A while ago we reported on what we saw as a violation of Erie v. Tompkins principles in Rimbert v. Eli Lilly, where a federal district court, ignoring three state intermediate court cases, predicted that New Mexico would expand tort liability by not following the learned intermediary rule.
Well, that judge recused himself for unstated reasons, and another judge got the case. Slip. op. at 6. What a difference a change in judges can make.
Well, the learned intermediary rule issue is probably moot now – although we’d love to see it asserted as an alternative ground for affirmance on appeal – since last week the case (a Prozac suicide claim) was thrown out on Daubert grounds. This is the case about the guy (who had all kinds of problems: served with divorce papers, relationships with at least one child, compulsive gambling and associated financial problems, obesity, diabetes, impotence, two packs per day smoking, financial woes), who not only shot himself and his wife of 42 years – but also shot his dog. All these facts are on pages 3-4 of the opinion.
His surviving children blamed Prozac, which the decedent had been taking for 7 to 8 weeks for a diagnosed case of “moderate depression.” They hired Grace E. Jackson, M.D., as their causation expert. She offered this opinion:
I believe it would be incorrect to suggest that Prozac was the necessary and sufficient cause of the deaths of [the decedents], and their pet dog (Ivy). However, in the context of the [murderer’s] pre-existing risk factors for suicide and homicide, and in the context of his diminished capacity to resist them, it is more likely than not true that Prozac converted a case of probable dysthymia into a case of agitated depression with obsessive and psychotic features. In this sense, Prozac was the definitive, contributive cause of the Rimbert tragedy.
Slip op. at 5 (emphasis original). And, since plaintiffs’ experts always seem to submit last-minute “amendments” to their reports in Daubert situations (whether they’re permitted to or not), Dr. Jackson gave classic “straw that broke the camel’s back” (without using that phrase) testimony, calling Prozac “the linchpin” and “the clincher” (both rather non-medical terms) “against a complex backdrop of personal and familiar circumstances.” Slip op. at 11.
These conclusions were quite questionable. Dr. Jackson basically made up risk factors for increased retention of Prozac – blaming a genetic predisposition she couldn’t say the deceased murderer had, that the murderer might have been taking other synergistic drugs that there was no evidence he took, and that other medications the murderer was taking might have had some sort of synergistic effect that was unknown to science, and that the murderer’s pre-existing medical conditions might have somehow increased the risk – although how something like diabetes could encourage violence was never really explained. Slip op. at 15-16.
And all this was doubly speculative anyway, since post-mortem blood tests didn’t show any increased levels of Prozac or its metabolites in the deceased murderer’s blood in the first place. Slip op. at 16. Of course, Dr. Jackson had speculative (“could have”) excuses for this too. Id.
That Dr. Jackson would do this isn’t very surprising because she “worked exclusively for Plaintiff’s counsel’s law firm” and (probably as a consequence) came to her conclusions “prior to doing any research or even knowing the facts of the case.” Slip op. at 18. The latter was demonstrated by an email exchange showing that, on no evidence at all, she told her employer that it was an “excellent case.” Slip op. at 18-19.
That’s a pretty good racket – hire your own experts as “full-time employee[s].” Slip op. at 18. Too bad that wasn’t before the Supreme Court in Levine. A showing of what really goes on in drug litigation may have made some of the justices more inclined to look favorably on preemption. But we digress.
The court chose to consider Dr. Jackson’s employee status and her backward reasoning as “credibility issues” for the jury. Slip op. at 19. There’s even more interesting credibility material if you’ve got Dr. Jackson in your case, see slip op. at 18-19 n.13, but we want to get to the meat of the opinion.
Dr. Jackson got booted for her methodological flaws with respect to general causation. The specific grounds:
(1) The studies Dr. Jackson relied upon had no controls, making it impossible to factor out Prozac as compared to the natural course of depression. That’s important because suicide is a huge risk of depression itself – one reason people take drugs like Prozac in the first place. Slip op. at 20-21.
(2) Animal studies don’t prove anything about humans – especially since animals don’t commit suicide (our comment). Some of the studies didn’t even involve Prozac, but other drugs – something Dr. Jackson misrepresented in her report. Others had results contrary to the point for which Dr. Jackson relied on them. The dosages were up to 100 times higher per weight in the studies than the dosage the murderer took in the case. Imagine a patient eating Prozac by the plateful. And in no case was there any evidence that the animals studied were comparable to people – especially important since we’re talking about brain-related effects here. The animal studies produced nothing more than an “untested hypothesis.” Slip op. at 22-25.
(3) Epidemiology – ain’t none in Dr. Jackson’s report. Slip op. at 26. That’s a problem, because there’s a whole bunch of (“myriad”) epidemiological evidence that Prozac doesn’t cause homocide, suicide, etc. Dr. Jackson (faithful employee that she was) ignored all of it. Slip op. at 26-27. And it gets worse – the FDA concluded that Prozac decreases suicide in persons over 65 (the murderer was 68). Slip op. at 27 & n.18. “As the best evidence of general causation, [epidemiology] must be addressed. Slip op. at 28. “To ignore completely [a] vast body of contradictory evidence. . .strikes at the heart of a proffered expert’s reliability. Slip op. at 28-29.
In other words, go away, Dr. Jackson.
(4) Cooking the books (“departure from standard methodology”) – Dr. Jackson, although purporting to embrace the Bradford Hill criteria (pp. 375-78), she “inexplicably did not apply” them. Slip op. at 29.
(5) “Chain-of-events causation.” That’s a fancy disguise for the old post hoc ergo propter hoc (“after this, therefore because of this”) fallacy. This breaks down events that follow one another into smaller units of alleged cause and effect. Anyway, “this type of indirect, chain-of-events causation [i]s not a generally accepted scientific methodology.” Slip op. at 31. That’s especially since Dr. Jackson “never attempted to publish the methodology she employed. . .in any peer reviewed journal.” Slip op. at 32. It was just something ginned up for litigation. Id. We’ have to add that this is hardly surprising, since Dr. Jackson was employed full-time by plaintiff’s counsel.
The court goes on to criticize the plaintiff’s response to the motion. That response:
- “[R]idiculed Defendant for” bringing the motion in the first place. Slip op. at 33.
- Relied on “other experts using other methodologies in other cases.” Id.
- Cited to “studies not relied upon by Dr. Jackson” after she had testified at her deposition that her report contained everything she relied upon. Slip op. at 33-34.
- “[C]hose not to call Dr. Jackson to testify” at the Daubert hearing. Slip op. at 33.
Dr. Jackson also got booted on specific causation, although the court spends a lot less time on this point. What analysis did Dr. Jackson claim she used to establish specific causation? No surprise there.
With apologies to Samuel Johnson, the real last refuge of a scoundrel is “differential diagnosis.” The court held:
[N]ot only did [Dr. Jackson] fail to provide objective reasons for eliminating this alternative explanation [the murderer’s “depression and myriad life stressors”] as highly improbable, but she also failed to demonstrate that the cause she identified (Prozac) was highly probable. Ultimately, almost everything in Dr. Jackson’s specific causation opinion is hypothetical and speculative, except for her conclusion.
Slip op. at 37. That was our impression at the beginning of this post when we first encountered these supposed explanations. See Slip op. at 15-16.
Well, the court thought so, too. Slip op. at 37-38. To that the court added Dr. Jackson’s “apparent failure to take into consideration seemingly relevant facts” and her failure “to explain the basis for her refusal to consider them.” Slip op. at 38. Those include things like, well, already having the supposed symptoms before ever taking the drug. Slip op. at 38-39 (discussing “restlessness”).
“A methodology that inexplicably ignores material facts and relies only on selective evidence does not lead to a reliable opinion.” Slip op. at 40.
That’s where it ends. We think the court was kind in limiting its order to excluding Dr. Jackson’s opinions.