Argued: October 18, 1999.
Decided: December 22, 2000
Appeal from the Order of the Superior Court of Pennsylvania, Philadelphia District, dated December 29, 1997 at Docket No. 3711 Philadelphia 1995, reversing the Order and Judgment of the Court of Common Pleas, Philadelphia County, dated September 15, 1995 at Docket No. 1027, September Term, 1982. ___ A.2d ___ (Pa.Super.Ct. 1997) 001 E.D. App. Dkt. 1999
Thomas R. Kline, Shanin Specter, Robert Ross, Philadelphia, for appellant, Jeffrey Blum.
Edward Madeira, Philadelphia, for appellee, Merrell Dow Pharmaceuticals, Inc.
Robert Toland, Orlando, FL., for appellee, PA Defense Inst.
Christopher Scott D'Angelo, Philadelphia, for appellee Product Liability Cncl.
Margaret S. Woodruff, Philadelphia, for appellee amicu curiae, Chemical Mfg.'s Ass'n.
Steven D. Johnson, Philadelphia, for appellee amicus curiae, Pharmaceutical Research.
Before FLAHERTY, CJ. and ZAPPALA, CAPPY, CASTILLE, NIGRO, NEWMAN and Saylor, J.J.
In this prescription drug product liability case, we granted allocatur to consider whether the Frye rule still governs the admissibility of expert scientific testimony in Pennsylvania or whether the Daubert rule has superseded it. The Frye rule bars novel scientific evidence until it has achieved "general acceptance" in the relevant scientific community. The Daubert approach relaxes this requirement to a degree and may permit evidence based on scientific theory not yet generally accepted so long as a variety of specified factors can be deemed to establish a high level of reliability. The trial court permitted plaintiffs' expert scientists to present testimony arguably admissible under Daubert, and the jury entered a verdict for plaintiffs. Superior Court reversed, holding that the Frye rule was the proper standard; that under Frye, the experts should not have been permitted to testify; that the verdict should be vacated and the case remanded for entry of judgment in favor of appellee Merrell Dow.
Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
The case involves the drug Bendectin, alleged to be an agent which causes the production of physical defects in developing embryos. The Blums alleged that their son Jeffrey Blum was born with a crippling clubbed feet injury due to the ingestion of Bendectin by his mother, Joan Blum, during his fetal gestation.
At trial, Blums presented expert scientific causation testimony, some of which was corroborated by defense experts, to establish that Bendectin, manufactured by Merrell Dow Pharmaceuticals, Inc., caused Jeffrey Blum's clubbed feet. The scientific testimony was based on data from epidemiological studies, chemical structure analysis, in vitro (test tube) and in vivo (animal) teratology studies, as well as unpublished "recalculations" performed on the findings of previously published studies. The trial court apparently admitted the testimony under the evidentiary rule embodied in Daubert, and the jury returned a verdict for the Blums and against Merrell Dow in the amount of $19.2 million, including $15 million in punitive damages. Merrell Dow moved unsuccessfully for judgment notwithstanding the verdict. On appeal, Superior Court held that it was error for the trial court to apply Daubert, a standard somewhat less exacting than that of Frye, after this court specifically deferred a decision on that question. Superior Court therefore vacated the judgment and remanded the case for entry of j.n.o.v. in favor of Merrell Dow.
This court adopted the Frye test in Commonwealth v. Topa, 471 Pa. 223, 369 A.2d 1277 (1977). It has been applied on numerous occasions since 1977. When Crews, supra, discussed and applied the test, it also discussed the Daubert test, but declined to apply it, stating that the question of "whether or not the rationale of Daubert will supersede or modify the Frye test in Pennsylvania is left to another day." Crews, 536 Pa. at 519 n. 2, 640 A.2d 400 n. 2.
As we recognized in Crews, Daubert relaxes, to some extent, the impediments to admission of novel scientific evidence. Under Daubert, expert scientific testimony is admissible if it fulfills two criteria. First, it must relate to "scientific knowledge," thereby establishing "a standard of evidentiary reliability," which "will be based upon scientific validity." To constitute scientific knowledge, the evidence must be grounded in the methods and procedures of science, based on more than subjective belief or unsupported speculation, and supported by appropriate scientific validation. The second criterion is that the evidence must be relevant; it must assist the trier of fact to understand the evidence or to determine a fact in issue. Under Daubert, in federal trials, "general acceptance" no longer constitutes a threshold necessity in determining admissibility. Other factors are whether the technique or methodology can be or has been tested, i.e., its "falsifiability, or refutability, or testability." Another consideration is whether the theory or technique has been subjected to peer review and publication. Other factors bearing on scientific validation are the known or potential rate of error, and the existence and maintenance of standards controlling the technique's operation. Crews, supra, 536 Pa. at 518 n. 2, 640 A.2d at 400 n. 2.
The essential difference between Frye and Daubert, is that Frye requires the scientific community to reach some consensus as to reliability then relies on such consensus to determine the admissibility of the challenged scientific evidence. Daubert, on the other hand, examining the same factors which lead to general acceptance in the scientific community, abandons the standard of "general acceptance" and substitutes a judicial evaluation and determination of scientific reliability.
In considering whether this court should follow the lead of the federal courts in moving from the Frye to the Daubert standard, we note that the parties were content to frame their arguments within the traditional framework established under Frye. Moreover, the primary evidence at trial supporting the conclusion that Bendictin caused appellant's birth defect, the testimony of Dr. Alan K. Done, was so flawed as to render his conclusions unreliable and therefore inadmissible under either Frye or Daubert, so a choice between the two standards is unnecessary to the resolution of this appeal. It would appear to be jurisprudentially unsound, therefore, to utilize this case as a vehicle to evaluate the wisdom of such a change.
Only the amicus brief for the Pennsylvania Defense Institute squarely states a position on whether the court should continue to adhere to Frye, arguing in the affirmative. Its arguments are, in summary: trial courts should not become the testing grounds for untried and untested scientific theories; Frye is more likely than Daubert to yield consistent results when applied by different jurists; there is an undeniable simplicity to Frye; by requiring evidence to be measured by an objective standard, i.e., what others in the area deem reliable, Frye prevents experts from providing self-serving support for their own opinions; and the Frye standard has been the law in the Commonwealth for more than twenty years.
Over thirty published epidemiological studies have found no statistically-significant association between Bendectin and limb defects; the FDA, after complete review in the 1980s, found that available evidence showed no basis for a conclusion that Bendectin causes or increases the risk of birth defects in humans. Nevertheless, Dr. Done, who is not an epidemiologist, discounted such conclusions on the basis of a selective review of the data from several of these studies, in which he detached the underlying data from the controls set up by the studies, and recalculated using simple math (as opposed to epidemiological principles). In effect, Dr. Done worked backwards through the science, from the statistical results back to the original mere associations that led to the studies in the first place. This procedure cannot be fairly described as generally accepted methodology for purposes of the Frye standard. See generally Blum v. Merrell Dow Pharmaceuticals, Inc., 705 A.2d 1314, 1322-25 (Pa.Super. 1997). With respect to application of the Daubert formulation, numerous federal courts have determined that Dr. Done's testimony is not admissible. See Lust v. Merrell Dow Pharmaceuticals, Inc., 89 F.3d 594, 596 (9th Cir. 1996) (quoting the trial court as follows: "Dr. Done has seen fit to `pick and choose' from the scientific landscape and present the Court with what he believes the final picture to look like. This is hardly scientific. Through Dr. Done's manipulation of the studies of others, he could suggest almost any result to the Court. The scientific world, however, does not operate with such a disjointed approach." The court added that "Done . . . was [in 1984] already a professional plaintiff's witness. It is not unreasonable to presume that Done's opinion on Clomid was influenced by a litigation-driven financial incentive." Id. at 597 (emphasis added). Referring to Done's testimony, the court continued: "When a scientist claims to rely on a method practiced by most scientists, yet presents conclusions that are shared by no other scientist, the [trial] court should be wary that the method has not been faithfully applied." Id. at 598.); Richardson by Richardson v. Richardson-Merrell, Inc., 857 F.2d 823, 829 (D.C. Cir. 1988) (holding that Dr. Done's conclusion was not capable of proving causation in human beings in the face of the overwhelming body of contradictory epidemiological evidence), cert. denied, 493 U.S. 882, 110 S.Ct. 218 (1989); Lynch v. Merrell-National Laboratories, Inc., 830 F.2d 1190, 1196 (1st Cir. 1987) (holding that Dr. Done's reanalysis was not an adequate epidemiological basis for his opinion); DeLuca v. Merrell Dow, 791 F. Supp. 1042, 1054 (D.N.J. 1992) (excluding Dr. Done's testimony in part because he had concluded that there was no other more likely cause of the birth defect, but provided no evidence ruling out any other possible causes, such as a previous abortion), aff'd, 6 F.3d 778 (3d Cir. 1993), cert. denied, 510 U.S. 1044, 114 S.Ct. 691 (1994). Dr. Done's expert testimony was also rejected in Ealy v. Richardson-Merrell, Inc., 897 F.2d 1159 (D.C. Cir. 1990); Lee v. Richardson-Merrell, Inc., 772 F. Supp. 1027, 1032 (W.D.Tenn. 1991), aff'd 961 F.2d 1577 (6th Cir. 1992); Richardson by Richardson v. Richardson-Merrell, 649 F. Supp. 799 (D.D.C. 1986), aff'd 857 F.2d 823 (D.C. Cir. 1988); Wade-Greaux v. Whitehall Laboratories, Inc., 874 F. Supp. 1441 (D.V.I.), aff'd 46 F.3d 1120 (3d Cir. 1994).
The record contains sufficient evidence for us to agree with Superior Court's conclusion that both Frye and Daubert would bar the challenged scientific testimony.
Superior Court was therefore correct in holding that the trial court erred in denying Merrell Dow's motion for j.n.o.v. Accordingly, we will affirm the judgment of Superior Court.
Messrs. Justice Cappy and Castille file dissenting opinions.
I respectfully dissent.
While I agree that the majority that the Frye standard is still the test by which the courts in this jurisdiction determine whether scientific evidence is admissible, I feel compelled to write in order to address the Superior Court's reasoning below. It is certainly not incumbent on this court to engage in a point-by-point analysis of a lower court's opinion. However, I believe that in this matter it is important to discuss the Superior Court's recitation of the Frye test as it has the potential to mislead the lower courts and the practicing bar. Specifically, I refer to the Superior Court's statement that there are "two ways to analyze the question of whether the causation testimony proffered . . . meets the Frye . . . standard. One focuses on whether the causal relationship is generally accepted by the scientific community, and the other on whether the methodology is generally accepted by the scientific community." Super. Ct. slip op. at 18.
Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
The Superior Court is correct that this court has long interpreted Frye as requiring that the methodology employed by the testifying scientist be generally accepted in the scientific community. See e.g. Commonwealth v. Blasioli, 713 A.2d 1117, 1119 (Pa. 1998). Yet, we have not stated that the conclusion reached by the scientist regarding causation must also be generally accepted in the scientific community.
As noted by the Superior Court, this additional step in the Frye test — requiring that the conclusion also be generally accepted by the scientific community — was added by the Commonwealth Court in McKenzie v. Westinghouse Electric Corp., 674 A.2d 1167 (Pa.Commw.Ct. 1996). I cannot find that this court, however, has endorsed this interpretation of the Frye test.
Furthermore, I believe that it would be imprudent for us to do so at some future date. The Frye standard is limited to an inquiry into whether the methodologies by which the scientist has reached her conclusions have been generally accepted in the scientific community. This is a sensible standard, one which imposes appropriate restrictions while stopping short of being needlessly inflexible. It restricts the scientific evidence which may be admitted as it ensures that the proffered evidence results from scientific research which has been conducted in a fashion that is generally recognized as being sound, and is not the fanciful creations of a renegade researcher. Yet, such a standard is not senselessly restrictive for it allows a scientist to testify as to new conclusions which have emerged during the course of properly conducted research. Thus, I would squarely reject that portion of the Superior Court's holding which would require that a scientist's conclusions, as well as the methodologies utilized in reaching those conclusions, are generally accepted in the medical community.
Next, I turn to examining the conclusion of the majority that the trial court erred in determining that the scientific evidence presented by Appellants was admissible. In my opinion, there exists great confusion as to precisely what standard the trial court utilized in making this determination. The trial court claimed it properly applied the Frye standard in making the determination that Appellants' scientific evidence was admissible. Tr. ct. slip op. at 50 n. 143. The Superior Court, on the other hand, concluded that while the trial court claimed that it was applying the Frye standard, it did not in fact do so. The Superior Court found that the trial court abandoned its role of keeping out junk science as a "gatekeeper," and instead improperly left in the hands of the jury the initial determination of whether the scientific evidence was reliable. Super. Ct. slip op. at 16. Finally, the majority of this court states that the trial court admitted the evidence under the federal rule announced in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). Majority slip op. at 2.
In a situation such as this — where there is great confusion as to what standard was employed by the trial court — I question whether it is possible for an appellate court to execute its appellate review within its narrow confines, and examine only whether the trial court abused its discretion in admitting the evidence. Rather, I believe that in situations such as these, an appellate court attempting to "review" the decision of the trial court would instead be performing its own, independent review of the evidence, arriving at its own Frye conclusion rather than reviewing the Frye determination made by the trial court. For, how can an appellate court be said to be reviewing the determination of the trial court where there is so much confusion as to what the trial court actually decided?
Thus, I propose that this matter be remanded for a new trial. As recognized by our sister states, the trial court judge's role as "gatekeeper" in applying the Frye test is a critical one. State v. Porter, 698 A.2d 739 (Conn. 1997); State v. Copeland, 922 P.2d 1304 (Wash. 1996). In my opinion, this crucial gatekeeper function is one which demands the particular skills and abilities of a trial court judge, and should not be commended to an appellate court in the first instance. Thus, in recognition of the special function played by the trial court judge in these determinations, I believe that where, as in this case, there is confusion as to which standard was applied, it would be prudent to remand the matter to the trial court for a new trial and therefore respectfully dissent.
I note that Appellee raised six other issues before the Superior Court, asserting that each of these claims entitled it to the grant of a new trial. Since my proposed disposition of the Frye issue would result in such relief being ordered, I believe that it would be unnecessary to remand this matter to the Superior Court for consideration of these claims.
I note that Appellants have raised an issue which was not addressed by the majority. Appellants claim that since Appellee did not raise the Frye objection in the first trial, then this issue is forever waived and could not have been raised in the second trial. This argument is specious. This court has long held that a new trial wipes the slate clean, and the second trial is held de novo. Arthur v. Kuchar, 682 A.2d 1250, 1255 (Pa. 1996); Commonwealth v. Oakes, 392 A.2d 1324, 1326 (Pa. 1978). Appellants, however, try to manufacture an inconsistency in our case law by pointing to the "holding" in Commonwealth v. Dobson, 405 A.2d 910 (Pa. 1979) that an issue not raised in the first trial is waived for purposes of the second. Dobson, however, is a plurality opinion and is thus of no precedential value. Furthermore, any validity that the Dobson plurality opinion ever had has been definitively erased by the recent Arthur opinion. Thus, I believe that this is a meritless issue.
CASTILLE, Justice, dissenting.
Like Justice Cappy, I agree with the majority that the Frye test should remain the general evidentiary standard for admitting expert scientific testimony in this Commonwealth. The test is an appropriate vehicle to prevent "junk science from improperly influencing the jury on matters not within their common knowledge and experience. However, I disagree with the majority's summary conclusion that, under the Frye standard, the trial court here erred in admitting appellants' expert testimony on causation. After reviewing the extensive record here, I am thoroughly convinced that the trial judge committed no such error. The majority's selective reference to other decided cases involving the plaintiffs' lead expert witness, rather than consideration of the actual record on the Frye question here, proves no error by the trial court. Accordingly, I respectfully dissent.
The majority states that the trial court admitted the offending testimony here under the test set forth in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). I do not believe that to be so. As Justice Cappy correctly notes, in point of fact, the trial court itself stated that it was applying the Frye test. Trial Court Slip Op. at 45 n. 130, 50 n. 143. The Superior Court recognized that the trial court applied Frye, while suggesting that the trial judge went on to criticize Frye's limitations on the fact-finder. 705 A.2d at 1321, 1323 (noting that trial judge found that plaintiffs' experts used same four methods and techniques as defense experts, and those were generally accepted; thus, the trial court concluded that "their opinion on causation met the Frye standard for admissibility"). The Superior Court held, not that the trial court erred in failing to apply Frye, but that its application was erroneous. Specifically, the Superior Court agreed with Merrell Dow that the way in which plaintiffs' experts "utilized" the four generally accepted techniques was itself a scientific "method" that had to be generally accepted before it could be heard. 705 A.2d at 1323-24.
The trial court's analysis of this difficult question was far more thoughtful and sophisticated than the majority, or the Superior Court, acknowledges. A fair reading of the trial court's exhaustive opinion reveals a court acutely aware that this was a close case under traditional Frye analysis, and concerned with the further limitation on scientific testimony suggested in the Commonwealth Court's rather novel opinion in McKenzie v. Westinghouse Electric Corp., 674 A.2d 1167 (Pa.Cmwlth. 1996), alloc. denied, 547 Pa. 733, 689 A.2d 237 (1997). McKenzie would require that the expert's opinion as to the causal relationship at issue, and not just the expert's methodology, must find general acceptance in the relevant scientific community before it may even be heard. In effect, under McKenzie, minority opinions and conclusions on causation may never be heard, no matter how sound the underlying methodology. The trial court felt bound by McKenzie but concluded that McKenzie was wrongly decided. See Trial Court Slip Op. at 77-82.
The trial court strongly believed that the evidence at issue here was admissible, and its opinion attempts to explain why that is so, or should be so, under any test — a not-at-all unreasonable approach for the court to take given the age of this twice-tried case, given that the court felt bound by McKenzie, and given that the dichotomy of Frye/Daubert is a question that this Court had made a point of not yet deciding. In addition, and perhaps of more importance, the trial court felt that the Frye issue in this case was "unique," i.e., that the expert testimony at issue did not lend itself to an easy application of Frye. The trial court went to great lengths to explain why it believed that this was so.
The lawsuit was filed in 1982 and was first tried in 1986.
The gravamen of Merrell Dow's Frye objection was that the plaintiffs' scientific experts' conclusions on causation (or, as Merrell Dow would have it, the manner in which they reached those conclusions) should themselves be viewed as a separate scientific "methodology" that must be "generally accepted" in the relevant scientific community before they may even be admitted under Frye. The trial court disagreed for two reasons: first, because it did not believe that conclusions on causation are a separate methodology needing general acceptance (a would-be classic Frye formulation rendered questionable by McKenzie); and second, because, at least in the litigation-driven Bendectin "scientific community" described to the court in this case, the notion of "general acceptance," or scientific "orthodoxy," if you will, on the question of causation was a questionable proposition to begin with. This was so because the trial court had heard extensive evidence concerning Merrell Dow's active and deliberate role, motivated by its litigation interests in defending lawsuits involving Bendectin, in actually creating and influencing the scientific orthodoxy that would then operate to suppress any contrary opinion that might harm its Bendectin litigation interest. As the trial court succinctly put it: "The testimony in this case demonstrates how "scientific consensus' can be created through purchased research and the manipulation of a "scientific' literature, funded as part of litigation defense, and choreographed by counsel." Trial Court Slip Op. at 46.
The court generally summarized the evidence of Merrell Dow's influence that supported this conclusion as follows:
Articles were intentionally inserted into peer review journals for use in court. Studies for publication in peer review journals were tailored to the needs of litigation, and paid for out of defense funds. Most significantly, for the integrity of a judicial system, `scientific' articles for publication in `peer review' journals were edited before publication by lawyers litigating the issues presented in the article. The testimony revealed that `follow-up' studies were solicited by the defendant through intermediaries, funded by the defendant.
* * *
The testimony demonstrated that articles were inserted in `peer review' journals, without review by independent authorities, but edited by lawyers; that `peer review' journals published, as valid, the results of `less than good studies', that articles were rejected for publication by prestigious journals before being published in the `peer review' journal of `Teratology.' The testimony exposed scientific literature created for purposes of legal defense. The testimony revealed a sycophantic relationship between `scientists' and their funding source: the defendant, Merrell Dow. The testimony revealed circularity of reasoning to prove preordained "scientific' conclusions, and the use of litigation defense funds for scientific research manipulation. Finally, the testimony revealed factual editing of supposedly scientific research literature by the very lawyers defending in litigation.
Id. at 67-68, 70-71. In such an instance, where the supposed "consensus" of "scientific opinion" in the relevant "scientific community" had been manipulated by the financial and litigation interests of an interested party, the trial court determined that cross-examination of contrary scientific conclusions, rather than their outright exclusion, was sufficient gatekeeping under Frye. Id. at 71-72.
A "teratogen" is an agent that causes the production of physical defects in the developing embryo. The "discipline" teratology refers to the study of the causes and effects of abnormal growth and development in the developing embryo. There is no degree program or certification process in the field of teratology. The journal Teratology, the official journal of the Teratology Society, was edited by Dr. Robert L. Brent, Merrell Dow's lead expert witness.
In short, the trial court did not, as the majority inaccurately suggests, ignore Frye in favor of Daubert. Its analysis derived from the specific and extensive record made in the case before it, as well as the unsettled status of the law since McKenzie, and was far more sophisticated than acknowledged by the majority opinion.
The majority never discusses why it believes appellants' expert causation testimony here was inadmissible, instead offering only that it agrees with the Superior Court's conclusion in that regard. The published opinion of the Superior Court, however, states that there are two ways to analyze admissibility under Frye, one focusing on whether the methodology employed by the testifying expert was generally accepted in the scientific community (a classic Frye analysis), and a second, more restrictive test, deriving from the Commonwealth Court's opinion in McKenzie, which requires that the expert testimony as to the specific causal relationship also find general acceptance in the relevant scientific community. The Superior Court held that, under either version of the Frye test, the trial court abused its discretion in admitting the expert testimony here. 705 A.2d at 1322-25.
The majority does not identify which of the Superior Court's independent analyses, and hence, which version of the Frye test, it is approving. Since McKenzie represents a novel extension of Frye never discussed or approved by this Court in the past, and it is difficult to conceive that the majority intends to radically change the law by such an oblique reference, I assume that the majority intends to adopt only the Superior Court's traditional Frye "methodology" analysis. I would specifically disapprove and overrule McKenzie. Like Justice Cappy, I believe that the Frye test in this Court's jurisprudence has only required, and should only require, that the methodology employed by the testifying scientist, and not his or her ultimate conclusions or opinions as to causation, be generally accepted by the relevant scientific community. See Commonwealth v. Blasioli, 552 Pa. 149, 713 A.2d 1117 (1998).
Having said this, I believe that the trial court's admission of appellants' expert causation testimony here should be affirmed both because the trial court properly concluded that the disputed testimony consisted of expert conclusions not subject to Frye's methodology screening and, even if those conclusions were a separate "methodology" generally subject to Frye, I do not believe that consensus (or "general acceptance") should be required in exceptional cases like this, where the scientific consensus derives largely from the proprietary influence and litigation interests of the adverse party. To explain my view on both points, a close consideration of the record in this case is necessary.
See Trial Court Slip Op. at 15 ("Dr. Done relied upon acceptable scientific methodology in reaching his opinion").
While pregnant with her son Jeffrey, appellant Joan Blum ingested Bendectin, a drug manufactured by Merrell Dow and promoted by it as a safe drug for pregnant women to control morning sickness. Jeffrey was born with clubfeet, a crippling condition that has already required multiple surgeries. The Blums sued Merrell Dow, alleging that Joan Blum's ingestion of Bendectin was a legal cause of Jeffrey's birth defect. The Blums pursued theories of product liability, negligence, fraud, and breach of warranty. A first trial ended in a jury verdict in favor of the Blums on January 20, 1987, awarding $1 million in compensatory damages and $1 million in punitive damages. On Merrell Dow's appeal, however, the Superior Court ordered a new trial because the verdict had been returned by an eleven-person jury, after the twelfth juror had become ill. 385 Pa. Super. 151, 560 A.2d 212 (1989). This Court affirmed, holding that Merrell Dow had been deprived of its constitutional right to a trial by jury. 534 Pa. 97, 626 A.2d 537 (1993).
After a seven-week retrial, a second jury found that Merrell Dow had acted negligently, failed to provide proper warnings, breached express and implied warranties, and acted fraudulently. The jury awarded $4 million to Jeffrey, $200,000 to his parents for medical expenses, and $15 million in punitive damages.
The question of causation was hotly contested at both trials and came down to the proverbial battle of the experts. There was no serious question as to the qualifications of the competing experts. Merrell Dow's experts categorically opined that Bendectin could not cause Jeffrey's birth defect (notwithstanding that they agreed that no scientific study could prove such an absolute conclusion), while the Blums' experts held contrary opinions. Each expert was subject to searching cross-examination and, obviously, the jury ultimately accepted the testimony of the Blums' experts.
Two of the Blums' experts, Drs. Done and Gross, were deceased by the time of the second trial. Their testimony from the first trial was read to the jury due to their unavailability.
Having lost the causation issue before a jury for the second time, Merrell Dow complained post-verdict that the jury never should have been permitted to hear the Blums' expert testimony that Bendectin caused Jeffrey's clubfeet and that, absent that evidence, Merrell Dow was entitled to judgment notwithstanding the verdict. The tack taken was to suggest that the expert plaintiffs' testimony on causation reflected a minority view, not "generally accepted" in the scientific community of researchers and experts who studied Bendectin's potential to cause birth defects in general, and clubfeet in particular. Further refining its analysis, Merrell Dow suggested that the manner by which the plaintiffs experts, and in particular their lead expert, Dr. Alan K. Done, arrived at a contrary conclusion on causation constituted a separate "scientific methodology" that was inadmissible because it was not "generally accepted" in the relevant scientific community. What Merrell Dow wanted, effectively, was a finding that, as a matter of law, its product does not cause birth defects. That finding would be accomplished in the guise of an evidentiary ruling, i.e. a holding that any expert who looked at the "scientific studies" — studies that Merrell Dow often subsidized and which concluded that there is no causal connection — and reached a conclusion contrary to Merrell Dow's experts, should not be permitted to testify because that conclusion must result from a methodology that is not "generally accepted."
Notably, Merrell Dow had made no such argument after the first trial.
By approving the Superior Court's conclusion, the majority apparently finds merit in Merrell Dow's analysis. Going even further, the majority takes pains to dismiss Dr. Done's testimony by citing to other cases where Dr. Done had testified, in particular by quoting one appellate court that labeled him a "professional plaintiffs witness," whose opinion "was influenced by a litigation-driven financial incentive" and whose testimony, precisely because it involved conclusions not shared by other scientists, should be viewed with caution. Majority Op. at 7 n. 5, quoting Lust v. Merrell Dow Pharmaceuticals, 89 F.3d 594, 598 (9th Cir. 1996). But the majority could just as easily cite to a case where the testimony of Merrell Dow's leading expert, Dr. Robert L. Brent, was found to be incredible. See Wells v. Ortho Pharmaceutical Corporation, 615 F.Supp. 262, 291 (N.D.Ga. 1985) (finding Dr. Brent incredible because, inter alia, the absolute terms in which he expressed his conclusions detracted from his believability; Brent's "testimony and manner suggested a degree of conviction in his own conclusions unwarranted in a discipline in which, according to other competent experts from both sides, explanations are only more or less probable. For these reasons, Dr. Brent lacked credibility as a witness"). The Wells court noted that Brent's slanted testimony also might be a result of his litigation interest:
It bears noting that Lust is the only case cited by the majority that questioned Dr. Done's objectivity. In Lee v. Richardson-Merrell, Inc., 772 F. Supp. 1027 (1991), Dr. Done did not even testify or participate. In Richardson v. Richardson-Merrell, Inc., 649 F. Supp. 799 (D.D.C. 1986), the court concluded that the quantity of the defendant's evidence, especially the FDA's approval of Bendectin, prevented Dr. Done from testifying concerning his contrary conclusion. Importantly, in that case, unlike here, the plaintiff did not offer evidence of Merrell Dow's fraudulent failure to disclose information to the FDA. Ealy v. Richardson — Merrell, Inc., 897 F.2d 1159 (D.C. Cir. 1990), simply followed the Richardson decision. Lynch v. Merrell-National Laboratories, 830 F.2d 1190 (1st Cir. 1957), likewise did not involve an allegation of fraud. Finally, Wade-Greaux v. Whitehall Laboratories, Inc., 874 F. Supp. 1441 (D.V.I. 1994), did not even involve Bendectin. More importantly, in Wade, Dr. Done testified that it was not clear that Primatene Mist caused birth defects.
[A]lthough Dr. Brent's credentials were most impressive, he was not a convincing witness. His criticisms of plaintiffs' attorneys and of expert witnesses who testify for plaintiffs in malformation lawsuits strongly suggest a distinct prejudice against plaintiffs and a corresponding bias in favor of defendants in such cases.
By preferring citation to unrelated cases — and selective citation at that — to the actual record presented to the trial court here, the majority proceeds upon what is at best a half-truth. The plaintiffs' experts certainly were impeachable, and were impeached, because of their "litigation driven" interest. But, unperceived by the majority opinion is the countervailing fact that Merrell Dow's experts — experts who testified to the "orthodoxy" that would render the plaintiffs' experts' "minority" conclusions inadmissible — themselves were impeachable, and were impeached, because of their own financial and litigation-driven biases.
What should be of central importance, but is ignored by the majority, is the Frye record here. That record shows not only that Merrell Dow's experts were as liable to criticism for having a "litigation driven" financial interest as were the Blums' experts, but also that much of the "science" in this area, held up by Merrell Dow as the objective, generally accepted scientific view that requires exclusion of the plaintiffs' experts' "contrary" conclusions, itself was a product of Merrell Dow's litigation-driven influence. For example:
• The lead defense expert, Dr. Brent, had been retained as an expert by Merrell Dow for some eighteen years, N.T. 6/8/94 at 2612, and thus had a financial incentive to testify favorably for Merrell Dow. Brent was described as the originator of the field of teratology, a scientific subspecialty studying the causes and effects of abnormal growth and development in the embryo. Brent also edited the would-be "peer review" journal Teratology, which published many of the works relied upon by Merrell Dow to establish the favorable consensus of the scientific community. Id. at 2602. The trial court noted that there was ample evidence of a "sycophantic relationship between Brent and the attorneys representing Merrell Dow," which necessarily called into question the scientific validity of materials published by Teratology. Trial Court Slip Op. at 40. Remarkably, Brent actually submitted draft "scientific" articles on Bendectin for review and approval by the attorneys representing Merrell Dow. As the trial court aptly noted, such a relationship "clearly affected the objectivity of his approach and the validity of his writing." Id. Somehow, this practice caused Brent no ethical qualms.
Brent also claimed a curious expertise in legal matters involving birth defect claims. For example, Brent had published an article in Teratology entitled, "Litigation-Produced Pain, Disease, and Suffering: An Experience With Congenital Malformation Lawsuits," in which he claimed a rather unique ability to determine who told the truth, and who lied, in congenital malformation lawsuits. He found that seventeen out of seventeen plaintiffs lied and 82.6% of plaintiffs' lawyers "distorted" the facts, while, in his fantastical opinion, only 25% of defendants distorted the facts and only one in twenty-one defense attorneys made any distortions. N.T. 6/8/94 at 2662, 2669-76. He also opined that there was a sycophantic relationship between the plaintiffs' medical experts in those cases and the plaintiffs' counsel. Id. at 2681. It is unlikely that Brent's obvious bias and eccentricities, and the irony of his indictment of the relationship of the plaintiffs and their experts in these cases, went unappreciated by the jury.
• Another defense expert, Dr. Samuel Shapiro, also had a direct financial incentive to testify favorably for Merrell Dow regarding the scientific "consensus" on whether Bendectin caused birth defects. Shapiro began by overstating his qualifications in the field of epidemiology, the area in which he was supposedly an expert — a misstatement he was forced to acknowledge. N.T. 5/27/94 at 1756-62, 2537-61. Shapiro was the head of the epidemiology department of Boston University. Merrell Dow approached Shapiro to study Bendectin while Shapiro's group was studying other drug compounds. At that time, Merrell Dow only offered Shapiro $5,000 to complete the study. After the favorable results of the study were released, Merrell Dow increased their support of Shapiro's studies to $1.8 million. N.T. 5/20/94 at 1043.
Shapiro testified to his unwavering belief that Bendectin could not cause birth defects, despite admitting that his study of the issue was flawed. In the study, he grouped together women who took the chemical compound found in Bendectin during the time of fetal limb formation and women who took the drug after fetal limb formation. N.T. 5/23/94 at 1297-98. By including both of these women in his study, he admittedly diluted the number of women who could possibly show any effect from use of the drug. Id. at 1298. Not only did subsequent scientific literature criticize his methodology, but Shapiro himself was critical of his group selections. Id. at 1298-1302. Shapiro testified that "[t]he major criticism that we had among ourselves is that we recognized that this would introduce some miscalculation; that if there were any drug that were causal that would result in an underestimate of the magnitude of the effect." Id. at 1301. Nevertheless, he stubbornly refused to attribute any significance to his underestimation: "If there were a causal relationship, that causal relationship would have been underestimated. If there were no causal relationship, which is what I believe, or none that could be demonstrated, I doubt if there could not have been any underestimates." N.T. 5/23/94 at 1317 (emphasis added). As the trial court aptly noted, the circularity of Shapiro's logic is obvious: "It demonstrates justification science not inquisitive science." Trial Court Slip Op. at 30.
The Superior Court's failure to recognize Shapiro's "litigation driven" role in this litigation made its Frye analysis fatally erroneous. The Superior Court finds fault with Dr. Done's conclusions because they were arrived at by "reanalyzing" the results of a single study, referred to by the Superior Court as the "Heinonen" study. 705 A.2d at 1324-25. It was Done's method of reanalyzing that study, to conclude that it helped prove causation, that the Superior Court felt was an "improper" utilization of data that rendered his conclusions inadmissible. What the Superior Court failed to realize, however, was that the so-called Heinonen study was inextricably connected to Dr. Shapiro. The study was published in the textbook, Birth Defects and Drugs in Pregnancy, authored by Shapiro, Heinonen and a Dr. Slone. It was the very study that the obviously biased Shapiro himself admitted was flawed. Dr. Done was not obliged to accept the results of that flawed study, produced by scientists connected to Merrell Dow, as gospel.
• Another defense expert, Dr. James Newberne, was directly connected to Merrell Dow, having formerly served as a vice president responsible for animal testing and drug safety. N.T. 5/26/94 at 1660. In this case, which included a theory of fraud, Newberne admitted that Merrell Dow had failed to report numerous musculoskeletal defects in data submitted to the United States Food and Drug Administration ("FDA") to obtain approval of Bendectin. As the trial court described, Newberne under-represented the incidence of clubfeet found in animal studies, overstated the number of animals studied, failed to disclose that an insufficient number of animals were studied and that the animals died due to improper care. Gross' Testimony at 1673a 12/8/86 at 73. Newberne also failed to report that the company's tests were "scientifically inadequate due to insufficient dosing levels." Trial Court Slip Op. at 2; Gross' Testimony, N.T. 12/8/86 at 71. The following exchange provides just a flavor of Newberne's remarkable testimony:
Q [the plaintiffs counsel]: Sir, it has been the pattern and practice of the Merrell Company, in reporting to the FDA, to pick and choose selective information over the past thirty years relating to the drug Bendectin, correct?
A: Yes, that is correct.
N.T. 6/1/94 at 2266.
Newberne also testified to the proprietary link between Merrell Dow and other studies that came to comprise the exclusionary "scientific consensus" concerning whether Bendectin caused birth defects. One study, conducted by a Dr. Roll in Germany, concluded in a 1982 report to the German Official Health Agency that Bendectin caused birth defects in rats. N.T. 6/1/94 at 2259. After learning of this unfavorable study, Newberne contacted Roll through an intermediary, sending a copy of the correspondence to Merrell Dow's lawyers. Following this contact, Roll undertook a second study using a different breed of rats. Id. at 2262. The second report concluded that there was not an increased risk of birth defects associated with Bendectin use. This was not surprising, for the rats chosen in the second study had an increased natural incidence of diaphragmatic hernia that masked any increased malformations caused by Bendectin. Trial Court Slip Op. at 33-34.
Finally, Newberne testified to a second instance where Merrell Dow directly influenced a "scientific" study. In the early 1980's, a Dr. Hendrickx performed an animal study that showed a significant increase in heart defects in monkeys from the use of Bendectin. N.T. 6/1/94 at 2313. Once again, Merrell Dow funded, with more than $300,000 from the company's litigation defense budget, a second study that achieved much more positive results for the company. Id. at 2314-15, 2321. When Hendrickx wrote to the company regarding funding for the study, he indicated that he would be willing to discuss or modify his proposal to "meet a common objective." Id. at 2320.
• Dr. Mark Hoekenga, Merrell Dow's former Vice President, Medical Liaison Worldwide, testified to facts that suggested that Merrell Dow had influenced the outcome of yet another study on Bendectin, this one conducted by a Dr. Smithell, which concluded that Bendectin was not the cause of birth defects. The study was rejected by the widely respected medical periodicals New England Journal of Medicine, British Medical Journal and Lancet, before, predictably enough, it was accepted by Brent's journal Teratology. Hoekenga Testimony at 142a 144a. Smithell's obvious bias and incentive were reflected in the fact that he was actively soliciting funds from Merrell Dow at the time of his study. Thus, Smithell wrote to the company that:
Much clearly depends upon the value of this publication to Merrell Dow National Labs. If it may save the company large sums of money, large sums in the California court (which is rather what I thought when we undertook this study), they may feel magnanimous. If with the passage of time, the study is of no great significance, I can only regard the [monetary] figure you suggest as generous and welcome.
Plaintiffs' Exhibit 200, Hoekenga Testimony at 165a. In another, equally remarkable letter, Smithell noted that he would "appreciate any gesture Merrell felt inclined to make," but imagined that if he could give Bendectin "a clean bill of health with regard to teratogenesis, this would be of substantial help in the courtrooms of California." Plaintiffs' Exhibit 185, Hoekenga Testimony at 170a.
• Finally, two other defense experts, Dr. Mark Klebanoff and Dr. Rochelle Tyl, published their findings on Bendectin in Teratology, the journal controlled by Dr. Brent. N.T. 5/18/94 at 890, 892-3. Those studies were undertaken after Merrell Dow had removed Bendectin from the market. Klebanoff's study had been rejected by the New England Journal of Medicine, yet Teratology published it. Id. Klebanoff met with Merrell Dow attorneys after his article had been published in Teratology to discuss aspects of it. Also invited to the meeting by Merrell Dow was one of Merrell Dow's consultants, Dr. Berendes, who was also a Division Director at the National Institute of Health and Dr. Klebanoff's superior. N.T. 5/20/94 at 1031-32.
In short, in rushing to dismiss Dr. Done's testimony as that of a biased, "professional plaintiff's witness," both the majority and the Superior Court overlook not only the bias of Merrell Dow's equally "professional defendant's witnesses," but also the record fact that the "scientific" orthodoxy on Bendectin held up to silence the appellants' experts was in many instances bought and paid for by Merrell Dow to further their litigation needs. The trial court found, and the record amply demonstrates, that, with untold millions of dollars at its disposal, and untold millions more at stake, Merrell Dow was able to create and influence a scientific subdiscipline devoted to result-driven studies that Merrell Dow could then cite to defeat lawsuits brought by those who alleged that their birth defects were caused by Merrell Dow's Bendectin.
The trial court, which labored under this entire sordid picture, was of the view that there was, in fact, no difference in methodology between the dueling experts. The only dispute involved their conclusions on causation. The record here — a record never discussed by the majority — overwhelmingly supports this finding. Indeed, even Dr. Brent admitted that the four scientific methodologies employed by appellants' experts — chemical structure analysis, in vitro analysis, in vivo animal studies and human epidemiological data — were the generally accepted methodologies utilized by the relevant scientific community to examine whether Bendectin caused birth defects such as Jeffrey Blum's clubfeet. Merrell Dow's experts employed the same methodologies. Trial Court Slip Op. at 6, 9, 47.
Chemical structure analysis determines whether a chemical compound fits into a class of compounds that have certain effects. In vitro studies determine the effect of chemical compounds on cells or parts of tissues. Trial Court Slip Op. at 42. In vivo animal studies determine how a chemical compound affects a living animal. Epidemiological studies consider whether causation may be inferred by comparing the incidence of a disease in a group of humans who have been exposed to the substance in question with the incidence in a group who have not been exposed.
The Superior Court recognized that the studies and data relied upon by the appellants' experts were universally accepted as "good science," but concluded that the way those experts utilized that science to draw conclusions on causation was not. 705 A.2d at 1323. The flaw in this reasoning is that it conclusively determines that no opinion other than the initial researcher's may ever be heard. In a situation, as here, where so much of the underlying research and interpretation was in the control of a party driven by a litigation incentive, such a holding would be, and is, absurd.
The absurdity is well-demonstrated here. Even if some justification was needed to explain why the appellants' experts interpreted the data differently, the explanation was forthcoming. The appellants' evidence demonstrated that the conclusions in the studies that Bendectin was not a teratogen — studies often subsidized by Merrell Dow — were based upon severe underestimates of the significance of Bendectin's association with birth defects. This resulted from both poor study designs as well as a fraudulent failure to report data. For example, in Merrell Dow's Bunde — Bowles epidemiological study, the women taking Bendectin in some cases were also designated as the control match for themselves. Furthermore, some women in the control group had taken Bendectin while some women in the Bendectin group had not taken Bendectin. Done Testimony at 2121a, 2151a. In another epidemiological study that received funding from Merrell Dow, the Jick study, the researcher calculated his results using only part of the data he collected — omitting data obtained between 1976 and 1977 — which obviously undermined the conclusions on causation suggested in that study. Id. at 2169a, 2998a. Furthermore, with respect to Merrell Dow's animal studies, the appellants' evidence showed that the company did not report certain malformations and did not analyze whether dead animal fetuses had malformations. N.T. 5/31/94 at 2094-95, 6/1/94 at 2266. Given the basic flaws in the methodology of the Merrell Dow studies testified to by qualified experts, the appellants' experts were not required to accept the conclusions in those studies.
Further evidence that the appellants' expert testimony on causation was not renegade science may be found in Dr. Done's testimony concerning the Shapiro Heinonen study, another study held up as proof of the generally accepted scientific view that Bendectin did not cause birth defects. The Shapiro Heinonen study examined the effects of ingestion of certain drug compounds (including the compound in Bendectin) during the first four months of pregnancy. Dr. Shapiro's statistical analysis of the data concluded that 63% more children had birth defects when their mothers ingested the chemical compound in Bendectin within the first four months of development than when they had not taken Bendectin. N.T. 5/23/94 at 1182. Dr. Shapiro reached this number by reducing the raw numbers of cases involving birth defects by accounting for hospital-by-hospital deviation and comparing the hospital standardized data against the control group. The effect of Shapiro's "standardization" was to reduce the instances of an apparent causal relationship between Bendectin and birth defects. Dr. Shapiro did not believe the incidence of birth defects detected in his interpretation of the data was statistically significant and, accordingly, concluded that Bendectin was not a teratogen.
The studies' data had been collected from various hospitals. Dr. Shapiro accounted for "hospital deviation" by considering the impact of a woman being treated at a particular hospital as relevant in determining what caused the malformations being studied.
Appellants' expert, Dr. Done, fully explained why he reached a different conclusion from this same data. First, the Shapiro-Heinonen study underestimated the probability of Bendectin's causing limb malformations because the study included women who took the drug beyond the period in pregnancy where limb bud development occurs. N.T. 5/23/94 at 1298. In addition, Dr. Done recalculated this data, the major difference being that he did not repeat Shapiro's decision to reduce the number of positive associations outright by excluding results from certain hospitals, i.e. Shapiro's "hospital standardization." When Dr. Done recalculated the data, he simply compared the raw number of positive associations received in Shapiro's data against the control group. This analysis of the data resulted in a finding that the likelihood of having a child with clubbed feet was 2.1 times greater if the chemical compound in Bendectin had been ingested during the first four months of pregnancy. N.T. 5/23/94 at 1476.
Far from being junk science, Dr. Done's recalculation of the data and his use of a non-hospital standardized comparison, while different from Dr. Shapiro's approach, were nevertheless the same "methods" of analysis that had been employed in other studies relied upon by Merrell Dow. In Merrell Dow's Bunde-Bowles study, for example, the company did not standardize the data for hospital deviation. See e.g., Defendant's Exhibit 251. Merrell Dow had also relied upon a study completed by a Dr. Paul Stolley that used the recalculation "methodology" to analyze a study completed by Dr. Smithell. See Defendant Exhibit 3057; Hoekenga Testimony at 152a. In fact, Merrell Dow vice president Dr. Hoekenga maintained that the recalculation performed by Dr. Stolley was relevant and reliable information and that Merrell Dow would have produced this type of information to the FDA. Id. at 152a.
In short, both Drs. Done and Shapiro utilized generally accepted, albeit different, methodologies to analyze the raw data. The only difference is that the scientists took into account different factors, which produced different results and led to different conclusions on causation. On the record here, the trial court did not err in concluding that the plaintiffs' expert's methodology was generally accepted.
The Superior Court's suggestion that Dr. Done's statistical analysis was not "generally accepted" also is not supported by this record. As the trial court noted, even among Merrell Dow's experts, there were differences of opinion on this point. For example, Dr. Newberne maintained that no statistical analysis was required to make sense of Merrell Dow's animal studies; whereas, Dr. Tyl specifically rejected such an approach. Trial Court Slip Op. at 63-66. As the trial court aptly noted, there is not even a "scientific consensus" among the defense experts as to the statistical methodologies that Merrell Dow claims disentitle Dr. Done from testifying. Id. at 63.
Thus, contrary to the Superior Court's finding, there is nothing in this record to suggest that the manner in which the appellants' expert witness interpreted or "utilized" the raw data was a "methodology" that is not "generally accepted." Rather, there is ample support for the trial court's conclusion that there was a single, accepted scientific methodology at issue here, with the parties' experts, predictably enough, differing only as to the conclusions on causation. As to this hotly contested issue, each side accused the other's experts of reaching conclusions that were biased by their litigation interests. This issue was a classic matter for the jury to resolve and the trial court properly left it to the jury.
Finally, I would note that, even if the appellants' experts' conclusions could be viewed as a separate methodology requiring a Frye analysis, I believe that the trial court properly admitted this testimony. As I have detailed above, the record here shows that Merrell Dow largely created the "generally accepted orthodoxy" that would freeze out viewpoints contrary to their litigation interests. Merrell Dow subsidized or otherwise influenced most of the studies that concluded that Bendectin does not cause birth defects. Merrell Dow's role in virtually creating, and then slanting, the "scientific community" should be a relevant factor in the Frye analysis. Accordingly, I would create a limited exception to Frye that would permit the introduction of expert opinions contrary to those opinions generally held by the "scientific community," when those opinions are a result of proprietary research influenced by an interested party.
There is something not a little offensive about an entity, creating a biased, litigation-driven scientific "orthodoxy," and then being permitted to silence any qualified expert holding a dissenting view on grounds of "unorthodoxy." Where the would-be relevant scientific community is a community beholden to the defendants' litigation interests, that biased community should not be permitted to squelch dissenting opposing opinions. The trial court here properly refused to allow that unjust result to occur.
Hence, I respectfully dissent.