Tumlinson, et al. v. Advanced Micro Devices, Inc.
Annotate this Case
Download PDF
IN THE SUPERIOR COURT OF THE STATE OF DELAWARE
IN AND FOR NEW CASTLE COUNTY
WENDOLYN TUMLINSON, JAKE
TUMLINSON, JILLVEH ONTIVEROS
and PARIS ONTIVEROS, by her natural
mother and next friend JILLVEH
ONTIVEROS,
Plaintiffs,
v.
ADVANCED MICRO DEVICES, INC.,
Defendant.
)
)
)
)
)
)
)
)
)
)
)
)
)
Supreme Court
No. 672, 2012D
Superior Court
New Castle County
C.A. No.: 08C-07-106 FSS
Submitted: August 16, 2013
Decided: October 15, 2013
On Remand to Decide Admissibility of Epidemiologist’s Opinion
under D.R.E. 702
SILVERMAN, J.
Plaintiffs allege that chemical exposure at Defendant’s semiconductor
factories caused birth defects in two of their children. Jake Tumlinson was born in
1987 and Paris Ontiveros was born in 1994, 21 and 14 years, respectively, before this
lawsuit was filed. Jake was born with anal atresia and stenosis, neurogenic bladder,
renal agenesis/hypoplasia, imperforate anus, and colo-vesicular fistula. These birth
defects, in combination, are referred to as VATER association. Paris was born with
pulmonic stenosis, congenital pulmonary valve atresia, ventricular septal defect, right
pulmonary hypoplasia, lower limb reduction defects and situs inversus with
dextrocardia, which, only for expedience, the court will henceforth refer to
collectively as situs inversus. Paris’s congenital heart problems are an important fact
for this litigation’s purposes. All these conditions are rare and, generally, their
origins are unknown.
Previously, this court held that Texas substantive law applies here.
Applying Texas’s heightened epidemiology evidence threshold, this court granted
Defendant’s motion to exclude the opinions of Plaintiffs’ expert, Dr. Linda Frazier.
The Supreme Court affirmed the choice of law, but remanded for further inquiry as
to the admissibility of Dr. Frazier’s opinion under Delaware procedural law,
specifically D.R.E. 702.
This court was instructed to determine the “expert
2
testimony’s reliability under Delaware law” to facilitate the “debate over what role
sufficiency plays in admissibility.”1
To prevail in a toxic tort case, plaintiff must show both general and
specific causation. General causation concerns whether a particular substance causes
anyone the specific harm alleged. Here, Plaintiffs must first prove that the chemicals
to which Plaintiffs’ parents were exposed at work can cause Plaintiffs’ birth defects,
including VATER association and situs inversus. Specific causation asks if the
substance(s) actually caused Plaintiffs’ injuries.2 Dr. Frazier opines about general and
specific causation and now the court is considering those opinions’ reliability.
Reliability is one factor for admissibility under D.R.E. 702. Later in litigation when
the merits of the case are addressed, the sufficiency of the admissible evidence to
meet Plaintiffs’ burden comes into play.
Dr. Frazier layers assumptions and assertions without adequately
explaining her sources or reasoning, but assuring the court that she has “faithfully
followed the requirements of good science.” Dr. Frazier’s hypothesis, developed for
this case, lacks the specificity of chemicals, exposure, or doses necessary to be
testible. The opinion has never been peer reviewed. Dr. Frazier pieces together
epidemiology, animal studies, in vivo studies, and in vitro studies that examine
1
2
Tumlinson v. Advanced Micro Devices, Inc., 2013 W L 4399 144 (D el. Aug. 16, 2013).
King v. Burlington Northern Santa Fe Railway Company, 277 Neb . 203 (2009 ).
3
tangential exposures and harms without any explanation as to the relative weight each
study deserves. She applies established methods, like dose-response, in novel ways,
again without necessary explanation. Finally, despite admitting these birth defects
occur in unexposed people, Dr. Frazier broadly eliminates any cause other than the
parents’ exposure. As explained below, the analytical gaps between the data and the
proferred opinions are simply too many and too wide.
I.
Expert opinion evidence under Delaware law is governed by Delaware
Rule of Evidence 702, which provides:
If scientific, technical or other specialized knowledge will
assist the trier of fact to understand the evidence or to
determine a fact in issue, a witness qualified as an expert
by knowledge, skill, experience, training or education may
testify thereto in the form of an opinion or otherwise, if (1)
the testimony is based upon sufficient facts or data, (2) the
testimony is the product of reliable principles and methods,
and (3) the witness has applied the principles and methods
reliably to the facts of the case.
Delaware uses Daubert v. Merrell Dow Pharmaceuticals, Inc.3 to apply Rule 702.4
Daubert requires the trial judge to act as a “gatekeeper,” determining
whether the proffered evidence is both “relevant” and “reliable.”5 Under Daubert,
“relevant” means the evidence relates to an issue and it will aid the fact finder. As
3
Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579 (1993).
M.G. Bancorporation, Inc. v. Le Beau, 737 A.2d 51 3, 521-522 (D el.1999).
5
Dau bert, 509 U.S. at 597.
4
4
mentioned above, the court already decided Dr. Frazier’s opinions were irrelevant as
the jury could not take them as proving under Texas law what they were meant to
prove - causation.
“Reliable” means “testimony must be supported by appropriate
validation—i.e., ‘good grounds,’ based on what is known.”6 In other words, an
“inference or assertion must be derived by the scientific method.”7 Daubert's nonexclusive criteria for determining reliability include: (1) whether the expert's theory
has or can be tested; (2) whether the theory has been subject to peer review; (3) the
known or potential error rate associated with the theory; and (4) the extent to which
the theory has been generally accepted in the scientific community.8
These criteria are applied to both the expert’s ultimate opinion as well
as foundational sources used in forming the opinion. While no individual criterion
is dispositive, peer review and pre-litigation research are principal ways for
demonstrating reliability.9 Where neither can be shown, the expert “must explain
precisely how they went about reaching their conclusions and point to some objective
source—a learned treatise, the policy statement of a professional association, a
published article in a reputable scientific journal or the like—to show that they have
6
Id. at 590.
Id.
8
Id. at 593–594.
9
Da ube rt v. Merrell D ow Pharm ., Inc. (Dau bert II), 43 F.3d 1311 , 1319 (9th Cir. 1995).
7
5
followed the scientific method, as it is practiced by (at least) a recognized minority
of scientists in their field.” 10
The reliability requirement, however, “must not be used as a tool by
which the court excludes all questionably reliable evidence. The [...] touchstone [...]
is helpfulness to the trier of fact.” 11 Judge Quillen, while thoroughly explaining
Daubert, observed:
Daubert is a two-sided coin. On the one side, it is
expansive, rejecting the exclusivity of the “general
acceptance” requirement; on the other side, it is restrictive,
with a focus on the Trial Judge's responsibility as a
gatekeeper on reliability.12
The trial judge must guard against speculation, but plaintiffs “do not have to
demonstrate to the judge by a preponderance of the evidence that the assessments of
their experts are correct, they only have to demonstrate by a preponderance of
evidence that their opinions are reliable.”13 As one court puts it: “In sum, while the
trial court acts as the evidentiary gatekeeper, it is not a goalkeeper.” 14
The party offering the expert has the burden of demonstrating the expert
is qualified and her opinion is relevant and reliable.15
Accordingly, while
Defendant’s lack of an opposing expert is less common and unhelpful, it is by no
10
Id.
United States v. Velasquez, 64 F.3d 844 , 850 (3d Cir. 1995 ) citing In re Paoli R.R. Yard PCB Litig., 35 F.3d 717,
744 (3d Cir. 1994).
12
Min ner v. Am . Mo rtgag e & Guar. Co., 791 A.2d 826, 841 (Del. Super. 2000).
13
In re P aoli, 35 F.3d at 744.
14
King, 277 Neb. at 227.
15
Bowen v. E.I. DuPont de Nemours & Co., Inc., 906 A.2d 78 7, 795 (De l. 2006).
11
6
means dispositive as Plaintiffs repeatedly imply. An opposing expert may help the
court focus on weaknesses and otherwise inform the court. But, even if the court
would have benefitted here from counter-opinion, its absence does not help Plaintiffs
establish in the slightest that Dr. Frazier’s opinions are scientifically reliable.
II.
Applying Daubert to different scientific fields has produced different
analyses and results. Accordingly, non-epidemiological precedent and cases about
different environmental toxins and exposure modalities are less helpful, and not
dispositive. The admissibility of the epidemiologist’s opinion here is a matter of first
impression in Delaware. Nevertheless, this reliability analysis starts with briefly
surveying Delaware cases applying Daubert in other contexts, then briefly surveying
other jurisdictions that have specifically considered similar epidemiological opinions.
Finally, within that analytical framework, the court will further consider Dr. Frazier’s
opinions, as the remand expects.
III.
Most Delaware cases applying Daubert involve substances with
established toxicity.16 Long, for example, holds that where clear scientific evidence
demonstrates that ephedra caused cardiac events, proper differential diagnosis alone
16
E.g., General Motors Corp. v. Grenier, 981 A.2d 531 (Del. 2009) (chrysotile, a form of asbestos, causes
meso theliom a); New Haverford Partnership v. Stroot, 772 A.2d 792 (Del. 200 1) (mold causes su bstantial health
issues, including respira tory and cognitive pro blems); Long v. Weider Nutritional Group, Inc., 2004 WL 1543226
(Del. Super. June 25, 2004) (clear scientific evidence links ephedra to cardiac events).
7
may prove specific causation and, therefore, epidemiology is unnecessary. 17 In
Grenier, after an elaborate Daubert proceeding like the one in this case, this court
properly admitted an expert’s opinion that friction chrysotile was probably
carcinogenic because it is so similar to unrefined chrysotile, a proven cause of
mesothelioma. The expert based his opinion “on his own research, published in a
peer-reviewed journal.”18 And, as to that, “his findings were consistent with the
findings published in other peer-reviewed papers.”19
Grenier also admitted
conflicting epidemiological testimony.20 The established reliability of the, albeit
conflicting, testimony in Grenier helps demonstrate how tenuous Dr. Frazier’s
opinions are. In summary as to Grenier, this case is far more like Richardson,21 a
contrary federal court decision, which was far easier to distinguish in Grenier than
it is here.22
McMullen involved a “Pediatric Condition Falsification” diagnosis - a
relatively new, but generally accepted condition. Where no published studies on
general causation existed in the medical community, McMullen admitted a well
structured, defined differential diagnosis where differential diagnosis is the only
17
Long, supra note 16.
Grenier, 981 A.2d at 536.
19
Id.
20
But see G renier, supra note 16 (Steele, C.J., dissenting) (testimony rejected because the hypothesis was untested,
not sub ject to p eer review, and lacked reliable m ethod ology).
21
Richardson by Richardson v. Richardson-Merrell, Inc., 857 F.2d 823, 824 (D.C. Cir. 1988)
22
Grenier, 981 A.2d at 538-539.
18
8
means of medically diagnosing PCF.23 Conversely, Scaife excluded a differential
diagnosis opinion that the prescription medicine Seroquel caused plaintiff’s diabetes
because it was unsupported by defined methodology or an identified biological
mechanism.24
Delaware also has cases where no specific toxin is implicated. For
example, in Minner, a “toxic building” case, an expert attempting to “make a
temporal connection between the Plaintiffs’ illnesses and the building without a
specific toxin identified [...] must demonstrate a deductive scientific process to
support her conclusion.”25
Minner reviewed 12 experts, detailing their opinions and methods.
Judge Quillen excluded certain experts entirely and specific portions of the admitted
experts’ testimonies, separating wheat from chaff. For example, in admitting expert
testimony that “volatile organic compounds, dusts and molds in the building” caused
Toxic Encephalopathy, Minner states that plaintiffs made a “bare showing” that the
experts methodology was reliable where a scholarly article “admits that it may be
difficult to identify a specific causative agent in the workplace and states chronic
solvent exposure can be associated with cognitive changes in an individual.”26 By
linking the precise harm with a category of substances, the court was satisfied that the
23
State v. McM ullen, 900 A.2d 10 3 (2006 ).
Scaife v. Astrazeneca LP, 2009 W L 1610 575 (D el. Super. June 9, 2009).
25
Minner v. Am. Mortgage & Guar. Co., 791 A.2d 82 6, 855 (De l. Super. 2000).
26
Id. at 858.
24
9
methodology was sufficient. Conversely, the court excluded opinion that exposure
in the building caused Fibromyalgia because the expert “does not follow a logical,
scientific, and deductive process to exclude other possible causative factors.”27 In
short, as the established scientific proof for causation decreases, an expert’s
methodology for forming her opinion must be increasingly detailed. Under Delaware
precedents alone, none admits an opinion including as many untested extensions of
published studies as Dr. Frazier’s here.
IV.
Epidemiology is the science of the relationship between human
behaviors and patterns, causes, and effects of diseases across the population.
“Epidemiology focuses on general causation.”28 Epidemiology observes exposure as
related to disease risk and, accordingly, “frequently, plaintiffs find epidemiological
studies indispensable in toxic tort cases when direct proof of causation is lacking.”29
Epidemiology can also be used to support specific causation when the association is
particularly strong and the setting and subjects of the study are substantially similar
to plaintiff.
As mentioned, while Delaware has not considered an epidemiological
opinion under Daubert to support a causation hypothesis, other jurisdictions have.
27
Id. at 855.
King, 277 Neb. at 213.
29
Id.
28
10
In its original decision here, the court relied on Havner30 and Garza,31 which
meticulously set out what it takes for an epidemiologist to establish causation in
Texas and how epidemiological opinion’s admissibility must be analyzed under Texas
law. Because this case will ultimately be decided under Texas law and Dr. Frazier
does not offer what it takes to prove causation in Texas even if her opinion is
admissible, the court did not consider whether the opinion is reliable under D.R.E.
702 and Daubert. Now that the court must consider reliability for D.R.E. 702
purposes, it must not only revisit Havner and Garza, which are non-binding authority
on the issue here, but also non-binding authority from other jurisdictions.
Havner, in rejecting the epidemiological opinion supporting Plaintiff’s
jury verdict, held that the studies on which an epidemiologist’s opinion is based must
show: “the relative risk would need to exceed 2.0, and the confidence interval could
not include 1.0, for the results to indicate more than a doubling of the risk and a
statistically significant association.” 32
Plaintiff must also prove that he is similar to the study’s subjects by
showing that he “was exposed to the same substance, that the exposure or dose levels
were comparable to or greater than those in the studies, that the exposure occurred
before the onset of injury, and that the timing of the onset of injury was consistent
30
Merrell Dow P harm., Inc. v. Havner, 953 S.W .2d 706 (T ex. 1997).
Merck & Co., Inc. v. Garza, 347 S.W .3d 256 (T ex. 2011).
32
Havner, 953 S.W.2d at 725.
31
11
with that experienced by those in the study.”33 In 2011, Garza reaffirmed and
emphasized this standard for admissible epidemiological opinion in Texas. Garza
made plain that the Havner standards are minimums. Clearing Havner and Garza’s
bar does not automatically justify admissibility.34 Again, for present purposes,
Havner and Garza are not controlling, but this court may adopt them.
As much as Havner clearly explains the statistical concepts and
techniques it applied in establishing the 2.0 minimum, it is not flawless. Its analysis
muddles general and specific causation.
Outside Texas, acceptance of Havner’s 2.0 relative risk bright-line rule
varies. Some jurisdictions follow it.35 Others accept the statistical significance
requirements as a measure of evidentiary sufficiency, but not as a threshold for
admissibility.36 And, others merely require a positive association, relying on the jury
to determine the significance of the studies after proper instruction.37
General causation does not consider the likelihood that a certain
exposure caused a certain harm.
Rather, it only considers the possibility.
Accordingly, as King holds, any epidemiological study showing a positive association
33
Id. at 720.
Garza, 347 S.W.3d at 265-266.
35
E.g. Siharath v. Sandoz Pharmaceuticals Corp., 131 F.Supp.2d 1347 (N.D. Ga. 2001) aff'd sub nom. Rider v.
Sandoz Pharm. Corp., 295 F.3d 119 4 (11th Cir. 200 2); Estate of George v. Vermont League of Cities and Towns,
993 A.2d 367 (V t. 2010).
36
E.g. Lofton v. M cNeil Con sum er & Sp ecialty Pharm aceuticals, 682 F.Supp.2d 662 (N .D. Texas 201 0) aff'd, 672
F.3d 372 (5th Cir. 2012).
37
E.g. King, supra , note 2; In re Joint E. & S. Dist. Asbestos Litig., 52 F.3d 112 4 (2d Cir. 1995 ); In re Prem pro
Products Liability Litigation, 738 F.Supp.2d 887, 892 (E.D.Ark.2010 ).
34
12
can be part of a reliable opinion regarding general causation. Recognizing that the
majority of courts equate a 2.0 minimum relative risk to plaintiff’s burden of proof,
King states that this requirement is too restrictive because no single epidemiological
study will prove plaintiff’s case.38 This less rigorous approach is followed by some
federal circuit courts.39 The Second Circuit explains, “we believe that it would be far
preferable [...] to instruct the jury on statistical significance and then let the jury
decide whether many studies over the 1.0 mark have any significance in
combination.” 40
As explained above, if epidemiology is offered to prove specific
causation, then the standard of proof and logic dictate that the study must show
plaintiff’s injuries, more likely than not, were caused by the exposure. To that extent,
a relative risk exceeding 2.0 makes sense because it means that, within that study’s
experimental group, there was more than a 50% chance that any individual defect was
caused by exposure as opposed to something else. Of course, that statistic can not be
extrapolated to plaintiff if he and his exposure are not fundamentally similar to the
study subjects.
On the other hand, where epidemiology is offered to support general
causation, whether the exposure could cause harm at all, a less stringent standard is,
38
King, 277 Neb. at 231.
See e.g. In re Joint E. & S. Dist. Asbestos Litig., supra note 25; In re H anfo rd Nuclea r Reservation Litig., 292
F.3d 112 4, 1137 (9th Cir. 2002 ).
40
In re Joint E. & S. Dist. Asbestos Litig., 52 F.3d at 1134.
39
13
by definition, weaker but potentially probative. If, for example, someone wanted to
determine if a coin was fair, it would not matter if heads came up at least twice as
often as tails. Any “statistically significant” deviation from the expected 50/50
distribution would support the theory that the coin was not fair. “Statistically
significant” essentially means that the unexpected thing happened often enough that
there is a minimal likelihood it occurred by chance. To use the coin example, if a
coin showed heads 600 times out of 1000 tosses, that result is unlikely to occur with
a fair coin and suggests, without proving, that the coin is not fair. That is so even
though the coin did not show heads twice as often as tails.
In epidemiology, certain standards govern statistical significance. “The
generally accepted significance level or confidence level in epidemiological studies
is 95%, meaning that if the study is repeated many times, the confidence interval
indicates the range of relative risk values that would result 95% of the time.”41
Reliability under Daubert is fundamentally based on accepted methodology in the
field. Accordingly, requiring results at the 95% confidence level is appropriate. It
also makes sense to require that the entire confidence interval show a positive
association—a relative risk over 1.0—to demonstrate that there is a positive
association expected at least 95% of the time. Beyond that, however, there is no
scientific reason to exclude a weaker positive association from a general causation
41
Havner, 953 S.W.2d at 723.
14
analysis. At that point, it is in the expert’s hands to demonstrate how a weak positive
association supports her opinion through an articulate, defined methodology.42
V.
Dr. Frazier relies on several epidemiological studies, some of which
satisfy Havner’s statistical requirements, some do not. And there has been much
back-and-forth about that. The reliability analysis here, however, turns less on
statistics and more on methodology because Dr. Frazier fails to sufficiently explain
her methods. That explanation is essential under all authorities.
Dr. Frazier is a well-qualified epidemiologist and occupational health
professional. That finding is essential under Daubert. But, just as the absence of
counter-opinion is not persuasive as to Dr. Frazier’s reliability, neither is Plaintiffs’
argument, ad auctoritatem, that Dr. Frazier’s qualifications alone make her opinions
reliable.
As discussed more thoroughly below, the methods used by an
epidemiologist to form an opinion as to causation substantially rely on the expert’s
judgment in selecting and weighing her sources. Accordingly, courts require the
expert to clearly define her methodology and application. Dr. Frazier fails to do that
at several points. No quantum of evidence can overcome that.
42
See e.g. King, 277 Neb. at 231-232.
15
VI.
Determining reliability under Daubert is a highly fact specific analysis
based on the circumstances of the injury and exposure suffered by plaintiffs as well
as the areas of science at issue. In their Supplemental Brief, Plaintiffs understate the
necessity of establishing an association between the substances and harm in question
before analyzing causation. Similarly, they understate the need to establish general
causation before addressing specific causation through differential diagnosis.
Conversely, Defendant overstates Plaintiffs’ burden. Plaintiffs are not required to
find a flawless, all-inclusive epidemiological study mirroring their precise
circumstances.43 Accordingly, neither side has been entirely helpful.
Critically analyzing Dr. Frazier’s opinions takes several steps. The first
is examining the testibility of her hypothesis, including specificity of substance, dose,
and harm. That speaks directly to the first and third Daubert criteria. Next, under the
second and fourth Daubert criteria, the court considers if the hypothesis has internal
indicia of reliability to the scientific community through pre-litigation testing and
peer review. An expert can also demonstrate reliability or bolster her opinion through
a clearly defined weight-of-the-evidence analysis. Lastly, the court must determine
if the hypothesis demonstrates specific causation as to Plaintiffs’ injuries by ruling
out other factors through differential diagnosis.
43
See, e.g., Grenier, supra note 16.
16
A.
Again, the first criterion suggested under Daubert for determining
reliability is whether Dr. Frazier’s hypothesis can be or has been tested.
Experimentation requires specificity, but neither the substance(s) nor dose element
for a testable hypothesis is established by Dr. Frazier. Obviously, no one expects
scientists to expose humans to harmful chemicals for a controlled, clinical
experiment. Experimentation, however, can also be performed empirically through
observation, as the epidemiological studies relied on in this case demonstrate.
Regardless of the experimental method, a hypothesis must address a specific question,
e.g. is exposure to X chemical(s) in Y dose for Z time likely to cause VATER
association? Dr. Frazier acknowledges that “in designing a proper epidemiologic
study it is important to properly define the characteristics of the group being studied.”
As causative agents, Dr. Frazier names 10 chemicals, but she repeatedly
asserts that other, unidentified toxins somehow contributed to the birth defects. Dr.
Frazier never opines, however, which toxins specifically, alone or in combination,
caused Plaintiffs’ very different birth defects. Rather, she asserts that it is an “and/or”
situation, where any or all of the chemicals, in undefined combinations caused the
birth defects.
Similarly, Dr. Frazier refuses to specify dosages, relying instead on
atmospheric concentration ranges as a surrogate. Yet, she also insists that the actual
17
dosages must be higher due to peak exposure episodes and dermal absorption.
Further, despite the dissimilarity in their respective work experiences, Dr. Frazier
claims that both adult Plaintiffs were exposed to the same quality and quantity of
toxins. And, she opines that somehow this nondescript exposure caused two
substantially different birth defects with little attempt to explain the differences.
Those opinions gloss over the differences between the specific environments
Plaintiffs’ parents worked in, the different work they performed, the different
chemicals they were exposed to, and the different exposure levels. Just because these
may have been similar, they cannot be lumped together for each parent, much less for
both.
Plaintiffs make much of cases where imprecision has been excused.44
Those cases discuss whether general causation can be assumed where neither the
specific dose required for human toxicity nor the specific dose plaintiffs received are
known. In each case, however, the substance in question is known to be harmful at
some exposure level and the plaintiff suffered the precise harm connected to that
exposure.
Comparatively here, Dr. Frazier has to account for the fact that specific
substances and doses are lacking where there is neither scientific consensus that these
44
See e.g. Hardyman v. Norfolk & W. Ry. Co., 243 F.3d 25 5 (6 th Cir. 20 01); Clausen v. M/V NEW CARISSA, 339
F.3d 104 9 (9th Cir. 2003); John's Heating Serv. v. Lamb, 46 P.3d 10 24 (Alaska 2002 ).
18
causation chemicals are toxic to humans nor a signature harm. Reliable scientific
opinion cannot be formulated by one assumption on top of another. Dr. Frazier’s
untested hypothesis that adult Plaintiffs’ exposure to chemicals at Defendant’s
facilities caused their children’s birth defects is untestable for Daubert purposes.
B.
Assuming without deciding that the lack of specificity is not fatal to Dr.
Frazier’s opinion, as it is, Daubert next asks if the expert’s hypothesis has been peer
reviewed. Peer-review’s importance speaks to the heart of Daubert. “[S]ubmission
to the scrutiny of the scientific community is a component of ‘good science,’ in part
because it increases the likelihood that substantive flaws in methodology will be
detected.” 45
In summary as to Dr. Frazier’s general causation opinion, she has found
reliable foundational studies suggesting an association between working in the
semiconductor industry and reproductive problems. Dr. Frazier, however, has not
asked her peers to review what she has made of those studies, not through an article,
case note, correspondence, or otherwise. Instead, Dr. Frazier contends, in effect, that
because her personal opinion was formed by synthesizing peer reviewed foundational
studies, that is as strong as if her opinion was peer reviewed. That notion is illogical
and against the scientific method itself.
45
Daubert, 509 U.S. at 593.
19
As mentioned above, as a corollary to considering whether an opinion
has been peer reviewed, courts also frequently consider whether the theory was
developed outside of litigation. Daubert II emphasizes the relative significance of
research conducted independent of litigation:
For one thing, experts whose findings flow from existing
research are less likely to have been biased toward a
particular conclusion by the promise of remuneration;
when an expert prepares reports and findings before being
hired as a witness, that record will limit the degree to
which he can tailor his testimony to serve a party's
interests. Then, too, independent research carries its own
indicia of reliability, as it is conducted, so to speak, in the
usual course of business and must normally satisfy a
variety of standards to attract funding and institutional
support. Finally, there is usually a limited number of
scientists actively conducting research on the very subject
that is germane to a particular case, which provides a
natural constraint on parties' ability to shop for experts who
will come to the desired conclusion.46
Here, Dr. Frazier’s findings were made for this litigation. Even if it is tangentially
related to research she has done, Dr. Frazier came to her opinion in response to
Plaintiffs’ request.
Peer review and research outside of litigation are the strongest indicia
of reliability. Even so, “that the expert failed to subject his method to peer-review
and to develop his opinion outside the litigation is not dispositive.”47 Where neither
46
47
Dau bert II, 43 F.3d at 1317.
Lust By & Through Lust v. Merrell Dow Pharm., Inc., 89 F.3d 594, 597 (9th Cir. 1996).
20
peer review nor prelitigation research can be shown, however, the expert’s burden to
prove reliability is higher. The expert “must explain precisely how they went about
reaching their conclusions and point to some objective source—a learned treatise, the
policy statement of a professional association, a published article in a reputable
scientific journal or the like—to show that they have followed the scientific method,
as it is practiced by (at least) a recognized minority of scientists in their field.”48 As
mentioned above and discussed next, the personal way Dr. Frazier came to her
opinions is not well-enough explained.
C.
Epidemiology alone does not prove causation; it only demonstrates an
association. Accordingly, even the most widely accepted and strongest epidemiology
is not a litmus test for toxic tort causation.49 Nor is a lack of epidemiology fatal to
plaintiff’s case.50 To bolster epidemiology, lacking or otherwise, the expert must
explain her methodology in detail and point to an objective, reputable source to
demonstrate that her methodology is accepted.51
The parties agree that epidemiologists generally rely on the “Bradford
Hill” factors to show that a causal relationship can be inferred from an association.
The factors are 1) temporal relationship, 2) strength of the association, 3) dose48
Dau bert II, 43 F.3d at 1319.
Havner, 953 S.W.2d at 718.
50
Siharath v. Sandoz Pharm. Corp., 131 F. Supp. 2d at 1358.
51
Dau bert II, 43 F.3d at 1318-9.
49
21
response relationship, 4) replication of the findings, 5) biological plausibility, 6)
consideration of alternative explanations, 7) cessation of exposure, 8) specificity of
the association, and 9) consistency with other knowledge.52 The enumerated factors
are neither all-inclusive nor are they in hierarchical order; the factors merely provide
a framework for establishing causation.
Experts can also evaluate data using a weight-of-the-evidence analysis,
including available epidemiology, to assess causation. Other forms of evidence
including toxicology, in vivo studies, in vitro studies, animal studies, and case studies
can be used together to show causation. As explained above, the reliability analysis
is performed on each bit of data used to formulate the ultimate opinion.
Bradford Hill and weight-of-the-evidence are discussed together because
both rely on gathering, comparing, and weighing a wide range of data from various
sources. And the same considerations go into the reliability of these foundational
sources, including statistical significance and “fit,” with either method. Both
techniques are used to methodically extrapolate from association to causation.
Bradford Hill contemplates beginning with epidemiology and using other sources as
support. Weight-of-the-evidence, on the other hand, allows an expert to fit all the
sources together like a puzzle. To some extent, relying on weight-of-the-evidence is
an admission that the available epidemiology is weak. Put another way, non-peer
52
King, 277 Neb . at 221 citing Reference Ma nual on Scientific Evidence 376 (Fed eral Judicial Center 2d ed. 200 0).
22
reviewed weight-of-the-evidence opinion made for litigation is most suspect
categorically.
When reliable studies from different scientific fields are weighed in a
defined, reliable manner, Bradford Hill or weight-of-the-evidence methodology may
be appropriate.53 Due to the expansive potential sources of data for those analyses,
establishing “fit” for each source is especially important when considering the
source’s reliability as part of a meta-analysis. That applies in spades here, where the
data comes from Taiwanese Fab workers, Dutch house painters, electronics workers,
the animal kingdom, and so on.54
“‘Fit’ is not always obvious, and scientific validity for one purpose is not
necessarily scientific validity for other, unrelated purposes.55
For example, in order for animal studies to be admissible
to prove causation in humans, there must be good grounds
to extrapolate from animals to humans, just as the
methodology of the studies must constitute good grounds
to reach conclusions about the animals themselves. Thus,
the requirement of reliability, or “good grounds,” extends
to each step in an expert's analysis all the way through the
step that connects the work of the expert to the particular
case.56
53
E.g. In re C han tix (Varenicline) Products Liab. Litig., 889 F. Supp. 2d 12 72 (N.D . Ala. 2012).
Tum linson v. Ad van ced Micro D evices, Inc., 2012 W L 1415 777 (D el. Super. Jan. 6, 2012).
55
Daubert, 509 U.S. at 591.
56
Paoli, 35 F.3d at 743.
54
23
Further, the Third Circuit explains, “not only must each stage of the expert's
testimony be reliable, but each stage must be evaluated practically and flexibly
without bright-line exclusionary (or inclusionary) rules.”57
The first part of Dr. Frazier’s opinion is epidemiological studies. For
epidemiology, the first question is whether the foundational study shows a
statistically significant association.58 Dr. Frazier asserts that statistical significance
is not part of the Bradford Hill or weight-of-the-evidence analysis. From a legal
standpoint that is incorrect.59 Regardless of whether the court adopts Havner’s
bright-line requirements or the more subjective ones in King, a study must
demonstrate some positive association in order to have value as part of a larger metaanalysis. And, the weaker the association, the less it can contribute. Dr. Frazier
seems to accept that.
As discussed in the original opinion, some of the epidemiology is
unreliable.60 Some fail to “fit” this case: the 2006 Hooiveld study involved different
chemicals; the SIA study examined spontaneous abortion. The JHU/IBM study
“observed no adverse effects on birth weight or gestational maturity, and no increased
risk of birth defects or childhood neoplasms associated with either maternal or
paternal work in the Clean Room Areas.” Other studies possibly may barely show
57
Heller v. Shaw Indus., Inc., 167 F.3d 14 6, 155 (3d C ir. 1999).
Id. at 220.
59
In re Joint E. & S. Dist. Asbestos Litig., 52 F.3d at 1134.
60
Id. at 6.
58
24
reliability, but suffer from significant infirmities. The 2008 Lin study and the 2009
Sung study involve foreign Fab workers, but fail to measure specific chemical
exposures or doses. The Sung study also does not report a relative risk or odds ratio
for its findings. The DEC study only found a slight positive association.
Dr. Frazier asserts that her using epidemiological and animal studies
with spontaneous abortion as their endpoint “fit” this case. The joint affidavit
conclusively states, “there is compelling evidence specifically linking spontaneous
abortion with birth defect.”
She also bases her opinion on “dose-response
relationship.” Dose-response means increasing the intensity or duration of exposure
increases the frequency or severity of an outcome. Dr. Frazier opines that if a large
exposure causes spontaneous abortion, the dose response relationship establishes that
a small exposure of any chemical causes birth defects. Therefore, in her opinion,
scientific papers tending to show chemical exposure causes spontaneous abortions
also prove that exposure in lower doses cause birth defects. That original reasoning
helps Dr. Frazier opine that exposure to assorted solvents and other chemicals in
unknown amounts probably caused one Plaintiff’s VATER association and the
other’s situs inversus.
The court accepts that dose-response is a legitimate scientific and
pharmacological concept. That does not mean, however, that Dr. Frazier is free to
reach her own conclusions about how dose-response applies to chemicals, exposures,
25
and birth defects either generally or specifically in this case. Her application of doseresponse here is untested, even in the animal studies. Further, it has not been
subjected to peer review and it was created for this litigation.
The way Dr. Frazier uses dose-response has a direct bearing on the
court’s gatekeeping function under Daubert. Dr. Frazier’s—if a big dose of a
chemical is known to cause abortion, then a smaller dose probably causes birth
defects—theory has a superficial ring to it. But how will a jury evaluate that? Even
if Defendants produce a counter-expert, the jury room is not a substitute for an actual
science laboratory. Accordingly, studies analyzing spontaneous abortion as an
endpoint do not satisfy the “fit” requirement of reliability where specific birth defects
are the endpoint in question. They also cannot be considered in Dr. Frazier’s weightof-the-evidence analysis.
Further, as to weight-of-the-evidence, there is no generally agreed upon
method for weighing different data.61 That is another reason why, to confirm the
opinion’s scientific reliability, an expert must detail her method of weighing the
importance and validity of each data source to assemble a cohesive picture,
particularly where the opinion is formed for litigation and not peer reviewed. Being
presented with only the experts' qualifications, their conclusions and their assurances
61
King, 277 Neb. at 221.
26
of reliability is not enough under Daubert.62 In Daubert II, opinion was excluded
where “plaintiffs’ experts have relied on animal studies, chemical structure analyses
and epidemiological data, [but] they neither explain the methodology the experts
followed to reach their conclusions nor point to any external source to validate that
methodology.”63
Dr. Frazier consistently failed to explain her process. For example, in the
Joint Affidavit, some epidemiological studies are discussed in detail. Dr. Frazier,
however, never explains why these studies are worth discussion while others are not,
nor how the other studies affected her conclusion, nor how each study was weighed
against the others. The hearing testimony did not fill-in the gaps.
Dr. Frazier invokes both Bradford Hill and weight-of-the-evidence
analysis to support her general causation hypothesis. As detailed above, both
methods require the scientist to articulate her thought process, evaluation methods,
and conclusions to establish reliability. As a matter of law, it is not enough if a
scientist merely invokes her qualification in order to establish her judgments as
reliable. Nor does baldly asserting in the joint affidavit “each of us has exercised our
judgment in this manner in arriving at our opinion” satisfy the burden.
62
63
Id.
Dau bert II, 43 F.3d at 1319.
27
For general causation, no study draws a conclusion between the
causation chemicals and VATER assocation or situs inversus. Similarly no study
other than the Lin study of male semiconductor workers in Taiwan draws a
conclusion between the supposed causation chemicals and cardiac malformations.
At best, this study suggests, but does not prove, a possible association, much less a
causal relationship. Further, as recognized by the Lin study’s authors, the study is
wrought with infirmities. There were only nine cases of heart anomalies across the
entire study population. The study only recorded birth defects that killed the children
before age 5. It ignored an unknown population of children living with
malformations. The study lacked adequate exposure measurements. The most the
study concludes, in the author’s words, is that “it is clear that many of [the
semiconductor chemicals] are potential reproductive toxins.” From this, Dr. Frazier
wants the jury to find that an undefined mix of 10 named chemicals “and others,” to
which Plaintiffs’ parents were exposed for unknown, different periods, could have
caused Plaintiffs’ birth defects.
D.
Once general causation is established, which Dr. Frazier does not do,
Plaintiffs must still demonstrate specific causation. To return to the coin analogy, to
prove that the 60% heads result was in fact due to an unfair coin, other possible
causes must be eliminated, such as starting orientation. Differential diagnosis is an
28
accepted method for addressing specific causation, but it too relies on the expert’s
judgment and experience. Accordingly, to be legally reliable, the expert must
describe how and why she ruled out other potential teratogens.
Assuming without deciding that Dr. Frazier’s opinion on general
causation were accepted by the scientific community, Dr. Frazier fails to apply
differential diagnosis in an articulate, methodical way. Mostly, she conflates her
differential diagnosis with her causation opinion. Dr. Frazier cannot differentiate
between the possible causes of Plaintiffs’ birth defects because no one knows what
really causes them. So, because Dr. Frazier concluded that the parents’ exposures can
generally cause VATER association and situs inversus, and there are no other known
causes to distinguish, Defendant must have caused Plaintiffs’ birth defects. But, that
only works if it has been proved that there is only one cause of those birth defects,
such as the way asbestos causes mesothelioma. Even Dr. Frazier does not call
VATER association or situs inversus signature diseases.
Viewed against that factual backdrop, it appears that Dr. Frazier’s
differential diagnosis for Jake Tumlinson sets up a strawman to knock down. For
Jake Tumlinson, Dr. Frazier rules out the mother’s obesity, a factor linked to birth
defects generally, by finding unsupported differences between Mrs. Tumlinson and
subjects in the studies linking maternal obesity and birth defects.
Without
explanation, Dr. Frazier rejects Mrs. Tumlinson’s obesity as a cause because she was
29
not diabetic, yet Dr. Frazier acknowledges that studies have found “maternal obesity
in the absence of diabetes was associated with birth defects.”
For Paris Ontiveros, Dr. Frazier merely asserts that “[t]here are no
alterative (sic) explanations for Paris Ontiveros’ malformations.” Yet, all the studies
Dr. Frazier relies on, as well as Dr. Frazier herself in the same affidavit as the above
assertion, acknowledge that, although rare, the birth defects occur in the population
at large, often with no more evidence of causation than here. She offers no reason for
ruling out these accepted, unexplained background causes.
For both children, she essential relies on circular logic to state that
because the parents were likely exposed to these chemicals prior to conception and
during gestation, the chemicals must have caused or predisposed the children to these
very rare birth defects. Saying, in effect, “I can think of one way Plaintiffs’ birth
defects could have been caused and I can’t think of another,” does not amount to a
differential diagnosis.
VII.
In conclusion, the court finds Dr. Frazier is well-qualified by training
and experience to offer opinions concerning epidemiology and related environmental
medicine. The court also acknowledges cause for her concern about the potential
reproductive health effects for men and women working in computer chip
30
manufacturing facilities. Clearly, as experts in the field agree, there is need for
further research.
Nevertheless, no scientific study has established a general link between
female workers, such as Jake Tumlinson’s mother, and his birth defects, including
VATER association. And only one study generally links male workers with,
potentially, one of Paris Ontiveros’s several birth defects.
Because the general cause of Plaintiffs’ birth defects is unknown, it
cannot be said that the work Plaintiffs’ parents did was the specific cause of their
birth defects. But even if it has been proven that working in a computer chip
manufacturing facility may cause birth defects like Plaintiffs’, Dr. Frazier’s opinion
that those defects were caused by their parents’ exposure to toxins at work, rather
than other causes, is unreliable. Therefore, as a matter of Delaware law, Dr. Frazier’s
opinions are inadmissible under D.R.E. 702.
cc:
Ian Connor Bifferato, Esquire
Richard S. Gebelein, Esquire
David W. deBruin, Esquire
Thomas F. Driscoll, III, Esquire
J. Zachary Haupt, Esquire
Frederick L. Cottrell, Esquire
Travis S. Hunter, Esquire
31
Some case metadata and case summaries were written with the help of AI, which can produce inaccuracies. You should read the full case before relying on it for legal research purposes.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.