Learning the wrong lessons from "an American tragedy": a critique of the Berger-Twerski informed choice proposal.

• Author: Bernstein, David E.
• Position: Response to Margaret Berger and Aaron Twerski, Michigan Law Review, vol. 104, p. 257, 2005

TABLE OF CONTENTS I. INFORMED CHOICE AND THE BENDECTIN TRAGEDY II. A BROADER CRITIQUE OF THE INFORMED CHOICE PROPOSAL A. The Proposal Invites Reliance on Unreliable Testimony B. Juries Are Not Competent to Determine Subtle Risk Assessment Issues C. Even Assuming Juror Competence, the Proposal Asks Too Much of Juries D. The Proposal Ignores the Problems Inherent to Multiple Trials E. The Proposal Fails to Consider the Potential Costs of Informed Choice Litigation F. The Informed Choice Proposal Would Lead to a Vast Surfeit of Warnings G. The Informed Choice Proposal May Be Barred by the Preemption Doctrine CONCLUSION Margaret Berger and Aaron Twerski are among the leading scholars in their respective fields of Evidence and Products Liability. I have benefited from their work on many occasions. (1) Precisely because of the deserved respect and esteem in which Berger and Twerski are held--not to mention the prominence of their forum, the Michigan Law Review--their proposal to create a new "informed choice" cause of action in pharmaceutical litigation is likely to receive sympathetic attention. Because I believe that their proposal is ill-conceived and dangerous, I feel compelled (with some trepidation) to write this response.

Berger and Twerski propose that courts recognize an informed choice cause of action that would allow plaintiffs claiming injury from pharmaceutical products to recover damages for deprivation of informed choice when (1) the causal relationship between the toxic agent and plaintiff's harm is unresolved at the time of litigation and will likely remain unresolved; (2) the drug is not therapeutic but rather its purpose is to avoid discomfort or to improve lifestyle; (3) it is almost certain that a patient made aware of the risk that is alleged to be associated with consumption of the drug would have refused to take it; and (4) defendant drug company was aware of the potential risk or should have undertaken reasonable testing to discover the risk and failed to provide the requisite information to the physician or patient. (2)

These guidelines, however, are rather vague. Whether they are meant to apply broadly or narrowly means the difference between a cause of action that would open a Pandora's Box of litigation and one that would be available only in limited, perhaps even extraordinary, circumstances. Apparently, Berger and Twerski intend the scope of the informed choice action to be broad indeed. So broad, in fact, that if adopted it could lead to an unprecedented wave of litigation against pharmaceutical manufacturers, including lawsuits involving products that are completely safe and effective.

Berger and Twerski suggest that the paradigmatic example illustrating the need for the informed choice cause of action is the failure of plaintiffs to recover damages from the maker of Bendectin. The plaintiffs contended that this morning sickness drug caused their children's birth defects. As demonstrated below, in Part I of this Essay, if the proposed informed choice tort's boundaries are broad enough to allow the Bendectin plaintiffs to recover damages, then they are extraordinarily, dangerously broad. Part II of this Essay argues that even if Berger and Twerski had chosen a better example that would allow for a much more limited interpretation of the scope of their proposal, the proposal still has significant weaknesses that render it a very bad idea.

1. INFORMED CHOICE AND THE BENDECTIN TRAGEDY

   Litigation claiming that Bendectin caused limb reduction and other birth defects began in the late 1970s and did not end until at least 2000. It involved thousands of plaintiffs and tens of millions of dollars in defense costs, and led to many pioneering judicial rulings excluding plaintiffs' scientific evidence. Most significant, Bendectin was the underlying subject of Daubert v. Merrell Dow Pharmaceuticals, (3) the Supreme Court case that ushered in the modern era in which courts subject questioned expert testimony to meaningful scrutiny to ensure its reliability.

   Perhaps because Berger and Twerski seek to "unmask" Daubert, they invoke the Bendectin litigation to justify their informed choice proposal. They suggest that although the Bendectin plaintiffs could not prove causation, the Bendectin plaintiffs could have met the criteria they lay out for an informed choice cause of action. If so, a review of the history of the Bendectin litigation reveals that their proposal is unjust, unworkable, and counterproductive.

   Criterion 1: The causal relationship between the toxic agent and plaintiff's harm is unresolved at the time of litigation and will likely remain unresolved.

   Neither pioneering Bendectin plaintiff Betty Mekdeci--whose "anguished cry" Berger and Twerski say they are responding to (4)--nor any of the subsequent Bendectin plaintiffs ever had sound reason to believe that Bendectin caused limb reduction birth defects, the main focus of the Bendectin litigation. In 1977, when Mekdeci brought her lawsuit, fourteen epidemiological studies of varying strength and quality had examined the relationship between Bendectin and birth defects and found no association. (5) While these studies were not powerful enough to rule out some connection between Bendectin and birth defects, they certainly provided no cause for alarm. Bendectin had been on the market since 1956 with no serious doubts raised regarding its safety in the scientific or medical community. Nor did Bendectin contain suspiciously toxic ingredients: one active ingredient of Bendectin was a simple B vitamin, and the other was an ingredient used in a popular over-the-counter sleeping pill.

   Meanwhile, Mekdeci's evidence that Bendectin did cause birth defects was "remarkably thin." (6) Many chemicals are known not to be teratogens in humans, so the mere fact that pregnant women ingested a pharmaceutical product such as Bendectin did not mean there was an inherent risk. Beyond the mere fact that she ingested Bendectin during pregnancy and later gave birth to a child with a limb reduction birth defect, Mekdeci's evidence of causation consisted primarily of eighty-six reports to the FDA of other women who had also given birth to children with limb reduction defects after taking Bendectin. (7) These reports are the direct source of Mekdeci's complaint, implicitly endorsed by Berger and Twerski, that Bendectin's manufacturer should have warned of a possible association with birth defects. (8)

   Berger and Twerski acknowledge that "[t]he mere fact that a child was born with a limb reduction to a mother who had ingested Bendectin did not necessarily point to Bendectin as the cause of the birth defect." (9) In fact, the mere fact that dozens or even hundreds of children were reported to have been born with limb reductions after their mothers ingested Bendectin doesn't, by itself, even suggest a risk. Approximately thirty million women took Bendectin, and by chance alone there would be ten thousand limb reduction defects among children born to these women. (10)

   Berger and Twerski apparently see the issue of whether Bendectin caused birth defects as "unresolved" at the time of litigation. As noted above, when the Bendectin litigation began, the relevant research was not strong enough to rule out the possibility that Bendectin caused a small increase in birth defects, but there was no reason to rule in that possibility either. There was never any valid scientific evidence supporting the proposition that Bendectin was a teratogen.

   As interest in the teratogenicity of Bendectin increased due to the litigation, evidence quickly piled up that Bendectin was safe No animal studies using doses equivalent or even substantially above human therapeutic doses showed teratogemclty. (11) Most epidemiological studies produced no statistically significant findings. (12) The few positive studies (13) each found an association with a different, unrelated birth defect, a pattern consistent with random chance or imperfections in the studies, but not with causation by Bendectin. (14) Meanwhile, other studies reported a negative association between Bendectin and specific birth defects. (15) Moreover, the results of specific studies showing an association between Bendectin and various unrelated birth defects were invariably not replicable. (16) By the early 1980s, there was a solid consensus in the medical community that Bendectin was not a teratogen. Nevertheless, the litigation continued.

   Berger and Twerski state that the manufacturer withdrew Bendectin from the market "due to widespread fears that it caused severe birth defects in the children whose mothers ingested the drug while pregnant." (17) As with other phantom risks, (18) however, the fears in question were the unreasonable fears of the lay public--stirred by irresponsible interest groups, (19) hired gun and delusional experts, credulous media coverage, (20) and plaintiffs' lawyers (21)--not the fears of the manufacturer, the FDA, or the scientific community. (22)

   Over time Bendectin became the most-studied drug used during pregnancy, and "the massive amount of data does not support a consistent statistical association between Bendectin usage in pregnancy and a particular syndrome or group of malformations." (23) Two meta-analyses of the data from all the epidemiological studies showed no association between Bendectin and birth defects. (24) The negative epidemiological data are supported by "ecological analyses" showing that the withdrawal of Bendectin from the U.S. market did not lead to a reduction in any category of birth defects. (25) A 2003 study concluded that the fact that the rate of birth defects remained constant after Merrell Dow withdrew Bendectin from the market is not consistent with the hypothesis that Bendectin is a teratogen. (26)

   A review of the relevant medical literature finds a consensus that Bendectin is not a teratogen. (27) Prominent teratologist Robert Brent concluded in 1995 that...