Wells v. SmithKline Beecham Corp.

Case: 09-50244 Document: 00511058145 Page: 1 Date Filed: 03/22/2010

IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT United States Court of Appeals
Fifth Circuit

FILED
March 22, 2010
No. 09-50244
Charles R. Fulbruge III
Clerk

MAX WELLS,

Plaintiff - Appellant

versus

S M IT H K L I N E B E E C H A M CORPORATION, doing b u s in e ss as
GlaxoSmithKline,

Defendant - Appellee

Appeal from the United States District Court
for the Western District of Texas

Before HIGGINBOTHAM, GARZA, and PRADO, Circuit Judges.
PATRICK E. HIGGINBOTHAM, Circuit Judge:
Max Wells invoked diversity jurisdiction and complained that
GlaxoSmithKline had failed to warn him that one of its drugs – Requip – could
make him gamble away millions. In support, Wells proffered testimony from
three experts who would opine that Requip can cause pathological gambling in
the general population. The company moved for summary judgment, arguing
that the experts’ testimony did not pass scientific muster. The district court
agreed – holding that even if the court admitted the expert testimony, the
opinions would not suffice to create a fact question about general causation,
 Case: 09-50244 Document: 00511058145 Page: 2 Date Filed: 03/22/2010

No. 09-50244

because the experts’ conclusions were not scientifically reliable. Wells appealed,
and we affirm on the alternative Daubert1 ground presented to us: because the
experts’ opinions are not scientifically reliable, the testimony is inadmissible –
leaving the liability cupboard bare.

I.
A.
Doctor Max Wells ran a successful pathology clinic but in 2000 fell ill with
Parkinson’s disease. Wells began taking Mirapex, a drug made by an unrelated
third-party manufacturer. Mirapex is a type of “dopamine agonist,” a class of
drugs that stimulates the dopamine receptors in the brain to alleviate symptoms
of Parkinson’s. Due to the neurological disease’s progression, Wells had to sell
part of his clinic in 2001, and in 2002 he retired. In 2005, Wells sold his
remaining interest in the clinic for a sizeable sum.
Since the 1970s, Wells had traveled regularly to Las Vegas. He enjoyed
gambling, and he claims to have kept his losses under control. After his
diagnosis, though, Wells began to lose more and more money. In 2004, he talked
to his doctor about an article he saw warning that Mirapex might cause problem
gambling. At that point Wells had already lost as much as $2 million. In
response, his doctor in October 2004 prescribed Wells a different dopamine
agonist – this time Requip (generically called ropinirole), the drug at issue in
this case. Requip’s label did not warn about gambling.
Wells’s gambling urges temporarily subsided but then came back with a
vengeance. From September 2005 (when he received final payment on the sale

1
Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993).

2
 Case: 09-50244 Document: 00511058145 Page: 3 Date Filed: 03/22/2010

No. 09-50244

of his clinic) until January 2006, Wells surrendered $10 million to the Nevada
gaming tables – including $4 million in January alone.2 He told his doctor about
the gambling problem, stopped taking Requip, and has not been back to Vegas.
B.
Wells sued GlaxoSmithKline, the manufacturer of Requip, alleging that
GSK 3 had failed to warn patients about the side effect of pathological gambling.4
For Wells to win under Texas law,5 he must show that the failure to warn caused
his injury.6 Causation has two levels, general and specific, and a plaintiff must
prove both. “General causation is whether a substance is capable of causing a
particular injury or condition in the general population, while specific causation
is whether a substance caused a particular individual’s injury.” 7 Sequence
matters: a plaintiff must establish general causation before moving to specific
causation. Without the predicate proof of general causation, the tort claim fails.8

2
Some of these losses came from online gambling.
3
GlaxoSmithKline refers to itself as GSK, a convention we adopt.
4
The parties interchangeably call the gambling problem “compulsive” and
“pathological,” but the medically proper term is “pathological.”
5
Both parties agree that Texas substantive law applies in this diversity case.
6
See Ackermann v. Wyeth Pharms., 526 F.3d 203, 208 (5th Cir. 2008) (invoking Texas
law); see also Wright v. Ford Motor Co., 508 F.3d 263, 275 (5th Cir. 2007) (same). “The
Supreme Court of Texas, like many courts, has equated the ‘more-likely-than-not’ causation
requirement to a more than 50% probability that a defendant’s wrongful conduct caused the
harm or injury.” Young v. Mem’l Hermann Hosp. Sys., 573 F.3d 233, 235 (5th Cir. 2009) (per
curiam) (citing Merrell Dow Pharms., Inc. v. Havner, 953 S.W.2d 706, 715–17 (Tex. 1997)).
7
Knight v. Kirby Inland Marine Inc., 482 F.3d 347, 351 (5th Cir. 2007) (citing Havner,
953 S.W.2d at 714) (quotation marks omitted).
8
See id.

3
 Case: 09-50244 Document: 00511058145 Page: 4 Date Filed: 03/22/2010

No. 09-50244

Wells engaged three expert witnesses to address causation, that Requip
can cause pathological gambling. In reaching their conclusions, the experts
relied upon: (1) published articles documenting case-specific correlations
between Requip and gambling; (2) a single unpublished study showing a nexus
between Parkinson’s medicines generally and gambling; (3) GSK’s internal data
revealing case-specific associations between Requip and gambling; and (4) the
fact that GSK has since changed the Requip label to warn about possible
gambling side-effects. Each of the three experts, though, conceded that there
exists no scientifically reliable evidence of a cause-and-effect relationship
between Requip and gambling.
The company moved for summary judgment, arguing: (1) that the expert
opinions did not meet Daubert’s admissibility requirements; or, alternatively, (2)
that the lack of scientifically reliable data supporting general causation
precluded recovery under Texas law. The district court assumed the expert
testimony’s admissibility, but found that the experts did not support Wells’s
claim with scientifically reliable evidence of causation as required by Texas law.
Wells appealed the summary judgment. Our review is de novo, applying
the same standards as the district court.9 We may affirm on any grounds
supported by the record.10

II.
A.
Daubert requires admissible expert testimony to be both reliable and

9
Riverwood Int’l Corp. v. Employers Ins. of Wausau, 420 F.3d 378, 382 (5th Cir. 2005).
10
Berquist v. Wash. Mut. Bank, 500 F.3d 344, 349 (5th Cir. 2007).

4
 Case: 09-50244 Document: 00511058145 Page: 5 Date Filed: 03/22/2010

No. 09-50244

relevant.11 “This entails a preliminary assessment of whether the reasoning or
methodology underlying the testimony is scientifically valid and of whether that
reasoning or methodology properly can be applied to the facts in issue.”12
Although there are “no certainties in science,” the expert must present
conclusions “ground[ed] in the methods and procedures of science.”13 In short,
the expert must “employ[ ] in the courtroom the same level of intellectual rigor
that characterizes the practice of an expert in the relevant field.”14
“[T]he Supreme Court [has] delineated certain factors to assist courts in
evaluating the foundation of a given expert’s testimony, though the Court
carefully emphasized the nonexhaustive nature of the listing.” 15 Suggested
considerations include: “whether the theory or technique the expert employs is
generally accepted; whether the theory has been subjected to peer review and
publication; whether the theory can and has been tested; whether the known or
potential rate of error is acceptable; and whether there are standards controlling
the technique’s operation.” 16 There is no formula, and the court must judge
admissibility based on the particular facts of the case.17

11
Knight, 482 F.3d at 352 (citing Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579,
589 (1993)).
12
Daubert, 509 U.S. at 592–93.
13
Id. at 590.
14
Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999).
15
Knight, 482 F.3d at 351 (citing Daubert, 509 U.S. at 593).
16
Id. (citing Daubert, 509 U.S. at 593).
17
See Kumho Tire Co., 526 U.S. at 150.

5
 Case: 09-50244 Document: 00511058145 Page: 6 Date Filed: 03/22/2010

No. 09-50244

B.
Wells proffered the testimony of three experts: Doctors Fong, Saklad, and
Kalechstein. None did more than baldly state that Requip can cause problem
gambling. At their depositions, the three admitted that no scientific basis
existed to confirm their conclusions.
Fong explained: “I think fundamentally there is something very compelling
and interesting going on here, but we don’t have enough data to suggest
causality . . . . I have not come to a conclusion that there is a causal relationship
between Requip and pathological gambling.” He further testified that – based
on the “scientific standard that [he] appl[ies] in [his] practice as a psychiatrist
and a research scientist” – the scientific literature did not establish a “cause-
and-effect relationship between Requip and pathological gambling.”
Saklad’s deposition went the same way. He was asked: “Applying the
scientific standard for cause and effect, has – has a cause-and-effect relationship
been established between taking Requip and having one or more impulse-control
disorders considered as a collective group?” He answered: “No.”
A bit more hesitant at first, Kalechstein ultimately fell in line with Fong
and Saklad: “I cannot say [Requip] directly causes [pathological gambling].
What I can say is that there is an association between the two variables. . . .
[T]hat’s different than talking about the issue of causation.”
Under the law of this circuit, these admissions drain the expert opinions
of probative force. In Black v. Food Lion, Inc., the plaintiff claimed that she
slipped in the defendant’s store and that the fall caused fibromyalgia
syndrome.18 The district court admitted the plaintiff’s proffered expert, and we

18
171 F.3d 308, 309 (5th Cir. 1999).

6
 Case: 09-50244 Document: 00511058145 Page: 7 Date Filed: 03/22/2010

No. 09-50244

reversed on Daubert grounds for abuse of discretion. We explained that the
expert’s testimony did “not bear the necessary indicia of intellectual rigor.”19
While “association” was well established, “experts have recognized that the
evidence that trauma actually causes fibromyalgia is insufficient to establish
causal relationships.”20 Given the experts’ concessions in this case, Black
controls – and we need not go any further. However, sensitive to the case-
specific nature of the Daubert inquiry, we will closely examine the experts’
methodology.
C.
The experts based their general causation conclusion primarily on the
scientific literature, which they claim shows an association between Requip and
problem gambling. The literature, though, does not provide the necessary
“scientific knowledge” upon which to base an opinion under Daubert.21 Doctor
Fong characterized all but one of the studies as “anecdotal evidence,” and each
expert conceded that the studies were not statistically significant epidemiology.
They were, in fact, case studies. Although, “[c]ase-control studies are not per se
inadmissible evidence on general causation,” 22 this court has frowned on
causative conclusions bereft of statistically significant epidemiological support.23

19
Id. at 312.
20
Id. (quotation marks omitted); see also id. at 313 n.2.
21
See Daubert, 509 U.S. at 589–90.
22
Knight v. Kirby Inland Marine Inc., 482 F.3d 347, 352 (5th Cir. 2007).
23
See, e.g., Burleson v. Tex. Dep’t of Criminal Justice, 393 F.3d 577, 586 (5th Cir. 2004)
(“Here, as in Allen, there are no epidemiological studies supporting a correlation between the
suggested causative agent and the type of cancer experienced by the plaintiff.”); Allen v. Pa.
Eng’g Corp., 102 F.3d 194, 197 (5th Cir. 1996) (“While appellants’ experts acknowledge the

7
 Case: 09-50244 Document: 00511058145 Page: 8 Date Filed: 03/22/2010

No. 09-50244

Only one study – the Weintraub Poster 24 – reached statistical significance.
The Poster suggests that Parkinson’s patients medicated with dopamine
agonists exhibit increased impulsive behavior, including pathological gambling.
But the study has other scientific problems making it insufficient as a basis for
expert opinion. First, “submission to the scrutiny of the scientific community is
a component of ‘good science,’” but the Weintraub Poster was never peer-
reviewed or published.25 Second, the study explains that its results “represented
a class association, as opposed to a specific medication, finding.” In other words,
the Weintraub Poster does not report a “controlled” test for Requip, a drug that
functions differently than other dopamine agonists.26 Finally, its authors
conceded that the very “nature of the study precluded determination of
causality.”
While “[w]e . . . understand that in epidemiology hardly any study is ever
conclusive, and we do not suggest that an expert must back his or her opinion

lack of statistically significant epidemiological evidence, they rely on certain studies as
‘suggestive’ of a link between EtO exposure and brain cancer. ‘Suggestiveness’ is not by the
experts’ own admission statistical significance . . . ; this basis for their scientific opinion must
be rejected.”).
24
D ANIEL WEINTRAUB ET AL ., DOPAM INERG IC THERAPY AND IM PULSE CONTROL
DISORDERS IN PARKINSON ’S DISEASE : A CROSS -SECTIONAL STUD Y OF OVER 3,000 PATIENTS
(2008) (summarizing an epidemiological study presented at a conference).
25
Daubert, 509 U.S. at 593.
26
See Knight, 482 F.3d at 353 (affirming rejection of a “study focused on organic
solvents as a class, including a wide-range of chemicals to which [the plaintiffs] were never
exposed”); see also Burleson, 393 F.3d at 585–86 (affirming rejection of a study showing that
exposure to a substance causes cancer but not showing that the specific type of exposure
suffered by the plaintiff could lead to his particular type of cancer); Allen, 102 F.3d at 197
(affirming rejection of a study showing that exposure to carcinogen caused a different type of
cancer than the one suffered by plaintiff).

8
 Case: 09-50244 Document: 00511058145 Page: 9 Date Filed: 03/22/2010

No. 09-50244

with published studies that unequivocally support his or her conclusions,” 27 here
“there is simply too great an analytical gap between the data and the opinion
proffered.”28 And the bases for the experts’ conclusions pass none of the
applicable Daubert tests: that Requip causes problem gambling is not generally
accepted, has not been subjected to peer review and publication,29 and is not
backed by studies meeting requisite scientific standards.30 Without the expert
testimony, Wells cannot prove general causation – and judgment must be
entered for GSK.

III.
Our conclusion that the trial court did not abuse its discretion is an
unremarkable sustaining of the district court’s gatekeeping role under Daubert.31

27
Knight, 482 F.3d at 354.
28
Gen. Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997).
29
One of Wells’s experts, Dr. Saklad, explained in his deposition that “causation
requires scientific literature that proves a statistically valid reproducible result which could
be used to establish causation. Such studies have not been done.”
30
It bears mentioning that, in addition to the literature, Wells’s experts purport to rely
on GSK’s internal documentation in reaching the conclusion that Requip causes problem
gambling. Specifically, GSK has, over the years and per the FDA’s requirements, collected
data on patients suffering increased gambling when taking Requip. This data shows a
relatively high number of self-reported spikes, but mining this data is not the scientific
method; rather, it is rife with bias and variability. Wells also points out that GSK has now
changed its Requip label to warn about possible gambling problems. This precaution, though,
did not grow out of scientific knowledge of causation; rather, GSK changed the label because
there might be an association. See 21 C.F.R. § 201.80(e) (“The labeling shall be revised to
include a warning as soon as there is reasonable evidence of an association of a serious hazard
with a drug; a causal relationship need not have been proved.”).
31
See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141 (1999) (describing Daubert as
“setting forth the trial judge’s general ‘gatekeeping’ obligation”).

9
 Case: 09-50244 Document: 00511058145 Page: 10 Date Filed: 03/22/2010

No. 09-50244

In finding the evidence scientifically unreliable – and thus insufficient to prove
causation under Texas law – it follows that the experts’ testimony was also
deficient under Daubert given its overlap with Texas questions of scientific
sufficiency.32
Perhaps Requip is a cause of problem gambling, but the scientific
knowledge is not yet there. Wells urges the law to lead science – a sequence not
countenanced by Daubert.33 And while the possibilities of their relationship
properly spark concerns sufficient to warrant caution, the courts must await its
result. AFFIRMED.

32
The Supreme Court of Texas has explained that “the same factors may be applied”
to “no evidence review of scientific evidence” as to review of admissibility of scientific
testimony. Merrell Dow Pharms., Inc. v. Havner, 953 S.W.2d 706, 714 (Tex. 1997) (referencing
E.I. du Pont de Nemours & Co. v. Robinson, 923 S.W.2d 549, 556–57 (Tex. 1995) (adopting the
federal Daubert standard in the Texas-law inquiry into the admissibility of expert testimony));
id. at 712 (explaining that scientific sufficiency in Texas “is determined by looking at
numerous factors including those set forth in . . . Daubert”).
33
See Rosen v. Ciba-Geigy Corp., 78 F.3d 316, 319 (7th Cir. 1996) (“Law lags science;
it does not lead it.”).

10