Travis Abbott v. E. I. du Pont de Nemours & Co.

                               RECOMMENDED FOR PUBLICATION
                               Pursuant to Sixth Circuit I.O.P. 32.1(b)
                                      File Name: 22a0261p.06

                   UNITED STATES COURT OF APPEALS
                                  FOR THE SIXTH CIRCUIT



                                                            ┐
 IN RE: E. I. DU PONT DE NEMOURS AND COMPANY C-8
                                                            │
 PERSONAL INJURY LITIGATION.
                                                            │
  ___________________________________________               │
 TRAVIS ABBOTT; JULIE ABBOTT,                                >        No. 21-3418
                                                            │
                                Plaintiffs-Appellees,       │
                                                            │
        v.                                                  │
                                                            │
                                                            │
 E. I. DU PONT DE NEMOURS AND COMPANY,                      │
                               Defendant-Appellant.         │
                                                            ┘

  Appeal from the United States District Court for the Southern District of Ohio at Columbus.
        Nos. 2:13-md-02433; 2:17-cv-00998—Edmund A. Sargus, Jr., District Judge.

                                    Argued: June 10, 2022

                            Decided and Filed: December 5, 2022

             Before: BATCHELDER, STRANCH, and DONALD, Circuit Judges.

                                     _________________

                                           COUNSEL

ARGUED: Damond R. Mace, SQUIRE PATTON BOGGS (US) LLP, Cleveland, Ohio, for
Appellant. Matthew W.H. Wessler, GUPTA WESSLER PLLC, Washington, D.C., for
Appellees. ON BRIEF: Damond R. Mace, Aneca E. Lasley, SQUIRE PATTON BOGGS (US)
LLP, Cleveland, Ohio, Lauren S. Kuley, Colter L. Paulson, SQUIRE PATTON BOGGS (US)
LLP, Cincinnati, Ohio, John A. Burlingame, SQUIRE PATTON BOGGS (US) LLP,
Washington, D.C., for Appellant. Matthew W.H. Wessler, GUPTA WESSLER PLLC,
Washington, D.C., Rachel Bloomekatz, BLOOMEKATZ LAW LLC, Columbus, Ohio, Jon C.
Conlin, F. Jerome Tapley, Elizabeth E. Chambers, Nina Towle Herring, Mitchell Theodore, Brett
Thompson, CORY WATSON, PC, Birmingham, Alabama, for Appellees.                     Brian D.
Schmalzbach, McGUIRE WOODS LLP, Richmond, Virginia, Mark A. Behrens, SHOOK,
HARDY & BACON, L.L.P. Washington, D.C., Sean P. Wajert, SHOOK, HARDY & BACON,
L.L.P., Philadelphia, Pennsylvania, Anne Marie Sferra, Christopher P. Gordon, BRICKER &
 No. 21-3418                      In re E. I. du Pont de Nemours & Co.                    Page 2


ECKLER LLP, Columbus, Ohio, Jeffrey R. White, AMERICAN ASSOCIATION FOR
JUSTICE, Washington, D.C., Alison Borochoff-Porte, POLLOCK COHEN LLP, New York,
New York, Gary A. Davis, DAVIS & WHITLOCK, P.C., Asheville, North Carolina, for Amici
Curiae.

       STRANCH, J., delivered the opinion of the court in which DONALD, J., joined in full,
and BATCHELDER, J., joined in part. BATCHELDER, J. (pp. 30–46), delivered a separate
opinion concurring in part and dissenting in part.

                                        _________________

                                             OPINION
                                        _________________

       JANE B. STRANCH, Circuit Judge. In the 1950s, E. I. du Pont de Nemours & Co.
(DuPont) began discharging vast quantities of C-8—a “forever” chemical that accumulates in the
human body and the environment—into the Ohio River, landfills, and the air surrounding its
plant in West Virginia, contaminating the communities’ water sources. By the 1960s, DuPont
learned that C-8 is toxic to animals and, by the 1980s, that it is potentially a human carcinogen.
Despite these and other warnings, DuPont’s discharges increased between 1984 and 2000.
By the early 2000s, evidence confirmed that C-8 caused several diseases among the members of
the communities drinking the contaminated water, which led to a class action lawsuit against
DuPont. The parties undertook negotiations and ultimately entered into a unique settlement
agreement in which DuPont promised to carry out treatment of the affected water and to fund a
scientific process that would inform the class members and communities about the dangers of
and harms from C-8 exposure. In service of that process, the class voted to make receipt of the
cash award contingent on a full medical examination to test for and collect data on C-8 exposure.
A panel of scientists then conducted an approximately seven-year epidemiological study of the
blood samples and medical records of over 69,000 affected community members, during which
litigation against DuPont was paused. The parties’ agreement limited the legal claims that could
be brought against DuPont based on the study’s determination of which diseases prevalent in the
communities were likely linked to C-8 exposure. The resulting cases were consolidated in a
multidistrict litigation (MDL).
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                        Page 3


        After two bellwether trials and a post-bellwether trial reached jury verdicts against
DuPont, the parties settled the remaining cases. That did not end all the C-8 litigation, as more
class members filed suit when they became sick or discovered the connection between their
diseases and C-8, including this case brought by Travis and Julie Abbott. At the Abbotts’ trial,
the district court applied collateral estoppel to specific issues that were unanimously resolved in
the three prior jury trials, excluded certain evidence from the trial based on the initial settlement
agreement, and rejected DuPont’s motion for a directed verdict on its statute-of-limitations
defense. The jury found for the Abbotts. On appeal, DuPont challenges those three district court
decisions. For the reasons that follow, we AFFIRM the judgment of the district court in full.

                                       I. BACKGROUND

        The Abbotts’ case has its roots in the 1950s, when DuPont began using C-8 to
manufacture Teflon© products at its Washington Works Plant in Parkersburg, West Virginia. C-
8, or perfluorooctanoic acid (PFOA), is a synthetic organic chemical that is soluble in water and
persists in both the human body and the environment. DuPont discharged C-8 into the air, the
Ohio River, and landfills without limits until the early 2000s, as explained below.

        DuPont learned in the 1960s that C-8 was toxic to animals and was reaching groundwater
in the communities surrounding its plant. By the late 1980s, DuPont internally considered the
chemical a possible human carcinogen and found that it stayed in the human bloodstream for
years. Despite warnings from its C-8 supplier on proper disposal and the availability of a
substitute, DuPont increased its C-8 discharges between 1984 and 2000. Documents obtained in
discovery in a 1998 case against DuPont revealed the contamination and kicked off a wave of
further litigation.

        A. The Leach Class Action and Settlement

        In the early 2000s, individuals who had consumed the contaminated water sued DuPont
in West Virginia state court in Leach v. E. I. du Pont de Nemours & Co., No. 01-C-698 (W. Va.
Cir. Ct.). They brought numerous claims under West Virginia common law, seeking equitable,
injunctive, and declaratory relief, and punitive and compensatory damages for alleged injuries
arising from C-8 exposure. In 2002, the West Virginia trial court certified a class of nearly
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                  Page 4


80,000 individuals “whose drinking water is or has been contaminated with” C-8 attributable to
DuPont’s C-8 discharges from the Washington Works Plant. (MDL R. 820-8, Leach Agreement,
PageID 11807)1 In 2005, the trial court approved the parties’ class-wide settlement agreement,
called the Leach Agreement in the later MDL proceedings. (See generally id.)

         The Leach Agreement fashioned unique measures to be undertaken over time to obtain
scientific and medical information in order to address the harms to the affected workers and
communities. For example, the parties agreed that DuPont would fund the design, installation,
operation, and maintenance of a water treatment project designed to “reduce the levels of C-8 in
the affected water supply to the lowest practicable levels as specified by the individual Public
Water Districts.” (Id., PageID 11821) The Leach Plaintiffs were also concerned about how the
members of the class were and would be harmed by C-8, so the class voted to make class
members’ receipt of the cash award reached in the settlement contingent on a full medical
examination.2 The medical data that resulted from those examinations were used in a broad
epidemiological study into the effects of C-8 on the community, which DuPont was required to
fund. (See MDL R. 2416-3, PageID 35731–32; MDL R. 820-8, PageID 11823) The community
health study was performed by the Science Panel, three independent epidemiologists jointly
selected by DuPont and the Plaintiffs, that carried out research on diseases among the
communities exposed to C-8 in the water districts around Washington Works. (MDL R. 820-8,
PageID 11823) The Leach Agreement also led to medical monitoring of diseases the Science
Panel deemed linked to C-8 for class members. (Id., PageID11826–27)

         The parties also agreed to a unique procedure that defined the parameters of legal actions
the Leach Plaintiffs could bring against DuPont based on the results of the epidemiological


         1
          The record contains documents filed in Abbott’s individual case, 2:17-cv-998 on the district court docket,
documents filed on the MDL docket, 2:13-md-2433, as well as documents filed in earlier individual cases against
DuPont. Where relevant, our opinion refers to documents filed on Abbott’s docket as “R.” and documents found on
the MDL docket as “MDL R.” Where documents from earlier individual cases are relevant, the case name is
included before the “R.” (e.g., “Bartlett R.” for documents from the Bartlett docket).
         2
           See Nathaniel Rich, The Lawyer Who Became DuPont’s Worst Nightmare, N.Y. Times (Jan. 6, 2016),
https://www.nytimes.com/2016/01/10/magazine/the-lawyer-who-became-duponts-worst-nightmare.html. A Leach
“Plaintiff” or “class member” is defined as those individuals who had consumed drinking water with 0.05 parts per
billion (ppb) or more “C-8 attributable to releases from Washington Works” from at least one of six specific public
water districts, private wells in those districts, or otherwise specified private wells. (MDL R. 820-8, PageID 11807)
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                         Page 5


study. For each disease studied, the Science Panel would ultimately issue either a “Probable
Link finding” or a “No Probable Link finding.” A “Probable Link” means, “based upon the
weight of the available scientific evidence, it is more likely than not that there is a link between
exposure to C-8 and a particular Human Disease among Class Members.” (Id., PageID 11805)
Once the Science Panel released its results, the right of individual class members to pursue their
personal injury and wrongful death claims against DuPont was limited to diseases with a
Probable Link finding. (Id., PageID 11811) In these lawsuits related to linked diseases, DuPont
agreed not to contest general causation—“that it is probable that exposure to C-8 is capable of
causing a particular Human Disease”—but it retained the right to contest specific causation and
assert any other defenses not barred by the Leach Agreement. (Id., PageID 11804, 11811) The
Agreement defined specific causation to mean “that it is probable that exposure to C-8 caused a
particular Human Disease in a specific individual.” (Id., PageID 11806) For diseases for which
the Science Panel reported a “No Probable Link finding” or found no association with C-8
exposure, class members would be forever barred from bringing claims for injury or death
against DuPont for C-8 exposure based on those diseases. (Id., PageID 11810) The Leach
Plaintiffs also agreed to refrain from seeking immediate relief—through a conditional release of
claims and a covenant not to sue DuPont for C-8 exposure—until the Science Panel completed
its study. (See id., PageID 11810–11)

       For seven years, the Science Panel engaged in the specified epidemiological study. In
one of the largest domestic epidemiological studies ever, over 69,000 class members provided
blood samples and medical records. (MDL R. 4306, Disp. Mot. Order No. 12 Denying JMOL on
Bartlett Claims, PageID 89502) In 2012, using this data and its own established protocols, the
Science Panel reported Probable Link findings as defined in the Leach Agreement for six
diseases: kidney cancer, testicular cancer, thyroid disease, ulcerative colitis, diagnosed high
cholesterol, and pregnancy-induced hypertension and preeclampsia. (MDL R. 5285, Disp. Mot.
Order on Issue Preclusion, PageID 128535) The Science Panel reached a No Probable Link
Finding for approximately 50 diseases; class members with those diseases were forever barred
from bringing claims against DuPont based on those diseases, even if later discovered facts and
science revealed a link to C-8. (Id.; MDL R. 820-8, PageID 11810)
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                   Page 6


         B. The MDL and Prior Appeal

         After the Science Panel’s Probable Link findings, the members of the Leach class with
linked diseases brought approximately 3,500 cases against DuPont pursuant to the Leach
Agreement. At DuPont’s request, the federal courts consolidated those cases in an MDL in the
Southern District of Ohio. The district court overseeing the MDL engaged in a months-long
process with the parties to identify 20 cases for discovery, then to narrow that list further for
bellwether trials. In guiding the parties’ selections, both “[t]he parties and the Court intend[ed]”
that the bellwether plaintiffs selected for initial discovery and ultimately trial “reflect a
representative sampling of cases which [would] provide meaningful information for the broader
population of cases.”3 Toward this end, the parties limited their initial plaintiff designations
according to specified parameters, and the court established a detailed procedure for selection of
the initial bellwether trials.        The parties were ordered to exchange lists of four proposed
plaintiffs, then each side was permitted to strike one of the other side’s selections. Ultimately,
the parties proposed and the court accepted six cases—three selected by the Plaintiffs’ Steering
Committee, three by DuPont—for bellwether trials. The district court overseeing the MDL also
oversaw the cases as they went to trial or settled.

         In the first bellwether trial—a case selected by DuPont—the jury awarded Carla Bartlett
$1.6 million in compensatory damages against DuPont for her state law tort claims related to
kidney cancer. See Bartlett v. DuPont, No. 13-cv-170. Five bellwether cases remained. The
next trial, Freeman v. Dupont, No. 13-cv-1103, a case selected by Plaintiffs, included a
negligence claim arising from Freeman’s testicular cancer and resulted in a jury verdict for
Freeman. DuPont settled the remaining bellwether cases with the Plaintiffs. The Plaintiffs’
Steering Committee then selected the first of the non-bellwether cases to go to trial in 2016.
Vigneron v. DuPont, No. 13-cv-136.                That case brought negligence claims, used the jury
instructions on negligence given at the Bartlett and Freeman trials, and resulted in a jury verdict
awarding $2 million in compensatory damages to the plaintiff.


         3
           The court’s and parties’ intentions were aligned with the broader purpose of bellwether trials, which serve
the “twin goals” of being “informative indicators of future trends and catalysts for an ultimate resolution.” Eldon E.
Fallon et al., Bellwether Trials in Multidistrict Litigation, 

, 2343 (2008).
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                         Page 7


        While DuPont continued litigation in the district court, it appealed the Bartlett case. In
that appeal, DuPont argued that the district court had interpreted the Leach Agreement in a way
that made the Bartlett trial and all other MDL cases fundamentally unfair. The district court had
determined that the bargain struck by the parties as set out in the language and defined terms in
the Leach Agreement barred any challenges to general causation. DuPont claimed that decision
was “[a] threshold contract interpretation error [that] eliminated the heart of a critical defense for
DuPont in each of the 3,500 cases” in the MDL and resulted in incorrect evidentiary rulings.
(MDL R. 5285, PageID 128547 (quoting Bartlett v. DuPont, No. 16-3310 (6th Cir.), DuPont
Appellant Br. at 1, 18))

        In February 2017, after oral argument but before we issued a decision in Bartlett, DuPont
announced a settlement with the remaining MDL cases, including Bartlett, and withdrew that
appeal. Although it halted further proceedings in Bartlett, the global settlement did not entirely
end the litigation. As the vast majority of the MDL cases wound down, some additional
Plaintiffs covered by the Leach Agreement, including Travis and Julie Abbott, filed cases.

        C. The Abbott Case

        Travis Abbott has lived and worked in and around Pomeroy, Ohio, since childhood.
Consequently, for 20 years—beginning at only 6 years old—Abbott was exposed to C-8
contaminated water at home and in his community. At age 16, Abbott found a mass in his left
testicle, and, after surgically removing his testicle, doctors diagnosed him with testicular cancer.
He did not experience a relapse until 10 years later when he was beginning to plan a family with
his wife, Julie, while still living in the Pomeroy region. In October 2015, Abbott sought medical
help for pain in his remaining testicle. A definitive diagnosis of testicular cancer came only after
doctors removed his testicle to conduct a pathology analysis. The spread of the cancer to his
lymph nodes required further surgery, and Abbott must take testosterone injections due to his
loss of both testicles.

        Travis and Julie Abbott sued DuPont in November 2017. The district court scheduled the
Abbotts’ case for a joint trial with that of another couple, the Swartzes, in early 2020. After
rejecting DuPont’s renewed challenges to the district court’s MDL rulings on the meaning of the
 No. 21-3418                      In re E. I. du Pont de Nemours & Co.                     Page 8


Leach Agreement, the district court granted partial summary judgment to the Abbotts on the
duty, breach, and foreseeability elements of Travis Abbott’s negligence claims based on
collateral estoppel.   The court further held that collateral estoppel precluded DuPont from
relitigating (1) the interpretation of the Leach Agreement and its application to evidentiary issues
and (2) the inapplicability of the Ohio Tort Reform Act (OTRA) to Travis Abbott’s claims.

       The month-long jury trial for the Abbott and Swartz cases began in January 2020. In
evidentiary rulings, the district court prohibited DuPont from offering evidence and testimony
that the court concluded would violate the Leach Agreement, including testimony asserting that
Travis Abbott’s level of C-8 exposure was insufficient to cause his cancers. The court instructed
jurors that 0.05 ppb was a threshold level for general causation, but that specific causation was
still at issue in the case. DuPont then presented testimony about the concentration of C-8 in
Abbott’s bloodstream and C-8’s half-life in the human body, along with expert opinions on
potential alternative causes of his cancers. The jury found for both Travis and Julie Abbott,
awarding them $40 million and $10 million in damages, respectively. The district court later
applied the Ohio Tort Reform Act to Julie Abbott’s award, reducing it to $250,000. Because the
jury did not agree on the Swartzes’ claims related to Mrs. Swartz’s kidney cancer, that case
concluded in a mistrial.

       This appeal in the Abbotts’ case followed.

                                           II. ANALYSIS

       DuPont raises several challenges to the district court’s decisions on appeal. First, it
challenges the order granting the MDL Plaintiffs’ motion for application of nonmutual offensive
collateral estoppel to duty, breach, general causation, and the inapplicability of the OTRA.
Based on that order, those issues were not submitted to the jury for its deliberations in the
Abbotts’ case. Next, DuPont argues that several of the district court’s evidentiary rulings related
to specific causation were erroneous. And finally, DuPont asserts that the district court abused
its discretion by entering a directed verdict denying DuPont’s statute of limitation defense. We
address each challenge in turn.
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                        Page 9


         A. Nonmutual Offensive Collateral Estoppel

         A district court has “broad discretion to determine” whether to apply collateral estoppel.
Parklane Hosiery Co. v. Shore, 

 (1979). We review de novo whether the
district court’s decision to do so was error. Abbott v. Michigan, 

 (6th Cir.
2007).

         In successive federal diversity actions, we apply state law to determine whether a prior
decision has preclusive effect, so long as the state rule is not “incompatible with federal
interests.” Prod. Sols. Int’l, Inc. v. Aldez Containers, LLC, 

, 457–58 (6th Cir. 2022)
(quoting Semtek Int’l Inc. v. Lockheed Martin Corp., 

 (2001)).

         Ohio courts generally apply issue preclusion when that issue “was actually and directly
litigated in the prior action” and “a court of competent jurisdiction” decided the issue, and “the
party against whom collateral estoppel is asserted was a party in privity with a party to the prior
action.” State ex rel. Jefferson v. Russo, 

 (Ohio 2020) (quoting Thompson v.
Wing, 

 (Ohio 1994)). The “fact or . . . point” in question must have been
“actually and necessarily litigated and determined” as part of a final judgment. Fort Frye Tchrs.
Ass’n, OEA/NEA v. State Emp. Rels. Bd., 

 (Ohio 1998); see State v.
Williams, 

 (Ohio 1996). And the party against whom estoppel is sought
must have had a “full and fair opportunity” to litigate the issue in the previous action. Walden v.
State, 

 (Ohio 1989) (quoting Hicks v. De La Cruz, 

(Ohio 1977)). In sum, Ohio’s standard is very similar to the federal one. See Smith v. S.E.C.,

, 362 (6th Cir. 1997) (en banc).

         DuPont initially claimed that Ohio law forbids the use of non-mutual offensive collateral
estoppel altogether.    While the “principle of mutuality” is generally a “prerequisite to the
application of collateral estoppel,” the Ohio Supreme Court has explicitly “recogniz[ed] the need
in certain instances for the flexibility and exceptions to such rule.” Goodson v. McDonough
Power Equip., Inc., 

 (Ohio 1983). Where a “party defendant clearly had his
day in court on the specific issue brought into litigation within the later proceeding, the non-party
plaintiff [can] rely upon the doctrine of collateral estoppel to preclude the relitigation of that
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                 Page 10


specific issue.” 

. Ohio is “willing to relax the [mutuality] rule where justice would
reasonably require it.”4 

.

         If Ohio’s requirements are met, the Supreme Court has offered four additional
considerations that may suggest caution in determining whether to apply offensive nonmutual
collateral estoppel against a party. Parklane Hosiery Co., 439 U.S. at 329–31; see Goodson, 

 & n.12 (discussing Parklane Hosiery factors); O’Nesti v. DeBartolo Realty Corp.,

, 809 (Ohio 2007) (same).                  First, courts should avoid applying nonmutual
offensive collateral estoppel where it would encourage “a ‘wait and see’ attitude” among
potential plaintiffs hoping “that the first action by another plaintiff will result in a favorable
judgment.” Parklane Hosiery Co., 

. Second, courts should not use the doctrine if
the defendant did not have a reason “to defend vigorously, particularly if future suits [were] not
foreseeable.” 

 Third, the doctrine should not apply “if the judgment relied upon as a basis for
the estoppel is itself inconsistent with one or more previous judgments in favor of the
defendant.” 

 Fourth and finally, courts should avoid the use of nonmutual offensive collateral
estoppel if the later action would give “the defendant procedural opportunities unavailable in the
first action that could readily cause a different result.” 

.

                 1. Application of Collateral Estoppel to the Negligence Claims

         As an initial matter, we address DuPont’s claim that our court has placed additional
constraints on the use of nonmutual offensive collateral estoppel in mass tort cases. DuPont
points to a footnote in In re Bendectin Products Liability Litigation, 

 (6th Cir.
1984), in which we noted that the Supreme Court’s decision in Parklane Hosiery “explicitly
stated that offensive collateral estoppel could not be used in mass tort litigation.” 

 at 305 n.11.
DuPont’s interpretation of Bendectin, however, is inconsistent with the Supreme Court’s clear


         4
           Even if mutuality were required, it is a “somewhat amorphous” concept under Ohio law. Brown v.
Dayton, 

 (Ohio 2000). A contractual relationship is not required; a “mutuality of interest,
including an identity of desired result,” may be sufficient. 

 “As a general matter, privity ‘is merely a word used
to say that the relationship between the one who is a party on the record and another is close enough to include that
other within the res judicata.’” 

 (quoting Thompson v. Wing, 

 (Ohio 1994)). Not only do
Plaintiffs share a contractual relationship with DuPont—the Leach Agreement—but they also share a mutuality of
interest and identity of desired result with all other plaintiffs in this MDL, who, like Abbott, are Leach class
members, allege injury due to drinking water contaminated with C-8, and seek the same result.
 No. 21-3418                          In re E. I. du Pont de Nemours & Co.                                   Page 11


pronouncement in Parklane Hosiery that “the preferable approach for dealing with” the fairness
concerns regarding offensive collateral estoppel “is not to preclude the use of offensive estoppel”
but instead to provide “broad discretion” to trial courts determining when it applies. Parklane
Hosiery Co., 

; see also City of Cleveland v. Cleveland Elec. Illuminating Co., 

 (6th Cir. 1984). Bendectin, an appeal of a district court’s class-certification
decision, focused on the requirements of Rule 23, and our opinion mentioned but did not hinge
on whether district courts could ever apply nonmutual offensive collateral estoppel in mass tort
cases. Bendectin, 749 F.2d at 304–05. No court has followed the Bendectin footnote beyond
agreeing that courts should not use offensive collateral estoppel in mass tort cases in ways
inconsistent with the Parklane Hosiery factors. See, e.g., In re Air Crash at Detroit Metro.
Airport, Detroit, Mich. on Aug. 16, 1987, 

, 324–25 (E.D. Mich. 1991)
(explaining that nonmutual offensive collateral estoppel could be used in mass tort cases if
consistent with the instruction in Parklane Hosiery and “should be developed on a case-by-case
basis”). Ohio has similarly instructed that offensive collateral estoppel is permissible in the mass
tort context where the Parklane Hosiery standards are applied.5 See Goodson, 

.

         Ohio’s collateral estoppel factors and the additional considerations delineated in Parklane
Hosiery provide the framework for the district court’s exercise of its broad discretion. We will
not place DuPont’s requested additional constraints on that discretion given the Supreme Court’s
clear instruction.

         In applying offensive collateral estoppel, the district court concluded that the three prior
jury trials—Bartlett, Freeman, and Vigneron—raised and litigated to a final conclusion the same
questions of duty, breach, and foreseeability raised in Travis Abbott’s negligence claims.
DuPont asserts that this use of nonmutual offensive collateral estoppel violated its due process
rights because duty, breach, and foreseeability in the three prior trials were factually distinct.


         5
          DuPont frames Goodson as demonstrating the Ohio Supreme Court’s disapproval of mass-tort collateral
estoppel. But the language DuPont quotes from Goodson that expresses caution about applying “a decision made by
one jury in the context of one set of facts” to “all subsequent cases involving separate underlying factual
circumstances” is specific to product liability litigation; it is not about mass-tort litigation generally. 

.
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                         Page 12


The Abbotts dispute that factual argument and counter that the use of collateral estoppel here
“serve[d] the core principles of judicial integrity and economy,” and the doctrine “was made for
a case like this one.” We apply Ohio law and Parklane Hosiery’s considerations in turn.

                        a. Ohio Law

       We begin by determining whether the “identical issue was actually decided in the former
case.” Goodson, 

. Factual differences do exist among the different cases, but
the question is whether any of those factual differences are legally significant—i.e., were crucial
to resolving the issues in the compared cases. See Smith v. Sushka, 

, 969–70 (6th
Cir. 1997) (quoting Monahan v. Eagle Picher Indus., Inc., 

 (Ohio 1984));
see also United States v. Stauffer Chem. Co., 

 (1984). DuPont claims that
duty, breach, and foreseeability were unique to each plaintiff given that each plaintiff was
differently situated. For example, it asserts, the Freeman and Vigneron Plaintiffs argued that
DuPont should have foreseen their injuries because the C-8 concentration in their water districts’
drinking water exceeded DuPont’s voluntary exposure guidelines. Travis Abbott’s water was
below these guidelines at relevant times. DuPont also contends that Abbott conceded that
DuPont was unaware of C-8 in his water supply before 2001, unlike the Plaintiffs in Freeman
and Vigneron who asserted that DuPont knew about, but did not warn them of, the C-8 in their
drinking water for over a decade. DuPont argues that these questions of duty, breach, and
foreseeability were so closely tied to the individual plaintiffs that preclusive effect is impossible.

       DuPont’s argument attempts to ignore the fundamental principle that the pertinent factual
issues for the negligence claims in each trial revolved around DuPont’s conduct and knowledge
in relation to the Leach class members. In Bartlett, Freeman, and Vigneron—the cases that
served as the basis for collateral estoppel—each jury received identical instructions on duty,
breach, and foreseeability. Each jury found that DuPont owed a duty to the class member,
breached that duty, and should have foreseen that injury would result from the alleged breach.
To illustrate, consider the jury instructions from the Bartlett case:
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                        Page 13


                                    NEGLIGENCE – DUTY
       To prove the existence of a duty, Mrs. Bartlett must show by a preponderance of
       the evidence that a reasonably prudent person would have foreseen that injury was
       likely to result to someone in Mrs. Bartlett’s position from DuPont’s conduct. In
       deciding whether reasonable prudence was used, you will consider whether
       DuPont should have foreseen, under the circumstances, that the likely result of an
       act or failure to act would cause injuries. The test for foreseeability is not whether
       DuPont should have foreseen the injuries exactly as it happened to Mrs. Bartlett.
       The test is whether under the circumstances a reasonably prudent corporation
       would have anticipated that an act or failure to act would likely cause injuries.

                                  NEGLIGENCE – BREACH
       If you find that DuPont owed Mrs. Bartlett a duty, you must next determine
       whether DuPont breached that duty. A corporation breaches a duty by failing to
       use ordinary care. As I have just instructed, ordinary care is the care that a
       reasonably careful corporation would use under the same or similar
       circumstances.
       If you decide that DuPont did not use ordinary care, then DuPont breached its
       duty of care to Mrs. Bartlett. If you decide that DuPont did use ordinary care,
       then DuPont did not breach its duty of care to Mrs. Bartlett. . . .

            NEGLIGENCE – PROXIMATE CAUSE – FORSEEABLE INJURY
       . . . . For Mrs. Bartlett’s injuries to be considered the natural and probable
       consequence of an act, Mrs. Bartlett must prove that DuPont should have foreseen
       or reasonably anticipated that injury would result from the alleged negligent act.
       The test for foreseeability is not whether DuPont should have foreseen the injury
       exactly as it happened to Mrs. Bartlett. Instead, the test is whether under the
       circumstances a reasonably careful person would have anticipated that an act or
       failure to act would likely result in or cause injuries.

(Bartlett R. 139, Bartlett Final Jury Instructions, PageID 6205–08)

       The instructions must and do reference each specific plaintiff, but their focus, and the
focus of the jury’s inquiry in each of the cases, was on DuPont’s conduct. The instructions state
that a duty exists when “a reasonably prudent person would have foreseen that injury was likely
to result to someone in Mrs. Bartlett’s position from DuPont’s conduct,” explaining that “[t]he
test for foreseeability is not whether DuPont should have foreseen the injuries exactly as it
happened to Mrs. Bartlett.” The instructions—and the law more generally—peg the duty to
whether “a reasonably prudent corporation would have anticipated” that its actions or inactions
would cause injury. Foreseeability in the context of the proximate cause jury instructions
 No. 21-3418                     In re E. I. du Pont de Nemours & Co.                         Page 14


similarly looks to DuPont’s actions. Put simply, these instructions turn on DuPont’s conduct,
not the particulars of Bartlett’s individual circumstances. To say otherwise and adopt DuPont’s
argument would make it virtually impossible to ever find preclusive effect in negligence claims.
The key concept applicable here is that DuPont’s conduct impacted the Plaintiffs in virtually
identical ways—contamination of their water supplies with a carcinogen. The district court was
correct to conclude that the “facts relating to DuPont’s negligence were virtually identical”
across the four trials.

        In sum, we are not persuaded by DuPont’s contention that near factual identity on the
Plaintiffs’ water district, location, exposure, timing, and toxicity is necessary and controlling—
instead of evidence of DuPont’s conduct. But even if that were the standard, we are not
convinced that the Plaintiffs here failed to cross that threshold. DuPont’s emphasis on the factual
differences between Travis Abbott’s case and those in the Freeman and Vigneron trials
overlooks the factual similarities between Abbott and the plaintiff in Bartlett. The record shows
that Abbott and Bartlett were exposed to more than 0.05 ppb of C-8 in the Tuppers Plains-
Chester Water District for overlapping periods of time. (R. 33-2, Expert Report, PageID 343–44;
MDL R. 2807-8, Expert Report, PageID 42884) Bartlett drank C-8 contaminated water in that
district from 1983 to 1989 and 1994 to 2004, while Travis Abbott was exposed from 1983 to
1998 and again from 2000 to 2004. The roughly ten years of corresponding use in the same
water district and similar exposure levels undercut DuPont’s claim that the juries were not
considering comparable facts relevant to duty, breach, and foreseeability. Nor does the record
support DuPont’s contention that its knowledge of contamination in the Vigneron and Freeman
cases sufficiently distinguishes the prior jury trials. DuPont argued in both Bartlett and Abbott
that it did not know it had contaminated their water and that the contamination did not exceed its
internal guidelines. (R. 188, Jan. 24, 2020 Trial Tr., PageID 7684) Nevertheless, the Bartlett
trial resulted in a jury verdict for Bartlett. The factual identity factor supports the district court’s
application of collateral estoppel.

        The next question is whether the resolution of the precluded issues was necessary to the
outcomes in the prior cases. Goodson, 

. There is little doubt that the jury
trials’ decisions on duty, breach, and foreseeability were necessary to each of the verdicts for the
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                      Page 15


earlier Plaintiffs on their negligence claims. See Menifee v. Ohio Welding Prods., Inc., 

 (Ohio 1984). Ohio applies the standard common law test for negligence claims,
which requires a finding on each of those elements. See 

       And finally, we consider whether the prior cases reached final judgment on the merits
and whether DuPont had a sufficient opportunity to litigate the issues in those cases. See
Walden, 

 (quoting Hicks, 

). As to actual litigation, the vast
size of the MDL and individual case dockets belie any argument to the contrary. The record is
clear that DuPont vigorously contested duty, breach, and foreseeability in all the prior trials.
That DuPont settled the Bartlett case after the jury verdict and judgment, while the case was
pending on appeal, does not change the preclusive effect of the district court’s decisions in that
case. See Watermark Senior Living Ret. Cmtys., Inc. v. Morrison Mgt. Specialists, Inc., 

, 426–28 (6th Cir. 2018); Coal. for Gov’t Procurement v. Fed. Prison Indus., Inc., 

, 484–85 (6th Cir. 2005); see also Restatement (Second) of Judgments § 13 cmt. g (1982)
(“[T]hat the parties were fully heard, that the court supported its decision with a reasoned
opinion, that the decision was subject to appeal or was in fact reviewed on appeal, are factors
supporting the conclusion that the decision is final for the purpose of preclusion.”). Thus, as to
the Ohio law that governs issue preclusion, we conclude that the district court’s analysis was
correct.

                       b. The Parklane Hosiery Considerations

       In Parklane Hosiery, the Supreme Court provided additional guidance as to the doctrine
of nonmutual offensive collateral estoppel. The unique parameters established by the Leach
Agreement and the resulting MDL play the key role in applying the Parklane factors here.

       We note first that the Leach Agreement created a limited, closed subset of possible
plaintiffs from the larger, original Leach class. That subset was comprised only of those who
had consumed contaminated water in specific water districts or wells for at least one year prior to
2005 and suffered from at least one of the six identified linked diseases, giving them sufficient
indicia of injury to move forward with individual suits against DuPont.
 No. 21-3418                        In re E. I. du Pont de Nemours & Co.                               Page 16


        The bargained-for exchange that the Leach Agreement established informs the
application of collateral estoppel here. Every class member agreed to release all claims related to
diseases without a Probable Link finding and not to sue DuPont until the Science Panel
completed its multiple-year study. DuPont agreed not to contest general causation.6 In light of
the benefits and concessions embodied in the Agreement, we disagree with our dissenting
colleague’s concern that it is fundamentally unfair to hold DuPont to the terms of the contract
that it negotiated and has received the benefit of, especially when DuPont has mounted multiple
challenges to the district court’s interpretation of the Agreement to no avail.                      See In re
Deepwater Horizon, 

 (5th Cir. 2014) (“There is nothing fundamentally
unreasonable about what BP accepted but now wishes it had not.”).

        Turning to the Parklane factors, we note as to the first factor that the MDL gave DuPont
a greater measure of power over case scheduling than in normal cases: few concerns about
Plaintiffs using a “wait-and-see” approach for another successful action are possible when
DuPont was able to select three of the six bellwether cases, including the first-tried case, Bartlett.
Second, the MDL structure presented DuPont with “every incentive,” Parklane Hosiery Co., 

, to defend itself vigorously in each of the early trials: the first two bellwether cases
tried were selected to inform the resolution of the 3,500 other pending cases, and DuPont knew
that the third trial could continue to influence the remaining litigation. Even after the global
settlement, DuPont was aware that cases could continue to be filed—cases that would necessarily
receive the same treatment as the MDL litigation. As to the third Parklane factor, there is no
concern about inconsistent verdicts with a previous judgment in favor of DuPont. 

 DuPont
was not successful at any trial.

        Importantly, the district court applied collateral estoppel only after three consistent jury
verdicts for the Plaintiffs in the only cases to proceed to trial—the first of which was a
bellwether selected by DuPont (Bartlett) and then another selected by the Plaintiff class
(Freeman). DuPont chose to settle the remaining bellwether cases with the Plaintiffs. As to the


        6
           Under Ohio law, “[t]he concept of foreseeability is an important part of all negligence claims, because
‘[t]he existence of a duty depends on the foreseeability of the injury.’” Cromer v. Children’s Hosp. Med. Ctr. of
Akron, 

 (Ohio 2015) (second alteration in Cromer) (quoting Menifee, 

).
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                Page 17


fourth Parklane factor, then, DuPont presented no evidence that it had any procedural
opportunities “that could readily cause a different result” in Abbott that were not available in the
earlier trials. Id. at 331. None of the Parklane Hosiery considerations weigh against application
of collateral estoppel in these circumstances.

        Thus, as to all the factors governing issue preclusion or collateral estoppel, DuPont has
received a full and fair opportunity for resolution of its issues—it had its day in court. DuPont’s
other objections—absence of advance notice of possible preclusive effect, the lack of
consideration of representativeness in bellwether selection, and alleged promises of no preclusive
effect—are not grounded in our collateral estoppel case law.7 At bottom, DuPont argues that we
should impose further rules constraining the use of nonmutual offensive collateral estoppel,
beyond the federal common law and the Supreme Court’s instructions in Parklane Hosiery.
DuPont does not offer any cases that create a notice requirement for collateral estoppel, nor does
it show that bellwether trials are prohibited from having such preclusive effect. See, e.g.,
Silvanch v. Celebrity Cruises, Inc., 

, 359–60 (2d Cir. 2003) (allowing an informal
bellwether case to have preclusive effect).

        In a similar vein, although both DuPont and our dissenting colleague emphasize the
applicability of In re Chevron U.S.A., Inc., 

 (5th Cir. 1997), that case involved a
proposed trial plan for a binding bellwether trial, which informed the Fifth Circuit’s stated
concerns about applying the trial’s outcomes to the full group of claimants. 

 at 1018–20; see
Zachary B. Savage, Scaling Up: Implementing Issue Preclusion in Mass Tort Litigation Through
Bellwether Trials, 

, 453–54, 456–57 (2013) (referencing Chevron and


        7
           We agree with our dissenting colleague, the Manual for Complex Litigation, and the Federal Judicial
Center that bellwether trials are most effective when “representative of the range of cases included in the MDL
proceeding.” Fed. Jud. Ctr., Bellwether Trials in MDL Proceedings 22 (2019); see Manual for Complex Litigation
§ 22.315 (4th ed. 2022). What makes a bellwether trial representative, however, is “litigation- and fact-specific.”
Bellwether Trials in MDL Proceedings at 22. Scholars have catalogued the many approaches that courts can take in
selecting bellwether plaintiffs: letting one party pick, requiring the parties to agree, allowing the parties to use
preemptory strikes against each other’s selections, leaving the decision entirely to the court, or some combination
thereof. See generally Fallon et al., supra; Loren H. Brown et al., Bellwether Trial Selection in Multi-District
Litigation, 

, 670–84 (2015). With the parties’ participation and cooperation, the district court
here engaged in a lengthy bellwether plaintiff selection process that used some of the same mechanisms that Judge
Fallon (who has overseen two MDLs involving over 30,000 claimants each) suggests are most effective. Fallon et
al., supra, at 2349–50, 2364–65.
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                  Page 18


explaining that binding bellwethers are “conceptually separate” from issue preclusion because
“the initial court running the bellwether determines its preclusive effect in advance of any
subsequent litigation”). Neither Parklane Hosiery—in which the Supreme Court offers the
clearest discussion on the limits and considerations for using offensive collateral estoppel—nor
the other case law DuPont cites suggests that DuPont’s asserted limitations on offensive
collateral estoppel exist. See Parklane Hosiery Co., 439 U.S. at 329–31.

         Even were we to imagine a fairness issue related to notice, the record does not support
DuPont’s arguments.           The district court did not promise that the general assumptions of
litigation—including that issue preclusion is possible—would not apply to the bellwether trials.
At most, the district court confirmed that the bellwether trials would not be “binding
bellwethers,” meaning that the results of those trials would not automatically be extrapolated to
non-bellwether plaintiffs.8 See Alexandra D. Lahav, Bellwether Trials, 

, 609–10 (2008). The Supreme Court has instructed the courts that the factors articulated in
Parklane offer the necessary constraints on the use of nonmutual offensive collateral estoppel.
We cannot and do not follow DuPont’s recommendation to create additional rules restricting the
use of the doctrine. We affirm the district court’s use of nonmutual offensive collateral estoppel
in this case.

                  2. Application of Collateral Estoppel to the Ohio Tort Reform Act

         DuPont also challenges the district court’s use of collateral estoppel to preclude the
application of the OTRA to Travis Abbott’s negligence claims, but the basis for that argument is
unclear. DuPont never asserted that the OTRA applied to Travis Abbott’s claims, and there
would be no grounds for such a contention. The OTRA cap on tort damages has a catastrophic
injury exception for those who lose “a bodily organ system,” Ohio Rev. Code
§ 2315.18(B)(3)(a), and no party has disputed that Travis Abbott’s loss of both his testicles


         8
          If a bellwether is “binding,” the parties designate a subset of overall cases, the results of which are to be
extrapolated to the broader whole. Generally, such a procedure requires that the parties “clearly memorialize” an
agreement to be bound in future trials, no matter the result, to avoid certain due process concerns. Dodge v. Cotter
Corp., 

 (10th Cir. 2000). That procedure was not employed in this MDL, where the parties
agreed that the bellwethers would be treated as ordinary trials whose results could be used to inform settlement or
the conduct of future trials.
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                       Page 19


qualifies his claims under this exception. The only OTRA challenge in the district court came
from the Abbotts, who argued that the law should not apply to Julie Abbott’s loss of consortium
claim. The district court disagreed and entered an amended judgment applying the OTRA to
reduce Julie Abbott’s $10 million jury award to the OTRA cap of $250,000. As the district court
did not apply issue preclusion on its reduction of Julie Abbott’s damages award and neither party
has objected to that reduction, the OTRA is not at issue on appeal.

       C. DuPont’s Evidentiary Challenges Related to Specific Causation

       In this appeal, DuPont frames its evidentiary challenges as three broad categories of
claims. First, DuPont argues that the district court erred in excluding expert testimony and
evidence on the dose-response relationship between C-8 blood levels and testicular cancer.
Second, it contends that the district court erred in allowing the Abbotts to offer expert testimony
on specific causation that relied—as authorized by the Leach Agreement—on the conclusion of
the Science Panel that the exposure threshold defining class membership was sufficient to cause
testicular cancer.   Finally, DuPont asserts that the district court erroneously excluded all
testimony on alternative causes of Travis Abbott’s cancer.

       These challenges are virtually identical to those DuPont raised in the Bartlett appeal that
was subsequently withdrawn due to the parties’ settlement. DuPont argued in Bartlett that the
district court erroneously interpreted general and specific causation pursuant to the Leach
Agreement. DuPont claimed it could not properly contest specific causation in Bartlett’s case
because it was prevented from offering evidence of Bartlett’s C-8 dose and the likelihood that
such a dose would cause kidney cancer. The district court’s order denying DuPont’s motion for
a new trial in Bartlett concluded that DuPont’s “position on causation conflate[d] the . . .
definitions . . . set forth in the Leach Agreement” and effectively sought to rewrite the provisions
about the Probable Link Findings in a way that would allow DuPont to challenge general
causation. The court explained that DuPont’s position would require plaintiffs not only to prove
their individual dose but also whether that particular dose was sufficient to cause the linked
disease. Allowing that standard would mean that “the Probable Link Findings may not apply to
a particular plaintiff, such as those plaintiffs who were in the lowest exposure groups.” In a
dispositive order covering all MDL cases, the court concluded that the parties’ bargain,
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                      Page 20


expressed in the unambiguous language of the Leach Agreement, is that Probable Link Findings
apply to any class member with a linked disease. Therefore, a plaintiff is “not required to come
forward with evidence proving that [her] individual dosage of C-8 [wa]s sufficient” to cause her
disease.

       Recognizing that DuPont’s evidentiary claims in the Abbotts’ case involved
interpretation of the Leach Agreement—an issue that was already decided in Bartlett, Freeman,
and Vigneron—the district court thoroughly explained that its decision on the proper
interpretation of the Agreement in those three previous cases was “final and binding.” That
interpretation foreclosed DuPont’s evidentiary arguments here and the district court therefore
rejected the claims.    That DuPont appealed Bartlett and its interpretation of the Leach
Agreement, but subsequently withdrew its appeal, had no effect on the finality of the prior three
decisions because “those previously appealable issues simply retained their finality for purposes
of collateral estoppel.” DuPont, the district court concluded, was precluded from raising these
same arguments yet again.

       In this appeal, DuPont did not challenge the aspect of the district court’s order applying
collateral estoppel to the interpretation of the Leach Agreement. It contested only the application
of collateral estoppel to elements of the negligence claims and the OTRA, and its fairness and
due process arguments were tailored to the tort claims, not the interpretation of the Leach
Agreement. Nor did DuPont contest the district court’s determinations that: the interpretation of
the Leach Agreement was necessary to the outcome of the proceedings in the three earlier cases;
the same relevant factual circumstances exist; the three cases reached a final judgment on the
merits that retained finality even after DuPont withdrew its appeal; or that DuPont had an
opportunity to litigate the proper interpretation of the Leach Agreement. Indeed, as the district
court explained, DuPont’s arguments about the Leach Agreement “have been made numerous
times to this Court, as well as before the Sixth Circuit.” By not challenging that aspect of the
district court’s collateral estoppel order in this appeal, the argument that the district court
improperly applied collateral estoppel to the contract interpretation issue is forfeited. See Guyan
Intern., Inc. v. Prof. Benefits Adm’rs, 

 (6th Cir. 2012). The district court’s
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                      Page 21


interpretation of the Agreement thus remains binding and, as explained above, is dispositive of
these evidentiary challenges.

       DuPont’s challenges, in any event, fail on their own merits. Evidentiary rulings are
reviewed for abuse of discretion, Hurt v. Com. Energy, Inc., 

 (6th Cir. 2020),
which “occurs when the district court relies on clearly erroneous findings of fact, improperly
applies the law, or uses an erroneous legal standard,” Innovation Ventures, LLC v. N2G Distrib.,
Inc., 

 (6th Cir. 2014) (quoting Mike’s Train House, Inc. v. Lionell, L.L.C., 

 (6th Cir. 2006)). We address each challenge in turn.

               1. Exclusion of DuPont’s Dose-Response Testimony and Evidence

       DuPont first challenges the district court’s evidentiary decisions related to the dose-
response relationship between testicular cancer and C-8 blood levels. It argues that the district
court excluded “expert opinions on Mr. Abbott’s dose and his specific resulting amount of
increased risk” based on an erroneous interpretation of the Leach Agreement’s general causation
provision. DuPont asserts that the Leach Agreement preserves its right to contest specific
causation, and the exclusion of expert testimony on dosage “gutted” that right. For the same
reason, DuPont argues that the court erred in allowing the Plaintiffs to tell the jury that 0.05 ppb
of C-8 is sufficient to cause Travis Abbott’s cancer.

       As discussed above, those arguments boil down to whether the district court properly
interpreted the definition of general causation in the Leach Agreement, which undisputedly
governs how the district court treated causation and dosage evidence. The Agreement gave the
Science Panel a clear charge: focus on an identified community and a particular chemical to
determine which diseases in the community are linked to C-8 exposure. Once the Science Panel
announced such a link, DuPont could not challenge general causation for that disease (“that it is
probable that exposure to C-8 is capable of causing” that particular disease) among class
members. (MDL R. 820-8, PageID 11804)             The Leach Agreement drastically limited the
persons authorized to bring suit against DuPont through two factors—the condition of class
membership (exposure to drinking water with 0.05 ppb of C-8 for at least a year) and satisfaction
of the Science Panel’s linked-disease finding (development of one of only six linked diseases).
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                        Page 22


The intersection of these two factors shows that the class bargained for and its members could
expect that satisfying the Science Panel’s linked-disease qualification would preclude the
introduction of evidence to suggest that the 0.05 ppb exposure level was insufficient to cause that
linked disease.

       The district court based its evidentiary decisions that DuPont now seeks to challenge on
the conclusion that DuPont’s proffered evidence would undermine the bargained-for exchange
memorialized in the Leach Agreement. Accepting DuPont’s position that it could introduce
evidence suggesting that exposure to more than 0.05 ppb of C-8 was necessary to cause testicular
cancer would have deprived Travis Abbott of DuPont’s agreement not to contest general
causation once the Science Panel found a probable link.             The tradeoff embodied in the
Agreement is that the No Probable Link Findings for 50 diseases applies to all class members
with any of those diseases, barring them from bringing suit against DuPont for non-linked
diseases regardless of how their individual dose and their related risk of disease were reported
and evaluated by the Science Panel. In other words, the vast majority of Leach class members
would not be allowed to challenge the Science Panel’s conclusions with dosage, individual
evidence, or scientific advances for any of the 50 non-linked diseases—the benefit that DuPont
now argues is its right to challenge for the six linked diseases. The district court did not abuse its
discretion in concluding that DuPont could not elicit or proffer evidence that undermined the
Leach Agreement’s general causation bargain, including evidence of specific dosage.

       As a factual matter, moreover, DuPont’s argument that the district court prohibited “all”
expert testimony and evidence on the dose-response relationship is incorrect. The court allowed
opinions on dose-response data when that evidence was consistent with the Leach Agreement
and the rules of evidence. The court’s limitations on expert testimony targeted testimony that
would have suggested Travis Abbott’s exposure was too low to cause his cancer, evidence that
violated the Leach Agreement. When such issues were not present, the district court allowed
DuPont to reference Abbott’s C-8 dose during the trial.

       DuPont makes the broader argument that the district court’s decision to allow testimony
and statements asserting that the class membership threshold of 0.05 ppb of C-8 was sufficient to
cause Abbott’s cancer was an abuse of discretion. This is, yet again, an attempt to challenge a
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                      Page 23


foreclosed issue—the district court’s interpretation of the Leach Agreement.         And even if
reconsideration of that interpretation were proper, the jury instructions were clear that specific
causation was an issue left to the jury. The instructions stated that the jury must decide whether
Abbott proved proximate cause (“an act or failure to act that was a substantial factor in bringing
about an injury and without which the injury would not have occurred”) to find for Abbott on his
negligence claim. The district court did not instruct the jury that exposure to C-8 at 0.05 ppb for
one year causes testicular cancer or that the 0.05 ppb represented a specific causation standard.
The jury instructions instead explained that the jury should “treat as proven in this case that C-8
is capable of causing kidney cancer and testicular cancer.” Notably, although the jury found for
the Abbotts, it did not reach the same verdict for the Swartzes.          This indicates the jury
understood that specific causation remained at issue in the Abbott/Swartz trial. The district court
did not abuse its discretion in denying DuPont’s challenges to dose-response evidence.

               2. DuPont’s Other Evidentiary Challenges

       DuPont next asserts that the district court should have excluded the Abbotts’ specific
causation expert because his testimony did not consider the Science Panel’s dose findings or
Travis Abbott’s specific dose. The expert in question, Dr. Pohar, used differential diagnosis
methodology to reach his conclusions. This methodology requires the physician to “consider[]
all relevant potential causes of the symptoms and then eliminate[] alternative causes based on a
physical examination, clinical tests, and a thorough case study.” Best v. Lowe’s Home Ctrs.,
Inc., 

 (6th Cir. 2009) (quoting Hardyman v. Norfolk & W. Ry. Co., 

 (6th Cir. 2001)). We have “recognize[d] differential diagnosis as ‘an appropriate
method for making a determination of causation for an individual instance of disease,” 

(quoting Hardyman, 

), and have held that “a medical opinion on causation based
upon a reliable differential diagnosis is sufficiently valid” under Rule of Evidence 702, 

(quoting Westberry v. Gislaved Gummi AB, 

 (4th Cir. 1999)). The record
supports that Dr. Pohar ruled out many other potential causes of testicular cancer to reach a
reasonable decision that C-8 exposure caused Abbott’s case. And Dr. Pohar did not err in ruling
in C-8 as a potential causal factor because he relied on the Science Panel’s determination that C-
8 exposure was a probable cause for a class member’s linked disease (here, testicular cancer)—a
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                      Page 24


determination to which both parties were bound under the plain terms of the Leach Agreement.
Allowing this testimony was not an abuse of discretion.

       DuPont also argues that the district court improperly excluded “all opinions that Mr.
Abbott’s cancer was more likely caused by his pre-exiting germ cell neoplasia in situ (GCNIS)
or idiopathic.” But the record shows that the district court allowed DuPont to present testimony
on alternative causes of Abbott’s cancer. A DuPont expert testified as to evidence that GCNIS
“nearly always” leads to testicular cancer. The district court did not allow that expert to testify
that GCNIS was the more likely cause because the expert was qualified only as a general
causation expert, not a specific causation expert. In fact, the court excluded him as a specific
causation expert because he did not rule in C-8 exposure as a possible cause, which the Leach
Agreement required. DuPont, therefore, did not put on a specific causation expert of its own.
Nevertheless, the district court allowed DuPont to offer the testimony from multiple experts that
most testicular cancer is idiopathic. DuPont was able to and did present evidence of alternate
causes for Abbott’s cancer but failed to present its own specific causation expert. The district
court did not improperly prohibit DuPont from arguing specific causation at trial.

       C. The Directed Verdict on DuPont’s Statute of Limitations Defense

       DuPont challenges the district court’s decision to reject its statute of limitations defense
as a matter of law. Before the trial, the district court denied DuPont’s motion for summary
judgment and strongly suggested that Abbott had filed his tort claim for his 2015 testicular
cancer within the two-year statute of limitations period. This finding arose from the court’s
conclusion that the earliest possible “triggering” date was when Abbott received a definitive
testicular cancer diagnosis less than two years before he filed his lawsuit. After the parties
presented their evidence to the jury, the district court rejected DuPont’s arguments that claims
related to the 1994 cancer were time-barred. Based on the overwhelming evidence that Travis
Abbott did not know about the connection between testicular cancer and C-8 pollution from
DuPont’s Washington Works plant and the significant inferences necessary for a jury to
conclude otherwise, the district court found that the Abbotts were entitled to judgment as a
matter of law on the statute of limitations issue. DuPont challenges two aspects of this decision.
First, Dupont argues that the statute of limitations for Abbott’s 2015 cancer ran before he
 No. 21-3418                       In re E. I. du Pont de Nemours & Co.                    Page 25


received a definitive diagnosis on November 16, 2015. Second, whether Abbott had notice that
DuPont was responsible for his bouts of testicular cancers was, according to the company, an
issue properly left to the jury.

        We review a district court’s decision on a motion for judgment as a matter of law de
novo, Hurt v. Commerce Energy, Inc., 

 (2020), and apply the applicable state-
law standards for evaluating such a motion in diversity cases, Morrison v. B. Braun Med. Inc.,

 (6th Cir. 2011). A directed verdict is proper under Ohio law when, “after
construing the evidence most strongly in favor of the party against whom the motion is directed,
reasonable minds could come to but one conclusion upon the evidence submitted.” Groob v.
KeyBank, 

 (Ohio 2006).

        Our de novo review focuses on when Travis Abbott knew that his cancer diagnoses could
be tied to DuPont’s C-8 pollution. Under Ohio law, a cause of action for bodily injury from
“exposure to hazardous or toxic chemicals” accrues when:

        the plaintiff is informed by competent medical authority that the plaintiff has an
        injury that is related to the exposure, or upon the date on which by the exercise of
        reasonable diligence the plaintiff should have known that the plaintiff has an
        injury that is related to the exposure, whichever date occurs first.

Ohio Rev. Code § 2305.10(B)(1). This law makes two pieces of knowledge critical to pinning
down the accrual date: (1) knowledge of the injury; (2) knowledge that the injury is tied to a
specific exposure. Norgard v. Brush Wellman, Inc., 

, 979–81 (Ohio 2002);
O’Stricker v. Jim Walter Corp., 

 (Ohio 1983). The Ohio Supreme Court has
cautioned that this discovery rule “must be specially tailored to the particular context to which it
is to be applied.” Norgard, 

. The Ohio Supreme Court has also emphasized
that “the underlying rationale for the statute of limitations” and public policy considerations
require “a liberal interpretation of the time of accrual” for claims alleging latent bodily injuries.
Liddell v. SCA Serv. of Ohio, Inc., 

 (Ohio 1994).

        The statute of limitations challenge to the claims for the 2015 cancer is straightforward to
resolve. DuPont asserts that the district court should have left to the jury to decide whether
Abbott knew of his 2015 testicular cancer at least by October 2015 when he received an
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                       Page 26


ultrasound and CT scan showing probable testicular cancer. DuPont assumes that Abbott was
aware of the link between testicular cancer and DuPont’s C-8 pollution, arguing that his suit filed
on November 14, 2017, therefore misses the statute of limitations by mere weeks. The district
court found as a matter of law that the earliest possible triggering date for Abbott’s 2015 cancer
was November 16, 2015, when Abbott received a definitive diagnosis of testicular cancer after
an orchiectomy.

        At trial, the Abbotts offered uncontroverted testimony from Travis Abbotts’ treating
physicians that the testicular cancer diagnosis was not finalized until November 16, 2015.
Although his doctors informed him earlier that the mass in his testicle was likely cancerous, the
diagnosis was not official until his providers had reviewed a pathology report on the removed
testicle.   Indeed, an earlier diagnosis without the pathology results, according to the
uncontroverted testimony, would have contravened the standard of care.

        Interpreting the statute of limitations as requiring Abbott to have sued DuPont before an
official diagnosis, moreover, would raise significant fairness issues. First, as the Abbotts argue,
it would implicitly require Abbott to have had earlier and greater certainty about his medical
diagnosis than his treating physicians had prior to November 16, 2015. Second, such a reading
would leave Leach class members with a difficult choice.          The Agreement prohibits class
members without a linked disease from suing DuPont. Had Travis Abbott sued without a
definitive diagnosis, DuPont would have had every incentive to argue that he was not a
qualifying class member under the Leach Agreement. DuPont’s position would leave Leach
class members with a choice of suing before a definitive diagnosis with the risk of dismissal for
lack of qualifying class membership or suing after with the risk of dismissal under the statute of
limitations. Neither the Leach Agreement nor the fundamental fairness concerns underlying
Ohio’s statute of limitations support such a result. Cf., e.g., Schmitz v. Nat’l Collegiate Athletic
Ass’n, 

 (Ohio 2018).

        The analysis is more fact-intensive for the 1994 cancer than that for the 2015 cancer.
There is no debate that Travis Abbott knew of his 1994 cancer well over two years before suing
DuPont. The issue is instead when, pursuant to Ohio’s discovery rule, Abbott became or should
have become aware of the link between that cancer and DuPont’s C-8 discharges.
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                      Page 27


       Abbott argues that the statute of limitations began to run as to his 1994 cancer when he
actually encountered information that did or should have made him aware of the link between C-
8 and the cancer, which he says occurred in October or November of 2015. DuPont suggests that
Abbott’s actual knowledge of a potential link between C-8 and his cancer is not what triggered
his claim’s accrual. Rather, there was media coverage of the link between C-8 and testicular
cancer and other notice sufficient to make a reasonable person in Abbott’s position aware that his
cancer was related to C-8 before the fall of 2015.

       Abbott’s witnesses consistently presented facts to the jury showing that despite media
coverage and some relatives’ independent lawsuits against DuPont, Abbott was not aware of the
connection between testicular cancer and DuPont’s C-8 pollution.         Abbott testified that he
learned about the connection between his cancer and C-8 only a few weeks before filing his
lawsuit, when his father told him about a TV ad discussing C-8’s tie to testicular cancer and his
administrative assistant suggested that he consult a lawyer on the issue. He further testified that
he had not heard about the Science Panel’s findings, did not receive notices about a link
between C-8 and testicular cancer, and did not subscribe to or read any newspapers discussing
the link to C-8.

       DuPont did not offer evidence at trial directly refuting the consistent evidence regarding
Abbott’s lack of knowledge about the link between his cancers and C-8. DuPont presented some
circumstantial evidence to support its assertion that Abbott did or should have known about the
connection to C-8.     For instance, when cross-examining Travis Abbott, DuPont elicited
testimony about his 2006 C-8 Health Project paperwork, which included questions about
testicular cancer and indicated that the Science Panel was looking into the disease’s relationship
to C-8 pollution (though, of course, in 2006, there was no established probable link between
testicular cancer and C-8). DuPont also provided evidence of local newspaper coverage of both
the Leach Settlement Agreement and the Science Panel’s decisions. At several points, DuPont
questioned Travis Abbott and other witnesses about the fact that some members of Abbott’s
extended family had already sued DuPont on claims related to other linked diseases.

       But Ohio’s discovery rule does not require that plaintiffs read the news or assume that
they have knowledge of their family’s legal affairs. Instead, “the statute of limitations begins to
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                  Page 28


run once the plaintiff acquires additional information of the defendant’s wrongful conduct” that
does or should put that plaintiff on notice that his injury is related to the conduct.9 Norgard,

; see Browning v. Burt, 

, 1006 (Ohio 1993) (statute of
limitations for negligent credentialing claim against hospital began to run when plaintiffs viewed
a television program making them aware that other ex-patients suffered from abnormalities
similar to theirs); Vaccariello v. Smith & Nephew Richards, Inc., No. 76594, 

,
at *5 (Ohio Ct. App. Aug. 3, 2000), aff’d, 

 (Ohio 2002) (statute of limitations for
bodily injury claim began to run when plaintiff viewed television program in which she learned
that a device implanted in her back could be the source of her injury). The question, then, is not
whether there was media coverage of C-8’s link to testicular cancer, or whether other members
of Abbott’s family brought claims against DuPont. The question is whether Abbott encountered
information that did or should have put him on notice that his cancer was connected to C-8.
DuPont did not present evidence that Abbott ever received such information. The district court
drew a helpful analogy to the McDonnell Douglas burden shifting framework common in
employment discrimination cases when denying DuPont’s motion for summary judgment.
DuPont first made a prima facie case that the statute of limitations barred Abbott’s claims
because the 1994 cancer and news about the link between C-8 and testicular cancer occurred
well over 2 years before Abbott filed suit. Abbott then rebutted that assumption by offering
proof that he did not know about the link between C-8 and testicular cancer. DuPont did not
meet its burden of challenging that rebuttal.

         The district court’s reasoning is sound. While the jury normally can make credibility
judgments, submitting the statute-of-limitations issue to them would have required them to
“draw inference upon inference upon inference” to find for DuPont. The record evidence
pointed in one direction: Abbott filed his claim less than two years after he became aware of the
         9
          Our dissenting colleague cites Flowers v. Walker, 

 (Ohio 1992), which found that
constructive knowledge of facts was sufficient to start the statute of limitations running for a medical malpractice
claim (not a bodily injury claim). Even if applicable, this standard is not inconsistent with the Ohio discovery rule
as the district court applied it: like a plaintiff bringing a bodily injury claim, a plaintiff bringing a medical
malpractice claim must encounter some information that does or should put them on notice. See Allenius v. Thomas,

 (Ohio 1989); see also Hambleton v. R.G. Barry Corp., 

 (Ohio 1984)
(finding that a party has constructive notice to trigger a statute of limitations if he “has knowledge of such facts as
would lead a fair and prudent man, using ordinary care and thoughtfulness, to make further inquiry” and “fails to do
so” (emphasis added)).
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                   Page 29


connection between C-8 and his testicular cancers. We therefore affirm the district court’s
judgment on the statute of limitations.

                                      III. CONCLUSION

       For the foregoing reasons, we AFFIRM the judgment of the district court in full.
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                  Page 30


                   _____________________________________________________

                      CONCURRING IN PART AND DISSENTING IN PART
                   _____________________________________________________

         ALICE M. BATCHELDER, Circuit Judge, concurring in part and dissenting in part.

         Throughout the last decade or so, this multidistrict litigation has generated more than ten
thousand record entries, two appeals, and five month-long jury trials. The district court has done
a commendable job, and the majority affirms the court’s relevant decisions in full, as they relate
to Travis and Julie Abbott.

         Respectfully, I must dissent. I would hold that, in mass-tort multidistrict litigation,
fundamental notions of due process require an additional safeguard before a court can issue a
collateral estoppel order against a defendant based upon a small number of potentially
unrepresentative bellwether trials. I would also hold that the general verdicts in the three early
trials lacked the specificity to bind the thousands of remaining cases. Finally, I would hold that
the district court erred, in part, by taking away from the jury DuPont’s statute-of-limitations
defense.

         For the reasons expressed below, I concur in Part II.B of the majority opinion but must
respectfully dissent from Parts II.A and II.C.

                                                          I.

         I’ll begin with Part II.A. I agree that nonmutual offensive collateral estoppel does not
necessarily violate due process in this context.1 Nowhere in Parklane Hosiery Co. v. Shore,

 (1979), did the Supreme Court create a categorical ban on that doctrine in mass-tort
litigation. The Court, instead, used “fairness” as its guide to determine when the doctrine is
appropriate. I also agree with the majority that the district court was not required to give DuPont
advance notice that the bellwether trials could later have preclusive effect.

         1
          The district court and the majority use the term “collateral estoppel,” also known as “issue preclusion.”
Brownback v. King, 

, 747 n.3 (2021). Although issue preclusion is the “more descriptive term,”
Yeager v. United States, 

, 119 n.4 (2009), I will refer to the doctrine as collateral estoppel for the sake
of consistency.
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                         Page 31


       That said, however, collateral estoppel was not appropriate in this case. The district court
used plaintiff-specific verdicts, based on general verdict forms, from three early trials—as to
which the court had told the parties from the outset that they would be informational and non-
binding— to preclude DuPont from contesting certain liability issues in thousands of potentially
different cases. For a court to apply offensive collateral estoppel against a defendant in a mass-
tort multidistrict litigation such as this, due process requires an inquiry into the
representativeness of the plaintiffs, as well as a faithful adherence to the collateral estoppel rules.
Because neither happened in this case, the district court’s sweeping estoppel order subverts
DuPont’s constitutional rights. I would reverse and remand.

                                                  A.

       It is foundational that all defendants, no matter how unsympathetic or heinous their
conduct, retain the full force of constitutional due-process protections. In my view, in the mass-
tort bellwether context, the Constitution requires that before a court issues a collateral estoppel
order it must assure that the cases estopped are reasonably representative of the first cases tried.
The district court here failed to do that—despite there being thousands of cases at stake—making
its estoppel order fundamentally unfair to DuPont in violation of due process.

       First, some background on the legal landscape. As the majority describes, in federal
diversity actions, state law determines whether collateral estoppel may render a prior decision
preclusive on an issue raised in a later case. In Ohio, issue preclusion applies when (1) the issue
“was actually and directly litigated in the prior action;” (2) a “court of competent jurisdiction”
decided the issue; (3) the “fact or . . . point” in question was “actually and necessarily litigated
and determined” as part of a final judgment; (4) “the party against whom collateral estoppel is
asserted was a party in privity with a party to the prior action;” and (5) the party against whom
estoppel is sought had a “full and fair opportunity” to litigate the issue in the prior action. State
ex rel. Jefferson v. Russo, 

 (Ohio 2020); Fort Frye Tchrs. Ass’n OEA/NEA v.
State Emp. Rels. Bd., 

 (Ohio 1998); State v. Williams, 

(Ohio 1996); Walden v. State, 

 (Ohio 1989). This doctrine conserves
judicial resources and protects against inconsistent decisions by recognizing that parties should
not be able to relitigate the same disagreement in perpetuity. See Taylor v. Sturgell, 553 U.S.
 No. 21-3418                     In re E. I. du Pont de Nemours & Co.                        Page 32


880, 892 (2008); San Remo Hotel, L.P. v. City & County of San Francisco, Cal., 

,
336–37 (2005).

       But like all doctrines, it has its limits. In Parklane, 

, the Supreme Court
considered whether a party could use what is called “nonmutual offensive collateral estoppel”—
that is, whether a plaintiff can seek to estop a defendant from relitigating an issue that the
defendant previously litigated and lost against a different plaintiff. The Court weighed the
downsides of allowing the doctrine but ultimately concluded that “the preferable approach . . . is
not to preclude the use of offensive collateral estoppel, but to grant trial courts broad discretion
when it should be applied.” 

. The Court held that, as a guiding principle, “offensive
estoppel” should not be applied where it “would be unfair to the defendant.” 

       As an example of when it might be unfair, the Court instructed district courts to “avoid
reward[ing]” a plaintiff “who could easily have joined in the earlier action” but chose not to “in
the hope that the first action by another plaintiff” resolved favorably. 

 at 330–31. The Court
noted that offensive estoppel may also be unfair where a defendant had “little incentive to defend
[an initial case] vigorously,” where there are inconsistent prior judgments, or where “the second
action affords the defendant procedural opportunities [that were] unavailable in the first action.”

 at 330–31.     In a footnote, the Court provided an example of when inconsistent prior
judgments would render estoppel unfair:

       In Professor Currie’s familiar example, a railroad collision injures 50 passengers
       all of whom bring separate actions against the railroad. After the railroad wins
       the first 25 suits, a plaintiff wins in suit 26. Professor Currie argues that offensive
       use of collateral estoppel should not be applied so as to allow plaintiffs 27 through
       50 automatically to recover.

 at 330 n.14 (citing Currie, Mutuality of Estoppel: Limits of the Bernhard Doctrine, 

, 304 (1957)). Finally, the Court noted that it did not exhaustively catalogue the factors
a district court should consider when reviewing for fairness. Id. at 331 (“The general rule should
be that . . . where, either for the reasons discussed above or for other reasons, the application of
offensive estoppel would be unfair to a defendant, a trial judge should not allow [its] use . . . .”).

       In other cases, the Court has limited collateral estoppel when the doctrine is sought
against the government, United States v. Mendoza, 

 (1984), or when there is an
 No. 21-3418                           In re E. I. du Pont de Nemours & Co.                                    Page 33


intervening change in the controlling facts or legal principles in a case, Herrera v. Wyoming, 

 (2019) (gathering examples). And “of course,” just as any other common law
doctrine, collateral estoppel is “subject to due process limitations.” Taylor, 553 U.S. at 891
(quoting Richards v. Jefferson County, 

 (1996)).

         Complicating matters here, though, is that multidistrict litigation (“MDL”) courts often
use a procedure called “bellwether trials” to help resolve mass-tort litigation, and, at least in
theory, the results of those trials can bind future cases under ordinary principles of collateral
estoppel. Bellwethers are preliminary trials meant to help the parties gather information, value
the cases, test legal theories, and, ultimately, reach a global settlement with minimal costs. See
Alexandra D. Lahav, Bellwether Trials, 

, 577–78 (2008). In practice,
their results are generally non-binding absent an agreement to the contrary between the parties.
See Eldon E. Fallon et al., Bellwether Trials in Multidistrict Litigation, 

,
2331 n.27, 2337 (2008). And for understandable reasons, there is usually a concerted effort to
ensure that bellwethers are representative of the larger group of MDL plaintiffs. That way, the
parties and the court can confidently and accurately draw inferences from them.2 See In re
Depuy Orthopaedics, Inc., 

, 348–49 (5th Cir. 2017) (“Bellwether trials are meant to
produce a sufficient number of representative verdicts and settlements to enable the parties and
the court to determine the nature and strength of the claims, whether they can be fairly developed
and litigated on a group basis, and what range of values the cases may have if resolution is
attempted on a group basis.” ); Zachary B. Savage, Scaling Up: Implementing Issue Preclusion
in Mass Tort Litigation Through Bellwether Trials, 

, 453 (2013)
(“[B]ellwether trials are distinct from ordinary trials because the transferee court selects cases
that are similar to the wider group of claims arising from the mass tort. These trials involve
similar facts, claims, or defenses as the wider group of cases, and are meant to help achieve


         2
           In the district court’s words, the bellwether trials here were meant to “produce a sufficient number of
representative verdicts and settlements to enable the parties and the court to determine the nature and strengths of
the claims . . . and what range of values the cases may have.” [MDL 5285, PageID# 128541 (quoting The Manual
for Complex Litigation, § 22.315)]. At another point, when the parties informed the court that they had settled
several of the bellwethers, the court stated that it “was, to put it mildly, surprised” because “[f]or over three years the
parties had taken the position that the purpose of the bellwether trials was to gather information regarding the
valuation of cases.” [MDL 4624, PageID# 100953].
 No. 21-3418                        In re E. I. du Pont de Nemours & Co.                               Page 34


global resolution of the litigation.”); see also Grundy v. FCA US LLC, No. 2:20-CV-11231, 

, at *1 (E.D. Mich. Nov. 22, 2021).

        Now to the merits. In my view, due process requires an additional safeguard before a
court can declare mass-tort preclusion on an issue of liability against a defendant: the court must
ensure that the sample of bellwether plaintiffs is reasonably representative of the rest.3 The Fifth
Circuit’s decision in In re Chevron U.S.A., Inc., 

 (5th Cir. 1997), is
instructive. In Chevron, a mass-tort case involving an oil spill, the district court planned to
conduct a “unitary trial” of 30 bellwether cases, with each side selecting 15 cases out of the
3,000 cases pending. 

 at 1019–21. The trial would be preclusive on both “general liability”
and “general causation” for the remaining 2,970 plaintiffs. 

 at 1019–20. The Fifth Circuit
rejected this plan on due-process grounds because it contravened “fundamental fairness” to
impose widespread liability against Chevron based on the results of a non-representative sample
of plaintiffs. 

 at 1019–21. The Fifth Circuit explained that, without a sufficient number of
representative trials, the district court’s trial plan lacked “the minimum level of reliability.” 

at 1020–21; see 

 (noting that the “core element” of bellwether trials is
“representativeness”).       The court held: “[B]efore a trial court may utilize results from a
bellwether trial for a purpose that extends beyond the individual cases tried, it must, prior to any
extrapolation, find that the cases tried are representative of the larger group of cases or claims
from which they are selected.” 

.

        The Sixth Circuit has not had the occasion to address the due-process restrictions on
bellwethers. But in the class-action context, we have described bellwethers as “a small number
of . . . plaintiffs, who can adequately represent the class, test their claims and legal theories first,
before proceeding with the rest of the class.” Abrams v. Nucor Steel Marion, Inc., 

, 977 n.2 (6th Cir. 2017). It appears that scholars agree with Chevron that bellwether trials
should be representative, and that a small sample size of bellwether trials has the potential to
prematurely “lock in” outlier jury findings. See, e.g., Fallon, supra, at 2344 (noting that a
“bellwether trial is most effective when it can accurately inform future trends and effectuate an

        3
          Only two of the three cases that were tried before Abbott-Swartz were bellwether trials. But for ease of
reference, I will refer to all three as “bellwether trials” and the plaintiffs in them as “bellwether plaintiffs.”
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                        Page 35


ultimate culmination to the litigation,” and that parties should catalogue the entire universe of
MDL cases to minimize the risk of trying an anomalous bellwether case); Savage, supra, at 463–
64 (arguing that collateral estoppel should apply in mass tort litigation but only after “defendants
[have] lost a substantial number of bellwether trials,” and that the court must “ensure[] that the
bellwether trials involved a wide range of plaintiffs—not just the most sympathetic ones”);
Byron G. Stier, Another Jackpot (in)justice: Verdict Variability and Issue Preclusion in Mass
Torts, 

, 739–43 (2009) (exploring “the possibility that the first verdict [in a
multidistrict litigation] would be inconsistent with subsequent verdicts”); see also Meiring de
Villiers, Technology Risk and Issue Preclusion: A Legal and Policy Critique, 9 Cornell J.L. &
Pub. Pol’y 523, 524 (2000) (“Liberal application of collateral estoppel in product liability . . . has
been criticized for putting the survival of entire industries at risk based on a single, possibly
erroneous, judgment.”); Ann. Manual Complex Lit. § 22.315 (4th ed. 2022) (bellwethers are
meant “to produce reliable information about other mass tort cases, [so] the specific plaintiffs
and their claims should be representative of the range of cases”).

       I would adopt Chevron’s approach and find that it is fundamentally unfair for a small,
non-representative sample of bellwether plaintiffs to bind a defendant in thousands of future
cases. Parklane makes it clear that even when the collateral estoppel requirements are met
(which I question here), the invocation of the doctrine should not be allowed if it would be
unfair. Parklane, 

; see Homaidan v. Sallie Mae, Inc., 

, 600 (2d Cir.
2021) (“Nonmutual offensive collateral estoppel . . . cannot be applied if it would be unfair to the
defendant.”); Merial, Inc. v. Sergeant’s Pet Care Prod., Inc., 

, 406 (6th Cir.
2020) (“Courts also must ask, whether it would be otherwise unfair under the circumstances to
permit the use of collateral estoppel.” (quotation marks omitted)); Marlene Indus. Corp. v.
N.L.R.B., 

 (6th Cir. 1983) (collateral estoppel must be “qualified or rejected
when [its] application would contravene an overriding public policy or result in manifest
injustice” (citation omitted)); Jack Faucett Assocs., Inc. v. Am. Tel. & Tel. Co., 

 (D.C. Cir. 1984) (where nonmutual offensive collateral estoppel is requested, “‘fairness’
gains special importance”).
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                 Page 36


         Here, the district court made no effort to ensure representativeness in its estoppel order.
Neither of the two bellwether trials (Bartlett or Freeman) appears to be representative of the
thousands of then-remaining cases. Nor do they appear to be representative of Abbott’s case.
Although there was some surface-level discussion of “representativeness” very early on in the
MDL proceedings, [MDL 30, PageID# 172; MDL 34, PageID# 218–19; MDL 194, PageID#
3694], the district court ultimately allowed the parties to each select three of their strongest cases
for bellwether trials. And the third trial, Vigneron, was chosen by plaintiffs after the court
clearly rejected any requirement that it be representative. [MDL 4461, PageID# 96026–27].
Indeed, the court instructed plaintiffs to choose one of the “most severely impacted plaintiffs” to
go first. [MDL. 4624, PageID# 100962; see also MDL 4461, PageID# 96026–27; MDL 4535-2,
PageID# 98584 (plaintiffs’ brief arguing that “the representativeness of the trial selections should
be of no moment”)].

         The parties argue about whether there are outcome-determinative differences between the
bellwether plaintiffs and Abbott. But as Judge Jones stated in her concurrence in Chevron, “the
determination of reliable representative plaintiffs is difficult in a toxic exposure case”––it
“involves such questions as quantity, geographic proximity, and temporal exposure to the toxic
substance, comparative lifestyles, and physical manifestations of exposure.” Chevron, 

 (Jones, J., concurring). The fact remains that the district court here explored none of
these questions in its estoppel order, despite having allowed the parties to cherry-pick “faces
from the crowd of plaintiffs.” 

 Moreover, the estoppel order impacts more than the Abbotts’
case—it binds DuPont in countless other cases, too. And it continues to do so as new cases are
filed.4 [6th Cir. R. 69].

         I am mindful that the fairness inquiry could be “potentially disruptive” if liberally applied
and that collateral estoppel remains a useful trial-management device when used in appropriate
cases. Merial, 806 F. App’x at 414. But in tension with those concerns is the fundamental and


         4
          I am not confident that, in a toxic-tort MDL case involving thousands of plaintiffs, a small group of
bellwether trials can ever be reasonably representative of the larger group. But the difficulty of ensuring
representativeness is no reason to do away with the doctrine of collateral estoppel. It merely underscores that, as a
practical matter, in rare cases such as this, collateral estoppel will usually be unfair because a court cannot
confidently extrapolate findings relevant to, and preclusive upon, the remaining group of cases.
 No. 21-3418                         In re E. I. du Pont de Nemours & Co.                                  Page 37


“essential prerequisite of due process” that a party have a full and fair “opportunity to be heard.”
Richards, 

 n.4. And it is Statistics 101 that a small, unrepresentative sample
cannot yield reliable inferences as to a larger group. Because the district court here failed to
assess the representativeness of the bellwether plaintiffs, the court’s far-reaching estoppel order
deprived DuPont of its constitutional right to have an “individual assessment of liability and
damages in each case.” Chevron, 

 (Jones, J., concurring).

         This is not to say that DuPont’s three losses were outliers. It may very well be that, if
given the chance to contest duty, breach, and foreseeability in each successive case, DuPont
would still lose. But maybe not. Out of the 3,500 pending MDL cases, only three were tried.
And about 75,000 potential lawsuits remained at the time of the estoppel order. Thus, it was too
early, and the cases are perhaps too disparate, to tell.

         The Abbotts claim that only binding bellwethers (where the parties agree in advance that
the trials will be preclusive) must be representative. [Appellee’s Br. 35]. But the Abbotts do not
provide any case support or justification for the claim that binding bellwethers require due-
process protections, but potentially binding informational bellwethers do not. The distinction
makes no difference. Before a district court allows a bellwether trial to be preclusive on
thousands of other MDL cases—whether by binding bellwether (before trial) or by informational
bellwether (after trial)—due process requires an inquiry into representativeness.5 The district
court’s concern for efficiency, while understandable, does not outweigh these overarching due-
process concerns. See In re Nat’l Prescription Opiate Litig., 

, 844–45 (6th Cir.
2020) (emphasizing that “enhancing the efficiency of the MDL as a whole” is not reason to
disregard “the same legal rules that apply in other cases,” and that “a party’s rights in one case
[cannot] be impinged to create efficiencies in the MDL generally”); Chevron, 

(Jones, J., concurring) (“Judges must be sensitive to stay within our proper bounds of



         5
          The only case cited by the majority to support giving bellwethers preclusive effect is Silivanch v. Celebrity
Cruises, Inc., 

 (2d Cir. 2003). But that case is a classic example of a binding bellwether, where the
parties agreed in advance that the bellwether would decide certain issues in the remaining cases. 

.
Moreover, Silivanch involved fewer than twenty-two plaintiffs, and the court’s discussion of bellwethers was by
way of background. The court did not in any way speak to the propriety of giving bellwethers (much less informal
bellwethers) preclusive effect.
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                        Page 38


adjudicating individual disputes.     We are not authorized by the Constitution or statutes to
legislate solutions to cases in pursuit of efficiency and expeditiousness.”).

       In the end, the district court has done something that no other circuit court has, to my
knowledge, allowed. It is one thing for a district court to bind a defendant in a single case after a
handful of informational bellwether trials involving similarly-situated plaintiffs. It is quite
another for a court to do so in thousands of future cases and without considering whether those
cases involve legally divergent facts. And for a court to change course after it told the parties
from the outset that the bellwethers would be informational and non-binding.              [MDL 34,
PageID# 218–19; see MDL 3973, PageID# 68182; MDL 4184, PageID# 80083; MDL 4382,
PageID# 93365–66; MDL 4624, PageID# 100947].

       In light of the “unique potential for unfairness” at play here, Jean Alexander Cosms., Inc.
v. L’Oreal USA, Inc., 

 (3d Cir. 2006), I would remand this case so that the
district court can assess in the first instance the representativeness of the bellwether plaintiffs
before applying collateral estoppel. See Bifolck v. Philip Morris USA Inc., 

, 84–
85 (2d Cir. 2019) (remanding case for determination of whether application of nonmutual
offensive collateral estoppel would be unfair).

                                                  B.

       I also cannot conclude that the black-letter requirements of collateral estoppel were
satisfied in this case. Recall that collateral estoppel requires, among other things, that the precise
issue raised in the later case was “actually and directly litigated in the prior action.” Russo, 

. . Because the three bellwether trials here used general verdict forms and resulted
in plaintiff-specific verdicts, the precise issues of duty, breach, and foreseeability raised in
Abbott have not been actually litigated and forever decided.

       The bellwether trials’ general verdict forms are insufficient. “[A] jury speaks only
though its verdict,” and therefore general verdicts often lack the specificity required to create
widespread issue preclusion.        Yeager v. United States, 

, 121–23 (2009)
(emphasizing that “speculation” and “conjecture” have no place in the issue-preclusion analysis);
United Access Techs. v. CenturyTel Broadband Servs., 

 (Fed. Cir. 2015)
 No. 21-3418                    In re E. I. du Pont de Nemours & Co.                       Page 39


(“When there are several possible grounds on which a jury could have based its general verdict
and the record does not make clear which ground the jury relied on, collateral estoppel does not
attach to any of the possible theories.”); S.E.L. Maduro (Fla.), Inc. v. M/V Antonio de Gastaneta,

 (11th Cir. 1987) (“If the jury could have premised its verdict on one or
more of several issues, then collateral estoppel does not act as a bar to future litigation of the
issues.”); Steen v. John Hancock Mut. Life Ins. Co., 

 (9th Cir. 1997)
(“Collateral estoppel is inappropriate if there is any doubt as to whether an issue was actually
litigated in a prior proceeding.”); see also In re Piercy, 

, 924–25 (6th Cir. 2021)
(noting that collateral estoppel determinations cannot be based on gaps in verdict forms); Black
v. Ryder/P.I.E. Nationwide, 

, 581–82 (6th Cir. 1994) (clear error where court
“engage[d] in pure speculation regarding the basis for the general verdict in the earlier case”).

       The verdict forms here asked the jury: “Do you find in favor of [the plaintiff] on his
negligence claim?” [See, e.g., Freeman, No. 2:13-cv-1103, R. 97, PageID# 1011]. Nothing
more. Unlike a detailed special verdict, this type of general verdict does not provide insight into
what the jury did, and did not, decide. It leaves the court with questions about what theories of
negligence formed the basis for the jury’s verdict, and what acts or omissions the jury believed
were foreseeable by DuPont. And when we have reasonable doubt as to what the first cases
found, we “err on the side of construing [those] prior ambiguous findings or holdings narrowly”
for purposes of collateral estoppel. United States v. United Techs. Corp., 

 (6th
Cir. 2015); see Merial, 806 F. App’x at 413 (denying collateral estoppel, in the alternative, “on
the basis of lack of clarity”); In re Braniff Airways, Inc., 

 (5th Cir. 1986)
(“[I]f reasonable doubt exists as to what was decided in the first action, the doctrine of res
judicata should not be applied.”); Harris v. Jacobs, 

 (9th Cir.1980) (“If there is
doubt on this score, collateral estoppel will not be applied.”).

       In Dodge v. Cotter Corp., 

 (10th Cir. 2000), the Tenth Circuit reversed the
district court’s application of collateral estoppel in part because the verdict form did not specify
which theories the jury relied upon in finding negligence. 

 at 1197–98. Similarly, in Hardy v.
Johns-Manville Sales Corp., 

 (5th Cir. 1982), the Fifth Circuit rejected the
application of collateral estoppel where a general verdict form was “ambiguous as to certain key
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                        Page 40


issues,” including “what the . . . jury decided about when a duty to warn attached.” 

 at 343–
44. The verdict forms here suffer from the same flaw.

       The Abbotts do not respond to this argument. And the district court’s cited authority is
distinguishable. See Adams v. United States, 

 (D. Idaho Oct. 29, 2010). In
Adams, the court informed the parties that the bellwether trial would have “preclusive effect,”
selected a “representative sample” of plaintiffs, and used a 47-question special verdict form to
avoid ambiguity on the specific issues being decided. 

 at *1–3.

       Importantly, the three bellwether trials here also involved distinct, plaintiff-specific facts
that bear heavily on negligence. These include each plaintiff’s susceptibility and location and
the length and timing of his or her exposure to C-8, as well as DuPont’s response and its
knowledge about which locations were exposed to C-8 (and at what levels) and about the
scientific developments regarding C-8 over the last fifty years. Each of these factual variations
can affect the duty and foreseeability elements of negligence. See Mussivand v. David, 

 (Ohio 1989) (“The existence of a duty will depend on the foreseeability of the
injury to appellee.”); Abrams v. Worthington, 

 (Ohio Ct. App. 2006) (“In
Ohio, the existence of a duty depends upon the foreseeability of injury to the plaintiff.”); Conte
v. Gen. Housewares Corp., 

, 636 (6th Cir. 2000) (same); Est. of Ciotto v. Hinkle,

, 1019–20 (Ohio Ct. App. 2019) (“[F]oreseeability defin[es] the scope and
extent of the duty.” (quotation marks omitted)); see also Palsgraf v. Long Island R.R., 

, 100–01 (N.Y. 1928) (“The risk reasonably to be perceived defines the duty to be obeyed . . . .
[A plaintiff] must show that the act as to him had possibilities of danger so many and apparent as
to entitle him to be protected . . . .”). Any combination of these factual differences could lead a
jury to find that a particular plaintiff’s injuries were not reasonably foreseeable and, therefore,
that DuPont did not owe or breach a duty of care.

       Consider location. Abbott grew up and lived in and around Pomeroy, Ohio, which is
56.9 river miles away from DuPont’s Washington Works plant in Washington, West Virginia.
[Abbott, No. 2:17-cv-998, R. 192, PageID# 8292; R.33-2, PageID# 342–43]. During that time,
his water was sourced from wells ranging anywhere from 14 to 56 miles away from source of C-
8 emissions. [Abbott R. 192, PageID# 8292; R.33-2, PageID# 342–43; R. 254-3, PageID#
 No. 21-3418                     In re E. I. du Pont de Nemours & Co.                     Page 41


14012]. While Bartlett sometimes drank water from the same source as Abbott (for the years she
lived in and around Tupper Plains, Ohio), [Bartlett, No. 2:13-cv-170, R. 131, PageID# 4530,
4613–14], Freeman and Vigneron drank water sourced from wells much closer to DuPont, only
about 1,500 feet away from the Washington Works plant, [Appellant’s Br. 29]. Beyond that, as a
general matter, the post-2017 plaintiffs, including Abbott and Swartz, appear to have lived
farther away from DuPont’s plant than the plaintiffs in the earlier trials. [See MDL 5208,
PageID# 125922–24, 125934–38]. Swartz, for example, lived outside the water districts listed in
the Leach Agreement and premised her negligence claim on periodic exposure to C-8, claiming
that she occasionally drank contaminated water when visiting the homes of others and during a
one-year part-time job. [MDL 5208, PageID# 125922–24; MDL 5278, PageID# 128443–44;
Swartz, No. 2:18-cv-136, R. 51-10, PageID# 1411].

        If divergent facts in later cases could lead juries to reach different conclusions, then
collateral estoppel is inappropriate. See CHKRS, LLC v. City of Dublin, 

, 491–92
(6th Cir. 2021) (finding collateral estoppel inapplicable in a contract dispute because the issue
was not “identical” to one that had already been litigated, and noting that an issue cannot be
defined “at too high a level of generality” such that it “overlooks the changed facts across”
cases); Est. of Van Dyke by Van Dyke v. GlaxoSmithKline, 

, at *5 (D. Wyo.
Nov. 1, 2006) (declining to apply collateral estoppel where each case involved “different facts,
doses, time frames, diagnoses, warnings and research”); Dopson-Troutt v. Novartis Pharms.
Corp., 

, at *2 (M.D. Fla. Sept. 20, 2013) (rejecting collateral estoppel in a
toxic exposure case because the plaintiff failed to address whether “the scientific knowledge
relevant in [the second case] . . . would be different” from that in the prior trials).

        While there are undoubtedly some similarities among Abbott, Bartlett, Freeman, and
Vigneron, there are also plenty of legally significant factual differences that I cannot overlook.
United States v. Stauffer Chem. Co., 

 (1984) (factual differences must be of
“no legal significance whatever” for collateral estoppel to apply). But even if I did agree that the
differences in Abbott have no legal significance at all, I am equally concerned about the
thousands of other potentially differently situated plaintiffs who stand to benefit from the court’s
estoppel order. The district court erred by disregarding these differences.
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                       Page 42


       The verdicts in those early cases did not, as the district court held, “ma[k]e clear that the
duty DuPont breached was to the entire communities surrounding its Washington Works plant
and not just to specific customers of individual water districts.” [MDL 5285, PageID# 128574
(emphasis added)]. The juries were asked only about negligence with respect to the particular
plaintiff, or someone in the position of the plaintiff. [See, e.g., Freeman R. 97, PageID# 1011
(Verdict Form); R. 102, PageID# 1050 (Jury Instructions) (in deliberating on “the existence of a
duty,” consider whether a “reasonable prudent person would have foreseen at the relevant time
that injury was likely to result to someone in Mr. Freeman’s position”); Vigneron, No. 2:13-cv-
136, R. 195, PageID# 8617 (same); Bartlett R. 139, PageID# 6205 (same)]. The juries were
never instructed about a “community” theory of negligence. And even if they had been, the
general verdict form would still have left it unclear if that theory served as the basis for their
decision. See Loughridge v. Chiles Power Supply Co., 

 (10th Cir. 2005)
(district court may not, after a verdict, “embrace[] [a theory] not addressed by the jury” because
that would be stepping “into the impermissible realm of speculation as to what the jury actually
determined”).

       I also anticipate that this preclusive effect essentially guts the utility of informational
bellwether trials. After today, it seems that parties can do nothing—other than not conduct
bellwethers at all—to prevent an informational bellwether from becoming binding. Parties can’t
purposefully select unrepresentative plaintiffs to go first, nor can they purposefully use general
verdict forms so that preclusion does not attach. See Savage, supra, at 464 (suggesting as much);
[Appellant’s Br. 24]. I suggest that the age of bellwethers will come to an end, as any residual
benefit of conducting one will be outweighed by its now-endorsed preclusive consequences.

       For all these reasons, the Abbotts cannot show that the three bellwether trials “actually”
decided “the precise issues” of duty, breach, and foreseeability for all future MDL cases.
I would vacate the district court’s estoppel order and remand.

                                                II.

       I am pleased to concur in Part II.B of the majority’s opinion, which concludes that the
district court properly excluded portions of DuPont’s expert testimony, properly admitted
 No. 21-3418                          In re E. I. du Pont de Nemours & Co.                                    Page 43


Abbott’s specific causation expert, and did not exclude all testimony on the potential alternative
causes of Abbott’s cancer.

         I write separately on this point to press one caveat: evidence of a plaintiff’s specific
dosage or level of exposure to a contaminant is relevant to specific causation, and such evidence
does not undermine DuPont’s concession on general causation so long as the evidence is used in
a way that questions the likelihood—and not the capability—of harm.6

         According to DuPont, the Science Panel found “that the amount of risk varies greatly
with dose, and that some of [the Panel’s] data showed that only ‘very high’ blood levels of C8
materially increased an individual’s risk.” [Appellant’s Br. at 34; see MDL 2813-4, PageID#
46017, 46019; Abbott R. 259-1, PageID# 18432–34]. That makes sense. Increased exposure
generally means increased risk of harm. And not every person drinking water contaminated with
C-8 over the course of fifty years and in different locations will have the same exposure levels.
That’s why in toxic exposure cases like this, relative risk analysis is often the meat and potatoes
of expert opinions. Pluck v. BP Oil Pipeline Co., 

, 676–77 (6th Cir. 2011).
Therefore, DuPont could have, in theory, elicited expert testimony pointing out that Abbott
apparently had a low dosage level of C-8 in his bloodstream and that it was therefore unlikely
that C-8 caused his cancer. That testimony would have squarely addressed specific causation,
not general causation, and would have been admissible.

         But that’s not quite what DuPont tried to do in the district court. Despite DuPont’s
characterizations on appeal, from early on DuPont has consistently and repeatedly insisted that it
could point out that a plaintiff’s C-8 levels could be so low that C-8 was incapable of causing his
or her cancer. [See MDL 1679, PageID# 22980–81; MDL 3972, PageID# 68167–74; MDL
4079, PageID# 71853–55; MDL 4226, PageID# 81635; MDL 4777, PageID# 108871–72,
108895; MDL. 5285, PageID# 128552; MDL 5294, PageID# 128750–58; R. 5305, 128936–39].
As the district court and the majority correctly describe, the Leach Agreement takes that kind of

         6
           The Abbotts briefly argue (and the majority finds) that DuPont forfeited this argument because it “chose
not to appeal the district court’s collateral estoppel decision on this issue.” [Appellee’s Br. at 18; see id. at 41].
Even if the Abbotts provided a developed forfeiture argument (they do not), their argument fails. The district court
did not rely on the estoppel order in making its evidentiary rulings; rather, it relied on its interpretation of the Leach
Agreement and the expert opinions specific to Abbott and Swartz.
 No. 21-3418                          In re E. I. du Pont de Nemours & Co.                                    Page 44


testimony off the table. And even though the court prevented DuPont’s proffered causation
expert from opining that it was “more likely” that Abbott’s cancer had alternative causes—which
would appear permissible—there was still no error. DuPont’s expert refused to rule in C-8 as a
possible cause of Abbott’s cancer, as the Leach Agreement required, and therefore his testimony
was properly excluded. [MDL 5301, PageID# 128860–66; see MDL 4079, PageID# 71861
(“DuPont has contractually agreed that its experts must rule in C-8 as a possible cause of [a class
member’s linked disease].”); Abbott R. 65, PageID# 2065].

         All in all, DuPont retained the right to call attention to a plaintiff’s C-8 levels in order to
contest whether C-8 likely caused that plaintiff’s cancer. But, as explained, DuPont tried to do
more than that at the district court. For that reason, I concur.7

                                                          III.

         I must dissent in part from Part II.C of the majority’s opinion as well. Abbott did not file
his federal lawsuit for his 1994 cancer until 2017, and there are good arguments that his claim is
time-barred by Ohio’s statute of limitations. The district court erred by taking that issue away
from the jury.

         As the majority explains, Ohio has a two-year statute of limitations on personal injury
claims. Ohio Rev. Code § 2305.10(A). It reads:

         [A] cause of action for bodily injury . . . that is caused by exposure to hazardous
         or toxic chemicals . . . accrues upon the date on which . . . by the exercise of
         reasonable diligence the plaintiff should have known that the plaintiff has an
         injury that is related to the exposure.

Id. § 2305.10(B)(1). For the Abbotts’ claim to be timely, it must have been brought no later than
November 14, 2015, unless tolled.



         7
          It does to some extent concern me, however, that the district court prohibited the jury from reviewing the
Science Panel’s findings, even though the jury asked to see them. [Abbott R. 187-1, PageID# 6857 (“Can we
use/see the 2012 Science Panel Report?”); see also id. at 6857 (jury question stating that one juror refused to
“consider as fact that the Science Panel determined drinking water containing .05 ppb . . . linked to testicular . . .
cancer”]. The jury in the first bellwether trial, Bartlett, also asked to see the Science Panel’s report, but the district
court refused to provide it then as well. [Bartlett R. 146, PageID# 6496 (“Can we see the scientific report that
determined 0.05 ppb”)].
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                     Page 45


       But in granting the Abbotts’ motion for judgment as a matter of law, the district court
erroneously required DuPont to present evidence of Abbott’s actual notice of his injury. [Abbott
R. 205, PageID# 10857–59; see generally id. at 10837–59; R. 241, PageID# 1222]. The court
relied on its prior summary judgment decision, which stated that “constructive notice is not the
applicable test.” [MDL 5304, PageID# 128912; see Abbott R. 205, PageID# 10843]. The court
held that DuPont did not present sufficient evidence that Abbott had “actually encountered”
information to put him on notice of the potential link between C-8 and his testicular cancer.
[MDL 5304, PageID# 128913].

       That was wrong. The “should have known” language in the Ohio Revised Code confirms
that sufficient evidence of constructive notice can be enough to start the clock. Ohio Rev. Code
§ 2305.10(B)(1); see Twee Jonge Gezellen, Ltd. v. Owens-Illinois, Inc., 

(6th Cir. 2007) (noting that, under Ohio law, the statute of limitations begins when a plaintiff
“discovered or should have discovered” both his “injury” and that his injury “was the result of
[defendant’s wrongful conduct]” (emphasis added)); Norgard v. Brush Wellman, Inc., 

 (Ohio 2002) (“[T]he discovery rule . . . provides that a cause of action does not
arise until the plaintiff discovers, or by the exercise of reasonable diligence should have
discovered, that he or she was injured by the wrongful conduct of the defendant.” (emphasis
added)); Flowers v. Walker, 

 (Ohio 1992) (“[C]onstructive knowledge of
facts, rather than actual knowledge of their legal significance, is enough to start the statute of
limitations running under the discovery rule.” (emphasis in original)). The Abbotts do not cite a
case that says otherwise.

       Evidence of Abbott’s constructive notice included (1) that, through local media coverage,
it was widely publicized—often on the front page of newspapers—that C-8 was linked to
testicular cancer; (2) that Abbott released his testicular cancer information to the “C-8 Health
Project” in 2006 to determine if his health had been affected by drinking water containing C-8;
(3) that by early 2015, his grandparents, his secretary, and about 3,500 individuals in his
surrounding area had sued DuPont for their linked diseases; and (4) that Abbott was a high
school principal at a school that held a public meeting about the Science Panel’s findings.
[Appellant’s Br. 50–53; Appellant’s Reply B. 23]. Abbott, of course, denies that he ever knew
 No. 21-3418                   In re E. I. du Pont de Nemours & Co.                       Page 46


C-8 could cause testicular cancer until two weeks before he filed suit. But self-serving denials
are not enough to take a triable issue away from a jury when contradictory circumstantial
evidence exists.

       Though perhaps weak, there was sufficient evidence of Abbott’s constructive notice as it
relates to his 1994 cancer to give the statute-of-limitations issue to the jury. See Anderson v.
Liberty Lobby, Inc., 

 (1986) (“[T]he weighing of the evidence, and the drawing
of legitimate inferences from the facts are jury functions, not those of a judge, whether he is
ruling on a motion for summary judgment or for a directed verdict.”); see Groob v. KeyBank,

 (Ohio 2006) (directed verdict is proper when, “after construing the
evidence most strongly in favor of the party against whom the motion is directed, reasonable
minds could come to but one conclusion upon the evidence submitted” (quotation marks
omitted)). In other words, a reasonable jury could have found that a person in Abbott’s position
exercising reasonable diligence would have been on notice about his injury and his claim against
DuPont prior to November 14, 2015 (two years before he filed suit).

       With regard to Abbott’s 2015 cancer, however, I agree with the majority that the statute
of limitations did not begin to run until after a pathologist removed and examined Abbott’s mass
and confirmed that it was cancerous.

                                               IV.

       For the foregoing reasons, I would reverse the district court’s grant of collateral estoppel,
as well as the court’s grant of judgment as a matter of law as it relates to Abbott’s 1994 cancer.
I respectfully dissent from those portions of the majority opinion and judgment.