TORTINI

For your delectation and delight, desultory dicta on the law of delicts.

Belgian Waffles and Post Hoc Ergo Propter Hoc

June 22nd, 2017

In language that could well be a Sokol hoax on the vacuousness of post-modernist non-thinking, the Court of Justice of the European Union issued a press release to announce its judgment in N.W. v. Sanofi Pasteur MSD, Case C-621/15 (Luxembourg, 21 June 2017). With European hypersensitivity to public disclosure, and in recognition of the right to be forgotten, the plaintiffs are known only as the “W” family. Mr. J. W. received Sanofi’s hepatitis B vaccine between late 1998 and the middle of 1999. In the summer of 1999, Mr. W. began to experience symptoms, which led to a diagnosis of multiple sclerosis over a year later, in November 2000. J.W. and his family sued Sanofi Pasteur in 2006. J.W. died in 2012.

The Ws filed their case in Paris, where the courts found for Sanofi Pasteur. The Cour d’Appel de Paris, impressed by the lack of scientific consensus to support W’s causal claim, held that the plaintiffs had failed to demonstrate causality. The Cour d’Appel de Paris court dismissed the W’s case. The dismissal was remarkable in the context of credulous French courts that had routinely recognized such claims. See C. Rougé-Maillart, N. Guillaume, N. Jousset, and M, Penneau, “Recognition by French courts of compensation for post-vaccination multiple sclerosis: the consequences with regard to expert practice,”47 Med. Sci. Law 185 (2007) (summarizing the course of M.S. and hepatitis B vaccination litigation in France). The Ws appealed further to the French Cour de Cassation (Court of Cassation), which punted to the EU Court of Justice, to ask whether the EU directive on liability for defective products required a different result than handed down by the Cour d’Appel.

The “Court of Justice” stated that the claimants had the burden of proof, but dubiously framed the causation issue as a choice between “certain and irrefutable evidence” and evidence of a “sufficiently high degree of probability.” What gives rise to sufficiently high degree of probability? In bold type, the EU press release announces that:

Where there is a lack of scientific consensus, the proof of the defect of the vaccine and of a causal link between the defect and the damage suffered may be made out by serious, specific and consistent evidence.”

But what is “serious, specific, and consistent evidence”? Scouring both the press release and the full decision of the Court provides one answer:

Anecdotes. Unspecified number of case reports of multiple sclerosis occurring in patients after vaccination, without regard to an anticipated or expected incidence of the disease in the vaccinated population suffices. The Court of Justice sums up the Ws’ case as a showing that:

The temporal proximity between the administering of a vaccine and the occurrence of a disease, the lack of personal and familial history of the person vaccinated and the existence of a significant number of reported cases of the disease occurring following such vaccines being administered may, where applicable, constitute sufficient evidence to make out such proof.”

The significant number of cases is never quantified or even described. The “sufficiently high probability” is never quantified or described. Presumably, the judges in Brussels can count, and the probability is some number greater than zero, but less than 1. The Court of Justice follows many other lay courts in falsely dichotomizing scientific disputes as involving either “certain, irrefutable” evidence or something less, and good enough for government (judicial) work. Even its representation of the Ws’ evidence as “serious, specific and consistent” and its suggestion of a “sufficiently high” probability are false, at least without spelling out the evidentiary base of the anecdotal evidence that is elevated to legally sufficient in the eyes of the Court of Justice. The Court fails to acknowledge, as a Court of Justice should, that large, high quality epidemiologic studies fail to find associations between hepatitis B vaccination and multiple sclerosis. See, e.g., Annette Langer-Gould, Lei Qian, Sara Y. Tartof, PhD; Sonu M. Brara, Steve J. Jacobsen, Brandon E. Beaber, Lina S. Sy, Chun Chao, Rulin Hechter, Hung Fu Tseng, “Vaccines and the Risk of Multiple Sclerosis and Other Central Nervous System Demyelinating Diseases,” 71 J. Am. Med. Ass’n Neurol. 1506 (2014); Miguel A. Hernán & Susan S. Jick, “Hepatitis B vaccination and multiple sclerosis: the jury is still out,” 15 Pharmacoepidemiology & Drug Safety 653 (2006).

The European right to be forgotten has apparently been extended to scientific evidence. There may be reasons more legitimate than racist xenophobia to exit the European Union.

The Slemp Case, Part 2 – Openings

June 10th, 2017

The Slemp Case, Part I – Jury Verdict for Plaintiff – 10 Initial Observations” (May 13, 2017)

The legal community is still trying to grasp the enormity of the $110M verdict against Johnson & Johnson, in the Slemp case. J & J says it will appeal, but resolution of legal issues will not necessarily clarify what happened factually in the Slemp case. Some legal commentators have attempted superficial analyses that try to correlate case outcomes with how cases are selected for trial in Missouri. In the five talc trials to date, plaintiffs have prevailed (with fulsome verdicts) in both cases in which plaintiffs’ counsel selected the case for trial. See Amy Rubenstein and Malerie Ma Roddy, “Talc Talk: 1 Of These Verdicts Is Not Like The Others,” Law360 (June 1, 2017). In the three other cases, selected by defense counsel, the defense has lost two of the three cases, again with outlandish verdicts. No statistical analysis of these correlations will suggest that the selection process is correlated with verdict outcome. If there is no general causation, then selection of plaintiff for trial should not correlate with outcome. More important, the Missouri verdicts cannot be reconciled with the ruling by Judge Johnson in the New Jersey talc cases. Carl v. Johnson & Johnson, No. ATL-L-6546-14, 2016 WL 4580145 (N.J. Super. Ct. Law Div., Atl. Cty., Sept. 2, 2016); see also “New Jersey Kemps Ovarian Cancer – Talc Cases” (Sept. 16, 2016).

A manufacturer is legally held to the standard of having expert knowledge of the hazards of a product, and warning about those hazards that are not common knowledge. The conflicts noted above, and the exculpatory views of various professional groups and federal and international agencies should mean, in a sane system of products liability law, that a manufacturer would have no liability in the ovarian cancer – talc cases. A recent review concluded:

While mechanistic, pathology, and animal studies do not support evidence for the carcinogenicity of talc on the ovarian epithelium 329, epidemiological studies have indicated an association with talc use and increased OC [ovarian cancer] risk.”

Brett M. Reid, Jennifer B. Permuth, Thomas A. Sellers, “Epidemiology of ovarian cancer: a review,” Cancer Biol. Med. 14 (2017). The authors went on, however, to note that the association was not consistently found among studies, and that the IARC had rejected the causal contention as having been shown. How on this evidence, can a manufacturer be held liable for not warning of a causal connection? And how could a manufacturer be found to have acted maliciously or outrageously, with substantial scientific support?

What is needed is careful, detailed evaluation of the actual evidence at trial. The International Agency for Research on Cancer (IARC), the FDA, the NTP, and virtually every other agency and professional group that has addressed the question whether talc causes ovarian cancer, have declared that there is no causal connection established. Have the plantiffs in these cases hit a treasure trove of data not seen by the scientific and regulatory community? Or have plaintiffs exploited the weaknesses of the jury system, and advanced arguments and evidence that would never move a panel of disinterested scientists?

Meaningful analyses of the talc trials are not likely to happen from hipshot commentary. Fortunately, Courtroom View Network videotaped the Slemp trial from openings to return of verdict, which has created a rich resource for lawyers, legal analysts, political scientists, scientists, and regulators to judge the efficacy and content of courtroom presentations of complex scientific evidence.

Less is More, Except When Less is Just Less

There are two common, glib pronouncements you can often hear uttered in defense counsel confabulations. The first is “Less is More.” The second is “Let’s not get into weeds.” These generalizations cannot be tested with jury research in which both sides’ presentations are often no longer than 60 minutes, or so. Actual research of trial research rarely can move beyond anecdotal observations.

In Slemp, the defense case went took up two days. The plaintiffs’ case took 12 days. The plaintiff presented multiple medical expert witnesses, including two epidemiologists who have been involved in studying talc and ovarian cancer, and publishing on the issue, for decades. The defense presented just one expert witness, Dr. Warner Huh, a gynecologic oncologist. Dr. Huh claimed to have epidemiologic expertise by virtue of his work on studies on cervical cancer screening. Dr. Huh, as we will see, never explained how, when, and where he received training in epidemiologic study design and biostatistics. This defense strategy on expert witnesses requires careful analysis. Furthermore, the plaintiffs’ counsel presented a minimally qualified regulatory expert witness to serve primarily as a document reader. The defense effectively cross-examined this witness by showing his ignorance and selectivity in document reading and presentation. J & J, however, never called a fact witness, or corporate witness, to give viva voce testimony, that rebutted the innuendo, defamations, and characterizations of the company by plaintiffs’ counsel.

The Opening “Statements”

Perhaps it is best to begin at the beginning. Voir dire is not available at Courtroom View Network, but the opening statements are on line. There is a widespread myth that Hans Zeisel’s research established that most cases are won in the openings. Zeisel debunked that reading of his work, without disagreeing that the first impressions of opening statement can be powerful. See Hans Zeisel, “A Jury Hoax: The Superpower of the Opening Statement,” 14 Litigation 17 (Summer 1988). Inquiring minds might want to know how the openings statements went for the parties involved in Slemp.

Counsel for the plaintiff and for the two defendants (J & J, Imerys) all gave strong arguments that went well beyond stating what the evidence will show. All counsel worked hard to establish ethos and pathos, but plaintiff’s counsel excelled at creating the appearance of scientific logos, even when there was none. Defense counsel, on the other hand, tried to avoid talking about epidemiology for the most part. When the defense did discuss epidemiology, they made some disturbing, unnecessary mistakes.

Plaintiffs’ Opening

The plaintiff’s opening was noteworthy for its fear mongering. There are some authors who seem to want to take credit for a so-called reptile strategy, but fear mongering has been part of the dark side of rhetoric since at least the dawn of recorded history.1 Edmund Burke captured the sum and substance of the reptile strategy, which was so much on display in 18th century politics:

No passion so effectually robs the mind of all its powers of acting and reasoning as fear.”

Edmund Burke, A Philosophical Inquiry into The Origin of Our Ideas of The Sublime And Beautiful – With Several Other Additions at Sect. II. Terror (1757). Plaintiff’s counsel argued that all women, everywhere [including the women jurors and the male jurors’ female kin] are threatened by the evil corporate behemoth that sells baby powder. Women, everywhere, are developing ovarian cancer by the millions because they have used talcum powder. Only this jury can stop the carnage because regulators have ignored the situation. Regulators and the scientific establishment are venal, and J & J has bought them off. Steve Bannon would be proud.

The plaintiff’s counsel argued that J & J’s talcum powder contains not one, not two, but three carcinogens: talc, asbestos, and heavy metals. Talc, of course, is the focus of the claim and the trial, but what about the other two? Plaintiffs’ counsel did not advert to any evidence or opinion that heavy metals cause ovarian cancer; nor did he even slow down to say what heavy metals he was claiming were present. The evidence that asbestos causes ovarian cancer is perhaps marginally credible, but the causal conclusion is still doubtful. The studies that suggest an association are generally poorly done and heavily confounded. The real issue, however, with the asbestos claim, other than its effectiveness in instilling fear and knee-jerk reactions among lay jurors, is that it obscures an important issue whether the tremolite that sometimes accompanies talc in serpentine rock deposits is actually tremolite asbestos. Tremolite, as a mineral, has numerous crystallogaphic “habits,” including acicular fibers, angular particles, plates, etc.2 Some time ago, the federal government adopted a convention of counting anything as a “fiber,” if it were greater than 5 μm in length, and it had an aspect ratio of at least three to one. Another agency, the U.S. EPA adopted a minimal 5:1 aspect ratio, but both federal regulatory definitions disregard both mineralogical and biological considerations for what is a “fiber” with biological effects. Pace Melanie3, a thing is not a phallic system if it is just longer than it is wide, and a tremolite cleavage fragment is not a fiber, even if it is three times longer than wide.

There were other notable rhetorical moves in Plaintiff’s opening statement. In most other litigations, Plaintiffs run away from the need to rule out random error in studies that their expert witnesses proffer as support for causal claims. See, e.g., In re Zoloft (Sertraline Hydrochloride) Products Liability Litigation, U.S. Court of Appeals, 3rd Circuit, No. 16-2247 (June 2, 2017). In doing so, Plaintiffs often distort the consensus views about statistical significance, from the American Statistical Association4 and other groups. In the Slemp case, however, Plaintiff’s counsel swung to an opposite extreme by over-endorsing statistical significance as the apparent end-all and be-all for assessing causality. Study validity, whether internal or external, received no serious mention; bias in the studies relied upon was not discussed in any meaningful way.

To highlight the disingenuousness of the Plaintiffs’ opening with respect to failing to consider study validity and bias, and its over-endorsement of statistical significance, the Plaintiff later in the trial flashed sound bites from a report, commissioned by J & J, on the claimed causal connection. The report was by done other than Professor Kenneth Rothman, along with others. Professor Rothman has been in the forefront of criticizing the use of statistical significance testing as a bright-line criterion for causation. Rothman, along with Jonathan Samet (hardly a defense friendly epidemiologist) and Harris Pastides, reported, in 2000, that there was a “statistically significant” association, but demurred on causation because of the problem of study bias and validity:

A weighted average of the results from epidemiologic studies to date measuring the relation between talc and ovarian cancer risk gives an overall relative risk of 1.31, with a 95% confidence interval of 1.21-1.41. Bias and causation are competing explanations for the weak positive association observed. This weak association could be an underestimate of a stronger association if there are errors in measuring talc exposure that apply uniformly to all study subjects (non-differential misclassification). On the other hand, non-differential misclassification does not bias an association that is null to begin with, so postulating non-differential misclassification cannot shed light on whether the association results from a causal relation or not. Most of the published studies are interview-based case-control studies, subject to recall bias, which can readily give rise to associations of this magnitude. The evidence from these studies regarding recall bias is mixed. Uncontrolled confounding can also easily explain associations this weak; although no single confounding factor would seem to account for the overall effect, the combined effect of several such unidentified confounders could do so. In considering these competing explanations of bias and causation, the evidence in favor of a causal explanation is only the overall weak association of a relative risk of 1.31. The lack of a plausible biologic mechanism, on the other hand, weighs against a causal interpretation. Also weighing against a causal explanation is the dose-response pattern among talc users, which is an inverse trend for both duration of use and frequency of use. A causal relation would predict a positive trend, not an inverse trend. Based on these considerations, we suggest that the evidence to date does not indicate that talc can be ‘reasonably anticipated to be a human carcinogen’.”

Kenneth J. Rothman Harris Pastides Jonathan Samet, “Interpretation of Epidemiologic Studies on Talc and Ovarian Cancer,” (Nov. 28, 2000).

From a rhetorical perspective, one of the more interesting moves in Plaintiff’s opening was a pivot from causation to association. Without ever really discussing the standards or factors for inferring causation, Plaintiff’s counsel invoked regulatory standards and avoided addressing himself to the Missouri standard of causation. The standards for whether to warn and for determining cause in fact were conflated and confused, in what seemed liked a deliberate effort.

Defense Openings

J & J’s Opening

The defense strategies were apparent enough. The defense emphasized that Ms. Slemp had a serous borderline tumor (SBT). The emphasis appeared to be a plea of confession and avoidance. In other words, maybe there is something going on with talc and ovarian cancer, but this is an SBT, and it is different, so you do not need to worry too much about the more general claim that talc causes ovarian cancer. SBTs are a subset of epithelial ovarian tumors, often characterized as semi-malignant, with a more favorable outcome than other ovarian epithelial tumors. The defense also strived to shift the spotlight to Ms. Slemp and her strikingly poor health, preëxisting her cancer diagnosis, as well as her massive obesity and her heavy smoking, both risk factors for ovarian cancer.

On the positive side, J & J’s counsel anticipated and warned the jury that plaintiffs’ expert witnesses would be seriously challenged on their post-hoc analyses and their failing to share their causal conclusions with the scientific community. J & J’s counsel did engage on the general causation claim, but mostly to argue that most governmental agencies and professional organizations have refused to accept the causal claim. To the limited extent that J & J got into the “weeds,” it identified the Bradford Hill factors as the generally accepted decision procedure for reaching causal conclusions. So far, so good (except for the insistence upon referring to Sir Austin as Sir Bradford, as though Bradford was the man’s first name). What defense counsel did not say, astonishingly, is that Sir Austin’s nine factors require a necessary predicate that is satisfied when:

Our observations reveal an association between two variables, perfectly clear-cut and beyond what we would care to attribute to the play of chance.”

Austin Bradford Hill, “The Environment and Disease: Association or Causation?” 58 Proc. Royal Soc’y Med. 295, 295 (1965). See “Woodside & Davis on the Bradford Hill Considerations” (Aug. 23, 2013). To be sure, there are any number of studies on talc and ovarian cancer that satisfy Hill’s requirement of an association not attributable to the play of chance, but what is lacking is the clear-cut association; that is, the associations that are “statistically significant” have not ruled out bias and confounding. The defense opening did introduce the concept of recall bias, which plays an important part in undermining the validity of the available case-control studies. What the defense did not say, however, is that misclassification biases in cohort studies do not always bias towards the null, and even if they did, then the latest cohort study (with a hazard ratio below 1.0) would have had a risk ratio even lower than 1.0 than what was reported.

The defense did emphasize the absence of the Bradford Hill factors for plausibility and dose response. A more sophisticated analysis would have acknowledged what Sir Austin had over 60 years ago: plausibility or explanation is not necessary to infer causation. Not all the Bradford Hill factors are equal, and the defense diluted the analysis by falsely elevating one factor, plausibility, to the status of a necessary criterion.

The defense opening also had some dubious moves. It deprecated case-control studies as inferior to cohort studies, both as a general proposition and specifically with respect to talc. Case-control studies are generally harder to do, as the Plaintiffs’ bar learned when it based an entire mass tort litigation on a single case-control study, the so-called Yale Hemorrhagic Stroke Project.5 The Plaintiffs won the Daubert war6 in that litigation, but lost their jury trials because juries ultimately saw the methodological flaws that the MDL court disregarded. The general proposition that cohort studies are always superior to case-control studies, however, is doubtful. The defense did not need to stake out this claim, especially since it was not going to call an epidemiologist to testify.

Some of the claims that the defense committed to in its opening were as stunning as they were dubious. J & J’s counsel promised that he would show that talc has been proven safe. That claim is, however, beyond what the available science can show, especially with a plethora of statistically significant associations in case-control studies. J & J need only show that the plaintiffs’ claim has not been established, but it created an unnecessary burden of proving safety. The rhetorical value of the claim is obvious, but promising to show something that cannot be delivered seems like a recipe for disaster. $110 million is a disaster.

Then there was a defense claim that epidemiology cannot show that talc causes ovarian cancer. The claim was unclear as to whether epidemiology cannot establish general or specific causation, and vague as to whether the inability resulted from the weak, equivocal evidence in the instance of talc and ovarian cancer, or some deeper inability that stems from the nature of the evidence itself. Of course, given that there is an expected base rate of ovarian cancer in the general population, epidemiology will be required, even if it may not be sufficient. But if epidemiology alone is not sufficient, then what else is required? The defense never clarified its claim.

As for specific causation, the notion that epidemiology never informs individual patient predictions or causal assessments seems far fetched. If are dealing with a case of mesothelioma and occupational crocidolite exposure, with relative risks in the thousands, the attribution based upon the existence of a very high relative risk seems eminently reasonable. The same with maternal thalidomide use and an infant’s phocomelia, even though there is a baseline risk of phocomelia. Even in the case of a heavy smoker and lung cancer, with relative risks in the range of 20 to 40, inferring specific causation seems like a sound inference, especially in the absence of evidence that the risk is segregated in some subgroup of the population that suffers the outcome.

Imerys’s Opening

Counsel for Imerys also gave an opening statement, which started on common defense themes of the need to reserve judgment until all evidence is heard, and the need to consider context for statements. Imerys echoed the dubious claims that epidemiology can never inform inferences about individual patients, that epidemiology has determined that talc is safe, that cohort studies are always better than case-control studies, and that cohort studies are better because they have many thousand of women in them as opposed to “just a handful” in case-control studies.

Imerys, however, soon wandered into territory that affirmatively undermined J & J’s defense. First, it applauded itself for having warned of “possible causation,” which tended to concede the point to plaintiffs that there is a duty to warn of possibly caused outcomes. Second, Imerys appropriately urged its bulk supplier defense, which placed the spotlight on J & J’s alleged culpability.

The Imerys lawyer may have offended the Missouri jury by referring to talc as a mineral formed millions of years ago. A large percentage of Americans believe that the Earth was created less than 6,000 years ago. And yet, we still believe that allowing ordinary citizens to decide scientific issues is a good thing!

In its opening, Imerys also misstated the law to its own detriment. In discussing its obligations to warn, Imerys asserted that as a mining company it had to follow the rules established by OSHA. Actually, not. Mining companies are under the jurisdiction of the Mining Safety and Health Administration (MSHA), and OSHA Hazards Communication regulations did not apply to mining companies for the years involved in the Slemp case. See Memorandum from Patricia K. Clark, Director of OSHA Compliance Programs, to Regional Administrators re Hazard Communication and Mining Operations (Mar. 3, 1992).


1 See also Anthony Pratkanis & Elliot Aronson, Age of Propaganda: The Everyday Use and Abuse of Persuasion at 210 (2000) (“[A]ll other things being equal, the more frightened a person is by a communication, the more likely her or she is to take positive preventive action.”); H.L. Mencken, In Defence of Women § 13 “Women and the Emotions” (1923) (“The whole aim of practical politics is to keep the populace alarmed (and hence clamorous to be led to safety) by menacing it with an endless series of hobgoblins, most of them imaginary.”); Conor Boyack, Feardom: How Politicians Exploit Your Emotions and What You Can Do to Stop Them (2014).

2 John W. Anthony, Richard A. Bideaux, Kenneth W. Bladh, and Monte C. Nichols, The Handbook of Mineralogy, vol. II (1995).

3 Melanie Safka, “Psychotherapy” (“A thing is a phallic symbol if it’s longer than it’s wide”).

4 Ronald L. Wasserstein & Nicole A. Lazar, “The ASA’s Statement on p-Values: Context, Process, and Purpose,” The American Statistician, available online (Mar. 7, 2016), in-press at DOI:10.1080/00031305.2016.1154108, <http://dx.doi.org/10.1080/>.

5 Walter N. Kernan, Catherine M. Viscoli, et al., “Phenylpropanolamine and the Risk of Hemorrhagic Stroke,” 343 New Engl. J. Med. 1826 (2000).

6 See In re Phenylpropanolamine Prod. Liab. Litig., 289 F. 2d 1230 (2003).