814 F.2d 655
55 USLW 2568, 7 Fed.R.Serv.3d 568
Peggy L. WHEELAHAN; Donna J. Guba; Alexander M. Guba, Jr.,
M.D.; Priscilla K. Long; Foredonia Wilson; Louise
Lewellen; Jayne Tsuchiyama; Oteria T. Myers; Leon Myers;
Karen M. Buell; Richard Buell; Judy Perlov; S. Perlov;
Susan Eve Roza; Eli Roza, Plaintiffs-Appellants,
Sue Ellen Marder; Karen Lee Irvin; Shirley Gamble; C.
Don Gamble; Nancy Jean Tillery; Mary Ann Heininger; Lori
Ann Kushner; Deborah Gustafson; David Gustafson; Robin R.
Reeder; Nadine L. Allen; George Allen III; Terry N.
G.D. SEARLE & Company, Defendant-Appellee.
United States Court of Appeals, Fourth Circuit.
Argued Feb. 3, 1987.
Decided March 16, 1987.
NOTICE: Fourth Circuit I.O.P. 36.6 states that citation of unpublished dispositions is disfavored except for establishing res judicata, estoppel, or the law of the case and requires service of copies of cited unpublished dispositions of the Fourth Circuit.
Appeal from the United States District Court for the District of Maryland, at Baltimore. Joseph H. Young, District Judge. (CA-82-3506-Y)
Howard Robert Erwin, Jr. (Michael A. Pretl; Pretly & Erwin, P.A., on brief), for appellants.
Paul Farrell Strain; Nell B. Strachan; Elizabeth C. Honeywell; Terri L. Turner; Venable, Baetjer and Howard; (Vicki A. Thompson; Ronald R. Marich, on brief), for appellee.
D.Md., 630 F.Supp. 1087
Before RUSSELL and ERVIN, Circuit Judges, and SPENCER, United States District Judge for the Eastern District of Virginia, sitting by designation.
This appeal is from a consolidated case in which seventeen women sued G. D. Searle & Co. (Searle) for injuries allegedly caused by the use of the Copper 7 intrauterine device, a contraceptive. The injuries suffered by these women included pelvic inflammatory disease (PID), ectopic pregnancy, and perforation of the uterus. The legal bases they asserted for recovery were strict liability, negligence, fraudulent misrepresentation, and breach of warranty.
The trial court solicited, from both the appellants and Searle, pretrial proposals for managing this complex litigation. Searle proposed either individual trials for each woman or several small consolidated trials, with a separate class for each type of injury. The appellants proposed a single consolidated trial for a smaller, yet representative class. The court rejected both proposals. Instead, the court ordered that the trials be bifurcated. In Phase I, the cases of all seventeen women would be tried jointly to determine in the abstract whether or not the Copper 7 could cause the kinds of injuries noted in the complaints. If legal causation was found, the cases would be separated for Phase II, in which each woman would have a separate trial to prove that her particular injury was actually caused by use of the Copper 7, and also to prove her damages. Both parties objected to this bifurcation order. The court dismissed the objections.
After three weeks of testimony from experts on both sides in Phase I, the jury deliberated for two and one half days but was unable to reach a verdict. The court dismissed the jury, and Searle moved for judgment after trial. The court granted this motion on the ground that the opinions given by the plaintiffs' experts lacked a sufficiently reliable scientific basis, and were therefore inadequate to meet the plaintiffs' burden of proof. The appellants offer three reasons for reversal, and Searle offers an alternative ground for affirmance.
The first ground asserted for reversal is that the trial court erred in ordering bifurcation on the question of causation. Under Fed. R. Civ. P. 42(b) the court may order a separate trial of any separable issue if this would further convenience, avoid prejudice, or be conducive to expedition and economy. The appellants argue that the issue of causation in this case was not so distinct and separable that it could be tried separately without injustice. See Gasoline Products Co. v. Champlin Refining Co., 283 U.S. 494, 500 (1931).
In support of this contention, the appellants argue that causation could be proved in either of two ways. First, it could be proved by expert testimony showing causation based on scientific and medical data. Second, it could be proved by testimony from the women's personal physicians who would show, through a process of elimination, that these women's injuries must have been caused by the Copper 7 because no other possible causative agent or condition was present. See In re Beverly Hills Fire Litigation, 695 F.2d 207, 219 (6th Cir. 1982) (causation may be proved by showing absence of other causes), cert. denied sub nom. Bryant Electric Co. v. Kiser. 461 U.S. 929 (1983). According to the appellants, the bifurcation ordered by the court obligated them to use the first method of proving causation and deprived them of their right to use the second method. By their understanding of the bifurcation order, the women were not permitted to include testimony from their personal physicians in Phase I of the trial. They argue, however, that by severing the testimony of the personal physicians from the determination of causation, the court was left with an issue that did not satisfy Rule 42(b) because it could not be tried separately without injustice.
We note first that bifurcation to determine causation in the abstract is not impermissible. See In re Richardson-Merrell, Inc. "Bendectin" Products Liability Litigation, 624 F. Supp. 1212, 1222 (S.D. Ohio 1985). In this case it is not clear to us that the bifurcation order actually precluded the appellants from offering the testimony of their personal physicians on the issue of causation in the abstract. At the oral conference at which he ordered bifurcation the trial judge stated, "We will have the experts, except the medical experts, testifying as to each individual.... [This issue] can be resolved as a battle of the experts." (Conference of October 25, 1985, transcript pp. 5-6). Although this clearly precluded the personal physicians from testifying that the Copper 7 caused a specific plaintiff's injuries, it did not prevent those physicians from testifying as experts based on their general clinical practice. Within the limited confines of Phase I, we believe that any physician's testimony would have been admissible if that physician could show, through his or her personal medical experience, that the Copper 7 was capable of causing injury. Causation, in that context, could have been shown using appellants' second method of proof--that is, showing that some women who used the Copper 7 suffered these injuries in the absence of all other possible causative agents or conditions. None of the women, however, attempted to offer such evidence.
Absent a showing that the order prevented the appellants from producing evidence relevant to the issue of causation in the abstract, we find no abuse of discretion in the trial court's order of bifurcation.
The second ground asserted for reversal concerns the sufficiency of the epidemiological (statistical) evidence presented by the plaintiffs. The evidentiary standard for judgment after trial is identical to the standard for judgment n.o.v. The court must determine whether the evidence, taken in the light most favorable to the nonmoving party, is sufficient to support a verdict in favor of the nonmoving party. If it is sufficient, judgment after trial must be denied. See Whalen v. Roanoke County Board of Supervisors, 769 F.2d 221, 227 (4th Cir. 1985), rev'd on other grounds, 797 F.2d 170 (4th Cir. 1986) (en banc).
The appellants attempted to show through epidemiological evidence that the risk of PID, ectopic pregnancy, and perforation of the uterus was greater in women who used the Copper 7 than in the general population. Their primary witness on this issue was Dr. Miriam Orleans, an epidemiologist, who reanalyzed PID data from Searle's own clinical trials of the Copper 7. Dr. Orleans found that the incidence of PID in these trials was about 6.7 percent, whereas the incidence of PID in the general population was 2.0 to 3.0 percent. Cross-examination of Dr. Orleans revealed, however, that her graduate assistant had improperly classified the raw data from which Dr. Orleans had worked. As Dr. Orleans herself recognized, no reliable epidemiological determinations could flow from the faulty classifications. Therefore, Dr. Orleans' testimony at this trial did not provide sufficient epidemiological evidence to meet the plaintiffs' burden of proof. Similarly, epidemiological evidence from other experts who relied on Dr. Orleans' study could not suffice to support a verdict for the plaintiffs.
The appellants also suggest that epidemiological evidence from Dr. Janet Daling, a defense witness, showed causation. Dr. Daling had studied the effect of prior use of copper IUDs (primarily the Copper 7) on subsequent primary tubal infertility, which is one of the possible consequences of PID. According to Dr. Daling, a woman who had never been pregnant and whose copper IUD had been inserted before 1975 was 30 percent more likely to suffer primary tubal infertility than a woman in the general population. In statistical terms, if the risk of primary tubal infertility in the general population was designated as 1.0, the risk of primary tubal infertility after using a copper IUD was 1.3.
The trial court rejected this evidence because it showed less than a two-fold increase in risk (i.e., a risk of 2.0) of PID from the use of the Copper 7. Based on In re Agent Orange Product Liability Litigation, 597 F. Supp. 740, 785 (E.D.N.Y. 1984), the court determined that a two-fold increase was necessary because it was equivalent to the required legal burden of proof--a showing of causation by a preponderance of the evidence. The appellants contend that requiring this threshold level of risk was erroneous. According to the appellants, a two-fold increase in risk may be appropriate when a plaintiff is using epidemiological evidence to prove that his or her injury was more likely than not caused by a particular agent, as was apparently the case in Agent Orange. In the present case, however, the plaintiffs in Phase 1 did not have to prove that their injuries probably were caused by the Copper 7; they merely had to show that their kinds of injuries could be caused by the Copper 7.
We need not determine whether there must be a two-fold increase in risk to prove, in the abstract, that an agent is merely capable of causing an injury, because the results of Dr. Daling's study were admittedly not statistically significant. The court cannot properly draw any conclusions about the increased risk when that increase is not statistically significant. Dr. Daling's epidemiological evidence therefore was insufficient to support a verdict in favor of the plaintiffs.
The third ground asserted for reversal concerns the sufficiency of the nonepidemiological evidence presented by the appellants. Although epidemiological evidence by itself can be sufficient to show causation, the absence of statistical support of causation is not fatal to plaintiffs' case. See Wells v. Ortho Pharmaceutical Corp., 788 F.2d 741, 745 (11th Cir. 1986) (plaintiff's burden of proving that injuries were caused by the product did not necessarily require production of scientific studies showing a statistically significant association between the product and the injury in a large population); Ferebee v. Chevron Chemical Co., 736 F.2d 1529 (D.C. Cir.), cert. denied, 469 U.S. 1062 (1984); Dore, A Commentary on the Use of Epidemiological Evidence in Demonstrating Cause-in-Fact, 7 Harv. Envtl. L. Rev. 429, 434 (1983) (courts may erroneously assume that a showing of no increased risk eliminates any possibility of causation but this may be related to difficulties inherent in epidemiology rather than the absence of a causal link). Causation thus can be proven by nonstatistical evidence such as tissue samples, standard tests, and patient examination.
Although the plaintiffs called several expert witnesses who offered nonstatistical evidence the trial court, after careful consideration, rejected these witnesses' testimony because they failed to show an adequate scientific basis for their conclusions. Dr. Haverkamp, for example, testified that the Copper 7 is "definitely related" to the development of PID, but he did not testify that the Copper 7 can cause PID. This clearly falls short of the requirement that the expert testify to a reasonable degree of medical certainty that the Copper 7 can cause PID. See Fitzgerald v. Manning, 679 F.2d 341, 350 (4th Cir. 1982); Merit Motors, Inc. v. Chrysler Corp., 569 F.2d 666 (D.C. Cir. 1977). Dr. Haverkamp also relied on medical literature related to PID, but in the study on which he based his conclusions, 38 percent of the participants wore the Dalkon Shield IUD, which is associated with a six-fold increase in PID; only one participant in the study used a Copper 7, and she did not develop PID. Dr. Haverkamp also testified to an association between the Copper 7 and ectopic pregnancy, but he offered no scientific basis for this conclusion, and he again failed to testify that the Copper 7 caused ectopic pregnancy. His testimony on causation related to perforation of the uterus was even more speculative, and he in fact testified that he did not know if the Copper 7 causes perforation.
Dr. Hatcher testified that the tailstring of the Copper 7 could retract into the uterus and thus complicate the diagnosis of perforation, but he did not offer evidence on the causation of perforation. He also suggested that because the tailstring breaches the cervical barrier, bacteria could be transferred into the uterus during insertion of the IUD. For both of these tailstring problems, however, Dr. Hatcher failed to distinguish between the Copper 7 and any other IUD.
Based on her own experiments, Dr. Fives-Taylor testified that bacteria have a greater tendency to adhere to the tailstring of a Copper 7 than they do to the tailstring of a Copper T, which is made of a different kind of plastic. Dr. Fives-Taylor failed to offer evidence, however, showing that bacteria adhering to the tailstring of the Copper 7 could pass through the cervical barrier and migrate into the uterus.
Finally, Dr. Perlmutter testified that all IUDs are associated with an increase in PID, but she could not offer an opinion about the Copper 7 on its own, despite the fact that her clinical practice included women who used the Copper 7.
An expert's scientific or medical opinion is sufficient to withstand judgment n.o.v. if the basic methodology employed to reach the conclusion is sound. Ferebee, suDra at 1535-36. For each of the plaintiffs' nonstatistical experts, however, there was a flaw at the threshold. They testified to mere association, but not to causation. They presented conclusions based in part on sound scientific evidence, but also based in part on speculation. They failed to distinguish the effects of the Copper 7 from the effects of other more pernicious IUDs. Based on these cumulative faults, we agree with the trial court's conclusion that the causation testimony of the plaintiffs' experts was without adequate scientific basis.
Searle suggests that regardless of our conclusions on bifurcation and adequacy of evidence, we should affirm the judgment because the plaintiffs failed to prove that Searle gave an inadequate warning. Searle invokes Comment K to the Restatement (Second) of Torts Sec. 402A, which provides that an unavoidably unsafe product is neither defective nor unreasonably dangerous if it is properly prepared and is accompanied by proper directions and warnings. Comment K generally applies to prescription drugs, and the Copper 7 has been classified by the Food and Drug Administration as a prescription drug because the copper in it is reactive.
In Singer v. Sterling Drug Co., 461 F.2d 288 (7th Cir.), cert. denied, 409 U.S. 878 (1972), the court rejected the contention that all prescription drugs are automatically excluded from strict liability merely by virtue of being prescription drugs. Other courts have withheld the protection of Comment K from IUDs that are not prescription drugs. See Coursen v. A. H. Robins Co., Inc., 764 F.2d 1329 (9th Cir. 1985); Worsham v. A. H. Robins Co., Inc.. 734 F.2d 676 (11th Cir. 1984).
The purpose of Comment K is to protect defendants who supply critically needed but potentially harmful products such as human blood or rabies vaccine. Given the broad range of contraceptive alternatives that are available, we strongly doubt that Comment K should in any case apply to the Copper 7. Even if it were applicable, however, Searle offered no evidence that the Copper 7 is unavoidably unsafe. See Beshada v. Johns-Manville Products Corp., 90 N.J. 191, 447 A.2d 539 (1980) (product that has some utility is still not reasonably safe if the same product could have been made or marketed more safely). We therefore reject Searle's alternate grounds for affirmance.
In summary, we hold that the trial court did not commit reversible error in ordering bifurcation on the issue of causation, and that such order of bifurcation did not deprive the plaintiffs of their opportunity to prove causation by a process of elimination. We further find that both the epidemiological and nonepidemiological evidence were insufficient to support a verdict for the plaintiffs. Therefore, the court's grant of judgment after trial is