Photo of Eric Alexander

A week ago, in a post-script to a post on Daubert decisions, we reported that the trial court in Hexum v. Eli Lilly & Co., No. 2:13-cv-02701-SVM-MAN, 2015 U.S. Dist. LEXIS 109737 (C.D. Cal. Aug. 18, 2015), had granted directed verdict at the end of plaintiffs’ case.  Our readers may have noticed that we (in both the singularly singular and common plural sense of the word) have a thing for proximate cause for failure to warn, which was why the defendant won in Hexum.  So, we decided to somewhat deeper dive on why the defendant drug manufacturer won and whether there are lessons to be learned.

We start with two caveats and one somewhat rhetorical question.  First, plaintiffs rested their case on the second day of trial.  How much evidence they could have presented was no doubt limited by the Daubert rulings and other evidentiary rulings, but they seem to have taken a pretty narrow view of the evidence they needed to present for their four remaining claims—negligent failure to warn, strict liability failure to warn, negligent misrepresentation, and fraud—all grounded in failure to warn the learned intermediary.  For instance, they chose not to present evidence from the physician who directed plaintiff (the one on the drug) to taper down and stop the drug.  Second, from the testimony presented in the order—which we assume is the most relevant—it does not look like plaintiffs pushed hard to get the “magic words” that most plaintiffs try to get from prescribing physicians.  The result was testimony that was more “fuzzy” on some key issues than we normally see in prescription drug product liability cases, even though this testimony was from the trial.  Assuming plaintiffs did not do any better with the prescribing physician in deposition than they did at trial—and one would expect they would have impeached him with inconsistent deposition testimony—then why did this case get past summary judgment?

But it did, so on to the directed verdict order.  The basic allegation was that the label was inadequate in how it warned about the risk of various (mostly transient central nervous system) symptoms upon discontinuation of the drug and the need to taper the drug down.  The plaintiff who took the drug—her husband also sued—experienced some of these a year after starting the drug when she moved away and talked to a new doctor about stopping the drug so that she could get pregnant.  The symptoms resolved without lasting effect, she had a successful pregnancy, and she managed the condition for which she had been prescribed the drug by switching to non-pharmaceutical therapies.  Id. at **13-15.  We suppose that we should mention that the drug was a selective serotonin and norepinephrine reuptake inhibitor and she was taking it for fibromyalgia, which was not an approved indication (and she knew it).  It apparently worked while she took it, so we have a hard time seeing why a suit was brought at all.  Plaintiffs also had no real theory that we can discern for what was allegedly lacking from the fairly thorough discussion in the label about discontinuation symptoms—at least after their expert case was gutted.  If the plaintiffs cannot articulate through admissible evidence some reason why the label’s actual warnings in the label about the very risk at issue were inadequate, then they should lose before they get to trial. Such a shortcoming also makes the proximate cause analysis unduly abstract:  the question should be whether the “stronger warning” urged by plaintiffs would have changed the prescribing physician’s behavior and avoided the injury not whether an unspecified “stronger warning” would have done the same.

While the lack of evidence of what the warnings should have been instead of what they were should have been another ground for directed verdict (or summary judgment), there is one basic fact pattern that gets to a defense win without even considering what the label did say or allegedly should have said—when the prescribing physician did not read the label.  That is what happened here and is the focus of the order.  We have posted many times about how warnings claims should fail unless the prescribing physician actually read the label that plaintiff contends should have provided additional warnings about whatever harm eventually (and allegedly) befell her.  What Hexum did was cut through less than clear testimony and hold plaintiffs to their burden.  The prescribing physician here did not recall reading the label before prescribing the drug to plaintiff, did not typically read labels “cover to cover,” and only recalled reviewing a third-party publication called Up To Date before prescribing the drug to plaintiff.  Id. at **8-9.  He acknowledged that labels came with the drug samples he received and that he would have reviewed the label if a sales representative asked him to do so, but plaintiffs did not close the loop on whether that happened with this drug before plaintiff’s prescription.  Id. at **9-11.  Some courts might have determined that all of this created a jury issue, but this court saw that “there is no proof beyond speculation that Wollaston read Cymbalta’s label before prescribing it to Hexum.”  Id. at *24.  Nor could it be inferred from the testimony about the doctor and patient’s discussion of discontinuation risk—yes, the patient admits that she was also warned, but more on that later—that the doctor had read the drug’s label before prescribing it to plaintiff.  Id. at **24-25.  Under the applicable and pretty good California law, this was a clear break in the causal chain for every claim that plaintiffs asserted.

In so holding, the court distinguished cases using a heeding presumption—which California does not have—to allow failure to warn claims to survive even when the prescriber did not read the label.  Id. at *21 n.4.  In our compilation on this issue, we do discuss the role heeding presumptions have played in cases from around the country.  To us, in addition to the heeding presumption being inapplicable to all prescription drug cases except those with warnings that remove all judgment by urging something like “don’t use the drug in someone exactly like this plaintiff,” it does not make sense that a heeding presumption would link up the broken causal chain from a doctor not reading the label.  You can’t heed what you don’t read.  Maybe this really requires a reading presumption, but that presumption would be rebutted by the doctor’s actual practice of having not read the label.

After concluding that plaintiffs lost because they did not establish that the prescribing physician read the drug’s label, the court considered plaintiff’s failure of proof on two possible alternative methods to prove causation that we do not think would have worked under the law anyway.  Plaintiffs often try to backdoor proximate cause by suggesting that a doctor who would have prescribed the drug anyway would have discussed the risk with the plaintiff (in some unspecified way compared to the actual discussion of the risk that occurred) and the plaintiff, blessed with the power of hindsight and a secondary profit motive, would have declined to take the drug (or been a more responsible patient in avoiding or minimizing the risk).  The plaintiff here, who admitted she had actually read the label’s discussion of discontinuation risk before taking the drug, testified that “I can’t guess and speculate what I would have done” if told the risk of discontinuation symptoms were higher than 1-2%—a purported contrast to the standard discussion in the label of clinical trial events occurring at 1% or greater and at a significantly higher rate than in placebo patients.  Id. at *29.  She also offered a lukewarm “I imagine not” when posed with the leading and admittedly speculative question about what she would have done if her doctor had told her the risks were “more than 44 to 78 percent.”  Id. at **29-30.  On top of the lack of predicate evidence about what her doctor would have told her (and maybe if anybody ever said the label should have included such percentages), this was “too ambiguous and speculative to allow a jury to find that if given stronger warnings Hexum would never have taken Cymbalta.”  Id. at *30.  This is all pretty far afield when the prescriber did not read the label, but we have a hard time seeing this sort of chain ever being sufficient to satisfy proximate cause.  At a minimum, there would need to be clear testimony from the prescribing physician on how the discussion with the plaintiff would have differed if presented with the specific alternative label and the plaintiff would have to answer questions tied to that testimony, not to generalities about a “stronger warning.”  The court also slammed plaintiffs for failing to present testimony from “the tapering physician” as a way to make a failure to warn claim.  Id. at **32-33.  But the prescription drug manufacturer does not owe a duty to warn every physician in any specialty who may later treat a patient on its drug.  It owes a duty to the physicians who prescribe its drug.  A breach of a non-existent duty cannot create liability.

Like we said, we have a thing when it comes to proximate cause for failure to warn.  Hexum was a nice win and may help some other court reject the sort of cobbled together speculation that some plaintiffs offer as proof of proximate cause for failure to warn with prescription medical products.  We would take that.