Finding "Grist For The Trier Of Fact" Under FRE 702

In civil RICO case, admitting expert regression analysis in regarding whether fraudulentl marketing off-label prescription drugs to doctors because the defendant's "attacks" on the expert's methodology "were all grist for the trier of fact" to resolve, rather than exclusion by the court under FRE 702; circuit considers the admissibility of regression analysis, in In re Neurontin Marketing and Sales Practices Litigation, __ F.3d __ (1st Cir. April 3, 2013) (Nos. 11–1904, 11–2096)

There is a fine line parties and the trial court must tread regarding the use of expert opinion testimony. On one level, an expert opinion may be excluded as unreliable under FRE 702. On the other, admission of the expert evidence does not mean that the jury needs to give it much, if any, weight. Earlier this month the First Circuit described the difference between opinion that is unreliable as opposed to opinion testimony that is simply disputable. The case provides a quick review of the presentation of statistical regression analysis in the presentation of a case.

In the case, medical providers such as Kaiser, sought to recover funds they had used to pay patient's prescriptions for off-label uses of one of the defendant's products. The off-label use had been promoted by defendant Pfizer in its fraudulent marketing campaigns. After the defendant suffered an adverse jury verdict in the case, their appeal raised a number of challenges to the admission of an expert's statistical analysis of causation in the case. The expert's qualifications apparently were not challenged, as she served as a professor at a national school of public health. However, the defendant did challenge this expert's methodology for assessing whether there was a "causal connection between the [defendant's] fraudulent marketing" and the quantity of prescriptions made for off-label uses among the plaintiff's patients.

Apparently this inquiry posed a "chicken and egg"-type problem in terms of what caused what. The expert traced the relation between expenditures for the drug's off-label marketing and the increase in prescriptions written for off-label uses. Her methodology and results:

explained the difference between correlation and causation and ... established causation by performing a regression analysis on sales information against promotional spending, while controlling for other variables. Her analysis excluded the many off-label prescriptions by physicians who received legitimate on-label promotion. She concluded that the “percentage[s] of Neurontin prescriptions that were caused by Pfizer's fraudulent marketing of Neurontin” were, by off-label indication, as follows: 99.4% of prescriptions for bipolar disorder; 70% of prescriptions for neuropathic pain; 27.9% of prescriptions for migraine; and 37.5% of prescriptions for doses over 1800 mg/day. Thus, three out of ten Neurontin prescriptions written by neurologists for migraine would not have been written or filled but for the alleged misconduct."
In re Neurontin Marketing and Sales Practices Litigation, __ F.3d at __ (citations omitted).

The First Circuit had little issue with the expert's use of regression theory, noting it was "a well recognized and scientifically valid approach to understanding statistical data, and courts have long permitted parties to use statistical data to establish causal relationships." The circuit noted a number of cases demonstrating this finding, including
  • Wards Cove Packing Co., Inc. v. Atonio, 490 U.S. 642, 657–58 (1989) (holding that under Title VII of the Civil Rights Act of 1964, “specific causation” is shown and a “prima facie case” is “establish[ed]” when plaintiff identifies a specific employment practice linked to a statistical disparity)
  • Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 994 (1988) (opinion of O'Connor, J.) (explaining that, to establish a prima facie case under Title VII, “[o]nce the employment practice at issue has been identified, causation must be proved; that is, the plaintiff must offer statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their membership in a protected group”)
  • Duren v. Missouri, 439 U.S. 357, 366–67 (1979) (permitting petitioner to establish prima facie violation of fair cross-section requirement of Sixth and Fourteenth Amendments by using “statistics and other evidence” to show that “the underrepresentation of women, generally and on his venire, was due to their systematic exclusion in the jury-selection process”)
  • Times–Picayune Pub. Co. v. United States, 345 U.S. 594, 621 (1953) (in antitrust case, looking to “economic statistics” to determine whether “demonstrably deleterious effects on competition may be inferred”)
  • In re High Fructose Corn Syrup Antitrust Litig., 295 F.3d 651, 660–61 (7th Cir. 2002) (permitting use of regression analysis to show causation in antitrust case)
  • Conwood Co., L.P. v. U.S. Tobacco Co., 290 F.3d 768, 794 (6th Cir.2002) (finding regression analysis “to be admissible on the issue of causation” in antitrust case (emphasis omitted) (quoting Jahn v. Equine Servs. PSC, 233 F.3d 382, 390 (6th Cir. 2000))).

The circuit rejected a variety of other challenges the defendant made to the expert's methodology, finding that where the expert had a choice the expert used the "gold standard" of data sets and that the expert's assumptions accompanying its regression analysis were appropriate. It also noted the deficiency of the defense argument against the expert testimony. One ground advanced by the defense was its instance that the data to be analyzed should have been based on physician reports as to whether they did or did not "rely on marketing when prescribing" the drug. The circuit rejected this, noting that there was little incentive for doctors to truthfully report that they made medical decisions based on a factor other than knowledge of drug efficacy. Another ground advanced was provided by the defense expert witness on statistics. This witness "testified that when he re-ran [plaintiff expert]'s regression analysis with different assumptions, he did not find a statistically significant relationship between Pfizer's promotion of [the suspect drug] and prescriptions" made for that drug. As the defense expert, except for testifying that he could not duplicate the plaintiff's expert's analysis after changing its assumptions, the defense expert "did not present his own causation or damages model," and so the trial "court rejected [the defense expert's] criticisms and accepted" those of the plaintiff's expert.

This First Circuit's approach to the use of an economic model verified by regression emphasized the role of the trial court gatekeeper as not to resolve all challenges to expert analysis. But the circuit recognized that even the most compelling problems were not solved at the gatekeeping stage. Rather, these persistent assumptions, qualifications and limitations "presented a question for the jury." The circuit noted that "'[s]o long as an expert's scientific testimony rests upon "good grounds, based on what is known," it should be tested by the adversarial process, rather than excluded for fear that jurors will not be able to handle the scientific complexities.” This model of considering expert testimony was "well accepted in the antitrust context from which RICO has drawn many of its causation principles," so that it was clear that the plaintiff's expert's testimony and the defense objections to it in this case "were all grist for the trier of fact; they warranted 'test[ing] by the adversarial process, rather than exclu[sion].'” In re Neurontin Marketing and Sales Practices Litigation, __ F.3d at __ (quoting Milward, 639 F.3d at 15).


Subscribe Now To The Federal Evidence Review

** Less Than $25 Per Month ** Limited Time Offer **

subscribe today button

Federal Rules of Evidence