As noted in part three of these notes on Wells, the court regarded the epidemiologic studies as “inconclusive” on whether spermicides cause birth defects. Somehow, the inconclusiveness of some studies undermined the defense’s expert witnesses, but not the plaintiffs’ witnesses:
“In finding the studies offered by defendant inconclusive on the issue of causation, the Court did not need to consider as substantive evidence the studies offered by plaintiff to suggest causation. Of course, the Oechsli study, the Smith study, the Buttar study on teratogenicity, and the Jick study all support this finding. As discussed below, the Court did consider the Oechsli and Smith studies in making its decision on the issue of failure to warn.”
Wells v. Ortho Pharmaceutical Corp., 615 F. Supp. 262, 292 & n.38 (N.D. Ga. 1985), aff’d and rev’d in part on other grounds, 788 F.2d 741 (11th Cir.), cert. denied, 479 U.S.950 (1986). The trial court’s “of course” hardly made its statement correct; of course.
Additional FDA Reviews and Opinions
In reaching its verdict on state of the art knowledge in 1980, the court ignored the FDA monograph that addressed safety, efficacy, and labeling of non-ionic spermicidal products. The court continued to ignore this monograph with respect to medical causation.
In 1983, the FDA’s Fertility and Maternal Health Drugs Advisory Committee (FMHDAC), concluded that non-ionic surfactant spermicides needed no additional warning. 615 F. Supp. at 278. One of Ortho’s witnesses, Mr. Armond Welch, worked for the FDA from 1946 through 1980, and was involved in regulation of the product. Mr. Welch pointed out that the FDA approved a contraceptive sponge device, designed to be used with non-ionic spermicides, after review of a New Drug Application, in the 1970’s. Id. at 280.
Defense expert witness, Dr. Stolley, reviewed the transcript of the December 1983 meeting of the FMHDAC, and described the Committee’s assessment of spermicides and congenital abnormalities as “comprehensive and thoughtful.” Dr. Stolley explained that there was no indication that the Committee had been unduly influenced by Ortho. This factual assertion was never factually challenged. Given the number of members of the Committee, their diverse backgrounds and political views, it is difficult to see how Ortho could have biased the Committee’s consideration of the issues; yet the plaintiffs’ witnesses made sweeping, subjective, and defamatory statements, which the trial court credited. One of plaintiffs’ expert witnesses, for instance, Dr. Mitchell, testified that the FMHDAC conclusion was irrelevant because, in his opinion:
“Those people are readily swayed, I think, by other interests, including pharmaceutical manufacturers, other people to whom they must report and who[m] they represent. I’m not sure what the weight of one of these advisory committees should ultimately be.”
Id. at 278. FDA Committee members are “readily swayed,” but trial judges (and jurors) have scientific insight to avoid being misled!
Dr. Mitchell offered no factual basis for his slanderous subjective personal opinion; yet Judge Shoob thought he had a credible demeanor! The contrast between what Judge Shoob found to discredit and accredit witnesses showed a remarkable bias in favor of plaintiffs’ expert witnesses. In evaluating one of the defense experts, Judge Shoob found that Dr. Robert Brent was “biased” because he had published disparaging remarks about plaintiffs’ attorneys and expert witnesses who testify for plaintiffs in birth defects litigation. Id. at 291. Similarly, Ortho’s company witnesses were not taken seriously because they were employees of the defendant. Id. at 278. Plaintiffs’ witnesses were not discredited for making disparaging remarks against industry or the FDA. Dr. Mitchell’s wholesale dismissal of an entire FDA committee and its work product, however, was accepted at face value, without reflecting upon Dr. Mitchell’s obvious bias and overzealous advocacy.
Indeed, Judge Shoob corroborated Dr. Brent’s view of plaintiffs’ counsel when he dismissed several counts of the complaint for complete lack of evidence. One dismissed count contained a claim for negligent failure to warn that the product does not always prevent conception. The uncontradicted evidence at trial showed that Ortho’s product did carry such a warning, that the plaintiff Ms. Maihafer had read the warning before use, and that she knew that the spermicide when used with a diaphragm would not be 100% effective. Id. at 291.
Dr. Robert Brent, a leader in the field of the study of birth defects, served as a voting member in the FMHDAC’s inquiry into teratogenicity, without disclosing his litigation work. Id. at 291. Rather than accrediting Dr. Brent’s testimony, this credential was turned against him by the trial court. Apparently, Dr. Brent assisted the defense prior to the meeting, but did not consider himself to have been “retained” until March 1984. Id. at 290. The trial court used Dr. Brent’s apparent equivocation between the date of his retention and the date of his first involvement to discredit both Dr. Brent’s testimony and the work of the FMHDAC. Id. at 294. Regrettably, the court allowed trial atmospherics about Dr. Brent to obscure the participation of many experts, from various perspectives, who participated in the several FDA findings and reports.
Plaintiffs’ witnesses took the animal evidence, which failed to support teratogenicity, and by personal opinion alone interpreted such evidence to support the claim that Ortho’s product was teratogenic in humans. One study of rats concluded:
“The results suggest that single vaginal application of [nonoxynol-9] is embryolethal and fetocidal but nonteratogenic in the rat at a dose approximately ten times higher than that recommended for controlling conception in women.”
Buttar, “Assessment of the Embryotoxic and Teratogenic Potential of Nonoxynol-9 in Rats Upon Vaginal Administration,” 2 The Toxicologist 39, 40 (1982). Other animal studies suggested absorption of the non-ionic spermicides. See, e.g., Buttar, “Transvaginal Absorption of Spermicides,” 13 Toxicology Letters 211 (1982); Chvapil, “Studies on Nonoxynol-9. II. Intravaginal Absorption, Distribution, Metabolism and Excretion in Rats and Rabbits, 22 Contraception 325 (1980).” The plaintiffs’ witnesses took absorption and lethality to mean teratogenicity, in the face of the studies’ failure to find malformations. Hand waving and puffery became scientific evidence, somehow, to satisfy a burden of proof. 615 F. Supp. at 273, 276.
The Jick Study
The plaintiffs relied heavily upon a study conducted and published by Hershel Jick, and colleagues. Hershel Jick, Alexander M. Walker, Kenneth J. Rothman, Judith R. Hunter, Lewis B. Holmes, Richard N. Watkins, Diane C. D’Ewart, Anne Danford, and Sue Madsen, “Vaginal Spermicides and Congenital Disorders,” 245 J. Am. Med. Ass’n 1329 (1981). This study makes Justice Sotomayor’s Matrixx Initiatives dictum clearly erroneous; the Jick study found a statistically significant difference between the rate of a composite of several birth defects in women who did, and those who did not, use spermicides.
But the Jick study had serious threats to its internal and external validity, which the trial court chose to ignore. For one thing, the statistically significant result came only for a composite of malformations, which included birth defects that had no biologically plausible relationship to one another in terms of their potential relationship to a fetal exposure to a teratogen. Defense expert, Dr. Stolley, for instance, described the diverse birth defects selected for inclusion in the Jick study to be “biologically implausible.” 615 F. Supp. at 284.
The Jick study did not report a statistically significant outcome with respect to the limb reduction deficit, which was the outcome of interest for the Wells plaintiffs. Further undermining the meaning of any confidence intervals or p-values from the Jick study was the multiple testing and comparisons that took place within the Jick cohort. By looking for many malformation outcomes, without a pre-specified aim of the study, the Jick paper was an exploratory, hypothesis-generating study, which did not support scientific, causal conclusions.
Perhaps the biggest problem with the Jick study, was its use of a measure of “exposure” to spermicides, a prescription by a woman any time up to 11 months before conception. Therefore, women and their malformed children were counted as cases without regard to whether the mothers had actually used spermicides at the time of conception or at any time during their pregnancies. As Dr. Stolley explained, Jick had defined “exposure” so inaccurately, that the data from the Jick study was “almost uninterpretable.” Id. at 284. The Jick study may have raised a hypothesis about spermicides, but it could not test that hypothesis in a meaningful way with the prescription data it used. Studies after the Jick paper failed to show an association between spermicides and birth defects, or limb reduction deficits specifically.
The defense called Dr. Richard Watkins, one of the authors of the Jick paper to testify narrowly on the validity problems in the study. Dr. Watkins emphasized that the study was too small to yield statistically stable results, that presumed exposure in the study did not mean actual exposure, and that the lumping of multiple outcomes together was scientifically inappropriate. Id. at 281. According to Dr. Watkins, he made these criticisms to Dr. Jick, who directed him to have nothing more to do with the paper.
For better or worse, however, the Jick study was published with Dr. Watkins as an author. The trial court did not explain whether Dr. Watkins ever requested his name to be removed from the manuscript, but the trial court clearly formed a jaundiced view of Dr. Watkins, based upon the contradiction between his testimony and the paper that bore his name.
The Watkins’ problem was compounded by Ortho’s request to have him investigate exposure, and Dr. Watkins’ (unsurprising) finding that some of the women classified as users had not, in fact, used spermicides at the time of conception or afterwards. Id. at 282. The court viewed Dr. Watkins’ taking publication credit, but then having challenged the validity of the study for the first time in a court room, rather than in the scientific community or its publications, as “severely” eroding his credibility. Id. Judge Shoob explained that:
“[b]ecause he had participated in the 1981 Jick study, Dr. Watkins might have been an ideal witness to comment on the validity of its conclusions. Dr. Watkins’ testimony on cross-examination, however, severely eroded his credibility. It is perplexing that a physician would risk his professional reputation by signing his name to a study about which he had serious reservations, especially when he knew the article would be published in a widely-read journal. Moreover, the dispute between Doctors Jick and Watkins, when Dr. Jick requested that Dr. Watkins discontinue work on the study, creates some question about Dr. Watkins’ impartiality.
Finally, that Dr. Watkins chose this courtroom as the first public forum in which to repudiate a study that he had helped conduct nearly four years earlier creates further doubt about his credibility. For these reasons, the Court found Dr. Watkins’ testimony not credible.”
Id. (internal citations omitted). Judge Shoob’s harsh condemnation of Dr. Watkins stands in stark contrast with his glowing approval of the credibility of plaintiffs’ witnesses, who had chosen the courtroom for their first declaration that spermicides cause birth defects. More to the point, however, is that whatever the merits of Judge Shoob’s credibility determinations, they did nothing to make the Jick study more persuasive or to erase the serious validity concerns raised by the defense.
Plaintiffs’ witnesses appeared not to consider all the evidence in reaching conclusions about causation; they relied heavily upon the Oechsli unpublished study, and Jick, but they seemed to ignore Cordero (1983), Warburton, Shapiro (1982), Mills (1985), Harlap (1980), as well as another unpublished study by Harlap.[i] 615 F. Supp. at 284 & n.26. The Court nonetheless found that plaintiffs had carried their burden of proving Katie Wells’ limb reduction deficits. Id. at 292.
Judge Shoob’s verdict has long since been reversed by the court of scientific review, which of course was small solace for Ortho. See James L. Mills, “Spermicides and Birth Defects,” in Kenneth R. Foster, David E. Bernstein, and Peter W. Huber, eds., Phantom Risk: Scientific Inference and the Law 87 (MIT Press 1993) (by time of the FMHDAC meeting in 1983, “the vast majority of evidence found no association between spermicide use and birth defects”); David F. Goldsmith & Susan G. Rose, “Establishing Causation with Epidemiology,” in Tee L. Guidotti & Susan G. Rose, eds., Science on the Witness Stand: Evaluating Scientific Evidence in Law, Adjudication, and Policy 57, 70 (OEM Press 2001) (“Weak science and inappropriate verdicts – Spermicide and birth defects”); see also “Wells v. Ortho Pharmaceutical Corp. Reconsidered – Part 1,” at footnote 1 (Nov. 12, 2012).
After Judge Shoob’s verdict, the FDA issued a notice, “Data Do Not Support Association Between Spermicides, Birth Defects,” in the FDA Drug Bulletin (1986). Dr. Watkins published his re-appraisal and criticism of the Jick study, which had been a lightning rod for Judge Shoob’s scorn. Dr. Watkins reported that many of the women who had had malformed children and who were counted as users of spermicides actually planned their pregnancies and were not using spermicides at all. Richard N. Watkins, “Vaginal spermicides and congenital disorders: the validity of a study,” 256 J. Am. Med. Ass’n 3095, 3096 (1986) (noting that the Jick study’s “definition of exposure to spermicide near the time of conception was grossly inaccurate”). Watkins wrote further that Jick (1981) was “unsupported by more complete evidence from its subjects.” Id; see also Lewis B. Holmes, “Vaginal spermicides and congenital disorders: the validity of a study – A Reply” 256 J. Am. Med. Ass’n 3096 (1986) (noting that the Jick article should never have been published).
“Somewhat like statements in a law review article written by a judge, or a judge’s comments in a lecture, dicta can be used as a vehicle for offering to the bench and bar that judge’s views on an issue, for whatever those views are worth.” McDonald’s Corp. v. Robertson, 147 F.3d 1301, 1315-1316 (11th Cir. 1998). Dicta can also be used to manipulate the path of the law by encouraging lower court’s to treat the dicta as holding. The appellate court that issues the dicta retains “plausible deniability” if the dicta proves improvident in future cases. The importance of dicta increases with the hierarchical level of the court. Dicta in Supreme Court cases is given great attention because of that Court’s role in setting policy and declaring the law beyond the narrow confines of the dispute between the parties. See, e.g., Judith M. Stinson, Why Dicta Becomes Holding and Why it Matters,” 76 Brook. L. Rev. 219 (2010). For all these reasons, Justice Sotomayor’s dicta about statistical significance and the Wells case should be taken seriously, and should be seriously rejected.
(to be continued)
[i] Harlap, “Congenital Abnormalities in the Offspring of Women Who Used Oral and Other Contraceptives around the Time of Conception” (unpublished study supported by the Center for Population Research, National Institute of Child Health and Development).