finding that admission of scientific expert testimony depends on four inquiries including "the expert's reasonableness in relying on his facts and data under Rule 703, the relevancy of the expert's testimony under Rules 401-403, and the sufficiency of the expert's methodology under Daubert"Summary of this case from Lithuanian Commerce Corp., Ltd. v. Hosiery
Civil Action No. SA-93-CA-0145.
March 11, 1997.
Michael T. Gallagher, Price Ainsworth, Fisher, Gallager, Perrin Lewis, Houston, Texas, Ralph I. Knowles, Jr., and Leslie J. Bryan, Doffermyre, Shields, Canfield Knowles, Atlanta, Georgia, for Plaintiff.
Vaughn Crawford, Snell Wilmer, Phoenix, Arizona, D. Mitchell McFarland, Liddell, Sapp, Zivley, Hill Laboon, L.L.P., Houston, Texas, and Daniel D. Rodarte, Rodarte Geringer, Los Angeles, California, for Defendants.
ORDER GRANTING MOTION TO EXCLUDE AND GRANTING IN PART MOTION FOR JUDGMENT AS A MATTER OF LAW
On this date, the Court considered the status of the above-styled and numbered cause. On February 10, 1997, the Court entered a preliminary order excluding the testimony of Dr. Shanna Swan and Dr. Luis Espinoza, expert witnesses for the Plaintiff. At the time of that preliminary ruling, the Court advised all parties that it would enter a more formal order on the exclusion at a later date. Pursuant to that earlier advisory, the Court now clarifies its holding and that holding's effect.
I. Background and Procedural Posture
The Plaintiff in the present case received two Heyer-Schulte implants in 1977. She claims that these implants caused her to develop Sjogren's Syndrome, an inflammatory disorder with the symptoms of dry eyes, dry mouth, and dry vagina. See Pre-Trial Order at 2. In order to establish her claim, the Plaintiff must show both general and specific causation — that is, that breast implants are capable of causing the condition she complains of, and that her breast implants were the cause-in-fact of her specific condition. See Brock v. Merrell Dow Pharmaceuticals, Inc., 874 F.2d 307, 313 (5th Cir.) (holding that inability of Plaintiff to establish that a drug could cause complained-of birth defects merited judgment as a matter of law), modified, 884 F.2d 166 (5th Cir. 1989); Doddy v. Oxy, 101 F.3d 448, 463 (5th Cir. 1996) (holding that, under the Fifth Circuit's interpretation of Texas law, a Plaintiff complaining of an injury must show that the defendant was the cause-in-fact of that injury). To meet her burden, the Plaintiff offers the testimony of two scientific experts, epidemiologist Dr. Shanna Swan and rheumatologist Dr. Luis Espinoza. The Defendants have challenged the testimony of both experts under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), asserting that the experts' methods lack sufficient reliability to help the jury.
II. Standards for the admissibility of scientific expert testimony
From the time of the scientific revolution in the late seventeenth century until the present, courts have been concerned with scientific methodology — both in the presentation of evidence conforming to the rules of reason and in the methods of the proceedings themselves. See Harold J. Berman and Charles J. Reid, Jr., The Transformation of English Legal Science: From Hale to Blackstone, 45 Emory Law J. 438, 497 (1996) (relating the evolution of English legal science to the scientific revolutions of the seventeenth century). While it is the task of a jury to decide what weight to give evidence in its deliberations, it is the role of the district court under Daubert to ensure that scientific expert testimony is of sufficient validity to warrant its admission into evidence. Daubert, 509 U.S. at 597, 113 S.Ct. at 2799.
A four-part inquiry governs the admission of scientific expert testimony in the Fifth Circuit. See Christophersen v. Allied-Signal Corp., 939 F.2d 1106, 1110 (5th Cir. 1991). This inquiry focuses on the expert's qualifications under Federal Rule of Evidence 701, the expert's reasonableness in relying on his facts or data under Rule 703, the relevancy of the expert's testimony under Rules 401-403, and the sufficiency of the expert's methodology under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. at 592-95, 113 S.Ct. at 2796-97.
Christophersen, of course, was decided before Daubert, and relied on a methodological inquiry under the Frye test. Because Daubert has a similar focus on methodology, the Court finds that Christophersen's four-part inquiry still valid.
With regard to this last criterion, some explanation is necessary. In order to be admissible, a scientific expert's testimony must be supported by appropriate validation — that is, it must have a reliable basis in the knowledge and experience of the expert's discipline. Daubert, 509 U.S. at 591, 113 S.Ct. at 2795-96. The hallmark of acceptable testimony turns on whether the scientific conclusion is testable and has been tested. Lesser, but still important, considerations are whether the scientific conclusion has been published in a peer reviewed journal, the amount of error associated with the expert's technique, and whether the theory is generally accepted in the scientific community. See Daubert, 509 U.S. at 592-95, 113 S.Ct. at 2795-96.
However, the Fifth Circuit has warned courts that the focus upon methodology has some limits. An extremely probing methodological review may be appropriate only when the expert relies on particularly objectionable or unconventional scientific theories. See Carroll v. Morgan, 17 F.3d 787, 789-90 (5th Cir. 1994). Moreover, Daubert is read by the Fifth Circuit to permit expert testimony on specific causation issues when the testimony is grounded in the methods and procedures of medical science. Id. In other words, a medical doctor testifying with regard to issues of causation may base his testimony upon a review of medical records (or his own exam), the doctor's experience, and a broad review of the literature. Id. See also Allen v. Pennsylvania Eng'g Corp., 102 F.3d 194, 196 (5th Cir. 1996); United States v. 14.38 Acres of Land, 80 F.3d 1074 (5th Cir. 1996); United States v. Posado, 57 F.3d 428, 433 (5th Cir. 1995).
This Court takes very seriously the role entrusted to it by Daubert and Christophersen. See 509 U.S. at 597, 113 S.Ct. at 2798-99; See also 939 F.2d at 1110. Indeed, the Court has wide discretion in matters pertaining to the admission of scientific expert testimony — and with this discretion, the Court believes, comes a responsibility to make clear the basis of its rulings. See Christophersen, 939 F.2d at 1109 (holding that review of a district court's decisions regarding admissibility of scientific expert testimony is under a clearly erroneous standard). Accordingly, the Court will describe in some detail the process behind its decision.
III. Dr. Shanna Swan
In the main, the Plaintiff wishes to have Dr. Swan testify regarding the association described in the scientific literature between breast implants and Sjogren's Syndrome or its symptoms (sicca symptoms). Such testimony should be excluded because it is based on facts not reasonably relied upon by an expert (Rule 703), it is based upon an improper methodology (Daubert), and it fails to pass the tests of relevancy (Rules 401-403). In short, there is a panoply of reasons why this testimony must be excluded.
A. A Description of the Scientific literature
The starting place for the Court's analysis is a review of two articles in the scientific literature. In Self-Reported Breast Implants and Connective-Tissue Diseases in Female Health Professionals: a Retrospective Cohort Study, 275 J.A.M.A. 616 (1996), Dr. Charles Hennekens and others conducted a large and very thorough epidemiology study (the Hennekens study). Through the use of questionnaires, the authors of the article interviewed 395,543 women in the health care field, 10,830 of whom had breast implants. 275 J.A.M.A. at 617. After review, the authors opted for a two-tailed statistical analysis to determine relative risk and the 95% confidence interval. Id.
The authors found that, compared with women who did not report breast implants, the relative risk of any connective tissue disease among those with breast implants was 1.24 (with a 95% confidence interval of 1.08 to 1.41, and significance to the 98.5% level). For Sjogren's Syndrome, the authors found a relative risk of 1.49, with a 95% confidence interval of .97 to 2.28, and a significance only to the 93.3% level. Because the significance level for Sjogren's Syndrome was below 95%, the authors characterized the significance as "borderline." 275 J.A.M.A. at 616. The report suggested that further research was necessary before firm conclusions could be drawn, especially because of the potential for an over-reporting bias of classic connective-tissue diseases. Id.
The second important study is Silicone Breast Prostheses and Rheumatic Symptoms: A Follow Up Study, 53 Ann. Rheum.Dis. 194 (1994), by Dr. Erik Giltay et al (the Giltay paper). This study interviewed by questionnaire 445 women, 235 with breast implants and 210 without, and found that women with breast implants reported more rheumatic complaints after implantation, calculating a relative risk for the symptom of burning eyes at 2.43, with a 95% confidence interval from 1.29 to 4.57. Id. at 194-96. The Giltay study, however, found no evidence of an increased prevalence of common rheumatic diseases among women who received implants, and it suggested that further study was necessary before any conclusion could be drawn regarding an implant-Sjogren's link. Id. Furthermore, the authors, like the authors of Hennekens study, found that bias in their sampling technique was a possible alternative cause of all their findings. Id. at 196.
These are the two most important papers to the Plaintiff's case. Indeed, Dr. Swan, the Plaintiff's own epidemiologist, found that all other studies, though perhaps relevant, were not reasonably relied upon to form opinions regarding causation. See Declaration under perjury of Dr. Shanna Swan at Paragraph 13; Transcript of Court Proceedings, February 10, 1997, 77:7-79:18 (hereinafter "Transcript"). The Court agrees with this analysis, and therefore will limit its attention to these two articles.
B. Rule 703 and Daubert applied to Dr. Swan's Testimony
When confronted with a proffer of expert testimony based upon particular studies, the Court must as an initial matter determine whether the studies could be reasonably relied upon by the expert. See Christophersen, 939 F.2d at 1110, 1114; Fed.R.Evid. 703. Concurrently, the Court must determine whether a scientific expert's method of using the study is sufficiently reliable to warrant admission. Daubert, 509 U.S. at 591, 113 S.Ct. at 2796. Therefore, in the present case, the Court must decide both whether the studies in question are reasonably relied upon, and whether Dr. Swan's method of using those studies has the mark of a scientific method. Daubert, 509 U.S. at 592-95, 113 S.Ct. at 2796-97; Christophersen, 939 F.2d at 1110, 1115-16; Fed.R.Evid. 703.
i) Application of Brock to the studies as written, with no reanalysis
With regard to the Rule 703 standard, the Fifth Circuit has held that epidemiological studies with lower-end confidence intervals less than one are not reasonably relied upon by experts to form opinions regarding causation. Brock v. Merrell Dow Pharmaceuticals, Inc., 874 F.2d 307, 311 (5th Cir. 1989). In the present case, the Hennekens study has a lower-end confidence interval less than one for the relative risk linking breast implants to Sjogren's Syndrome. Therefore, under the holding of Brock, it is unreasonable as a matter of law for Dr. Swan to rely on the Hennekens study to draw any conclusions regarding a Sjogren's-implant link. Accordingly, the testimony based on this paper should be excluded under Federal Rule of Evidence 703.
The Giltay paper does not directly address the issue of a Sjogren's-implant link, and therefore it is not addressed in this section.
The Plaintiff has argued that Brock was decided before Daubert, implying that after Daubert, Brock is of little utility. The Court must disagree. While the Fifth Circuit has not directly addressed the effect of Daubert on Brock, it is worth noting that Daubert cited Brock with approval. 509 U.S. at 595, 113 S.Ct. at 2798. Furthermore, other district courts in the Fifth Circuit have held Brock still valid in the post -Daubert era. LeBlanc v. Merrell Dow Pharmaceuticals, 932 F. Supp. 782, 784 (E.D.La. 1996). The Court has no choice but to conclude that Brock is still good law.
Moreover, even if Brock did not apply, the Court would still find it unreasonable for an epidemiologist to rely on the Hennekens study to conclude that a link exists between breast implants and Sjogren's Syndrome. As noted in the Reference Manual on Scientific Evidence, an observed association between exposure to a product and a condition may reflect a true cause-effect relationship or a spurious finding. See Reference Manual on Scientific Evidence 157 (1994) (supplement to Charles Allen Wright et al., Federal Practice and Procedure). To distinguish between these alternatives, it is necessary first to consider the effect of confounding factors. Id. Once this consideration is complete, the observed association must be judged against the Koch and Henle postulates to determine the likelihood of a causal link. Id. at 160-64. Only after these steps are taken can a conclusion be drawn regarding whether the association represents a valid observation and whether a causal relationship is implicated. Id. at 157-58.
Indeed, the Plaintiff's own expert recognizes this fact. See Section III C, infra.
The authors of the Hennekens study themselves listed several confounding factors they felt weakened their study. 275 J.A.M.A. at 619-21. Because of these possible sources of bias, the authors did not draw any conclusion regarding an implant-Sjogren's link other than that they observed an association of borderline statistical significance. Id. at 616. In short, the confounding factors combined with the low level of statistical significance make it unreasonable for an epidemiologist to rely on the Hennekens study to draw conclusions about the validity of an association between breast implants and Sjogren's Syndrome. Id. at 621 (describing the difficulties of differentiating, on basis of the Hennekens study, between low risk and no risk of an implant-caused disease). The paper's methodology is not sufficient to show that an observed association between implants and Sjogren's Syndrome is anything but a spurious finding. See Reference Manual on Scientific Evidence, supra, at 157-59.
In addition, the Koch and Henle factors as applied to the Hennekens study militate against causally connecting breast implants and Sjogren's Syndrome. To find causation from an association, an expert must look to the observed association's strength, the temporal relationship, the association's consistency, the biologic plausibility, the possibility of alternative explanations, the specificity of the association, and the dose-response relationship. Id. at 161. For the Hennekens study, the strength of the association is low, the consistency of the association is also low (i.e., the study suggested that more tests were necessary before meaningful conclusions could be drawn), alternative explanations are likely, the association is not specific, and there is no dose-response relationship. 275 J.A.M.A. at 620 (describing no clear trend of increasing relative risk over time). Thus, even if Brock were not the law, the Court would still find that a reasonable epidemiologist would not rely on the Hennekens study to conclude that a causal relationship existed between breast implants and Sjogren's Syndrome. This testimony must therefore be excluded. See Fed.R.Evid. 703.
ii) Reanalysis of the Hennekens Study
The Plaintiff also offers to have Dr. Swan testify as to her reanalysis of the Hennekens study. Basically, what Dr. Swan has done is substitute a one-tailed statistical test for the two-tailed test used by the Hennekens study. Doing this establishes statistical significance at about the 96% level, with a lower-end confidence interval greater than one — thereby, according to the Plaintiff, allowing the Hennekens study to meet the Brock standards. The Court must disagree. Brock clearly contemplated the use of two-tailed statistical tests, as is evidenced by its discussion of minimum and maximum confidence intervals. See Brock, 874 F.2d at 311-12. Moreover, in the epidemiology context, it seems plain that courts have preferred to rely upon two-tailed statistical tests. See Reference Manual on Scientific Evidence, supra, at 152-56. For both these reasons, the Court concludes that the reanalysis does not satisfy Brock, and therefore should be excluded.
One-tailed statistical tests, by their very nature, have only one-sided confidence intervals. Conversely, two-tailed tests have two-sided confidence intervals.
Furthermore, the Court notes that Dr. Swan has not seen the data collected by the Hennekens study, and because she has not seen the data, she cannot say with certainty whether her use of a one-tail statistical test is appropriate. See David Freedman et al., Statistics 495-96 (1978) (providing that use of a one-tail statistical test is appropriate only when a researcher knows what sort of data he will collect). By choosing a one-tailed statistical test, Dr. Swan assumes a priori that the data tends to show that breast implants have negative health effects on women — an assumption that the authors of the Hennekens study did not feel comfortable making when they looked at the data. In essence, Dr. Swan has made a hypothesis about the Hennekens data that, although testable, has not been tested — a weighty consideration under Daubert. See 509 U.S. at 593-94, 113 S.Ct. at 2796-97. In addition, her reanalysis of the Hennekens data is unpublished, not generally accepted, and contains within it an a priori assumption that introduces an unknown and potentially devastating amount of error. For all these reasons, the reanalysis should be excluded under Daubert. Finally, it should be noted that this statistical reanalysis amounts to little more than playing with numbers: changing from a one-tail to a two-tail test makes it easier for the Hennekens study to reach the Brock standards, but does not change the court's analysis of bias or the Koch and Henle factors. Thus, even with a one-tail test, the Hennekens study does not satisfy the standards of Federal Rule of Evidence 703.
The Plaintiff further wishes to have Dr. Swan reanalyze the Hennekens study using larger P-value, thereby making the numbers statistically significant. Once again, the reanalysis should be excluded under Daubert: the larger P-value introduces too much possibility of error, the use of the larger P-value is not generally accepted the breast implant context, has not been published in that context, and has not been peer reviewed in that context. Furthermore, the larger P-value will once again not address the Koch and Henle factors, nor will it change the problems of confounding factors causing a false association. Thus, even if a larger P-value were appropriate, the evidence should still be excluded under Rule 703.
iii) Testimony regarding sicca symptoms
The Plaintiff also wishes to have Dr. Swan testify regarding a causal link between breast implants and the symptoms of Sjogren's Syndrome (sicca symptoms), but not between implants and Sjogren's Syndrome itself. Testimony in this regard faces special problems, as a plaintiff must show in a products liability case that the injuries complained of were caused in fact by the product in question. C.F. Doddy, 101 F.3d at 463 (citing Texas cases). Cause-in-fact means that the Plaintiff must show that "but for" the product, the injury would not have taken place. Doddy, 101 F.3d at 463. In the present case, the Plaintiff admits that she has Sjogren's Syndrome, a condition that causes dry eyes, dry mouth, and dry vagina. Pre-Trial Order at 2. Given that the Plaintiff has Sjogren's Syndrome, it is impossible for the Plaintiff to establish that "but for" her breast implants, she would not have sicca symptoms — after all, the very symptoms she wishes to blame her breast implants for are also caused by her inflammatory disorder. Thus, if the Plaintiff cannot show that breast implants caused her Sjogren's Syndrome, then the Plaintiff cannot establish that her dry eyes were caused in fact by the breast implants. See Doddy, 101 F.3d at 463.
The Plaintiff argues that Sjogren's Syndrome is just a descriptive term for a constellation of symptoms. Therefore, the Plaintiff concludes, she should be able to show that breast implants cause her individual sicca symptoms, even if she cannot establish that her breast implants cause Sjogren's Syndrome. This argument is specious. If Sjogren's Syndrome is not a classic disease but rather just a set of symptoms, then failure of the Plaintiff to establish that breast implants cause "Sjogren's Syndrome" amounts to a failure by the Plaintiff to establish that breast implants cause the symptoms of that condition.
The Plaintiff also has not introduced evidence that a person with Sjogren's Syndrome would have their afflictions made worse by breast implants. Furthermore, the Court cannot simply assume that the implants, if they cause sicca symptoms, would necessarily make worse the Sjogren's Syndrome. Consider: the Plaintiff's eyes produce no lubricating tears because of her Sjogren's Syndrome. If the Plaintiff's eyes produce no tears whatsoever, then even if her breast implants do interfere with tear production, she could not establish that her implants somehow lessened her tear-producing ability. The situation is very analogous to giving a blind person a drug that, as a side effect, causes vision loss: as the person is already blind, a Court cannot simply assume (without medical testimony) that the drug made the person's blindness "worse."
In addition to this problem of causation, the Court also notes problems regarding the adequacy of the Giltay study. Giltay et al. sampled 445 women, 210 of whom had breast implants. The small size of the study leads the Court to conclude that it is not reasonably relied upon by experts to form conclusions regarding an association between breast implants and sicca symptoms. Furthermore, even if the Giltay paper had a sufficient sample size, the Court would exclude testimony based on it for reasons similar to those described for the Hennekens study — the Giltay study does not sufficiently satisfy the Koch and Henle factors nor does its methodology suggest that sampling bias is not an adequate explanation for the results. See 53 Ann.Rheum.Dis. at 194-96. For all these reasons, testimony based upon the Giltay study should be excluded as per Federal Rule of Evidence 703.
Adequacy of a sample size is an important consideration in assessing the validity of a study, and the adequacy of a sample size must be viewed in terms of the specified level of statistical significance, the chance of missing a real effect, the estimated magnitude of disease risk, and the background rate of disease. These last two factors are the most critical. Reference Manual on Scientific Evidence, supra, at 140-44. In general, when there is a high background rate of a specific condition and the increased risk is small, a researcher needs a fairly large sample size to ensure that his results are not simply due to chance. Id. Dr. Swan testified that, to adequately and accurately find the relative risk associated with a rare disease like Sjogren's Syndrome, a sample size of one hundred thousand persons would be required. Transcript at 74:11-20. The Court agrees with that analysis, but notes that if those numbers are true, then to study a condition with an even higher background rate — such a dry eyes — requires higher numbers. Reference Manual on Scientific Evidence, supra, at 141.
C. Rule 401-403 exclusion of Dr. Swan's Testimony
The Court also must exclude the proffered testimony under Federal Rule of Evidence 403. Dr. Swan testifies that she believes there to be insufficient evidence to show any causal link between breast implants and Sjogren's Syndrome or its symptoms. Transcript, 95:16-20. She also testifies that there is only limited evidence for the proposition that breast implants have any correlation with Sjogren's Syndrome, and informs that Court that more studies are needed. Transcript at 94:14-21. This suggests that her testimony is of exceptionally low probative value: what Dr. Swan in essence tells the jury is that there is insufficient evidence to conclude that breast implants cause Sjogren's Syndrome or any Sjogren's-like ailment. Id. Such testimony does not do much to advance the Plaintiff's case, as all it does is barely support the proposition that the studies relied upon may show a correlation between Sjogren's Syndrome and breast implants. Conversely, the potential for this testimony to have undue prejudice is extremely high: it is expert testimony on a confusing subject, and its nuances are more likely to overwhelm and confuse the jury than to help it reach a decision. Accordingly, Dr. Swan's testimony should be excluded under Federal Rule of Evidence 403, in addition to the factors already discussed. See Christophersen, 939 F.2d at 1112.
The Court also must consider the effect that the exclusion of Dr. Espinoza has upon this case. Without Dr. Espinoza, the plaintiff can offer no evidence that would establish that her breast implants specifically caused her disease. Because the Plaintiff cannot establish specific causation, Dr. Swan's testimony on epidemiology also could be excluded under Rule 401.
IV. Testimony of Dr. Luis Espinoza
The Plaintiff also offers the testimony of Dr. Luis Espinoza, a noted rheumatologist. Dr. Espinoza, according to the Plaintiff, will testify both as to general causation (like Dr. Swan) and as to specific causation. As detailed below, this testimony must be excluded because it satisfies neither Rule 703 nor Daubert.
A. General Causation
Dr. Espinoza cannot testify that breast implants generally cause Sjogren's Syndrome. The literature Dr. Espinoza wishes to rely upon — particularly the Giltay Paper and the Hennekens study — is not reasonably supportive of the proposition that breast implants are associated with Sjogren's Syndrome, much less for the proposition that breast implants cause Sjogren's Syndrome. See Section III, B, supra, Therefore, testimony based upon these studies should be excluded under Brock and under Federal Rule of Evidence 703. Id. Furthermore, as Dr. Espinoza himself admits, his own studies have not been published or peer reviewed, his sampling technique is fraught with bias, his theory regarding general causation is not generally accepted in the medical community, and he has not tested his theories. See Transcript at 15:19-15:10; Deposition of Luis Espinoza, 63:18-64:22. Therefore, his testimony regarding general causation should be excluded under Daubert as well as Rule 703.
In addition, testimony based on the Giltay paper should be excluded, as it does not address whether breast implants cause Sjogren's Syndrome.
Finally, it should be noted that the Plaintiff's own expert on epidemiology has testified that the scientific studies at this point are insufficient to establish that breast implants cause Sjogren's Syndrome. See Section III, C, supra. If the Plaintiff's own experts hold that position, then Dr. Espinoza's testimony on causation must be of low probative value. Therefore, the Court could exclude this testimony under Federal Rule of Evidence 403 as well.
B. Specific Causation or "Cause in Fact"
The Plaintiff also wishes Dr. Espinoza to testify that the Plaintiff's breast implants caused her Sjogren's Syndrome. As noted above, no published epidemiological study has found a statistically significant increased risk between breast implants and Sjogren's Syndrome. See Section III, A-C, supra. Those studies that do exist are inadequate to draw any connection between Sjogren's Syndrome and breast implants, as the Plaintiff's own expert recognizes. See Section III, C, supra. While epidemiological evidence is not a necessary element in every toxic tort case, it is certainly a very important element, especially when there is no evidence of the biological mechanism which links the product to the complained-of condition. See Brock v. Merrell Dow, 874 F.2d at 312-13. In the absence of any evidence regarding general causation, the Court will not permit Dr. Espinoza to testify as to specific causation. Id.; Fed.R.Evid. 402.
Moreover, even if the Court were to find that breast implants can cause Sjogren's Syndrome, the Court would still exclude Dr. Espinoza's testimony on specific causation because Dr. Espinoza's reasoning fails the test of parsimony. Dr. Espinoza has found that the Plaintiff possesses anomalous antibody levels in her blood chemistry; this anomaly leads him, based upon his observations with other women, to conclude that the Plaintiff's condition is implant-related. Transcript at 16:13-17:24. However, the witness admits that if the Plaintiff did not have breast implants but had the exact same symptoms and blood chemistry, then his diagnosis would have been non-implant-caused Sjogren's Syndrome. Transcript at 16:13-17:6; Espinoza Deposition at 37. Essentially, this is a bit like saying that if a person has a scratchy throat, runny nose, and a nasty cough, that person has a cold; if, on the otherhand, that person has a scratchy throat, runny nose, nasty cough, and wears a watch, they have a watch-induced cold. Such reasoning is extremely suspect, which has prompted other courts to reject it as unscientific in the absence of convincing epidemiology evidence. Brock v. Merrell Dow Pharmaceuticals, Inc., 874 F.2d 307, 310 n. 11 (5th Cir. 1989), citing In re Agent Orange Prod. Liab. Litig., 611 F. Supp. 1223, 1237-38 (E.D.N.Y. 1984) (Weinstein, J., presiding) (rejecting testimony that, if a person was exposed to Agent Orange and had a disease, that disease was caused by Agent Orange, but if a person had a disease with no exposure to Agent Orange, that person had a naturally occurring condition).
Essentia non sunt multiplicanda praeter necessitatem. This principle, known as the rule of parsimony, is also known as Occam's Razor. See 20 Encyclopedia Britannica 868 (11th ed. 1911). The Court here applies the rule of parsimony not to establish a new standard of proof for the Plaintiff to surmount, but rather as a device to be used after a careful and proper weighing of the evidence under the standards of Daubert, Christophersen, and Federal Rule of Evidence 703. In general terms, this is the same sort of reasoning as applied by Judge Weinstein In re Agent Orange Prod. Liab. Litig., 611 F. Supp. 1223, 1237-38 (E.D.N.Y. 1984). The Fifth Circuit has ruled that Occam's razor may be used once a court has engaged in a proper weighing of testimony and evidence. See Justiss Oil Co. v. Kerr-McGee Ref. Corp., 75 F.3d 1057, 1060-61 (5th Cir. 1996).
To be fair to Dr. Espinoza, not all philosophers (or jurists) endorse a wide application of Occam's Razor. See Clark v. Taylor, 163 F.2d 940 (2d Cir. 1947); Kenneth Burke, A Grammar of Motives 324 (1945). See also Aristotle, Nichomachean Ethics, X, ch. 8, 1179(a), cited in Clark v. Taylor, 163 F.2d 940, 950-51 n. 11 (1947). However, given the general lack of evidence linking breast implants to Sjogren's Syndrome, the rule of parsimony offers a convenient shorthand for illustrating some of the flaws in Dr. Espinoza's scientific reasoning, especially after the Court has weighed the evidence.
To illustrate why such reasoning should be excluded, imagine a very simplified example of epidemiology and Dr. Espinoza's methods. Assume that Dr. Espinoza saw a set of 100,000 women without breast implants, 10,000 of whom possess symptoms and blood chemistry identical to that of the Plaintiff. Faced with these symptoms, Dr. Espinoza would diagnose all 10,000 of them with non-implant Sjogren's Syndrome. Transcript at 16:13-18. Further, assume that Dr. Espinoza saw a set of 100,000 virtually identical women with breast implants, 16,000 of whom suffered from symptoms identical to that of the Plaintiff. Dr. Espinoza would diagnose all 16,000 as having implant-related Sjogren's Syndrome. Transcript at 17:2-24.
Because Dr. Espinoza has not followed the rule of parsimony, his method of diagnosis introduces an unacceptable rate of error. Consider: in our example, women with breast implants get Sjogren's Syndrome 1.6 times as often as women without breast implants. But a woman with breast implants and Sjogren's Syndrome does not know if her Sjogren's Syndrome is associated with her breast implants or with a natural event. In our example, it is fair to say that about 10,000 out of the 16,000 women with implants and Sjogren's Syndrome came down with their symptoms for reasons other than breast implants — or, in other words, 10,000 of the women in the breast implant group exhibit the symptoms of Sjogren's Syndrome for the same unknown reason that 10,000 of the women in the non-implant group have the disease. Conversely, about 6,000 women in our example would be able to attribute their condition to the increased risk associated with breast implants. However, Dr. Espinoza would diagnose all the afflicted women with breast implants as having implant-related Sjogren's Syndrome. Given the 1.6 relative risk described in this example, Dr. Espinoza will be wrong in his diagnosis of causation 62.5% of the time. Such a high rate of error forces the Court to conclude that this testimony regarding specific causation must be excluded under Daubert. 509 U.S. at 594, 113 S.Ct. at 2796.
It is interesting to note that a higher relative risk makes the error rate associated with this technique drop considerably. Thus, with a relative risk of 2.6, the error rate associated with the technique is only approximately 38.5%. With a relative risk of 9, the error rate of Dr. Espinoza's diagnosis will be about 11%. And with a relative risk of 15, the error rate of this sort of diagnosis would only be about 6.7%.
Dr. Espinoza's testimony regarding specific causation also fails for a second reason. The Fifth Circuit has held that medical testimony on the issue of specific causation satisfies Daubert if the doctor bases his opinion on an exam (or a review of the medical record), his experience, and a broad reading of the medical literature. See Carroll v. Morgan, 17 F.3d at 789-90. In the present case, Dr. Espinoza has examined the Plaintiff, and has wide experience. Transcript at 19:10-12. However, Dr. Espinoza has not based his diagnosis on a wide reading of medical literature. Instead, Dr. Espinoza has based his diagnosis of causation on studies that do not support a finding that breast implants cause Sjogren's Syndrome. See Section III, A-C, supra. Thus, unlike the doctor in Carroll who could point to specific passages in several different medical textbooks for support of his diagnosis, Dr. Espinoza can only point to a few limited and insufficient studies. See Section II, A-B, supra. Because Dr. Espinoza bases his diagnosis on only a few studies rather than a broad literature review, because he bases his diagnosis on studies that do not support his position, and because he bases his diagnosis on studies that are not reasonably relied upon for the propositions he asserts, the Court must exclude this testimony under Daubert and Federal Rule of Evidence 703. See Carroll v. Morgan, 17 F.3d at 789-90.
Indeed, Dr. Espinoza has only testified that he has relied upon the Hennekens and Giltay papers in reaching his conclusion. While he may have relied upon others, he has not made those facts or data clear to the Court.
Finally, the Court notes that grave problems may exist in Dr. Espinoza's methods of using epidemiological research. During the proffer of Dr. Espinoza's testimony, the following rather amazing exchange took place between Plaintiff's counsel and Dr. Espinoza:
DR. ESPINOZA: I have read the [Hennekens] paper, yes.
COUNSEL: Yes . . . and did you see a higher relative rate of incidence of disease in all three categories for women with breast implants?
DR. ESPINOZA: Correct.
COUNSEL: And again, whatever the statisticians, whatever spin they may want to put on it, does that have significance to you as a medical doctor in determining disease causation?
DR. ESPINOZA: Yes, it has.
Transcript at 158. A fair reading of this interchange is that Dr. Espinoza does not care about the level of statistical significance associated with a relative risk — whatever the statistical "spin" on the relative risk found in the Hennekens study, Dr. Espinoza finds the relative risk significant as a medical doctor. If the Court takes Dr. Espinoza at his word, then it seems that Dr. Espinoza is not interested in the level of statistical significance associated with the Hennekens study, but rather is only interested in the reported relative risk — a number that cannot be adequately analyzed without taking into account the statistical "spin." See Reference Manual on Scientific Evidence at 152-166. Testimony based on a method that seemingly ignores statistical significance should be excluded under Daubert and Fed.R.Evid. 403.
As stated previously, the Court takes seriously the gate-keeping role assigned to it under Daubert. As a general matter, the Plaintiff's attempt to rely on current scientific literature to establish causation of the Plaintiff's case has been doomed to failure by that literature's infancy. As the Plaintiff's own expert has testified, the studies are insufficient to show that breast implants cause Sjogren's Syndrome or sicca symptoms. Testimony using these studies to link Sjogren's and breast implants is excludable under Rule 703 and Daubert. Testimony using these studies to establish specific causation should, for similar reasons, not be admitted.
Due to this Court's exclusion of expert testimony, the Plaintiff can present no evidence that breast implants cause Sjogren's Syndrome or sicca symptoms. Accordingly, the Court will grant the Defendant's motion for judgment as a matter of law with regard to Plaintiff's claims on this issue. See Fed.R.Civ.Pro. 50(a) (judgment as a matter of law permitted when there is no legally sufficient evidentiary basis for a reasonable jury to find for the party). Therefore, it is hereby ORDERED that the Defendant's motions to exclude Dr. Shanna Swan and Dr. Luis Espinoza are GRANTED. It is further ORDERED that the Defendant's motion for judgment as a matter of law is GRANTED IN PART, such that Plaintiff's claims relating to Sjogren's syndrome or its symptoms are DISMISSED pursuant to Rule 50 of the Federal Rules of Civil Procedure.