In re: Mirena IUS Levonorgestrel-Related Products Liability Litigation (No. II)
William M. Janssen*
WHY IT MADE THE LIST
Amidst great fanfare, the Kingdom of Sweden launched its mighty warship, the Vasa, into Stockholm harbor on August 10, 1628. The ship had taken two years and hundreds of craftsmen to build. It was three-fourths the length of a modern football field, more than 150 feet tall, with sixty-four cannons manned by 300 soldiers operating from two gun decks, one built atop the other. The Vasa was elaborately ornamented, befitting its intended status as a powerful flagship of the Swedish navy. It had roughly 700 sculptures and decorations, painted in vibrant reds, blues, and yellows; some sculptures were even gilded in gold leaf. “Camouflage and melting into the background wasn’t anything for Vasa. Instead, the exact opposite applied: the ship would be impossible to miss.” The ship was to be “a gigantic billboard for Sweden and [its King] Gustav II Adolf.”
There must have been someone responsible for the cannons, and someone else responsible for the gunports; someone was in charge of the masts and the sails, and someone else in charge of the rudder. Almost certainly there also was someone responsible for the sculpting and decorations, and someone in charge of the painting and gold-leafing. But who was responsible for making sure Vasa could sail?
Late in the afternoon, Vasa cast off from the shore. Soon, a simple gust of wind caused Vasa to tilt hard to its port side, water gushed through its open gun-port doors, and in minutes the ship sank to the seafloor—a 1,300-meter maiden voyage ended stunningly within sight of the very shipyard where it had been built. As it turns out, Vasa was too tall and heavy for its small hull below the waterline creating a calamitously unstable center of gravity. In other words, the problem with Vasa was that it couldn’t sail.
The Vasa lesson was not well remembered by the claimants in In re Mirena IUS Levonorgestrel-Related Products Liability Litigation (No. II). There, a multidistrict litigation (“MDL”) of centralized products liability lawsuits pending throughout the country sought damages for intrauterine device contraceptive users who suffered from a rare health disorder known as idiopathic intracranial hypertension. Unfortunately, their claims had the Vasa problem—it just couldn’t sail. The MDL claimants could not establish that the intrauterine device at issue had the capacity (general causation) to cause this hypertension disorder. What followed was an exhaustively detailed federal district court opinion granting MDL-wide summary judgment. The eloquently comprehensive opinion is now a template for other courts’ treatment of the threshold products liability issue of “general causation,” a distinction that warrants its placement among the top food and drug cases of 2019.
The Mirena MDLs
Mirena is an intrauterine system (“IUS,” also called an “IUD”—intrauterine device) used to prevent pregnancy. The system is comprised of a “T”-shaped polyethylene frame connected to a reservoir (containing a mixture of levonorgestrel and silicone) mounted around a vertical stem looped with a polyethylene removal thread. Levonorgestrel is a synthetic progestin. The placement of the Mirena system in a patient’s uterus, along with its gradual release of levonorgestrel, operates to prevent pregnancy. The system’s developer, manufacturer, and distributor—various Bayer healthcare companies—claim that the use of Mirena is “over 99% effective” in preventing pregnancy for up to five years. The system was approved by FDA in 2009 and is now included on the World Health Organization Model List of Essential Medicines.
Two different multidistrict litigations were initiated against Mirena. The first, MDL No. 2434 (centralized in 2013), contended that the device triggered post-insertion uterine perforation in patients and asserted product claims on behalf of about 1,300 affected plaintiffs. The presiding judge in this first Mirena MDL ruled that plaintiffs’ expert witnesses were unreliable under the federal courts’ Daubert standard for expert gatekeeping, thereby foreclosing those witnesses from offering competent causation testimony for plaintiffs’ claims. Without such evidence, plaintiffs’ claims did not survive Bayer’s omnibus motion for summary judgment, which the district court granted in 2016. The United States Court of Appeals for the Second Circuit affirmed both the Daubert expert exclusion order and the ensuing omnibus summary judgment ruling in late 2017. Certiorari was sought but denied by the U.S. Supreme Court.
The second group of litigations, MDL No. 2767, alleged that Mirena triggered the onset of a different medical condition, idiopathic intracranial hypertension (“IIH,” otherwise known as pseudotumor cerebri (“PTC”)). IIH is an uncommon disease (about one case per year per 100,000 people) involving an increase in intracranial pressure caused by excessive cerebrospinal fluid in the skull. Symptoms can include headaches, a hearing disorder, and a swelling of the optic nerves; in severe cases, IIH can lead to blindness. IIH is a diagnosis by exclusion, and is reported to occur twenty times more frequently in obese or overweight women of child-bearing age. It is treated behaviorally by encouraging weight loss; in more serious cases, lumbar punctures can drain excess cerebrospinal fluid. More than 900 plaintiffs filed IIH product liability claims against Mirena.
MDL centralization was sought, but at first refused, by the Judicial Panel on Multidistrict Litigation back in 2014. On a second application three years later, the Panel reversed course and granted the centralization plaintiffs requested, over defendants’ objections, explaining its change of view:
While we previously expressed concern that individualized causation issues might predominate in this litigation, the records in the many actions filed since then demonstrate that discovery and pretrial motions concerning the issue of general causation have been, or will be, at the center of all actions—that is, whether the hormonal component in Mirena is capable of causing intracranial hypertension.
The Panel selected Judge Paul A. Engelmayer of the Southern District of New York to preside over the new MDL.
Court Rulings on the New Mirena MDL
Just as the JPML had anticipated, general causation became the fulcrum issue in the now centralized Mirena IIH lawsuits. Causation, of course, “is a required element in every products liability case.” Satisfying that requirement implicates two showings: “General causation is whether a substance is capable of causing a particular injury or condition in the general population, while specific causation is whether a substance caused a particular individual’s injury.”
Plaintiffs nominated seven expert witnesses on general causation; Bayer nominated twelve of its own. In late 2018, after Daubert briefing and oral argument, Judge Engelmayer issued an order much like the one issued in the first Mirena MDL. Each of the plaintiffs’ seven expert causation witnesses was excluded. The court found that none had reached their conclusions through personal experiments, laboratory work, or new epidemiological assessment; several had based their conclusions on mere case reports or earlier-published epidemiological studies which either had not found IIH causation or had repudiated an earlier IIH causation finding; and several simply proffered “a thesis as to how, biologically, use of Mirena may cause IIH.” Accordingly, Judge Engelmayer ruled that each expert’s opinions were methodologically unreliable under the Daubert standard and, thus, inadmissible.
Tracing the familiar path staked in the first Mirena MDL litigation, Bayer then moved for summary judgment, positing that the court’s exclusion of all plaintiff IIH general causation experts left no remaining genuine dispute of material fact. Specifically, Bayer argued that summary judgment was categorically required in a pharmaceutical products case where, as here, the plaintiff lacks expert evidence of general causation, and, alternatively, even if non-expert evidence of general causation could theoretically suffice, the MDL plaintiffs lacked it. The MDL plaintiffs countered by arguing: (1) general causation was not a required element, and proof of specific causation alone could suffice; (2) if general causation was required, various, non-expert forms of evidence could be stitched together in a manner that would allow a factfinder to find general causation; (3) the court had erred in a preliminary order that directed plaintiffs to informally identify corporate admissions by Bayer on which they intended to rely in opposing summary judgment; and (4) summary judgment for Bayer prior to plaintiff-specific discovery was unconstitutional.
The court began its analysis of the parties’ competing arguments by dispatching the threshold fight over whether general causation evidence is even needed in pharmaceutical products cases. After noting the plaintiffs’ failure to cite convincing authority for this proposition, the court concluded: “plaintiffs’ portrait of state law as absolving a products-liability plaintiff from a need to establish general causation—the capacity of the product in question to cause the injury alleged—is simply wrong.” To the contrary, wrote the court, “all relevant jurisdictions require some evidence of general causation in products liability cases involving complex products liability (or medical) issues.” Judge Engelmayer then explained why this must be so:
To assure reliable outcomes in a circumstance where the origins of an injury are not obvious or within the scope of a lay juror’s everyday experience, and to avoid the risk that juries would equate correlation (the fact that a given plaintiff used a product and developed injuries) with causation, it is imperative that the factfinder be presented evidence that the product was capable of causing the injury of which a plaintiff complains.
Accordingly, to survive Bayer’s summary judgment motion, plaintiffs were obligated to produce admissible evidence from which a reasonable jury could find that Mirena was capable of causing IIH.
Bayer invited the court to make short work of the issue. Without expert evidence of general causation (as the plaintiffs were now left in the wake of the court’s Daubert order), Bayer pitched that summary judgment was inescapable as a matter of law. The court found “much force” to this contention, since it is “well established” that expert causation testimony is necessary where that issue lies outside the knowledge of lay jurors, as it generally does in pharmaceutical and medical device cases. But the court also noted that its controlling appellate authority, the U.S. Court of Appeals for the Second Circuit, had never definitively resolved the question of “expert-less” general causation; so, the court chose the surer course of assuming, for purposes of Bayer’s pending summary judgment motion, the possibility that general causation in complicated medical products liability cases could conceivably be shown through non-expert means.
But no such alternative proof existed. First, the court rebuffed plaintiffs’ contention that the differential diagnoses of their respective individual doctors could supply all the causation proof the law required. Such doctors’ opinions, ruled the court, were formed solely from examining their patients, and thus could not “qualify as a reliable judgment, consistent with Daubert, of the capacity of the product in question (here, Mirena) to cause the condition in question (here, IIH).” Quoting an earlier federal Daubert ruling, the court noted: “[E]vidence of specific causation is irrelevant without evidence of general causation.”
Second, the court brushed aside plaintiffs’ suggestion that “snippets” of their excluded expert witnesses’ opinions (presumably, short of those witnesses’ ultimate, and now excluded, pronouncements of general causation) could be packaged together, presented in that form to the jury, and thereby supply the factfinder with adequate bases for finding general causation. Such a proposal, the court ruled, was “unsustainable”: “the propositions here that plaintiffs propose to use as building blocks for a lay finding as to Mirena’s capacity to cause IIH are complex, scientific in nature, and highly technical. They patently require expertise to decode and apply.” Laying so complicated a “scientific inquiry before a jury in the absence of a reliable expert opinion . . . [and then asking] the jury to reach a conclusion from it as to general causation, is an invitation to speculation, mischief, and error.”
Third, the court spurned plaintiffs’ argument that Bayer’s own general causation expert witnesses could fill the gap because they could tutor the jury on the methodological criteria that a qualified, admissible expert witness would employ in reaching a general causation conclusion. But being schooled on multi-factor epidemiological standards, the court concluded, “is not a mere box-checking exercise. Sophisticated judgments instead must be made about the existence and probative value of each constituent scientific factor and its relations to the others.” This is no proper jury function:
Plaintiffs’ notion that a jury, taught by Bayer’s experts what the [required general causation] criteria are, could then reliably apply them badly misapprehends—indeed, it trivializes—this mode of epidemiological inquiry. If qualified expert epidemiologists can misapply the . . . factors, a lay jury certainly cannot be counted on to individually evaluate and collectively weigh these factors in a suitably scientifically rigorous manner, one that avoids the risk of conflating correlation with causation.
Fourth, the court rejected plaintiffs’ position that a certain retrospective case-control study published in 2019 might be placed before the jury in a manner that would allow them to find general causation. Acknowledging that this study did, in fact, find a statistically significant association between Mirena and IIH, the court then observed that the study itself expressly disclaimed any finding of causation because the association it found could be explained by Mirena’s use by patients with attributes that already present an enhanced risk of developing IIH (namely, obesity and recent weight gain).
Fifth, the court dismissed the few final arguments posed by the MDL plaintiffs. Contrary to plaintiffs’ insistence otherwise, Bayer had never corporately admitted that Mirena caused IIH, nor could FDA-approved labels for other, different levonorgestrel products qualify as general causation proof. And the granting of summary judgment prior to patient-specific discovery did not violate the U.S. Constitution’s Seventh Amendment because in the absence of any evidence from which a jury might reliably find general causation, there was no genuine dispute as to material fact and summary judgment was warranted.
The capacity of a product to inflict a claimed harm really should not be a controversial prerequisite for product liability victory. Indeed, even in its most pioneering infancy, strict liability theory never doubted the necessity of proving causation. That certainty was borne of a well-settled heritage; a hundred years ago, the eminent Judge Benjamin Cardozo admonished against overlooking this causation truism: “We must be on our guard . . . against confusing the question of negligence with that of the causal connection between the negligence and the injury . . . . ‘Proof of negligence in the air, so to speak, will not do.’” Today, notwithstanding variations among states, causation remains a “core” element on which all jurisdictions agree. That element, in turn, is comprised of two showings: is the product capable of causing this particular harm (general causation), and did it inflict that harm on me (specific causation)?
In the world of pharmaceuticals and Daubert gatekeeping, however, this first undertaking (general causation) can prove to be a challenging—and often unattainable—climb. As Judge Posner aptly wrote some years back: “Law lags science; it does not lead it.” Scientific guesswork, he noted, “even of the inspired sort,” has no place in the courts. So, general causation may obligate plaintiffs to produce as evidence something that may well be evading science: proof of a product’s toxic capacity. This is not, however, a failing of the justice system. Rather, it is an inescapable attribute of justice. “[T]he classic test for determining cause in ‘fact’ directs the ‘factfinder’ to compare what did occur with what would have occurred if hypothetical, contrary-to-fact conditions had existed . . . . An act or omission is not regarded as a cause of an event if the particular event would have occurred without it.” Because the bedrock anchor of our compensatory civil law is the responsibility to others, “are-you-responsible-for-what-happened” operates as an irreducible first principle of justice. If it is not possible for you to have been responsible (that is, if general causation is absent), then justice commands your exculpation.
Judge Engelmayer’s opinion spans thirty-five pages in the Federal Supplement-Third to explain how he resolved what he called “the gateway issue of general causation” in the Mirena IIH litigations. Its careful, methodical, and scholarly treatment comes with a feature that not all judicial opinions possess: readability. Walking the reader slowly through an easy-to-follow recounting of each step in his painstaking analysis of complex principles presented in a readily appreciated manner, Judge Engelmayer’s opinion now serves as the veritable template for proper general causation decisionmaking.
What will likely make Judge Engelmayer’s opinion so influential in the law of general causation is the care with which it was crafted.
Noting that the Mirena IIH plaintiffs had “not adduced any . . . products liability cases disclaiming the need to prove general causation,” the court could have reached a preemptive, summary conclusion on the question, but did not; instead, the court chose to parse through a formidable inventory of national case law to reach what, by then, readers would accept as the well-supported conclusion that “all relevant jurisdictions require some evidence of general causation in products liability cases involving complex products liability (or medical) issues.” Pages later, Judge Engelmayer was still completing his meticulous explanation why “proof of general causation plays a vital role in complex cases where the capacity of a product to cause a species of injury is not intuitively obvious.”
The opinion also contributes to the nation’s general causation jurisprudence by its careful and well-explained discussion of the Mirena plaintiffs’ reliance on proof of specific causation as a sort of back-filler for their lack of general causation evidence. That back-filling role, Judge Engelmayer explained, is one that specific causation just cannot fill. Plaintiffs proposed to offer the testimony of the myriad of doctors who treated them and then engaged in differential diagnosis to surmise that Mirena caused their respective IIH. “There is of course a proper place for testimony about the causes of an individual plaintiff’s symptoms,” wrote the court. “That is the essence of the specific causation inquiry undertaken if there is competent evidence of general causation.” But that “if” matters greatly. At the risk of restating the obvious: unless Mirena can first be reliably shown to be a true candidate for causing a patient’s IIH, the product has no business being included in a differential diagnosis assessment at all. Numerous prior court decisions have treated this general-causation-first sequencing as self-evident.
Interestingly, not all courts have shut this sequencing door quite so tightly as Judge Engelmayer and those prior decisions have. Some courts, albeit tepidly, have reserved on the question, refusing to pronounce unequivocally that a physician’s patient-specific differential diagnosis could never qualify as Daubert-proper general causation evidence. If this reservation is simply based on a “never-say-never” philosophy in opinion-writing, it is a safe judicial approach but not an especially helpful one. To surmise that a differential diagnosis opinion could supply proper general causation proof could be, perhaps, just a shorthand way of acknowledging that the same expert witness could carry both proofs. That is not outside the realm of possibility, to be sure. An expert epidemiologist with a Daubert-qualifying history of performing validating general causation science could be summoned to examine a patient and, armed with that expert’s own, sound, scientifically-established experience, conduct a differential diagnosis, and then offer an expert opinion that might be accepted as establishing both general and specific causation. But that is not because the differential diagnosis itself carries both proofs; it is because the expert, independent of her differential diagnosis, can separately reach both conclusions. Thus, courts are not necessarily wrong in this “never-say-never” hedge. But it certainly invites the mischief of lower courts misreading this conclusion as justification for allowing general causation to go to the jury on the basis of a physician’s testimony that he did a really super, darn good patient exam.
Judge Engelmayer’s opinion does not stumble into that misguidance. His opinion sharply focuses the inquiry for readers: pharmaceutical personal injury plaintiffs need Daubert-qualifying proof that the product they blame can cause the type of injury that they suffered and then Daubert-qualifying proof that it did so. As the opinion crisply summarizes the point later in its discussion: “Simply put, a jury’s role is not to engage in impermissible ‘scientific guesswork.’”
Finally, the opinion well serves the general causation jurisprudence in one other way. Judge Engelmayer seems to have come quite close to ruling that an absence of expert evidence of general causation necessarily compels the entry of summary judgment in pharmaceutical cases. Perhaps convinced of the wisdom of the “never-say-never” approach here, he stepped back from that leap. What he did next, however, verifies the harmlessness of that choice. For the better part of the next seventeen pages, the opinion carefully weighed the full array of the Mirena plaintiffs’ non-expert substitutes for proving general causation. After a searching consideration, each was rejected. Thus, the court’s “never-say-never” pause led to an exhaustive assessment that answered the essential question that general causation always poses: was there a reliable factual basis on which a jury could find that Mirena causes IIH, or would jurors be left to guess?
The opinion in In re Mirena IUS Levonorgestrel-Related Products Liability Litigation (No. 11) gives general causation the importance, time, and care it deserves. It is a model for similar analyses by other jurists and thus represents a formidable contribution to the jurisprudence in this area.
* William M. Janssen is a professor of law at the Charleston School of Law in Charleston, South Carolina, where he teaches products liability, mass torts, civil procedure, and constitutional law.
 The Sculptures of Vasa, Vasa Museet, https://www.vasamuseet.se/en/vasa-history/art (last visited Mar. 29, 2020). See generally Timeline, Vasa Museet, https://www.vasamuseet.se/en/vasa-history/timeline (last visited Mar. 29, 2020). The tale of the Vasa and of the museum where the now-raised ship is displayed is wonderfully told by the curators of the Vasa Museum, one of Scandinavia’s most visited tourist sites.
 See Timeline, supra note 1.
 See The Disaster, Vasa Museet, https://www.vasamuseet.se/en/vasa-history/disaster (last visited Mar. 29, 2020).
 387 F. Supp. 3d 323 (S.D.N.Y. 2019) [hereinafter IIH MDL Summary Judgment].
 See About Mirena—What is Mirena, Bayer, https://www.mirena-us.com/about-mirena/ (last visited Mar. 7, 2020).
 See Bayer HealthCare Pharms., Inc., Mirena Package Insert §§ 11.1 & 12.1 (FDA approved June 8, 2017), available at https://labeling.bayerhealthcare.com/html/products/pi/Mirena_PI.pdf (last visited Mar. 7, 2020) (noting mechanism of action is believed to be “thickening the cervical mucus preventing passage of sperm into the uterus, inhibition of sperm capacitation or survival, and alteration of the endometrium”).
 See id. In October 2019, Bayer submitted a Supplemental New Drug Application to extend Mirena’s indication for up to six years, based on the results of a phase 3 extension trial. See Bayer Submits Supplemental New Drug Application to FDA for Intrauterine Device (IUD) Mirena® (levonorgestrel-releasing intrauterine system) 52 mg, Cision PR Newswire, https://www.prnewswire.com/news-releases/bayer-submits-supplemental-new-drug-application-to-fda-for-intrauterine-device-iud-mirena-levonorgestrel-releasing-intrauterine-system-52-mg-300942050.html (last visited Mar. 7, 2020).
 See About Mirena—Q&A, Bayer, https://www.mirena-us.com/q-and-a/ (last visited Mar. 7, 2020).
 See World Health Organization, Model List of Essential Medicines 22.1.3 (21st list 2019) (Intrauterine devices–levonorgestrel-releasing intrauterine device–intrauterine system with reservoir containing 52 mg of levonorgestrel).
 See In re Mirena IUD Prod. Liab. Litig., 202 F. Supp. 3d 304 (S.D.N.Y. 2016), aff’d, 713 F. App’x 11, 13 (2d Cir. 2017). See generally In re Mirena IUD Prod. Liab. Litig., 938 F. Supp. 2d 1355 (J.P.M.L. 2013) (granting MDL centralization).
 See In re Mirena IUD Prod. Liab. Litig., 169 F. Supp. 3d 396 (S.D.N.Y. 2016) (discussing Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579 (1993)).
 See In re Mirena IUD Prod. Liab. Litig., 713 F. App’x 11 (2d Cir. 2017).
 See id.
 Mirena MDL v. Bayer Healthcare Pharm., Inc., __ U.S. __, 138 S. Ct. 1299 (2018).
 See IIH MDL Summary Judgment, supra note 4, at 327.
 See id. at 331.
 See id. at 337 n.2.
 See In re Mirena IUS Levonorgestrel-Related Prods. Liab. Litig., 38 F. Supp. 3d 1380 (J.P.M.L. 2014).
 See In re Mirena IUS Levonorgestrel-Related Prod. Liab. Litig. (No. II), 249 F. Supp. 3d 1357, 1359 (J.P.M.L. 2017). The Panel added that “[i]ssues concerning general causation, the background science, and Mirena’s labeling and regulatory history with respect to the alleged injury will be common to all actions,” and, thus, MDL centralization would “eliminate duplicative discovery, prevent inconsistent pretrial rulings on Daubert and other issues, and conserve the resources of the parties, their counsel, and the judiciary.” Id. at 1361.
 IIH MDL Summary Judgment, supra note 4, at 336 (quoting In re Bausch & Lomb Inc. Contact Lens Sol. Prods. Liab. Litig., 693 F. Supp. 2d 515, 520 (D.S.C. 2010)).
 See id. at 333–34 (summarizing Daubert ruling reported at In re Mirena IUS Levonorgestrel-Related Prod. Liab. Litig. (No. II), 341 F. Supp. 3d 213, 233 (S.D.N.Y. 2018)). The court chose not to rule on plaintiffs’ pending motions to exclude Bayer’s expert general causation witnesses, surmising that those motions might be rendered moot by Bayer’s anticipated summary judgment motion. See id. at 334.
 See id. at 335.
 See id. at 337–39.
 Id. at 337.
 Id. at 339.
 See id. at 340. See also id. at 336 (quoting Celotex Corp. v. Catrett, 477 U.S. 317, 322-23 (1986) (“there can be ‘no genuine issue as to any material fact,’ since a complete failure of proof concerning an essential element of the nonmoving party’s case necessarily renders all other facts immaterial”)).
 See id. at 341–43.
 See id. at 343.
 See id. at 340.
 See id. (quoting In re Rezulin Prods. Liab. Litig., 441 F. Supp. 2d 567, 578 (S.D.N.Y. 2006)).
 See id. at 344. See also id. (“At root, plaintiffs seek to exhume excluded testimony of an intrinsically expert nature and to invite a lay jury to derive from it the very proposition that the Court precluded plaintiffs’ experts from offering.”).
 See id. at 346.
 See id. at 348.
 See id. at 349–50.
 See id. at 351–53.
 See id. at 353–57.
 See id. at 357–58.
 See, e.g., Greenman v. Yuba Power Prod., Inc., 377 P.2d 897, 900 (Cal. 1963) (“A manufacturer is strictly liable in tort when an article he places on the market, knowing that it is to be used without inspection for defects, proves to have a defect that causes injury to a human being.”) (emphasis added).
 See Martin v. Herzog, 126 N.E. 814, 816 (N.Y. 1920) (Cardozo, J.).
 See David G. Owen, Products Liability Law § 5.3, at 257 (3d ed. 2015).
 See id. § 11.2, at 721 (“Proof of causation in toxic substance cases always involves, at least implicitly, two separate forms of causal proof: (1) general causation and (2) specific causation. To establish “general causation,” a plaintiff must establish that the suspect agent is capable of causing the particular injury or illness suffered by the plaintiff.”) (footnotes omitted).
 Rosen v. Ciba-Geigy Corp., 78 F.3d 316, 319 (7th Cir. 1996) (Posner, J.).
 Id. (“the courtroom is not the place for scientific guesswork, even of the inspired sort . . . . There may be evidence to back up [an expert’s] claim . . . ”).
 Prosser and Keeton on the Law of Torts § 41, at 265 (5th ed. 1984).
 IIH MDL Summary Judgment, supra note 4, at 358.
 This is not to disparage his colleague Judge Cathy Seibel’s impressive treatment of the general causation issue in the first Mirena MDL, a treatment which Judge Engelmayer credited as informing his IIH decision. See id. at 329–30.
 Id. at 337.
 Id. at 339.
 Id. at 340.
 See, e.g., In re Lipitor (Atorvastatin Calcium) Mktg., Sales Practices & Prod. Liab. Litig. (No II) MDL 2502, 892 F.3d 624, 643–44 (4th Cir. 2018) (“differential diagnosis is . . . accomplished by ‘determining the possible causes for the patient’s symptoms and then eliminating each of these potential causes until reaching one that cannot be ruled out or determining which of those that cannot be excluded is the most likely’”) (emphasis added; citations omitted); Kirk v. Schaeffler Grp. USA, Inc., 887 F.3d 376, 392 (8th Cir. 2018) (“A differential etiology rules in plausible causes and then systematically rules out less plausible causes until a most plausible cause emerges.”) (emphasis added); McManaway v. KBR, Inc., 852 F.3d 444, 454–55 (5th Cir. 2017) (“a differential diagnosis is only relevant after general causation has been reliably established ‘because a differential diagnosis presumes that chemical X can cause condition Y generally, but does not itself so prove’”) (emphasis added; citation omitted); Johnson v. Arkema, Inc., 685 F.3d 452, 468–69 (5th Cir. 2012) (“an expert may not rely on a differential diagnosis to circumvent the requirement of general causation;” “before courts can admit an expert’s differential diagnosis, which, by its nature, only addresses the issue of specific causation, the expert must first demonstrate that the chemical at issue is actually capable of harming individuals in the general population,” rejecting a differential diagnosis “based on the presumption” of general causation); Huerta v. BioScrip Pharmacy Servs., Inc., 429 F. App’x 768, 773 (10th Cir. 2011) (“a differential diagnosis can be admissible if the district court concludes that it is reliable and if general causation has been established”) (emphasis added); Kilpatrick v. Breg, Inc., 613 F.3d 1329, 1342 (11th Cir. 2010) (differential diagnosis “involves a process of compiling, or ruling in, a comprehensive list of possible causes that are generally capable of causing the illness or disease at issue, and then systematically and scientifically ruling out specific causes until a final, suspected cause remains”; “It assumes the existence of general causation, and focuses instead on specific causation. The expert must show through reliable evidence that the remaining cause ruled in as actually being capable of causing the condition.”).
 See, e.g., Sims v. Kia Motors of Am., Inc., 839 F.3d 393, 402 (5th Cir. 2016) (noting that court has never held “that differential diagnosis may never satisfy Daubert. Rather we observe that the district court has broad discretion to make the fact-specific inquiry in a given case as to whether such an approach is sufficiently reliable, especially in the absence of evidence ‘ruling in’ an expert’s conclusion.”); Ruggiero v. Warner-Lambert Co., 424 F.3d 249, 254 (2d Cir. 2005) (“We cannot say that a differential diagnosis may never provide a sufficient basis for an opinion as to general causation. There may be instances where, because of the rigor of differential diagnosis performed, the expert’s training and experience, the type of illness or injury at issue, or some other case-specific circumstance, a differential diagnosis is sufficient to support an expert’s opinion in support of both general and specific causation.”); C.W. ex rel. Wood v. Textron, Inc., 807 F.3d 827, 838–39 (7th Cir. 2015) (“We disagree with the district court’s categorical exclusion of differential etiology as a method to establish general causation . . . . [T]here may be a case where a rigorous differential etiology is sufficient to help prove, if not prove altogether, both general and specific causation.”).
 See Edward J. Imwinkelried, The Admissibility and Legal Sufficiency of Testimony About Differential Diagnosis (Etiology): Of Under—and Over—Estimations, 56 Baylor L. Rev. 391, 406 (2004) (“Differential etiology analysis is not formal scientific proof. On the general causation issue, an opinion derived from such analysis seems to be at most an educated guess.”) (footnotes omitted); id. at 414 (“Standing by itself, the differential etiology opinion will almost always be insufficient to establish general causation. Although etiological analysis may identify the factor as the most likely cause in an individual case, without more that analysis raises only a suspicion on the issue of general causation.”).
 IIH MDL Summary Judgment, supra note 4, at 344.
 See id. at 341 (“It is well established that ‘expert testimony is required to establish causation’ where the issue of causation is ‘beyond the knowledge of lay jurors.’ Other courts, surveying the law of the 50 states and territories, have concluded that each jurisdiction typically adheres to this principle.”).
 See also In re Lipitor (Atorvastatin Calcium) Mktg., Sales Practices & Prod. Liab. Litig. (No II) MDL 2502, 892 F.3d 624, 647 (4th Cir. 2018) (“There may be cases involving complex issues in which a party admission standing alone can suffice to avoid summary judgment. But we would expect those cases to be rare indeed.”).