Sunday, April 15, 2007
A while back, I posted about a dispute regarding the impact of litigation on research, in particular in the context of the neurological effects of welding rods. One of the attorneys involved (who was involved in the correspondence in Neurology, sent the following. As I noted in the previous post (and consistent with my article linked to in that post), I generally think the dangers of such discovery are overstated.
Dear Professor Childs:
You recently called your readers’ attention to a publication by Dr. Brad Racette, entitled "The impact of litigation on neurologic research," published in 67 Neurology 2124 (2006). Dr. Racette, along with his legal counsel, decries the untoward effect of litigation on research generally, and on his research specifically. As you noted, Dr. Racette, along with his co-authors and legal counsel, and I engaged in a further exchange of views in the Correspondence section of Neurology, but the 400-word limit on letters to the editor prevented a full exploration of the issues. Dr. Racette’s response suggests that the federal court ruled that the subpoena prosecuted was over broad and burdensome. Most readers will discern the obvious equivocation in that a quashed subpoena would have placed no burden upon Dr. Racette. To the contrary, the subpoena at issue was served with a proposed protective order and an invitation to discuss how the subpoena and the order might be modified to protect legitimate interests. Those invitations were rebuffed, and the Court ultimate enforced the subpoena in large part; hence Dr. Racette’s belated special pleading in the official journal of the American Academy of Neurology.
Indeed, there is much more I could say about Dr. Racette’s activities in conducting the screenings of Alabama workers, which "medico-legal" activities gave rise to his published papers. See Racette, et al., Prevalence of parkinsonism and relationship to exposure in a large sample of Alabama welders," 64 Neurology 230 (2005); Racette, et al., "A rapid method for mass screening for parkinsonism," 27 Neurotoxicology 357 (2006). Unfortunately, a cloak of secrecy has been imposed, in the form of Court-imposed protective orders, and I am not at liberty to share all that I know with people who have not signed a mandated non-disclosure affidavit.
It might be helpful, however, to restate why the discovery provisions of state and federal rules of civil procedure are needed at all in the context of scientific studies, whether or not the studies enjoyed sponsorship by parties to litigation. After all, scientific studies are interpretations of analyses, often statistical, of data, which in turn represent multiple levels of hearsay. Consider a typical epidemiologic study. Patients give self-reports of symptoms. Physicians evaluate signs. Diagnostic tests are conducted, and their results reported and interpreted. Consultations may be sought. Diagnoses may be reached, and conveyed to patients, who in turn self-report their medical histories to subsequent treating physicians. Opportunities for error and systematic bias exist at every stage of the process of collecting diagnostic information and attempting to relate it to occupational, environmental, pharmaceutical, or lifestyle exposures.
In our legal system, expert witnesses are generally permitted to hold forth with their opinions, based upon studies they did not conduct. Moreover, expert witnesses need not even be able to describe and vouch for the reliability of the many steps involved in conducting the studies upon which they rely. All they need do is accredit the study as one that they believe experts in their field would reasonably rely upon, and they are then permitted to build upon, and draw inferences from, the multi-layered hearsay study as though it were a proven fact. Typically, the expert witness tries to accredit the study by virtue of the authors’ reputations, the journal’s prestige, the claimed rigor of the peer-review process, and the lack of challenge to the study in the scientific literature. This method of accrediting studies can be highly effective in front of unsophisticated judges and jurors, who do not appreciate the severe limitations in the peer-review process. At best, all these lines of accreditation are weak, indirect proxies for the truth.
As the coat of arms of the Royal Society proclaims: "Nullius in verba." See http://www.royalsoc.ac.uk/page.asp?id=6186 <visited on April 14, 2007>. Or as the late President Reagan said in folksier style: "Trust but verify." The Korean cloning scandal, cold fusion, and other painful examples illustrate that scientists do an imperfect job of policing their community and their journals. Maybe the truth prevails in the fullness of time, but time is not a luxury that a litigant, plaintiff or defendant, caught in the jaws of the U.S. mass tort system, enjoys.
The example of the Hemorrhagic Stroke Project ("HSP"; also known as the Yale Stroke Study) is illustrative of how a study, with all the indicia of reliability, can go badly astray. Because of continuing concern over a possible relationship between the use of medications containing phenylpropanolamine ("PPA"), the FDA suggested that the manufacturers sponsor study to look for an association between recent use of PPA and hemorrhagic stroke. The manufacturers complied, and the sponsored a study, which enjoyed an independent scientific advisory board and FDA oversight. The manufacturers had no control over the conduct of the study, and the investigators retained scientific and editorial control over the work.
The HSP was a multi-centered case-control study that looked for hemorrhagic strokes in persons within an age range, and then examined whether any of them had used PPA-contained medications within a specified time frame of the onset of the stroke. The principal investigators of the HSP were highly accomplished scientists, but most of the work was done by their juniors, eager to establish themselves in tenured faculty positions. Notwithstanding the prestige of the principal investigators, the university (Yale) where the analysis was done, the oversight of disinterested scientists, open presentation of the results to an FDA Advisory Committee, the study was flawed by errors that led directly to the finding and reporting of an association in the paper, which was a featured publication in the New England Journal of Medicine. When the defendants in the PPA litigation sought discovery into the underlying data, plaintiffs’ counsel ridiculed the defendants’ efforts to discredit the "industry’s own study." Although plaintiffs were able to get past reliability challenges in federal Multidistrict Litigation, the problems with the HSP proved well-nigh insurmountable – defendants won virtually all the cases tried to verdict in the PPA litigation. After one of the first defense wins, plaintiffs sought a new trial on grounds that they were prejudiced by the extent to which the defendant attacked the validity of the HSP. At oral argument, the trial judge, Hon. Anthony J. Mohr, rejected this challenge and expressed his own skepticism about the integrity of the HSP. See, e.g., O’Neill v. Novartis Conumer Healthcare, Superior Court for California, County of Los Angeles, Transcript of Post-Trial Hearing at 47 (Mar. 18, 2004)("Yale gets a black eye"), aff’d, ___Cal. Rptr. 3d ___, 2007 WL 586606 (Feb. 27, 2007).
The skepticism about a study such as the HSP is likely to be heightened when the study had been sponsored by an interested stakeholder. It is not only industry that has sponsored studies designed to be brought to bear on public health policy issues. Organized labor has frequently sponsored such studies, as evidenced by the collaboration between the asbestos insulators’ union and the late Irving Selikoff. More recently, plaintiffs’ counsel have directly and indirectly sponsored studies with the goal of deploying them in litigation. Several blatantly fraudulent studies, funded by plaintiffs’ counsel or by plaintiffs’ support groups, for the silicone gel breast implant litigation, helped fuel the celebrated Daubert exclusions in that litigation. The fraudulent and unreliable nature of many of the silicone studies would not have been detected if not for the subpoena-supported inquiry into the underlying data of the questionable studies.
In both science and law, studies can not be properly explored for reliability without full access to the underlying methods and data. Most journals impose severe page limitations, which prevents the sort of openness of access to data, called for by the National Academies of Science and the National Institutes for Health. Some journals now require that underlying data be submitted upon acceptance for peer review or for publication. Some of the more methodologically savvy journals have even started to make important aspects of the study, such as questionnaires and raw data sets, available as supplemental content on their websites.
Perhaps I should take solace from the following observation. Although Dr. Racette’s work in Alabama was sponsored by plaintiffs’ counsel, plaintiffs’ counsel themselves have rejected the strictures called for by Dr. Racette. In the litigation over alleged neurologic effects of welding fume exposure, plaintiffs’ counsel aggressively have sought the data underlying studies sponsored by defendants. Dr. Racette’s public relations efforts took his special pleading to the pages of Business Week, where he complained about having to sit for three depositions. See Arnst, "Keeping Lawyers Out of the Lab," Business Week (Feb. 5, 2007). (All these depositions involved the treatment and care of specific patients, and they were in no way different from what treating physicians routinely encounter. Inquiries into the research sponsored by plaintiffs’ counsel was deemed off limits by Dr. Racette’s counsel.) I have to say that I agree with the spokesperson for the American Association for Justice, née ATLA, who was quoted in this Business Week article as not being moved by Dr. Racette’s lament. Prohibitions against the discovery of underlying data can serve only to perpetuate the false allure of scientific infallibility, while denying litigants the right to confront and cross-examine the crucial expert witness opinion testimony that is built upon the shaky foundations of poorly conducted research. We should reject the proposal to create a privilege or a barrier that would deny litigants access to the crucial determinants of the validity of the evidence offered against them.
McCarter & English, LLP
N.B. The views expressed above are my own personal views, and not necessarily shared by my firm or my clients. I think it is likely obvious, but I will say it anyway: I have typically represented defendants in product liability cases, and I have been on both sides of demands for underlying study data.