and thus, "data exists such that—at the appropriate time—we can link coupons with pharmacy records and determine their effect on whether an individual Class member was injured." Id.

Defendants say Prof. Rosenthal's methodology is wrong. For support, they rely on their own expert's analysis. Defendants' expert, Prof. Hughes, asserts that he tried to match all the transactions listed in the produced PBM claims data with Mylan's coupon data using Prof. Rosenthal's proposed methodology. Doc. 1850-1 at 10-11 (Hughes Decl. ¶ 20). And, according to Prof. Hughes, he was able to match fewer than 16% of the transactions listed in the Mylan coupon data with the PBM claims data. Id. Thus, defendants assert, Prof. Rosenthal's methodology fails to account for a large number of consumers who paid nothing for the EpiPen using a coupon.

Defendants' criticisms about Prof. Rosenthal's methodology go to the weight of her opinion but not its admissibility. With his analysis, Prof. Hughes recognizes that the methodology allows for some identification of coupon users—even if less than 16%. And, plaintiffs assert, discovery is continuing. So, to the extent discovery produces additional identifying information, Prof. Rosenthal may be able to use her methodology to identify more coupon users. As our Circuit has recognized, Daubert doesn't require the court to decide whether Prof. Rosenthal's opinions are "undisputably correct." Etherton v. Owners Ins. Co., 829 F.3d 1209, 1217 (10th Cir. 2016) (citation and internal quotation marks omitted). Instead, the court must determine whether "the method employed by the expert in reaching the conclusion is scientifically sound and that the opinion is based on facts which sufficiently satisfy Rule 702's reliability requirements." Id. at 1217-18 (citations and internal quotation marks omitted). So, the court's Daubert inquiry should "`focus . . . solely on [the expert's] principles and methodology, not on the conclusions that they generate.'" Id. at 1217 (quoting Daubert, 509 U.S. at 595). On that issue, the court finds that Prof. Rosenthal's methodology for identifying EpiPen coupon users sufficiently is reliable that the court properly may consider her opinions on class certification.

ii. Consumers Who Had Flat Co-Pays

For the second type of consumers—i.e., ones who had flat co-pays—defendants assert that Prof. Rosenthal's methodology cannot determine reliably who falls into this category. Consumers who paid a single, flat co-pay for their EpiPens fall outside the class definition because they would have made the same copayment whether they purchased a branded EpiPen or a generic product. Thus, defendants contend, plaintiffs must show a reliable method for identifying these consumers and excluding them from the class.

Prof. Rosenthal opines that she can identify co-pay consumers reliably using data maintained by PBMs. She explains that, in two other pharmaceutical cases, "[d]eclarations submitted by PBMs . . . specifically identified and confirmed that data are available through the PBMs sufficient not only to identify class members but also to determine whether or not they had a fixed copay or other coinsurance arrangements." Doc. 1711-1 at 15-16 (Rosenthal Reply Report ¶ 20). Prof. Rosenthal suggests that "[t]hese data must be available in order to allow insurers and their PBMs the ability to process prescription claims at the retail level." Id. And, "[t]hese PBM declarations confirm the obvious that these data can be provided if requested." Id. Specifically, Prof. Rosenthal cites one PBM declaration that identifies the types of data "generally maintained in an industry standard format created by the National Council for Prescription Drug Programs [("NCPDP")] . . . ." Id. Prof. Rosenthal also opines that the PBM declarations from these other cases "are entirely consistent with the data produced by multiple PBMs in this matter." Id. at 16 (Rosenthal Reply Report ¶ 21).

Defendants contend, in response, that the Wellbutrin court rejected a similar opinion by Prof. Rosenthal. In re Wellbutrin XL Antitrust Litig., 308 F.R.D. 134, 150-51 (E.D. Pa. 2015). In that case, the court found that the class "ha[d] 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" because its "evidence in support of ascertainability consist[ed] mainly of conclusory statements by its experts that records exist that could be used to ascertain the class and the existence of NCPDP standards." Id. at 151 (citations and internal quotation marks omitted). Wellbutrin held that "[t]his evidence is not enough to show by a preponderance of the evidence that the class is ascertainable." Id. Importantly, the court reached this decision when it decided whether the Wellbutrin plaintiffs had shouldered their burden to establish that class certification was warranted under Fed. R. Civ. P. 23(b)(3).11 Id. But the court didn't exclude Prof. Rosenthal's opinions as unreliable under Daubert.12

In short, Wellbutrin's conclusion doesn't apply here because plaintiffs have made a stronger showing than plaintiffs did there. Prof. Rosenthal's report provides a direct, specific, and principled explanation how she can avoid the Wellbutrin problem. Prof. Rosenthal's opinion here doesn't rely on conclusory statements. Instead, she has offered a reliable method for identifying these consumers using PBM data. Defendants' challenges about whether the PBM data actually is available in this case—as opposed to other cases—go to the weight of her opinions. The court declines to exclude her opinions for this reason.

iii. Consumers Who Paid Cash

Finally, defendants assert that Prof. Rosenthal offers no reliable methodology for identifying the EpiPen consumers who paid cash for their EpiPen purchases. In her Reply Report, Prof. Rosenthal responds to the concerns described in Prof. Hughes's Expert Report about the availability of data for cash payers. Doc. 1711-1 at 29 (Rosenthal Reply Report ¶ 48). Prof. Rosenthal identifies two means to gather information about cash-paying EpiPen consumers. "First, the individuals will have records on the transactions." Id. "Second, pharmacy records will contain information on these transactions . . . ." Id. She points out that "Plaintiffs have in fact produced a large sample of pharmacy records for cash payers in the form of the IQVIA Xponent data." Id. But, she concedes that the IQVIA Xponent data "used in [her] report do not have patient identifiers (IQVIA protects these for good reason)." Id. Nevertheless, Prof. Rosenthal opines that "such identifiers do in fact exist at the dispensing pharmacies and presumably the Court could compel their production if it were deemed necessary." Id.

Again, defendants cite the Wellbutrin case as support for excluding Prof. Rosenthal's methodology for identifying consumers who paid cash for EpiPens. Defendants assert that the Wellbutrin court was "not persuaded by [Prof. Rosenthal's] conclusory statements" that "records at the retail pharmacy and PBM level . . . can be used to identify class members" when she had not "examined or analyzed these pharmaceutical records." In re Wellbutrin XL Antitrust Litig., 308 F.R.D. at 150. But, as already discussed, the Wellbutrin court considered Prof. Rosenthal's opinions in the context of determining whether class certification was appropriate. Id. at 150-51. It didn't exclude her opinion about identifying consumers with pharmaceutical records under Daubert as unreliable. Instead, it found that her opinions didn't support class certification under Rule 23(b)(3).

The court is persuaded that defendants have come forward with legitimate and even substantiated questions about Prof. Rosenthal's opinions for cash payers. But the governing law doesn't use that threshold as the standard for excluding an expert's opinion. Instead, the standard inquires whether Prof. Rosenthal has marshaled "sufficient facts to validate [her opinions] in the eyes of the law . . . ." Pioneer Ctrs., 858 F.3d at 1342 (citation and internal quotation marks omitted). Mindful that "rejection of expert testimony is the exception rather than the rule," Fed. R. Evid. 702 advisory committee's notes to 2000 amendments, the court exercises its discretion and declines to exclude Prof. Rosenthal's methodology for identifying consumers who paid cash for EpiPens. On balance, Prof. Rosenthal offers a reasonably reliable method for identifying these consumers using either their own or pharmacy records. In the end, merits discovery may substantiate Prof. Hughes's view. The task of gathering pharmacy records may prove overwhelming because it requires collecting data from about 65,000 pharmacies, more than 23,000 of which are independent. Doc. 1850 at 16. Also, defendants question whether class members will have records showing both that they purchased an EpiPen and that they paid more than nothing for their purchase (meaning that insurance, coupons, or patient assistance didn't cover the entire price of the products). Doc. 1884 at 11. These arguments test the weight of Prof. Rosenthal's opinion, but they don't preclude its admission from the court's consideration of the class certification motion.13

The court declines to exclude Prof. Rosenthal's ascertainability opinion under Daubert.14

c. Probability Opinion

Next, defendants argue that the court should exclude Prof. Rosenthal's probability opinion. This opinion "estimate[s] the percentage" of putative class members "who might not have suffered an overcharge from the particular misconduct alleged by Plaintiffs." Doc. 1711-1 at 29-30 (Rosenthal Reply Report ¶ 50). Defendants give four reasons for excluding Prof. Rosenthal's probability opinion: (1) the opinion depends on unreasonable assumptions that the record doesn't support; (2) the opinion lacks support in the relevant field and case law; (3) the opinion is incapable of measuring classwide injury; and (4) it is improper rebuttal testimony. The court addresses each of defendants' arguments, in turn, below.

i. Unsupported Assumptions

Defendants argue the court must exclude Prof. Rosenthal's probability opinion as unreliable because it is not "based upon conclusions from established evidentiary facts." United States v. Rice, 52 F.3d 843, 847 (10th Cir. 1995); see also Cochrane v. Schneider Nat'l Carriers, Inc., 980 F.Supp. 374, 378 (D. Kan. 1997) (excluding expert opinion that was "based on unjustified assumptions" and "therefore will not assist the trier of fact"). To support this argument, defendants direct the court to three assumptions that Prof. Rosenthal makes in her analysis.

First, defendants assert that Prof. Rosenthal wrongly assumes the same probability for the rate that each third party payor ("TPP") will reimburse the purchase of a branded EpiPen or 2-Pak. See Doc. 1711-1 at 31-33 (Rosenthal Reply Report ¶¶ 54, 56). Defendants contend her assumption is not supported by the record. For support, they cite portions of Mylan's expert's report that identify differences among TPPs that, the competing expert contends, Prof. Rosenthal failed to consider when forming her opinions. Doc. 1850 at 19. These differences include variations in formularies, benefit design, negotiating power, and rebate payment. Id.

Second, defendants contend that Prof. Rosenthal uses an arbitrary and unfounded estimate of 95% to calculate the rate consumers would have switched from the EpiPen to a generic EAI device. Doc. 1850 at 20; see also Doc. 1711-1 at 30-31 (Rosenthal Reply Report ¶ 53, n.59). Again relying on Mylan's expert's opinion, defendants argue Prof. Rosenthal's generic conversion assumption is just wrong. Doc. 1850 at 20. They assert her assumption is flawed because she didn't consider other evidence of lower generic switching rates, including evidence of the Teva generic's performance and forecasts. Id. Defendants also refute Prof. Rosenthal's assertion that her methodology "can accommodate a range of yardsticks." Doc. 1711-1 at 19 (Rosenthal Reply Report ¶ 27). They contend that any reduction in her generic conversion assumption still doesn't account for the probability that the consumers are brand loyalists who would have continued purchasing the EpiPen even if a generic EAI was available. Doc. 1850 at 21.

Third, defendants assert that Prof. Rosenthal's 2-Pak probability calculations are unreliable. Using a Mylan slide deck from May 2010, Prof. Rosenthal estimates that 65% of patients would have preferred to purchase a 2-Pak instead of a single EpiPen. Doc. 1711-1 at 32-33 (Rosenthal Reply Report ¶ 55, n.61). Defendants argue that her assumption is flawed because it ignores evidence showing that, after May 2010, doctors were recommending and patients were preferring to purchase EpiPens in packages of two.

All of defendants' criticisms about the assumptions Prof. Rosenthal uses to support her probability opinion go to the weight of her opinion testimony, but not its admissibility. Prof. Rosenthal's Reply Report adequately explains why she made the assumptions her probability analysis made and why, she contends, those assumptions are proper under the facts of the case. Doc. 1711-1 at 30-33 (Rosenthal Reply Report ¶¶ 53-56). And, her Reply Report rebuts the criticisms that Mylan's expert makes about those assumptions. Id. at 19-20, 24-25 (Rosenthal Reply Report ¶¶ 27-29, 37-39). Defendants' challenges to those assumptions don't show that Prof. Rosenthal's probability opinion is so unreliable that the court must exclude it. See, e.g., In re Lidoderm Antitrust Litig., No. 14-md-02521-WHO, 2017 WL 679367, at *28 (N.D. Cal. Feb. 21, 2017) (denying motion to exclude expert opinion because it was based on "reasonable assumptions and evidence, and supported by reasoned principles as well as academic scholarship" and while "some of those assumptions [were] disputed," those disputes did not "make [the expert's] reliance on them improper").

Instead, defendants' challenges more properly are presented through the "presentation of contrary evidence" on class certification because these are "the traditional and appropriate means of attacking shaky but admissible evidence." Daubert, 509 U.S. at 596 (citation omitted). And, indeed, defendants have asserted these challenges in their Opposition to the class certification motion. See, e.g., Doc. 1636 at 41-43 (criticizing Prof. Rosenthal's failure to account for consumers who would have purchased a 2-Pak even if a single pack was available), 53-55 (challenging Prof. Rosenthal's assumption that 95% of consumers would have switched to a generic), 60-61 (explaining how brand loyalty and rebates may result in no injury to a consumer).15 The court declines to exclude Prof. Rosenthal's probability opinion based on the assumptions she made to support it.

ii. Unsupported in the Field or Case Law

Next, defendants assert that Prof. Rosenthal's probability opinion is unreliable because her methodology has no support in the field or the case law. Defendants concede that Prof. Rosenthal's Reply Report opines that "health economists routinely use probabilities to estimate the distribution of harms and benefits to populations from a range of policy and clinical interventions." Doc. 1711-1 at 30 (Rosenthal Reply Report ¶ 51). And she cites two articles as support for this assertion. Id. (Rosenthal Reply Report ¶ 51, n.57). Relying on an opinion from Mylan's expert, defendants argue that these two articles don't use the same methodology that Prof. Rosenthal has used. Doc. 1850 at 22. And thus, they contend, plaintiffs can't show that her methodology has support in the field. Plaintiffs respond that Mylan's expert merely tries to distinguish Prof. Rosenthal's methodology from the ones discussed in the two cited articles. But, they contend, Mylan's expert doesn't refute Prof. Rosenthal's basic premise that health economists routinely use probabilities to analyze the distribution of harms and benefits to a population.

Also, defendants assert that no court ever has accepted Prof. Rosenthal's probability analysis. For support, defendants assert that "plaintiffs in one recent case submitted Prof. Rosenthal's probability analysis as the leading piece of evidence that they argued warranted consideration of the First Circuit's denial of class certification. But the district court rejected the analysis and plaintiffs' motion." Doc. 1850 at 23 n.8 (citing Electronic Order, In re Asacol Antitrust Litig., No. 15-cv-12730-DJC (D. Mass. Apr. 11, 2019), ECF No. 772). As plaintiffs correctly point out, the Massachusetts order defendants cite is a three-paragraph text order denying a motion for leave to file a renewed motion for class certification because it "would be futile" and "would involve re-litigation of matters previously decided." Electronic Order, In re Asacol Antitrust Litig., No. 15-cv-12730-DJC (D. Mass. Apr. 11, 2019), ECF No. 772. The text order says nothing about Prof. Rosenthal or that the court is rejecting her analysis. Id.

In contrast, plaintiffs assert that the First Circuit considered a probability analysis "similar" to the one offered here by Prof. Rosenthal, along with other summary judgment evidence, and concluded that a TPP plaintiff had submitted sufficient causation proof to survive summary judgment. Doc. 1860 at 23 (citing In re Celexa & Lexapro Mktg. & Sales Practices Litig., 915 F.3d 1, 13-14 (1st Cir. 2019) (holding district court erred by granting defendant's summary judgment motion)). Specifically, Prof. Rosenthal had opined in Celexa & Lexapro that a defendant pharmaceutical company's "spending on promotions in general correlated positively with sales." In re Celexa & Lexapro Mktg. & Sales Practices Litig., 915 F.3d at 12. She concluded that the defendant's off-label promotions had produced 76% of Celexa prescriptions and 54% of Lexapro prescriptions. Id. at 13. And "Dr. Rosenthal estimated that if [the TPP 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. The First Circuit noted that the record showed the TPP plaintiff likely paid for more than five prescriptions, and "[s]o the odds that [the TPP plaintiff] was not harmed if the drugs were, indeed, ineffective was likely infinitesimal (assuming the prescriptions were independent of one another)." Id. Thus, the First Circuit concluded, the TPP plaintiff "could establish causation and injury at least for any TPP who paid for more than a handful of different patients' prescriptions." Id. at 14.

The First Circuit also recognized that "there is room for reasonable disagreement on the merits of Dr. Rosenthal['s]" assumptions, but it recognized that a jury could accept them "as reasonable." Id. at 13. And while the court noted that the district court had not yet conducted a Daubert analysis of the experts' opinions, id., the Circuit nevertheless didn't find any flaws with Dr. Rosenthal's analysis. To the contrary, the court considered her opinion when reversing the summary judgment order. Id. at 13-14.

After examining the parties' competing arguments, the court is unpersuaded by defendants' argument that Prof. Rosenthal's methodology lacks sufficient support in the field to require excluding her opinions on class certification. Defendants' disagreements about her methodology go to the weight of her opinion, which the court can consider when deciding the class certification motion.16

iii. Incapable of Measuring Classwide Injury

Defendants next argue that Prof. Rosenthal's probability opinion is unreliable because it is not capable of measuring classwide injury. Defendants assert that plaintiffs have the burden to establish classwide injury which, they contend, requires a showing of actual injury to all or substantially all class members. Instead, defendants assert, Prof. Rosenthal's probability opinion simply estimates the percentage of putative class members "who might not have suffered an overcharge from the particular misconduct alleged by Plaintiffs." Doc. 1711-1 at 29-30 (Rosenthal Reply Report ¶ 50) (emphasis added). Thus, defendants contend, Prof. Rosenthal's opinion is not reliable because it doesn't establish the actual percentage of consumers who, in fact, sustained an injury.

Plaintiffs respond that they aren't required to establish the percentage of consumers who actually sustained injuries to prevail on class certification. Instead, they contend, class certification just requires that plaintiffs show "there is no basis to conclude that a `great number' of members could not have been harmed as alleged by plaintiffs." In re Syngenta AG MIR 162 Corn Litig., 2016 WL 5371856, at *4 (quoting Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 824 (7th Cir. 2012)). And, plaintiffs argue, Prof. Rosenthal's analysis sufficiently demonstrates that few consumers could not have sustained harm from defendants' alleged misconduct. Doc. 1711-1 at 29-33 (Rosenthal Reply Report ¶¶ 50-56). In their Reply, defendants argue that this isn't the correct standard for determining predominance on a Rule 23 motion. And, even if it were, Prof. Rosenthal's opinion doesn't satisfy that standard.

These arguments go to the merits of the class certification motion. The court will decide when ruling the class certification motion whether plaintiffs have shouldered their burden to establish classwide injury sufficient to certify a class under Rule 23. And it will consider whether Prof. Rosenthal's expert opinions constitute sufficient evidence to support class certification under the governing case law.17 But defendants' arguments here don't establish that Prof. Rosenthal's opinion about the percentage of uninjured class members is so unreliable that the court must exclude it under Daubert.

Also, defendants argue, Prof. Rosenthal's model cannot assess which class members, if any, were harmed by each individual theory of liability. Defendants assert that plaintiffs are required to "tie each theory of antitrust impact to an exact calculation of damages." Comcast Corp. v. Behrend, 569 U.S. 27, 37 (2013). Defendants assert that Prof. Rosenthal's analysis flunks this test because she shows "how combining multiple sources of impact reduces the probability of a class member remaining uninjured." Doc. 1711-1 at 34 (Rosenthal Reply Report ¶ 58). Plaintiffs respond that Prof. Rosenthal properly considered multiple theories of harm in her analysis because class plaintiffs have proposed certifying a class that includes multiple theories of liability, including the alleged generic delay and improper switch to the 2-Pak.

The court recognizes that Comcast differs from this case. In Comcast, plaintiffs alleged four different theories of antitrust impact, but the district court "accepted [just one of the theories] of antitrust impact as capable of classwide proof and rejected the rest." 569 U.S. at 31. The Supreme Court held that the district court erred by certifying a class based on an expert's regression model that "assumed the validity of all four theories of antitrust impact initially advanced" by plaintiffs, even though the district court had rejected three of those theories. Id. at 36. Thus, the Court found that the model "falls far short of establishing that damages are capable of measurement on a classwide basis," and "[w]ithout presenting another methodology," plaintiffs could not prove "Rule 23(b)(3) predominance." Id. at 34. Instead, the Court reasoned that "[q]uestions of individual damage calculations will inevitability overwhelm questions common to the class." Id.

That's not the posture of this case. Plaintiffs have asserted several theories of liability. And, at this stage, the court hasn't excluded any of the theories that Prof. Rosenthal uses in her analysis. Plaintiffs assert that Prof. Rosenthal's theory is consistent with Comcast because, in that case, the Supreme Court explained that "a model purporting to serve as evidence of damages in this class action must measure only those damages attributable to that theory." Id. at 35. And, because plaintiffs have identified several theories that purportedly injured the putative class members—thus producing damages—it was proper for Prof. Rosenthal to combine the theories in her analysis. See Butler v. Sears, Roebuck & Co., 727 F.3d 796, 800 (7th Cir. 2013) (recognizing that "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"). Also, plaintiffs distinguish Prof. Rosenthal's probability opinion from the model excluded in Comcast because her probability opinion isn't a damages measurement. Prof. Rosenthal provided her damages opinions elsewhere in her expert report, and there, she separated her damage opinions among the various theories.

The court agrees that defendants' arguments about Comcast don't preclude the court from considering Prof. Rosenthal's probability opinion on class certification.

iv. Improper Rebuttal Testimony

Last, defendants assert that Prof. Rosenthal's probability methodology is improper rebuttal testimony. Similar to their arguments directed at the ascertainability opinion, above, defendants argue that Prof. Rosenthal should have offered a model to assess the possibility of uninjured class members in her opening report. Thus, defendants contend, the probability opinion she offers in her Reply Report constitutes improper rebuttal testimony.

Plaintiffs respond that, after Prof. Rosenthal submitted her initial report, defendants filed their Opposition to the class certification motion. It argued that plaintiffs' proposed class definitions contain too many uninjured class members. And one of Mylan's experts (Dr. John H. Johnson, IV) issued a report criticizing the conclusions announced in Prof. Rosenthal's initial report. Plaintiffs assert that Prof. Rosenthal's Reply Report properly responds to the arguments raised by defendants' Opposition and Mylan's expert. The court agrees with them.

Prof. Rosenthal's Reply Report explicitly recites that she offers her rebuttal opinions in response to Dr. Johnson's criticisms about her analysis, including his assertions that Prof. Rosenthal's analysis fails to account for uninjured class members. See Doc. 1711-1 at 3; see also id. at 29-37 (Rosenthal Reply Report ¶¶ 50-60). Also, the Reply Report cites specific portions of Dr. Johnson's report and gives Prof. Rosenthal's reasons for disagreeing with his conclusions. See id. at 20-24 (Rosenthal Reply Report ¶¶ 30-36). Exercising its discretion, the court declines to exclude Prof. Rosenthal's probability opinion as improper rebuttal testimony.

For all these reasons, the court rejects defendants' request that the court exclude Prof. Rosenthal's probability opinion under Daubert.18

d. Professor Rosenthal's Opinions that Rely On Professor Elhauge's Opinions

Finally, defendants assert that the court should exclude Prof. Rosenthal's opinions that rely on the opinions of another one of the putative class plaintiffs' experts. Prof. Rosenthal reaches some of her conclusions by relying on opinions offered by Professor Einer Elhauge. See, e.g., Doc. 1500-3 at 33 (Rosenthal Expert Report ¶ 79), Doc. 1711-1 at 18-19, 37 (Rosenthal Reply Report ¶¶ 26, 61). Defendants contend that Prof. Elhauge's opinions are unreliable and inadmissible under Daubert. And, by separate motion, they ask the court to exclude his opinions from the court's consideration of the class certification motion. See Doc. 1604.

The court addresses defendants' motion to exclude Prof. Elhauge's opinions in the next section of this Order. As discussed below, the court rejects defendants' premise—that the court should exclude any of Prof. Elhauge's opinions that Prof. Rosenthal has relied on when forming her opinions. Consequently, the court rejects defendants' final argument for excluding Prof. Rosenthal's opinions.

For the reasons explained, the court denies defendants' motion to exclude Prof. Rosenthal's expert opinions.

2. Motion to Exclude Expert Opinions of Professor Einer Elhauge

Defendants next ask the court to exclude the expert opinions of Prof. Einer Elhauge. Prof. Elhauge opines that Mylan entered into exclusionary contracts with PBMs and its EpiPens4Schools program which harmed the entire putative class by foreclosing a part of the market to Auvi-Q (a rival EAI device manufactured and sold by Sanofi-Aventis U.S. LLC ("Sanofi")). Doc. 1500-2 at 6 (Elhauge Expert Report ¶ 3). According to Prof. Elhauge, Mylan's market foreclosure, in turn, decreased Auvi-Q sales and prevented Auvi-Q from competing with Mylan, thereby reducing Mylan's incentives to lower its prices across the market and across the putative class.19 Id.

Defendants assert three arguments to support their request that the court exclude Prof. Elhauge's opinions. First, defendants contend that Prof. Elhauge is not qualified to render opinions based on econometric models. Second, defendants argue that Prof. Elhauge's econometric analyses are either unreliable or no longer relevant to the case. And finally, defendants assert that Prof. Elhauge's opinions rely on improper and unreliable rebuttal testimony that doesn't suffice to cure the defects in his initial analysis. The court addresses each argument, below.

a. Professor Elhauge's Qualifications

Defendants assert that Prof. Elhauge is not qualified to render his opinions because he is neither an economist nor an econometrician (an economist with highly specialized training in mathematics and statistics who builds economic models). Plaintiffs recognize that Prof. Elhauge is not an economist or an econometrician, but, they argue, Prof. Elhauge is well-qualified to testify as an expert in antitrust economics. Also, plaintiffs assert, Prof. Elhauge has extensive expertise applying economics and econometrics to antitrust issues. Thus, they argue, he is qualified to render the opinions he offers in this case. The court agrees with plaintiffs. Prof. Elhauge is qualified to opine on the topics that he offers here.

Prof. Elhauge is the Petrie Professor of Law at Harvard Law School, where he teaches antitrust law, health policy, and other subjects. Doc. 1500-2 at 6-7 (Elhauge Expert Report ¶ 4). Prof. Elhauge has authored or coauthored several books about antitrust law and economics. Id. Also, he has authored many articles involving economic analysis of antitrust and other legal issues, including articles on monopolization, bundled discounts, loyalty discounts, and reverse payment settlements. Id. Since 1991, Prof. Elhauge has written more than 40 legal articles that leading law reviews and bar journals have published. Id. at 81-85. Also, Prof. Elhauge has testified as an expert witness on antitrust economics in dozens of federal cases, as well as before Congress, arbitration panels, and international competition agencies. Id. at 7 (Elhauge Expert Report ¶ 5). His expert testimony has included testimony about reverse payment settlements, other horizontal agreements, vertical agreements, mergers, monopolization and exclusionary conduct, price discrimination, health economics, patent economics, and contract economics. Id. Courts have referred to Prof. Elhauge as a "`"highly qualified antitrust titan."'" See, e.g., Castro v. Sanofi Pasteur Inc., 134 F.Supp.3d 820, 830 (D.N.J. 2015) (quoting In re Mushroom Direct Purchaser Antitrust Litig., No. 06-0620, 2015 WL 5767415, at *4 (E.D. Pa. July 29, 2015) (quoting Natchitoches Par. Hosp. Serv. Dist. v. Tyco Int'l, Ltd., 247 F.R.D. 253, 273 (D. Mass. 2008))).

In his deposition testimony, Prof. Elhauge conceded that he is "not an expert in econometrics . . . ." Doc. 1861-2 at 12 (Elhauge Dep. 29:8-23). Also, he testified that he doesn't come up with "new, novel econometric methods." Id. But, Prof. Elhauge testified that he has "expertise . . . applying econometrics to antitrust issues." Id. And he considers himself an expert in "the application of econometrics to antitrust issues." Id. at 11-12 (Elhauge Dep. 28:25-29:3). That's what Prof. Elhauge says he did in this case—i.e., apply econometric methods to antitrust issues by performing regression analyses to estimate the effects of Mylan's exclusionary contracts with PBMs.20

At least two courts have found that Prof. Elhauge is qualified to perform this type of expert analysis, even though he is not an econometrician. See, e.g., In re Mushroom Direct Purchaser Antitrust Litig., No. 06-0620, 2015 WL 5767415, at *5 (E.D. Pa. July 29, 2015) (rejecting argument that Prof. Elhauge was not qualified to provide expert testimony, recognizing that "[t]here is no serious question that Prof. Elhauge possess[es] skill or knowledge in conducting regression analysis in antitrust cases that is greater than the average layman even if it were conceded that his expertise with multiple regression analysis as an econometric tool in general is not as deep as that of an econometrician," and concluding that his "lack of econometric/economic credentials" affected the weight but not the admissibility of his testimony (citations and internal quotation marks omitted)); see also Natchitoches Par. Hosp. Serv. Dist. v. Tyco Int'l, Ltd., No. 05-12024 PBS, 2009 WL 3053855, at *3 (D. Mass. Sept. 21, 2009) (finding that Prof. Elhauge was qualified to provide expert testimony because the court-appointed expert in econometrics "pinpointed no methodological flaws or technical errors in the econometric analysis that Professor Elhauge presented" and concluding that his "lack of econometric/economic credentials affects the weight, not the admissibility, of Professor Elhauge's testimony").

And, in a third case, the court found Prof. Elhauge's econometric analysis sufficiently reliable to deny the defendant's motion to exclude under Daubert. Castro v. Sanofi Pasteur Inc., 134 F.Supp.3d 820, 831-42 (D.N.J. 2015). Although the defendant there never challenged Prof. Elhauge's qualifications, the court summarized Prof. Elhauge's educational background, professional experience, publications, and expert testimony; and the court concluded that he is "eminently qualified" to provide expert opinions in the form of econometric analysis. Id. at 830.

For the same reasons, the court finds that Prof. Elhauge is qualified to render his expert opinions in this case.

b. The Reliability and Relevancy of Professor Elhauge's Econometric Analyses

Next, defendants argue that the court should exclude Prof. Elhauge's opinions about Auvi-Q market foreclosure and delayed generic entry because, defendants contend, Prof. Elhauge used flawed econometric analyses to support his opinions.

i. Auvi-Q Foreclosure Analysis

Prof. Elhauge opines that "Mylan's exclusionary contracts with PBMs . . . foreclose[ed] a part of the market to Auvi-Q . . . ." Doc. 1500-2 at 6 (Elhauge Expert Report ¶ 3). Defendants assert that this opinion is unreliable because Prof. Elhauge bases his conclusion on a flawed regression analysis that failed to account for "serial correlation." Doc. 1854 at 8, 14-16.

"Multiple-regression analysis is a statistical tool used to determine the relationship between an unknown variable (the `dependent' variable) and one or more `independent' variables that are thought to impact the dependent variable." In re Urethane Antitrust Litig., 768 F.3d 1245, 1260 (10th Cir. 2014) (citing Saks, Michael J., et al., Reference Manual on Scientific Evidence 179, 181 (2d ed. 2000)); see also Daniel L. Rubinfeld, "Reference Guide on Multiple Regression," in Reference Manual on Scientific Evidence 303, 305 (3d ed. 2011), https://www.fjc.gov/sites/default/files/2015/SciMan3D01.pdf (hereinafter "Reference Guide on Multiple Regression") (explaining that regression analysis "involves a variable to be explained— called the dependent variable—and additional explanatory variables that are thought to produce or be associated with changes in the dependent variable").

In Prof. Elhauge's regression model, the dependent variable is Auvi-Q's market share, and the two independent variables are: (1) the percentage of lives subject to PBM coverage restrictions on Auvi-Q, and (2) the price ratio of Auvi-Q and EpiPen. Doc. 1500-2 at 57-58 (Elhauge Expert Report ¶ 112). Thus, Prof. Elhauge's model tests whether either of the two independent variables affects the size of Auvi-Q's market share (the dependent variable).

Defendants contend, however, Prof. Elhauge's model is unreliable because it relies on underlying data that is "serially correlated." Specifically, Prof. Elhauge relied on Auvi-Q market share data for six PBMs over a 34-month period, thus producing 204 data points for Prof. Elhauge's regression analysis. Doc. 1500-2 at 56-57 (Elhauge Expert Report ¶ 110). But, defendants contend, these data points are not independent sample results. Instead, defendants argue, the data points are "serially correlated" because PBMs generally decide formulary placement on an annual basis. So, if a PBM had chosen to restrict Auvi-Q in January, it generally would restrict Auvi-Q in every month for the rest of the year. Defendants argue that Prof. Elhauge's regression model fails to control for the fact that PBMs generally don't adjust their formularies one month at time. Thus, defendants contend, his data points are "serially correlated," meaning that they aren't independent and thus make the variables appear far more correlated than they actually are. See Reference Guide on Multiple Regression, at 355 (defining "serial correlation" as "[t]he correlation of the values of regression errors of time"). And, according to Mylan's expert, when one controls for this serial correlation embedded in Prof. Elhauge's model, the model isn't capable of generating a statistically significant result. Thus, defendants contend, Prof. Elhauge's model is unreliable.

Plaintiffs respond by citing Prof. Elhauge's Expert Reply Report. In that Report, Prof. Elhauge recognizes that serial correlation is present in his foreclosure regression but, he explains, serial correlation "does not make the regression biased in favor of finding a significant foreclosing effect." Doc. 1711-2 at 40 (Elhauge Expert Reply Report ¶ 81). Instead, Prof. Elhauge asserts, the presence of serial correlation only affects the statistical significance of the results. Id. And, Prof. Elhauge explains his regression model does produce statistically significant results when one uses "clustered standard errors" to control for serial correlation. Id.; see also id. at 40-42 (Elhauge Expert Reply Report ¶¶ 80-84). Although "using clustered standard errors reduces the statistical power of the regression by effectively grouping all observations at the same PBM," one can "obtain statistical significance" by "includ[ing] more PBMs." Id. at 40 (Elhauge Expert Reply Report ¶ 81). Also, one can reduce serial correlation significantly by "adding PBM fixed effects to the regression." Id. at 40-41 (Elhauge Expert Reply Report ¶ 82).

The court finds that Prof. Elhauge's Reply Report sufficiently explains how to account for serial correlation in his model. Defendants' criticisms about his regression analysis— particularly the criticism asserted by Mylan's expert—go to the weight of Prof. Elhauge's opinions, but not their admissibility. See, e.g., Southwire Co. v. J.P. Morgan Chase & Co., 528 F.Supp.2d 908, 929 (W.D. Wis. 2007) (refusing to exclude expert opinion based on an econometric model because the expert "accounted for the existence of serial correlation and the need to correct for it" and thus "his method is not so inherently flawed as to preclude him from establishing whether defendants' alleged wrongdoing had any effect on the price of copper" and whether the expert "erred in applying his model and whether he failed to properly correct for serial correlation are jury questions not amenable to resolution on summary judgment"). Defendants can challenge—and have challenged—Prof. Elhauge's opinions about class certification by offering competing expert opinions.21 See Doc. 1636 at 49, 49 n.43 (Mylan Defs.' Opp'n to Class Pls.' Mot. for Class Certification). But the court declines to find that Prof. Elhauge's analysis is inherently unreliable such that the court should exclude it. The court thus denies defendants' Motion to Exclude Prof. Elhauge's opinions about Auvi-Q market foreclosure based on PBM exclusionary contracts.

ii. EpiPens4Schools Program Analysis

Prof. Elhauge also opines that "Mylan's exclusionary contracts with . . . its EpiPens4Schools program . . . foreclose[ed] a part of the market to Auvi-Q . . . ." Doc. 1500-2 at 6 (Elhauge Expert Report ¶ 3). Defendants assert that the court should exclude Prof. Elhauge's opinions for two reasons. First, defendants argue that Prof. Elhauge bases his opinion on an inaccurate assumption—i.e., that Mylan imposed exclusivity requirements in exchange for free EpiPens for schools through its EpiPens4Schools program. Defendants assert that the factual record doesn't support this assumption because, from 2013 to 2015, 96% of schools participating in the EpiPens4Schools Program were not subject to any exclusivity requirements. Defendants contend that plaintiffs' discovery of these facts has compelled them to abandon "any independent causally related damages based on the EpiPen4Schools program." Doc. 1590-3 at 2 (May 14, 2019 email from plaintiffs' liaison counsel).

Second, defendants contend, the court should exclude Prof. Elhauge's opinions because they are no longer relevant now that plaintiffs have abandoned this theory in their case. Plaintiffs concede that Prof. Elhauge's opinions about the EpiPens4Schools program no longer matter because plaintiffs aren't pursuing this theory. Doc. 1861 at 26. And, they explain, Prof. Elhauge's Expert Reply Report makes clear that he already has withdrawn his opinion about the EpiPens4Schools program. Doc. 1711-2 at 46 (Elhauge Expert Reply Report ¶ 90) (assuming "no foreclosing impact from the EpiPens4Schools program because [Prof. Elhauge] understand[s] that plaintiffs have abandoned that claim"). Thus, plaintiffs assert, the court shouldn't exclude an already-withdrawn opinion, but instead, the court should find that defendants' arguments are moot.

Defendants disagree that the issue is moot. Prof. Elhauge offered an opinion about the EpiPens4Schools program in his initial report. And, defendants contend, the correct procedural device is for the court to grant defendants' motion to exclude the opinion as uncontested so that no question exists about the status of this theory in further proceedings before this court or on appeal. The court agrees with defendants. Because plaintiffs concede that they have abandoned an independent claim for damages based on the EpiPen4Schools program, the court excludes Prof. Elhauge's opinions on this subject from its consideration at the certification stage because they no longer are relevant to this case. See Fertik v. Stevenson, No. 12-10795-PBS, 2016 WL 4148193, at *3 (D. Mass. Aug. 4, 2016) (prohibiting an expert from testifying about a "now-abandoned theory").

iii. Delayed Generic Entry Analysis

Finally, defendants seek to exclude Prof. Elhauge's opinions about delayed generic entry because, they contend, his opinions aren't reliable and have no probative value. Prof. Elhauge opines that "reverse payments in patent settlements . . . injure competition across the market (and thus across the class) when they result in generic entry dates that are later than the entry date that would have occurred without that payment in a no-payment settlement." Doc. 1500-2 at 5 (Elhauge Expert Report ¶ 2). Using various data points, Prof. Elhauge has developed a model that, he opines, can "determine the generic entry date in a no-payment settlement that would have been economically rational given [the parties'] actual bargaining power." Id. Specifically, Prof. Elhauge "illustrate[s] this classwide methodology, showing that with a $100 million reverse payment amount, the no-payment settlement entry date that would have been economically rational given their actual bargaining power would have been March 13, 2014, some 15.3 months earlier than the entry date in the actual settlement." Id. Defendants contend that Prof. Elhauge's opinions about delayed generic entry are unreliable for several reasons.

First, defendants argue, Prof. Elhauge's opinions are flawed because he predicts an entry date for a Teva generic EAI occurring before Teva even had secured FDA approval for its generic product. The FDA didn't approve the Teva generic until August 16, 2018. Doc. 1854 at 18 n.11 (citing FDA, FDA Approves First Generic Version of EpiPen (Aug. 16, 2018), https://www.fda.gov/news-events/press-announcements/fda-approves-first-generic-version-epipen). Defendants thus contend that no settlement could have prevented Teva from entering the market before the date when the FDA approved Teva's introduction of its EAI device.

The putative class plaintiffs respond that Teva's actual market entry date is irrelevant to Prof. Elhauge's analysis. Prof. Elhauge explained how he used factual data from the record to determine Teva's expectations for a generic entry date. Doc. 1500-2 at 39-40, 43 (Elhauge Expert Report ¶¶ 71, 78). And, Prof. Elhauge described how his analysis "addressed only the economically rational entry date in a but-for settlement agreement" and "expressly disclaimed offering any opinion on whether Teva would have actually been able to enter on that settlement entry date given any obstacles, such as the need for FDA approval." Doc. 1711-2 at 23 (Elhauge Reply Report ¶ 45) (emphasis in original). Prof. Elhauge further explained "the fact that after the settlement Teva may have actually struggled to obtain FDA approval does not bear on what the entry date would have been in the but-for settlement, which can only be affected by expectations at the time of settlement." Id. at 24 (Elhauge Reply Report ¶ 47). Prof. Elhauge recognized that FDA approval may affect plaintiffs' ability to prove "actual causation" if Teva struggled to secure FDA approval for reasons not caused by anticompetitive conduct. Id. But, Prof. Elhauge "simply stayed within [his] assignment of analyzing the entry date in a but-for settlement that the parties would have rationally agreed to," and he "[left] it to plaintiffs to establish through other evidence whether in the but-for world Teva would have actually obtained FDA approval and been able to enter by that but-for settlement date." Id. Prof. Elhauge asserts that "[s]taying within [his] assignment does nothing to undermine the reliability or classwide nature of the methodology [he] used to determine the economically rational entry date in the but-for settlement." Id. The court agrees.

Defendants also contend that Prof. Elhauge's model is flawed because he testified that the actual entry day and actual alleged payment don't matter, but instead, they are "illustrative" placeholders to show that one can apply this model to any set of facts. Doc. 1605-1 at 10 (Elhauge Dep. 64:16-25). But, defendants argue, the model can't demonstrate classwide harm unless it calculates an entry date before June 22, 2015—the date of the Teva settlement agreement. Defendants contend that, if Teva wasn't entering the market until after June 22, 2015 anyway, then defendants couldn't have caused any delay in Teva's generic entry because it entered the settlement before then. For this reason, defendants contend that Prof. Elhauge's model is flawed.

The putative class plaintiffs expose the weakness of this argument. As plaintiffs explain, to prove causation—i.e., that defendants' conduct delayed the entry of a Teva generic—plaintiffs just need to show that Teva would have entered the market sooner than it did but for the reverse payment settlement. For example, if plaintiffs prove that, but for a reverse payment settlement, Teva would have entered the market on June 23, 2015—the day after the settlement—plaintiffs still will have shown that the reverse payment settlement delayed generic entry by more than three years. Thus, defendants' argument on this front doesn't persuade the court that Prof. Elhauge's model is unreliable simply because it may calculate an entry date following the date of the settlement agreement.

Second, defendants assert that Prof. Elhauge's opinions are unreliable because his analysis simply assumes that defendants made a $100 million reverse payment settlement to Teva, but he has no factual basis for this assumption. The court disagrees. Prof. Elhauge has explained why he used the $100 million reverse payment settlement in his calculation. Prof. Elhauge says that he used a 15% patent strength estimate and the parties' profit projections to calculate "that reverse payment must have been at least $47.8 million to get Teva to agree to the actual settlement." Doc. 1711-2 at 19 (Elhauge Reply Report ¶ 37); see also Doc. 1500-2 at 43-44 (Elhauge Expert Report ¶ 80); Doc. 1497-2 at 43-44 (Sealed Elhauge Expert Report ¶ 80). Prof. Elhauge then "used a hypothetical reverse payment amount of $100 million to illustrate how [his] methodology would apply and produce a predicted but-for settlement entry date using classwide evidence and a classwide method." Doc. 1711-2 at 19 (Elhauge Reply Report ¶ 37). Prof. Elhauge explained that he did so "because there is still a significant amount of non-coordinated discovery remaining and because the point at this stage was simply to show a reliable classwide methodology, which can be illustrated by a reverse payment amount of $100 million as well as any other." Id. And, he explained, his model can estimate the but-for entry date for any given settlement amount ranging from $47.8 to $446 million. Doc. 1500-2 at 45-46 (Elhauge Expert Report ¶ 85, fig.12). Prof. Elhauge sufficiently has explained why he used the $100 million reverse payment settlement in his model. Defendants' argument on this point doesn't show that Prof. Elhauge's use of this data point at this stage for illustrative purposes renders his opinions unreliable.

Third, defendants argue Prof. Elhauge's analysis is flawed because he assumed a patent strength of 15% (meaning that Teva had a 85% chance of prevailing in its lawsuit challenging the EpiPen patent). Defendants assert that this assumption conflicts with the opinion of plaintiffs' patent litigation expert (Prof. Andrew Torrance) who opines that the EpiPen patent strength was as high as 30%. But, Prof. Elhauge explains why he chose 15% patent strength for his model. Prof. Elhauge recognized that "any patent strength from 0-30% was equally possible given Professor Torrance's opinion." Doc. 1711-2 at 19 (Elhauge Reply Report ¶ 36). "But if any patent strength from 0-30% is equally possible," Prof. Elhauge concluded "that means that the expected patent strength was 15%, and it is the overall expectation that drives the settlement analysis." Id. (emphasis in original). Thus, Prof. Elhauge asserts "15% is the appropriate estimate to use for an expected patent strength." Id. Prof. Elhauge also explains that his model can estimate delay across a range of assumed patent strengths. Doc. 1500-2 at 46-47 (Elhauge Expert Report ¶ 86). And, his Report includes a Figure showing "the effect of varying the patent strength up to as high as it could possibly have been while still being consistent with Mylan having rationally accepted the actual settlement and paid the illustrative $100 million reverse payment amount." Id.; see also id. (Elhauge Expert Report fig.13). Defendants criticize this model because, they contend, it ignores "the other equally likely values—including values that result in no generic delay." Doc. 1886 at 11 (citation and internal quotation marks omitted). But the court finds that this argument goes to the weight of Prof. Elhauge's opinion, not its admissibility.

Finally, defendants assert that Prof. Elhauge's analysis is based on an unproven model, never accepted by any court as reliable. But, as plaintiffs point out, at least three courts have approved Prof. Elhauge's methodology in pay-for-delay cases. See, e.g., In re Namenda Direct Purchaser Antitrust Litig., 331 F.Supp.3d 152, 174 (S.D.N.Y. 2018) (denying motion to exclude Prof. Elhauge's opinion that "it would have been economically rational for both parties to enter into a no-payment settlement in a but-for world by specific dates" and rejecting defendants' arguments that his opinions "are speculative, internally inconsistent, and contradicted by the evidence" because "Defendants' disagreements with Plaintiffs' expert are appropriate subjects for cross-examination"); In re Androgel Antitrust Litig. (No. II), No. 1:09-MD-2084-TWT, 2018 WL 2984873, at *17 (N.D. Ga. June 14, 2018) (concluding that Prof. Elhauge's reverse payment settlement opinion, as well as other evidence, sufficiently supported plaintiffs' alternative settlement theory to survive defendants' summary judgment motion and noting that "[a]ny criticism the Defendants have of the experts' methodologies or conclusions are best handled through cross-examination and the production of contrary evidence"); United Food & Commercial Workers Local 1776 & Participating Emp'rs Health & Welfare Fund v. Teikoku Pharma USA, 296 F.Supp.3d 1142, 1186-88 (N.D. Cal. 2017) (denying motion to exclude Prof. Elhauge's opinions in an antitrust lawsuit involving Lidoderm patches because, though "defendants [were] correct that Elhauge could point to no other case where an expert has employed the exact model he employs here, both the components of his model (estimating parties' bargaining strengths and expectations of patent strength) and the assumptions that go with it (the parties' own pre-settlement forecasts) are consistent with accepted economic theory and well-established principles"). See also Doc. 1648-1 at 8 (Elhauge Dep. 167:4-19) (explaining that Prof. Elhauge has used the same methodology in other cases to evaluate a but-for settlement date, including the Namenda, Lidoderm, and Androgel cases).

Defendants also contend that Prof. Elhauge's model is unreliable because it "would render class certification essentially `automatic' in every `pay for delay' case, regardless of the actual facts." Doc. 1854 at 9. At least two courts have rejected this very argument. See, e.g., In re Namenda Direct Purchaser Antitrust Litig., 331 F. Supp. 3d at 173 (rejecting defendants' argument that "Professor Elhauge presumes that any payment from a patentee to a prospective generic, regardless of its size or the reasons for making it, causes anticompetitive delay" because it was "a mischaracterization of [his] analysis"); In re Androgel Antitrust Litig., 2018 WL 2984873, at *17 (rejecting defendants' argument that Prof. Elhauge's "model[ ] assume[s] that a reverse payment always causes delay" and concluding that his "model[ ] assume[s] nothing of the sort" but, instead, Prof. Elhauge "use[d] [his] experience and knowledge in the field to conclude that reverse payments cause delay, and [he] confirm[ed] that conclusion in [his] model[ ] addressing the specific settlement at issue in this case").

Similarly here, Prof. Elhauge has used his experience and knowledge in the field to conclude that reverse payments cause generic delay. Doc. 1500-2 at 5 (Elhauge Expert Report ¶

2). Then, using "the parties' actual settlement, profit projections, and estimates about the patent litigation and reverse payment value," Prof. Elhauge "calculate[d] their actual bargaining strength and then use[d] that information to determine the generic entry date in a no-payment settlement that would have been economically rational given their actual bargaining power." Id. Defendants' criticisms of his methodology go to the weight the court should give his opinions, but they don't undermine the reliability of his opinions sufficient to exclude them. The court thus declines to exclude Prof. Elhauge's opinions about delayed generic entry as unreliable.

c. The Reliability of Professor Elhauge's Rebuttal Testimony

Last, defendants' motion argues that the court should exclude analyses that Prof. Elhauge presented in his Reply Report for two reasons. First, defendants argue, Prof. Elhauge's Reply Report contains improper rebuttal testimony. And second, defendants assert that the analyses in the Reply Report are unreliable.

i. Rebuttal Testimony

Defendants assert that Prof. Elhauge's Reply Report contains two "new" analyses that he should have disclosed in his initial report. Doc. 1854 at 20. According to defendants, these analyses are improper rebuttal testimony. Plaintiffs respond that Prof. Elhauge doesn't offer any new analyses in his Reply Report. Instead, plaintiffs contend, his Reply Report properly responds to the criticisms that Mylan's expert made about Prof. Elhauge's regression analyses. Thus, plaintiffs argue, the analyses in his Reply Report are proper rebuttal testimony.

As already discussed, "[r]ebuttal evidence is evidence which attempts to `disprove or contradict' the evidence to which it is contrasted." Tanberg v. Sholtis, 401 F.3d 1151, 1166 (10th Cir. 2005) (quoting Black's Law Dictionary 579 (7th ed. 1999)). "When a party opens the door to a topic, the admission of rebuttal evidence on that topic becomes permissible." Id. (citation omitted). "Permissible does not mean mandatory, however; the decision to admit or exclude rebuttal testimony remains within the trial court's sound discretion." Id. (citation omitted).

Defendants take issue with various modifications that Prof. Elhauge made to his regression analysis. Doc. 1711-2 at 36 (Elhauge Reply Report ¶ 70). Defendants assert that these modifications are substantive changes that produce entirely new damage figures, as depicted in a series of 24 new tables. Id. at 45-64. But, as Prof. Elhauge's Reply Report makes clear, he made the modifications to his regression analysis specifically in response to Mylan's expert's critiques. See id. at 36 (Elhauge Reply Report ¶ 70) (summarizing the criticisms made by Mylan's expert about Prof. Elhauge's model and explaining that "[n]one of these critiques are valid, but the results of the regression are similar even if one modifies the regression to address his critiques."). Prof. Elhauge's modifications thus rebut Mylan's expert's testimony on the same subject matter. So, the court finds, this modified regression is proper rebuttal testimony.

Also, defendants object to the analysis in Prof. Elhauge's Reply Report that purports to separate damages calculations for third party payors ("TPPs") and cash end-payors. Again, Prof. Elhauge offers this opinion in his Reply Report to rebut Mylan's expert's opinion that his method "cannot determine how overcharges differ between third party payors (TPPs) and cash end-payors." Doc. 1711-2 at 69 (Elhauge Reply Report ¶ 103). And, Prof. Elhauge explains, he calculated separate damages using Mylan's rebate data. Id. (Elhauge Reply Report ¶ 105). Plaintiffs assert that defendants did not produce this rebate data until after Prof. Elhauge completed his initial Report and his deposition. Defendants concede that "Plaintiffs used data that Mylan produced after Prof. Elhauge's initial Report was submitted" but, defendants contend, "those data were not required to perform Prof. Elhauge's analyses." Doc. 1886 at 12 n.11. Defendants assert that "Mylan had already produced the bulk of its data well in advance of Prof. Elhauge's initial Report, and those data were equally amenable to use in his analysis." Id. The court declines to accept defendants' conclusory assertion that Prof. Elhauge could have performed his regression with data Mylan already had produced. Because the parties agree that Prof. Elhauge relied on data that Mylan produced after he completed his initial Report and sat for his deposition,22 the court refuses to find that Prof. Elhauge's analysis is improper rebuttal testimony.

In sum, the court rejects defendants' arguments that Prof. Elhauge's Reply Report offers improper rebuttal testimony.23

ii. Reliability of Reply Report's Analyses

Next, defendants assert, the analyses in Prof. Elhauge's Reply Report are unreliable. They assert three reasons why the court should exclude them.

First, defendants assert that the Reply Report's analysis is unreliable because Prof. Elhauge omits the relative price of EpiPen and Auvi-Q from his regression model. Defendants assert that price is a "major factor" that influences Auvi-Q sales—the dependent variable in the regression analysis. Doc. 1854 at 22. And, "to be admissible, a regression analysis must control for the `major factors' that might influence the dependent variable." Reed Constr. Data Inc. v. McGraw-Hill Cos., 49 F.Supp.3d 385, 400 (S.D.N.Y. 2014) (citation omitted). By excluding price from his regression model, defendants argue, Prof. Elhauge has no way to determine whether the effect he observes in his regression results from Mylan's PBM contracts or, instead, the higher price of Auvi-Q.

Plaintiffs respond that one of Prof. Elhauge's rebuttal analyses simply added PBM fixed effects to the regression analysis in response to criticism by Mylan's expert. See Doc. 1711-2 at 40-41 (Elhauge Reply Report ¶ 82). But the regression never eliminated the price variable, and thus, the analysis still controlled for price. Id. Plaintiffs concede that another one of Prof. Elhauge's rebuttal analyses did drop the price variable, but, plaintiffs explain, the analysis continues to use time and PBM fixed effects to control for any differences in price across time or between PBMs. Id. at 35 (Elhauge Reply Report ¶ 68). Also, Prof. Elhauge specifically excluded the price variable in response to the criticisms lodged by Mylan's expert. Id. at 36 (Elhauge Reply Report ¶ 70).

The parties' competing arguments about Prof. Elhauge's exclusion of the price variable and whether he properly controlled for this factor in his regression go to the probative value of his analysis. See, e.g., Bazemore v. Friday, 478 U.S. 385, 400 (1986) ("Normally, failure to include variables will affect the analysis' probativeness, not its admissibility."); In re Urethane Antitrust Litig., 768 F.3d 1245, 1260-61 (10th Cir. 2014) ("Consequently, the exclusion of major variables or the inclusion of improper variables may diminish the probative value of a regression model. But such defects do not generally preclude admissibility, and courts allow use of a regression model as long as it includes the variables accounting for the major factors." (citing Bazemore, 478 U.S. at 400)). But defendants' attacks don't establish that Prof. Elhauge's analysis is inadmissible under Daubert.

Second, defendants argue that Prof. Elhauge's rebuttal analysis employs the wrong standard for statistical significance. More specifically, Prof. Elhauge uses a 0.10 threshold for statistical significance. But, defendants contend, the conventional level of statistical significance used by both the scientific community and the courts is 0.05. See Apsley v. Boeing Co., 691 F.3d 1184, 1198 (10th Cir. 2012) ("The significance level is typically placed at 5%."); see also Reference Guide on Multiple Regression, at 320 ("In most scientific work, the level of statistical significance required to reject the null hypothesis (i.e., to obtain a statistically significant result) is set conventionally at 0.05, or 5%.").

Plaintiffs respond that Prof. Elhauge used the 0.10 threshold in just one of his regressions—one that he offered to respond to Mylan's expert's criticisms about his failure to account for serial correlation. See Doc. 1711-2 at 40-41 (Elhauge Reply Report ¶ 82). And, while defendants assert that 0.05 is the accepted threshold, they don't cite any authority requiring an expert to use this standard. Indeed, the Reference Guide on Multiple Regression— the authority defendants cite in their motion—recognizes that:

Reference Guide on Multiple Regression, at 321 n.48.

Also, courts generally have found that challenges to statistical significance go to the weight, but not the admissibility, of a regression model. See, e.g., Kadas v. MCI Systemhouse Corp., 255 F.3d 359, 362-63 (7th Cir. 2001) (explaining that "[t]he 5 percent test is arbitrary" and "[t]he question whether a study is responsible and therefore admissible under the Daubert standard is different from the weight to be accorded to the significance of a particular correlation found by the study. It is for the judge to say, on the basis of the evidence of a trained statistician, whether a particular significance level, in the context of a particular study in a particular case, is too low to make the study worth the consideration of judge or jury."); Kurtz v. Kimberly-Clark Corp., 414 F.Supp.3d 317, 331 (E.D.N.Y. 2019) ("Regressions should not be excluded on the ground that they fail to meet arbitrary thresholds of statistical significance."); In re High-Tech Emp. Antitrust Litig., No. 11-CV-02509-LHK, 2014 WL 1351040, at *15 (N.D. Cal. Apr. 4, 2014) ("The Court finds that the fact that these two variables are not statistically significant at the 1%, 5%, and 10% levels goes to the weight, not the admissibility of [the] model."). Similarly, here, the court finds that defendants' arguments about the statistical significance of Prof. Elhauge's analysis go to the weight the court should give his opinions when deciding the class certification motion. But, they do not preclude the court from considering it on class certification.

Finally, defendants assert that one of Prof. Elhauge's rebuttal analyses uses the wrong test for statistical significance. Defendants explain that there are two tests for statistical significance: a two tailed test and a one-tailed test. A two-tailed test allows for "the effect [on the dependent variable] to be either positive or negative." Reference Guide on Multiple Regression, at 321. And a one-tailed test is used "when the expert believes, perhaps on the basis of other direct evidence presented at trial, that the alternative hypothesis is either positive or negative, but not both." Id.

Prof. Elhauge's Reply Report explains that he used a one-tailed test in his regression. He also explains why. Doc. 1711-2 at 41 (Elhauge Reply Report ¶ 82 n.146) ("A one-sided test is appropriate here because we are only interested in whether the regression shows a significant negative (foreclosing) effect of the formulary restrictions, and because there is no plausible theory under which the formulary restrictions would increase Auvi-Q's share."). Stated another way, Prof. Elhauge assumed that the formulary restrictions only could produce a negative effect on Auvi-Q's market share (i.e., the dependent variable). And thus, he used a one-tailed test for this regression.

Defendants assert that Prof. Elhauge's use of the one-tailed test makes his analysis unreliable because both the scientific community and the courts prefer the two-tailed test. See, e.g., Palmer v. Shultz, 815 F.2d 84, 95-96 (D.C. Cir. 1987) (concluding that "generally two-tailed tests are more appropriate in Title VII cases"); Smith v. City of Bos., 144 F.Supp.3d 177, 196 (D. Mass. 2015) (observing in a Title VII case, "[t]he weight of the case law appears to favor two-tailed tests."). See also Reference Guide on Multiple Regression, at 321 (explaining that use of the two-tailed test "is usually appropriate"). But, as plaintiffs note, defendants cite no authority that requires an expert to use a two-tailed test as a condition for finding the expert's opinions admissible under Daubert. And, plaintiffs contend, the expert should have discretion to decide whether to use a one-tailed or two-tailed test based on the regression that he is testing. Indeed, the Reference Guide on Multiple Regression explains:

Reference Guide on Multiple Regression, at 321. And the Guide cautions: "Because there is some arbitrariness involved in the choice of an alternative hypothesis, courts should avoid relying solely on sharply defined statistical tests." Id.

Plaintiffs assert that it was proper for Prof. Elhauge to use a one-tailed test in this particular regression because, he determined, it's not realistic to expect that defendants' exclusionary contracts ever would cause Auvi-Q's market share to increase. Defendants respond with the conclusion of Mylan's expert. He opines that—for at least two major PBMs—Prof. Elhauge's regression actually shows that formulary restrictions caused Auvi-Q's market share to increase. Thus, defendants assert, the data is capable of producing either a positive or negative effect. And, they contend, Prof. Elhauge should have used a two-tailed test in his regression.

The court refuses to exclude Prof. Elhauge's opinion simply because Mylan's expert disagrees with it. Prof. Elhauge has explained his reasons for using the one-tailed test. And even if a factfinder ultimately would reject his reasoning, his reasoning is, at worst, plausible. Defendants' attacks against Prof. Elhauge's decision don't show that his opinions are unreliable or preclude their admissibility. The court thus declines to exclude Prof. Elhauge's opinions based on his use of the one-tailed test for one of his regressions.

In sum, the court grants defendants' Motion to Exclude Prof. Elhauge's opinions in part and denies it in part. The court grants the motion to exclude his opinions about Auvi-Q market foreclosure based on Mylan's EpiPens4Schools program because plaintiffs since have abandoned this theory. And, the court denies defendants' motion in all other respects.

3. Motion to Exclude Opinions and Proposed Testimony of Professor Andrew K. Torrance

Finally, defendants ask the court to exclude opinions and proposed testimony of Prof. Andrew K. Torrance. Prof. Torrance offers opinions about the patent lawsuit that King Pharmaceuticals (a Pfizer subsidiary) filed against Teva alleging that Teva's generic version of the EpiPen infringed two of King's EpiPen patents ("Teva patent litigation"). Prof. Torrance opines about Teva's likelihood of success in the Teva patent litigation had the parties not settled the lawsuit. Also, Prof. Torrance opines about the costs and duration of the Teva patent litigation had the parties not settled.

Defendants base their motion to exclude on three arguments. First, defendants contend that Prof. Torrance fails to explain how he reached his conclusion that Teva had a greater than 70% chance of prevailing in the Teva patent litigation. Second, defendants assert that Prof. Torrance's opinions don't fit with the factual record. Third, defendants argue that Prof. Torrance didn't apply accepted methods and principles for analyzing patent invalidity. All these reasons, defendants assert, render Prof. Torrance's opinions unreliable. And thus, defendants assert, the court should exclude them. The court addresses each of the three arguments, separately, below.

a. Professor Torrance Adequately Explained How He Reached His Conclusions.

First, defendants argue that Prof. Torrance doesn't explain how he determined Teva's likelihood of success in the Teva patent litigation. Defendants assert that Prof. Torrance offers nothing but baseless speculation to support his conclusions. Thus, defendants contend, Prof. Torrance's opinions will not help decide whether it was reasonable for Teva to settle the Teva patent litigation.

Plaintiffs respond that Prof. Torrance bases his opinions on both: (1) his experience as a patent attorney, legal scholar, and law professor,24 and (2) his review of materials that he considered when forming his opinions. Doc. 1849-1 at 19, 21 (Torrance Dep. 240:8-16, 257:8-258:18). As plaintiffs explain in response, other courts have found similar expert testimony admissible under Daubert when the expert bases his opinions about the likelihood of success in litigation on his experience and his review of the relevant litigation materials. See, e.g., In re Namenda Direct Purchaser Antitrust Litig., 331 F.Supp.3d 152, 187-88 (S.D.N.Y. 2018) (denying motion to exclude opinions of a "patent attorney with extensive experience in the area of patent law" about "the likelihood [of success] in [a patent] lawsuit" because the expert "unquestionably has the expertise to evaluate the things he assessed—from expert reports to patent file folders—and to draw conclusions about who is more likely to win a patent lawsuit"); In re Solodyn (Minocycline Hydrochloride) Antitrust Litig., No. 14-md-02503, 2018 WL 563144, at *16 (D. Mass. Jan. 25, 2018) (refusing to exclude opinion of patent law expert who opined about the likelihood of success in patent litigation "based upon his own expertise and experience in the field of patent law"); United Food & Commercial Workers Local 1776 & Participating Emp'rs Health & Welfare Fund v. Teikoku Pharma USA, 296 F.Supp.3d 1142, 1186-87 (N.D. Cal. 2017) (denying motion seeking to exclude expert testimony about likely outcome of patent litigation when the expert had reached his conclusions by "applying his scientific background and knowledge to his review of the trial record"); In re Wellbutrin XL Antitrust Litig., 133 F.Supp.3d 734, 765-67 (E.D. Pa. 2015) (denying motion to exclude expert's opinion about the chances of prevailing in underlying patent litigation because the expert—a patent law professor—was "a qualified expert offering appropriate expert testimony" and had reached his conclusions after "review of the briefs, pleadings, [Abbreviated New Drug Application], and underlying patent at issue in the . . . litigation").

For the same reasons, the court refuses to exclude Prof. Torrance's opinions here. Prof. Torrance sufficiently explained how he reached his conclusions—using his experience in the field of patent law and his review of the relevant materials. See, e.g., F & H Coatings, LLC v. Acosta, 900 F.3d 1214, 1224 (10th Cir. 2018) (finding no error in admission of expert testimony because the expert's "experience and industry knowledge were sufficient . . . to find that his expert conclusions . . . were reliable").

Defendants also attack Prof. Torrance's conclusions by citing their own expert's critique of Prof. Torrance's report. Doc. 1848 at 9-10. Defendants' expert believes that Prof. Torrance's opinions are flawed because he failed to consider the trial court's decision denying Teva's pretrial motion for judgment as a matter of law. Id. Also, defendants' expert asserts that Prof. Torrance's conclusion that Teva had a relatively large probability of infringing one of the EpiPen patents doesn't comport with his opinion that Teva also had an above 70% chance of success in the Teva patent litigation. Id. at 10. These arguments go to the weight the court should give Prof. Torrance's opinion on class certification—not whether the court should exclude them altogether as unreliable.

The court rejects defendants' first argument supporting their motion to exclude Prof. Torrance's opinions under Daubert.

b. Professor Torrance's Opinions Sufficiently Fit with the Facts of the Case to Allow their Admission.

Second, defendants argue that Prof. Torrance's opinions about Teva's likelihood of success in the Teva patent litigation are untethered to the factual record because he relies on arguments that the parties never raised at trial in the Teva patent litigation, or ones that the parties waived during the course of the litigation. More specifically, defendants assert that Prof. Torrance's opinions are unreliable because he bases them on: (1) the patent prosecution history, even though the parties never raised this subject at trial in the Teva patent litigation; (2) an "unenforceability" defense based on inequitable conduct, even though Teva never raised this defense in the Teva patent litigation; and (3) an "obviousness" defense to patent infringement, even though Teva never asserted this defense in the Teva patent litigation.

Plaintiffs respond that Prof. Torrance doesn't offer opinions about Teva's likelihood of success at trial. Instead, Prof. Torrance opines about Teva's likelihood of prevailing "in any final adjudication by the Court of Appeals for the Federal Circuit." Doc. 1500-4 at 4-5 (Torrance Expert Report ¶ 3.a.). Thus, plaintiffs contend, defendants' criticisms that Prof. Torrance's opinions aren't tethered to the trial record don't matter for purposes of determining their reliability. And, even so, plaintiffs assert that the trial record supports Prof. Torrance's opinions.

The parties to the Teva patent litigation settled their dispute after they had tried their case to a district judge in a bench trial—but before the district judge had issued a decision. So, defendants argue, the trial record that the Federal Circuit would consider on appeal was complete. And, according to defendants, Prof. Torrance's opinions rely on arguments that Teva never asserted at trial or ones that it had waived at trial. Plaintiffs disagree.

First, plaintiffs explain that the patent's prosecution history is relevant because the Federal Circuit applies a two-step process when deciding patent infringement claims. Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1581 (Fed. Cir. 1996) (citing Markman v. Westview Instruments, Inc., 52 F.3d 967, 976 (Fed. Cir. 1995)). In the first step, the court determines "the proper construction of the asserted claim." Id. at 1581-82 (citing Markman, 52 F.3d at 976). In the second step, the court determines "whether the accused method or product infringes the asserted claim as properly construed." Id.

The Federal Circuit has explained that "[t]he first step, claim construction, is a matter of law, which [the Federal Circuit] review[s] de novo." Id. at 1582 (citing Markman, 52 F.3d at 979). And, when conducting this de novo review, the Federal Circuit looks "first to the intrinsic evidence of record, i.e., the patent itself, including the claims, the specification and, if in evidence, the prosecution history." Id. (emphasis added). So, plaintiffs assert, Prof. Torrance properly relied on the prosecution history when forming his opinions about Teva's likelihood of success in the Teva patent litigation because, even though the parties may not have cited the prosecution history during trial, the Federal Circuit still could rely on the prosecution history during its de novo review of claim construction on appeal. The court agrees with plaintiffs. Prof. Torrance's reliance on the prosecution history does not render his opinions unreliable.

Next, plaintiffs argue that the unenforceability defense based on inequitable conduct could apply in the Teva patent litigation because the parties had identified it as a contested issue of fact and law in the Intelliject litigation (a lawsuit related to the Teva patent litigation). Doc. 1837-8 at 3 (Pls.' Statement of Contested Issues of Fact and Law ¶ 13 ("Whether Intelliject has proven by clear and convincing evidence that the '432 Patent is unenforceable due to patent misuse?")). Plaintiffs explain that a finding of unenforceability for the '432 Patent in the related Intelliject litigation would operate as res judicata in the Teva patent litigation, if the cases hadn't ended with the parties' settlements—ones that plaintiffs describe as "anticompetitive pay-for-delay settlement[s]." Doc. 1862 at 18.

Defendants counter that the defense asserted in the Intelliject litigation (and cited by plaintiffs) was unenforceability based on patent misuse—not unenforceability based on inequitable conduct. Indeed, the issue recited in the Statement of Contested Issues of Fact and Law reads: "Whether Intelliject has proven by clear and convincing evidence that the '432 Patent is unenforceable due to patent misuse?" Doc. 1837-8 at 3 (Pls.' Statement of Contested Issues of Fact and Law ¶ 13) (emphasis added). And, the court agrees with defendants that an inequitable conduct defense differs from a patent misuse defense. See Resqnet.Com, Inc. v. Lansa, Inc., No. 01 Civ.3578(RWS), 2004 WL 1627170, at *4 (S.D.N.Y. July 21, 2004) ("Patent misuse and inequitable conduct are different defenses. Misuse is a defense based upon the idea that a patent properly procured has been broadened in scope impermissibly so as to be used to impose an unreasonable restraint in competition . . . [i]nequitable conduct, on the other hand, renders the patent unenforceable permanently, and involves concealing facts or being dishonest in the procurement of the patent"); see also Electro Source, LLC v. Nyko Techs., Inc., Nos. CV 01-10825 DT (BQRx), CV 02-520 DT (BQRx), 2002 WL 34536682, at *12 (C.D. Cal. Apr. 15, 2002) (explaining that "patent misuse [claims] differ from . . . inequitable conduct claims" because "[i]nequitable conduct relates to the means of obtaining a patent, and deems any patent obtained by such means unenforceable" while "[p]atent misuse relates to the enforcement of a patent after it has been obtained").

Prof. Torrance's report suggests that certain actions taken by the patent inventors before the United States Patent and Trademark Office ("USPTO") could amount to inequitable conduct. See Doc. 1500-4 at 21 (Torrance Expert Report ¶ 49 (opining that the inventor's failure to file timely an Information Disclosure Statement "[a]t best . . . suggests carelessness, and, at worst, a desire to postpone consideration by the examiner of potentially damaging prior art")); see also Doc. 1497-4 at 30 (Sealed Torrance Expert Report ¶ 72 (explaining that, based on Prof. Torrance's review of the patent application history, it suggests that the inventors did not submit some potentially relevant prior art to the USPTO during prosecution)). But neither Teva nor Intelliject raised inequitable conduct as an affirmative defense in the underlying patent litigation. Thus, defendants argue, the inequitable conduct defense was waived. So, according to defendants, the trial court never would have considered the inequitable conduct defense in rendering its decision in the Teva patent litigation. And, Teva never could assert this defense in an appeal to the Federal Circuit because it had waived the defense.

The court agrees with defendants that it is unlikely that the Federal Circuit would consider an inequitable conduct defense based on the Teva patent litigation record. But it's not impossible. Generally, "`a federal appellate court does not consider an issue not passed upon below.'" Golden Bridge Tech., Inc. v. Nokia, Inc., 527 F.3d 1318, 1322 (Fed. Cir. 2008) (quoting Singleton v. Wulff, 428 U.S. 106, 120 (1976)). However, appellate courts also have "the discretion to decide when to deviate from this general rule of waiver . . . ." Id. (citing Singleton, 428 U.S. at 121); see also Celgene Corp. v. Peter, 931 F.3d 1342, 1356 (Fed. Cir. 2019) (recognizing that the Federal Circuit has "discretion to reach issues raised for the first time on appeal"). Although "[d]eparting from the general rule of waiver is appropriate only in limited circumstances," the Federal Circuit may consider an issue not presented to the lower tribunal when there is "an intervening change in the law" or when "the `interest of justice' guides [the Federal Circuit] to consider the issue despite the fact that it was not raised below." Celegene, 931 F.3d at 1356 (citations omitted).

When considering whether to exclude Prof. Torrance's opinion, the court is guided by Rule 702's advisory committee's notes. They recognize that "rejection of expert testimony is the exception rather than the rule." Fed. R. Evid. 702 advisory committee's notes to 2000 amendments; see also In re Syngenta AG MIR 162 Corn Litig., No. 14-md-2591-JWL, 2017 WL 1738014, at *2 (D. Kan. May 4, 2017). While Daubert requires the court to serve as a gatekeeper for expert testimony, "[v]igorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof" remain "the traditional and appropriate means of attacking shaky but admissible evidence." Daubert, 509 U.S. at 596 (citation omitted). Ultimately, the "district court has `considerable leeway in deciding in a particular case how to go about determining whether particular expert testimony is reliable.'" In re Syngenta AG MIR 162 Corn Litig., 2017 WL 1738014, at *2 (quoting Kumho Tire, 526 U.S. at 152); see also Vining v. Enter. Fin. Grp., 148 F.3d 1206, 1218 (10th Cir. 1998) ("The admission of expert testimony is left to the sound discretion of the trial court.").

Also, Daubert doesn't require the court to decide whether Prof. Torrance's opinions are "undisputably correct." Etherton v. Owners Ins. Co., 829 F.3d 1209, 1217 (10th Cir. 2016) (citation and internal quotation marks omitted). Instead, the court must determine whether "the method employed by the expert to reach his conclusion is scientifically sound and that the opinion is based on facts which sufficiently satisfy Rule 702's reliability requirements." Id. at 1217-18 (citations and internal quotation marks omitted). Thus, when performing the Daubert inquiry, the court should "`focus . . . solely on [the expert's] principles and methodology, not on the conclusions that they generate.'" Id. at 1217 (quoting Daubert, 509 U.S. at 595).

The court is satisfied here that Prof. Torrance has used a reliable methodology to reach his opinions about Teva's likely success in the Teva patent litigation. Defendants' criticisms about his opinions—i.e., that Prof. Torrance relied on defenses that Teva never asserted in the Teva patent litigation—go to the weight the court should give them on class certification. But, the court refuses to exclude his opinions from consideration.

Finally, plaintiffs assert that the "obviousness" defense to the validity of the '432 patent was at issue in both the Teva patent litigation and the Intelliject litigation. See Doc. 1837-8 at 2-3 (Pls.' Statement of Contested Issues of Fact and Law ¶¶ 7-9, 11, 12); see also Doc. 1837-9 at 2-3 (Teva's Statement of Contested Issues of Fact and Law ¶¶ C.-E.). Defendants concede that Teva asserted the obviousness defense in the Statement of Contested Issues of Fact and Law. Doc. 1888 at 12. But, they argue, Teva did not pursue this defense at trial. Defendants contend that Teva based its invalidity defense at trial on "the legal doctrine of anticipation"—and not that "the '432 patent was invalid for obviousness." Id. For support, defendants cite a single passage from the trial transcript where counsel clarified that Teva was "asserting an anticipation defense based on the '429 patent," but "not asserting derivation" or "prior invention." Doc. 1888-5 at 4 (Tr. 403:6-13) (emphasis added). With this statement, counsel may have limited Teva's use of the invalidity defense to anticipation based on the '429 patent, but it didn't waive the other obviousness defenses that Teva asserted based on the '316 and '686 patents. Compare Doc. 1837-9 at 2 (Teva's Statement of Contested Issues of Fact and Law ¶ C. ("Whether one or more of the asserted claims . . . of the '432 patent, are invalid as anticipated or obvious . . . by . . . the '429 patent")) with id. at 2-3 (Teva's Statement of Contested Issues of Fact and Law ¶¶ D.-E. (asserting an invalid as obvious defense based on the '316 and '686 patents)). Based on this record, the court can't conclude that Teva waived its obviousness defense at trial such that the Federal Circuit would refuse to consider this defense on appeal. And, the court declines to find Prof. Torrance's opinions unreliable because he considered an obviousness defense when forming his conclusions about Teva's likelihood of success in a final adjudication before the Federal Circuit.

In sum, the court agrees with plaintiffs that the way in which Prof. Torrance considered the trial record and what the Federal Circuit could consider on appeal does not render his opinions unreliable or require their exclusion under Daubert. The court thus denies this aspect of defendants' Motion to Exclude Prof. Torrance's opinions.

c. Professor Torrance Adequately Applied Accepted Methods and Principles for Analyzing Patent Invalidity.

Last, defendants argue that Prof. Torrance's opinions are inadmissible because he did not analyze the validity of the '432 patent consistent with the governing law. Prof. Torrance opines that the '432 patent claims are "quite weak, and likely invalid" due to obviousness. Doc. 1500-4 at 31 (Torrance Expert Report ¶ 74); see also 35 U.S.C. § 103 ("A patent for a claimed invention may not be obtained . . . if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains."). The Supreme Court has established a framework for analyzing whether a patent is invalid for obviousness under 35 U.S.C. § 103. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 406 (2007). Under that framework, "`the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved.'" Id. (quoting Graham v. John Deere Co., 383 U.S. 1, 17 (1966)).

Defendants assert that Prof. Torrance failed to apply this legal framework because his report neither (1) explains how he construed the claims of the '432 patent, nor (2) analyzes the prior art. Defendants thus contend that Prof. Torrance's opinions are unreliable because the expert never applied the accepted standard for analyzing patent invalidity due to obviousness.

Plaintiffs respond that Prof. Torrance's report both construes the claims and analyzes the prior art. Also, plaintiffs assert that Prof. Torrance confirmed that he performed this work in his deposition. For support, plaintiffs cite Prof. Torrance's report where he recited: "To form my opinion, I analyzed a variety of information, materials, and documents, the most relevant of which are the litigated patents themselves, the prosecution histories of those patents, and the materials and records related to the litigations." Doc. 1500-4 at 13 (Torrance Expert Report ¶ 27).

With respect to the patent itself, Prof. Torrance described how he reviewed the patent and explained how one must interpret the claims of the patent to understand the invention it protects. Id. at 14 (Torrance Expert Report ¶¶ 29-30). Prof. Torrance also reviewed the patent's litigation history, including the trial court's Markman rulings. Id. at 16, 29 (Torrance Expert Report ¶¶ 37, 71). And, he concluded that "the claim constructions adopted by the judge would have disadvantaged King and Meridian more than [Teva and Intelliject]." Id. at 30 (Torrance Expert Report ¶ 71). Defendants respond that none of these passages explain how Prof. Torrance construed the claims. In his deposition, Prof. Torrance conceded that he included no "particular section on claim construction" in his report. Doc. 1849-1 at 17 (Torrance Dep. 231:24-232:16). But, Prof. Torrance explained, claim construction was something that he considered when forming his opinions and is found "throughout the report." Id. Plaintiffs don't provide any specific citations showing where Prof. Torrance construed the claims in the report.

With respect to the prior art, Prof. Torrance explains that he "reviewed the available prosecution histories of each of the patents at issue," taking a "particular interest [in] the variety of prior art references and legal grounds for rejection the examiners identified . . . ." Doc. 1500-4 at 15 (Torrance Expert Report ¶ 34); see also id. at 18 (Torrance Expert Report ¶ 42 (explaining that Prof. Torrance considered a number of factors when forming his conclusions including "the number and complexity of asserted patent claims in each patent" and "the volume of relevant prior art")). Based on his review, Prof. Torrance concluded the '432 patent's claimed invention included "key elements" such as "a lack of kickback and a second locked extended position of the needle cover" that make it "highly likely that there exists prior art relevant to auto-injectors that discloses these simple elements." Id. at 30-31 (Torrance Expert Report ¶ 74). Plaintiffs don't provide any specific citations showing exactly which prior art Prof. Torrance considered when forming his opinions. As Prof. Torrance testified, he "refer[s] to the prior art usually as a group" in his report. Doc. 1849-1 at 27 (Torrance Dep. 346:9-25). But also, Prof. Torrance testified about specific prior art that he considered in his report. Id. (Torrance Dep. 347:5-348:22 (referring to his discussion of the '686, '458, and '336 patents in ¶¶ 48.a., 51, 60, and 62 of his expert report)).

The court agrees with defendants. Prof. Torrance's report never explains how he construed the claims or how he compared those construed claims to the prior art before reaching his conclusion that the '432 patent is likely invalid due to obviousness. Instead, the report contains general references about his analysis of the claims and prior art. See, e.g., 1500-4 at 27, 29, 30-31, 36 (Torrance Expert Report ¶¶ 65, 66, 70, 74-76, 92). And, plaintiffs' only response to defendants' argument on this point is Prof. Torrance's vague explanation that the claim construction analysis is found "throughout the report." Doc. 1862 at 15 (citing Doc. 1849-1 at 27 (Torrance Dep. 232:15)).

Nevertheless, the court declines to exclude Prof. Torrance's opinion on this basis. Prof. Torrance correctly cites the law governing patent validity. See, e.g., Doc. 1500-4 at 12-13 (Torrance Expert Report ¶ 26 (citing KSR Int'l Co., 550 U.S. 398)). And, Prof. Torrance says he applied this law to reach his conclusions. See id. at 23-24, 27, 30-31, 36 (Torrance Expert Report ¶¶ 56, 57, 65, 74, 92). Defendants have the better end of the criticisms that the report lacks an explanation how Prof. Torrance reached his conclusions after applying the appropriate governing standard. But that doesn't mean the court should exclude his opinions. While closer than other decisions, these criticisms, the court finds, go to the weight of his opinions, not their admissibility. The court thus denies defendant's Motion to Exclude Prof. Torrance's opinions on this ground.

For all these reasons, the court denies defendants' Motion to Exclude Opinions and Proposed Testimony of Andrew K. Torrance.

III. Conclusion

This Order has ruled the five pending motions seeking to exclude certain experts' opinions from the court's consideration of the putative class members' Motion for Class Certification.

For reasons explained above, the court denies the two motions filed by the putative class plaintiffs—i.e., their (1) Motion to Strike the Partial Testimony of Dr. Michael Blaiss (Doc. 1584), and (2) Motion to Strike the Testimony of Prof. James Hughes (Doc. 1852).

The court grants in part and denies in part defendants' Motion to Exclude Expert Opinions of Professor Einer Elhauge (Doc. 1604). The court grants the motion to exclude his opinions about Auvi-Q market foreclosure based on Mylan's EpiPens4Schools program because plaintiffs since have abandoned this theory. In all other respects, the court denies defendants' Motion to Exclude Expert Opinions of Professor Einer Elhauge.

And, the court denies the other two motions filed by defendants—i.e., their (1) Motion to Exclude Opinions and Proposed Testimony of Andrew K. Torrance (Doc. 1847), and (2) Motion to Exclude Expert Opinions of Professor Meredith Rosenthal (Doc. 1602).

IT IS THEREFORE ORDERED BY THE COURT THAT the putative class plaintiffs' Motion to Strike the Partial Testimony of Dr. Michael Blaiss (Doc. 1584) is denied.

IT IS FURTHER ORDERED THAT defendants' Motion to Exclude Expert Opinions of Professor Meredith Rosenthal (Doc. 1602) is denied.

IT IS FURTHER ORDERED THAT defendants' Motion to Exclude Expert Opinions of Professor Einer Elhauge (Doc. 1604) is granted in part and denied in part, as specified in this Order.

IT IS FURTHER ORDERED THAT defendants' Motion to Exclude Opinions and Proposed Testimony of Andrew K. Torrance (Doc. 1847) is denied.

IT IS FURTHER ORDERED THAT the putative class plaintiffs' Motion to Strike the Testimony of Professor James Hughes (Doc. 1852) is denied.

IT IS SO ORDERED.