“I’m not a doctor, but I play one on TV” – shorthand for “I am pretending to know what I’m talking about, but I really don’t.” We have blogged frequently about experts inclined to such pretense and about courts’ varying levels of tolerance for experts who lack expertise or who render opinions lacking foundations and reached without reliable methodologies. Recently, in Hexum v. Eli Lilly & Co., 2015 U.S. Dist. LEXIS 106153 (C.D. Cal. Aug. 10, 2015) – a Cymbalta “discontinuation case” — the United States District Court for the Central District of California did a noteworthy job of precluding two experts from testifying about opinions they had no business rendering.
In Hexum, plaintiff claimed that she suffered severe “discontinuation” symptoms after ceasing to take Cymbalta, an SSRI manufactured by defendant Lilly, because Lilly’s Cymbalta label provided her treating physician with inadequate warnings about such adverse events. Before trial, Lilly moved to exclude the testimony of two of plaintiff’s experts: 1) a psychiatrist, rendering opinions, inter alia, about the adequacy of Cymbalta’s label; and 2) a “social psychologist,” offering opinions about how “a reader” would interpret various portions of the label.
Initially, Lilly asserted that the psychiatrist’s testimony should be excluded in its entirely because it lacked a reliable methodology but was reached, instead, by “cherry-pick[ing] through documents and data, periodically providing supplemental reports explaining how the “newly” discovered data supports his previously expressed opinions.” Hexum, 2015 U.S. Dist. LEXIS 106153 at *6. The Court held, “[W]hile [the psychiatrist’s] pattern of filing a report and later opining in supplements that the data supports his prior conclusions raises significant questions regarding [his] credibility, it does not independently render his opinions inadmissible.” Id. at *6-7. Declining to preclude the psychiatrist’s testimony in its entirety, the Court turned to his individual opinions.
In his report, the psychiatrist opined “that prescribing physicians rely on the accuracy of a drug’s label to educate themselves about a medication,” “rarely” performing their own studies or conducting their own research. Id. at *8. While he acknowledged “that doctors have access to other sources of information,” he stated that “treating physicians rely on the label as the ultimate authority on a drug’s safety and efficacy.” Id. at *8-9. He also opined “that Lilly obscured the results of [published clinical trials” by publishing their results in a journal published in the Netherlands and “not widely read in the United States.” Id. at *9. For this reason, the psychiatrist opined “that many American doctors would have been aware of the label but not of [the journal article]. Id. Finally, the doctor opined that Cymbalta’s label was misleading in a number of ways, including its use of the term “discontinuation symptoms” instead of “Cymbalta Withdrawal,” and several statements about adverse events associated with discontinuation of Cymbalta. Id. at *9-10.
The court held that the psychiatrist was passing off personal opinion as expert testimony. He “fail[ed] to cite any relevant authority supporting his opinions regarding physicians’ general prescribing practices, physicians’ reliance on the label, whether they would have been familiar with the Netherlands journal article, and whether Lilly intentionally obscured its clinical trials through its choice of journal. Thus, the court excluded the psychiatrist’s opinions on all of these subjects. Id. at *10-11.
The court did permit the psychiatrist “to opine that Cymbalta can cause discontinuation symptoms and to testify regarding the nature of those symptoms,” holding that the psychiatrist’s “expertise in clinical psychiatry, his experience prescribing antidepressants, and his publications about antidepressant withdrawal” qualified him to render these opinions He was also permitted to testify “regarding clinical data he examined,” and was permitted to opine that it was “his personal opinion as a clinical psychiatrist that Cymbalta’s label [did] not adequately reflect the clinical data regarding Cymbalta, but was not permitted “to speculate regarding what would have happened in a given study” if a different methodology were used. Id. at *11-12.
The court also held that the psychiatrist’s opinions regarding the severity of discontinuation symptoms were inadmissible because he “invent[ed] his own definitions and call[ed] the Cymbalta label inadequate on that basis. The only authority he cite[d] for these definitions [was] his own book,” generally cited without page numbers. Id. at *12.
Finally, the court considered the psychiatrist’s opinions about plaintiff’s “need for medical treatment for fibromyalgia and her potential response to” treatments other than Cymbalta. Id. at *14. The court held that the psychiatrist, who was not a rheumatologist, had never published anything on rheumatology, and had never treated fibromyalgia, was not qualified to render any of these opinions.
The Social Psychologist
This expert – a Ph. D. social psychologist – rendered various opinions about how “a reader” would interpret Cymbalta’s label, including the opinion that “a reader would interpret the phrase “the following symptoms occurred” to mean that the symptoms occurred in the aggregate, “not that each of [them] occurred.” The expert also opined that “Lilly could have altered the Cymbalta label to more accurately reflect the data” in a key study and that “neither the FDA guidance on “frequency” nor “class labeling” would have stood in the way.” Id. at *19. Lilly argued that the expert arrived at his conclusions without using a reliable methodology, and the court resoundingly agreed, notwithstanding the expert’s assertion that he “use[d] the same method that the FDA uses to evaluate promotional materials and product labels.” Id. The court held, “. . . Even assuming arguendo that [the expert’s] assertions regarding the FDA’s procedures are correct, Plaintiffs still fail to show that [his] conclusions are the result of reliable methods reliably applied. The FDA relies on analysis of this type performed by experts with sufficient training and experience,” while the expert’s training and experience were anything but sufficient. Id. at *21 (emphasis in original).
Similarly, the expert’s opinion was “inadmissible to the extent that he opine[d] regarding whether Cymbalta’s label [was] inadequate as a matter of semantic and linguistic interpretation. [He] is not offered as an expert in the English language or grammar. . . . Instead, he relies on his own examination of the label’s “precise language” and “how the information was communicated” – including the “broader context of which [sic] influenced the meaning conveyed.”
Moreover, the court noted that the expert opined “simply regarding how a generic “reader” would interpret” certain label statements [and did] not purport to opine specifically on how physicians would interpret” them. The court held, “To the extent that [the expert] purports to opine on how any reader would understand Cymbalta’s label, his opinion does not rely on any particular expertise and is thus unhelpful to the trier of fact. . . . Moreover, the relevant inquiry is how a prescribing doctor would understand the label – [the expert] is not a doctor and has not studied doctors’ interpretations of Cymbalta’s label. He is thus unqualified to render this opinion based on his own experience; it is also not the result of reliable methods.” Id. at *23-24.
And so, relying on the most basic of Daubert principles – an expert must be qualified, must arrive at his or her opinions through a reliable methodology, and can’t get his personal opinions admitted just because he insists they are correct – the court got it right. At the risk of sounding like a broken record (does that simile even exist anymore?), we commend this opinion to other courts, particularly those considering the admissibility of the opinions of a certain ubiquitous “regulatory” expert.
POSTSCRIPT: Eight days after the Court issued its opinion on Lilly’s expert motions, it granted a directed verdict for Lilly at the close of the plaintiff’s case at trial. See Hexum v. Eli Lilly and Company, 2015 U.S. Dist. LEXIS 109737 (C.D. Cal. Aug. 18, 2015). The Court held, “On the record, [plaintiff] fails to offer sufficient proof that [her doctor] read Cymbalta’s label before prescribing Cymbalta to [her]. . . . [Thus] Plaintiffs failed to carry their burden of proving that stronger warnings would have altered [the doctor’s] conduct.” Hexum, 2015 U.S. Dist. LEXIS 109737 at *7-8. Another so-basic-but-so-often-ignored principle – no warnings causation if the doctor didn’t read the label – and the Central District of California got it right again.