Linnen v. A.H. Robins Co.

11 Mass. L. Rptr. 205

• Court: Massachusetts Superior Court
• Decided: January 4, 2000
• Docket: No. 972307
• Status: Published

Opinion

Brassard, J.

On November 30, 1999 this matter was before the court for evidentiary hearings on the motion of defendant Wyeth-Ayers Laboratories (“Wyeth”), a division of American Home Products (“AHP”), to exclude the expert testimony of plaintiffs’ experts, Dr. Stuart Rich (“Dr. Rich”) and Dr. Robyn Barst (“Dr. Barst”).³ Plaintiffs Thomas and Mary Linnen allege that the death of their 30 year old daughter Mary Linnen (“Ms. Linnen”) in February 1997 from primary pulmonary hypertension (“PPH”) resulted from her short-term ingestion of the diet pills fenfluramine and phentermine (“fen/phen”) in May 1996. To support this allegation plaintiffs offer the expert opinion of Dr. Rich, a physician at the Rush Heart Institute, Rush-Presbyterian-St. Luke’s Medical Center in Chicago, Illinois. Dr. Rich is a Diplómate of the American Board of Internal Medicine in Cardiovascular Diseases and a board certified Diplómate of the American Board of Internal Medicine. Dr. Rich has conducted research and published numerous articles in medical journals regarding various aspects of pulmonary hypertension, and is a co-author of the International Primary Pulmonary Hypertension Study (“IPPHS”), the only epidemiological study with respect to the association between PPH and exposure to anorectic agents such as fenfluramine.⁴

AHP, the manufacturer of the fenfluramine prescribed for Ms. Linnen, moves to exclude the expert testimony of Drs. Rich and Barst. AHP seeks to exclude the testimony on the basis that it fails to point to any scientifically reliable evidence that fenfluramine caused or contributed to Ms. Linnen’s PPH. AHP argues that there is no evidence to support Dr. Rich’s testimony that the IPPHS established an association between PPH and the short-term ingestion of fenfluramine. Dr. Rich and one defense expert, Dr. Walter F. Stewart, Ph.D. (“Dr. Stewart”) submitted their proposed direct testimony in writing and were cross examined at the hearing by opposing counsel. Dr. Stewart, an epidemiologist, is an Adjunct Associate Professor at the Johns Hopkins School of Hygiene and Public Health in Baltimore, Maryland. Dr. Stewart is also president of Innovative Medical Research in Baltimore, Maryland.

For the following reasons, the motion to exclude the expert testimony of Drs. Rich and Barst and the results of the IPPHS is DENIED.

BACKGROUND

Ms. Linnen was diagnosed with PPH in November 1996 and died in February 1997. The issue in this case is whether her death was caused by her ingestion of fen/phen for approximately 3 weeks in May 1996.

Based primarily on the results of the IPPHS, Dr. Rich will testify that there is a significant association between Ms. Linnen’s PPH and her ingestion of fen/phen for approximately three weeks in May of 1996. Dr. Stewart responds that the epidemiological study co-authored by Dr. Rich itself establishes that there is no statistically significant association between PPH and exposure to fenfluramine for 90 days or less. Dr. Stewart also maintains that in his testimony Dr. Rich attempts to contradict that finding. AHP contends that Dr. Rich’s testimony does not satisfy the Daubert standard of scientific reliability, and moves to exclude both Dr. Rich’s testimony and reliance by Dr. Rich upon the results of the IPPHS to establish causation in this case.

Odds Ratio

An epidemiological study is considered statistically significant if it determines that there is a magnitude of association (“odds ratio") of 2.0 or greater between the use of a substance (here, anorectic drugs) and a medical condition (here, PPH). The IPPHS found an odds ratio of 1.8 with a 95% confidence interval of 0.5 to 5.7 for the use of anorexigens for less than three months. The major dispute between these experts is how to interpret that 1.8 figure. Dr. Stewart testified that “[bjecause the lower bound of the 95% confidence interval falls below the value of 1.0, the odds ratio of 1.8 is not statistically significant.” In Dr. Stewart’s view, an odds ratio of 1.8 establishes that a user such as Ms. Linnen, who took fenfluramine for less than three months, did not have a statistically significant chance of developing PPH. AHP contends that because the IPPHS, the only epidemiological study available, shows an insignificant statistical association between short-term ingestion of fenfluramine and PPH, the results of the study should not be admissible or relied upon to establish causation in this case.

The plaintiffs respond that 1) 1.8 is not an accurate measure for Ms. Linnen’s actual risk based on the sample size and the correct determination of Ms. Linnen’s category within the study and 2) even if the 1.8 figure were accurate, an epidemiological result that is not statistically significant by itself may prove causation when considered in connection with other criteria such as case studies and clinical experience.

According to Dr. Rich, the IPPHS was not intended to study causation, because causation between fenfluramine and PPH had already been established. Rather, he explained that the purpose of the IPPHS was to establish whether and how the odds ratio was affected by continued use. Dr. Rich noted that after three months of exposure to fenfluramine, the odds ratio increased from 1.8 to 23.1. He testified that this finding illustrates only that the risk of PPH increases dramatically with longer use, and that it would be a gross misinterpretation of the IPPHS to conclude, as defendants do, that the data show no significant risk of developing PPH from using fenfluramine for less than three months. Dr. Rich testified that IPPHS was not and could not be designed to determine the minimum threshold exposure required to cause PPH.

The Structure of the Study

The IPPHS included 95 cases for case control analysis, with 355 matched controls. Subjects were divided into categories depending on recency of use and onset of symptoms of PPH. Subjects who experienced the onset of symptoms within one month of exposure were considered “possible users.” No odds ratio was calculated for possible users.

Subjects who experienced symptoms after one month were categorized as “definite users.” Definite users were further analyzed as to “recent” and “past” exposure. Recent users had used anorectic drugs within one year of the onset of symptoms. Past users were those whose symptoms occurred after the drug exposure had been discontinued for more than a year. The IPPHS established that the use of anorectic agents, fenfluramine in particular, regardless of re-cency of use or length of exposure, carried a statistically significant odds ratio of 6.3 (95% confidence interval: 3.0-13.2). Odds ratios were calculated separately for recent and past users. For recent users, the odds ratio was 10.1 (95% confidence interval: 3.4-29.9).

Before applying the Daubert analysis to Dr. Rich’s testimony, the court reviewed the study on which Dr. Rich relied, the conclusions he drew from that study, and the challenges raised to his theories by the defense experts, both at the hearing and in comprehensive submissions to the court.

Dr. Stewart’s Testimony

Dr. Stewart contends that because Ms. Linnen took fenfluramine for less than one month, she must be categorized as a “possible user,” for whom no odds ratio was calculated. Dr. Stewart further testifies that the only odds ratio that can properly be applied to Ms. Linnen is that found for subjects who took anorectics for less than 90 days, or 1.8. Since an odds ratio of 1.8 is by definition not statistically significant, Dr. Stewart concludes that the IPPHS establishes that there is no statistical association between PPH and the short-term ingestion of fenfluramine.

According to Dr. Stewart, because the 1.8 odds ratio for a user such as Ms. Linnen is not statistically significant, the IPPHS cannot scientifically establish that the use of appetite suppressants for less than three months causes PPH. Because there are no other epidemiological studies that measure this association, he maintains that it is possible to state unequivocally that epidemiology has not established to a reasonable scientific certainty that exposure to fenfluramine for three months or less causes PPH.

Dr. Rich’s Testimony

Accurate Odds Ratio is 6.3, not 1.8

Dr. Rich disputes these assertions, contending that the odds ratio that most accurately reflects Mary Linnen’s experience is 6.3, the ratio found for “all users,” regardless of recency of use or length of exposure. Dr. Rich states that the odds ratio of 1.8, urged by AHP as the appropriate category for Ms. Linnen, cannot be read in isolation. First, Dr. Rich explains that the sample of subjects who took the drug for less than three months is too small to have statistical significance. Dr. Rich points out that the IPPHS was “powered” to study the effects of continued use, not short-term causation.

In addition, Dr. Rich maintains that in Ms. Linnen’s case, the analysis must include the "recent use” subcategory, with its odds ratio of 10.1, because Ms. Linnen developed symptoms within one year of exposure to the drug. Dr. Rich testified that an accurate odds ratio for Ms. Linnen would therefore require a post hoc subset analysis of patients who were exposed for less than three months but developed symptoms within a year of exposure (recent use). Because the odds ratio for use of less than 3 months is 1.8 and the odds ratio for recent use is 10.1, the correct figure is a combination of these figures, or approximately 2.8, which is statistically significant.

Dr. Stewart vigorously disputes this assertion, arguing that Ms. Linnen was categorized as a possible user (exposure for less than a month), and no odds ratio was calculated for possible users. Consequently, Dr. Stewart argues that the recent use subset, which was designed to measure the odds ratio for definite users (exposure for more than a month), may not properly be applied to determine Ms. Linnen’s odds ratio, and that the 1.8 odds ratio must stand alone.

Reliance on other clinical evidence

Dr. Rich testified, however, that even if the odds ratio of 1.8 is applicable, that ratio does not by itself establish the lack of an association. In addition to the problem of the small sample size, Dr. Rich testified that while an odds ratio of less than 2.0 is not statistically significant, it is clinically and diagnostically significant, and is routinely considered by physicians as evidence of causation. Odds ratios of less than 2.0 may be probative of causation when supported by other clinical evidence. See Vassallo v. Baxter Healthcare Corporation, 428 Mass. 1, 13 (1998). Dr. Rich testified that in Ms. Linnen’s case there is overwhelming clinical evidence, such as case studies, supporting a relationship between her PPH and her short-term exposure to anorectic agents.

Dr. Rich bases his conclusion on Ms. Linnen’s medical record and his own career experience with PPH including case reports, his research and clinical observations and what he, as the co-author of the IPPHS, considers an accurate interpretation of the data. Dr. Rich testified that, because the IPPHS was not designed to measure causation, an odds ratio of 1.8 does not mean that a patient who ingested anorectic agents for less than 90 days must have acquired PPH from some other cause. Dr. Rich maintains that the study was never meant to be interpreted in that manner.

DISCUSSION

Daubert-Lanigan Analysis

In determining the admissibility of scientific evidence, the court must act as a gatekeeper to determine whether the proposed evidence is reliable. See Daubert v. Merrell Dow, 509 U.S. 579 (1993); Commonwealth v. Lanigan, 419 Mass. 15 (1994). An expert is permitted to present opinions “that will assist the trier of fact to understand the evidence or to determine a fact in issue.” Daubert at 588. Because of their specialized knowledge, experts are subject to more liberal rules of evidence than other witnesses. They may testify to opinions and may base those opinions on facts and data not in evidence. Id. Moreover, a jury may tend to give special deference to the opinion of an expert. See Commonwealth v. Vitello, 376 Mass. 426, 453 (1978).

To ensure that expert opinions are sufficiently reliable to assist the trier of fact, an expert must be able to demonstrate that his conclusions are based on scientific knowledge, with “scientific” defined as implying “a grounding in the methods and procedures of science,” and “knowledge” defined as connoting “more than subjective belief or unsupported speculation.” Daubert at 590. To satisfy this standard, an expert witness must show that his opinions “have been arrived at in a scientifically sound and methodologically reliable fashion.” See Ruiz-Troche v. Pepsi-Cola of Puerto Rico, 161 F.3rd 77 (1st Cir. 1998). The issue addressed in this motion is whether Dr. Rich’s testimony satisfies that standard.

In determining whether expert opinions constitute reliable scientific testimony, a court may apply the nonexhaustive standards set forth in Daubert, supra. Daubert and its progeny teach that to determine whether expert testimony is admissible, the court must make a “preliminary assessment of whether the reasoning or methodology underlying an expert’s testimony is scientifically valid.” Daubert at 592-93.

In making such an assessment, the court must determine whether the reasoning or methodology properly can be applied to the facts in issue. Commonwealth v. Lanigan, 419 Mass. 15, 26 (1994), quoting Daubert at 592-93. There are factors that the court may consider in determining the reliability of a scientific opinion, including but not limited to: 1) whether the theory has been subjected to peer review and publication, 2) whether the theory has been tested, 3) whether the theory has a known error rate and 4) whether the theory has general acceptance in the general scientific community. Lanigan at 26.

There is no dispute that Dr. Rich has the experience, qualifications, education and training to offer expert testimony about the results of an epidemiological study of the association between pulmonary hypertension and certain anorectic agents. As the co-author of the IPPHS, Dr. Rich is qualified to testify with respect to the methodology underlying the study and what he considers the proper interpretation of its results.

Furthermore, Dr. Rich’s opinions are based on sound scientific methodology. Daubert at 592. The IPPHS is recognized as a reliable epidemiological study which has been accepted in the scientific community. The results of the study were published in an article in the New England Journal of Medicine entitled “Appetite Suppressant Drugs and the Risk of Primary Pulmonary Hypertension” on August 28, 1996. The article received favorable peer review prior to publication. Peer review and publication is important in determining if a scientific theory is reliable because it “increases the likelihood that substantive flaws in methodology will be detected.” Daubert at 593.

The fact that experts offer differing interpretations of the data contained in the IPPHS does not render one expert’s testimony inadmissible. As long as the opinions are based on sound methodology, vigorous cross examination and presentation of contrary evidence are the appropriate means of attacking disputed scientific evidence. Daubert at 596.

In this case, the experts dispute whether the IPPHS established a significant association between PPH and short-term exposure to fenfluramine. Because Dr. Rich’s testimony has a reliable basis in the knowledge and methods of his discipline, his expert testimony (including reliance on the IPPHS) is admissible. Id. at 592. Under the circumstances of this case, it is for the jury to determine the value of competing expert testimony.

ORDER

For the foregoing reasons, defendant Wyeth-Ayers’ motions to exclude the testimony of Drs. Rich and Barst and to exclude expert testimony relying on the IPPHS are DENIED.

3 Dr. Rich testified for the plaintiffs with the understanding of the parties that the court's ruling on his testimony would also apply to the testimony of Dr. Barst.

4 The IPPHS, an international epidemiological study conducted in five European countries, measured the statistical association between certain anorectic agents and the development of PPH. The results, released in July 1996 and published in the New England Journal oj Medicine in August 1996, showed a statistically significant association between PPH and the ingestion of fenfluramine for longer than 90 days.