Avoiding the Adverse Effects of Causality Assessments

As with many product manufacturers, pharmaceutical companies monitor adverse events allegedly associated with their products. Pharmaceutical companies are subject to extensive governmental regulations both in the United States and abroad. Under these regulations, pharmaceutical companies must compile and submit adverse drug reaction reports to the U.S. Food and Drug Administration ('FDA') and foreign regulatory bodies on an ongoing basis. See 21 CFR 314.80 (2007); The Rules Governing Medicinal Products in the European Union, Volume 9A (2007) available at http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-9/pdf/vol9_2007-07_upd07.pdf. Adverse drug reaction reports are essentially voluntarily submitted anecdotal reports of a complication or event that the reporter may believe to be associated with a drug. The source of information regarding the underlying event may be a health care provider, but is just as likely to be a patient, family member, news report, or plaintiff's attorney. Accordingly, the reporters are self-selecting and, to a certain extent, are skewed toward an assessment that the drug was somehow associated with the adverse event. After all, it is unlikely that someone would contact a pharmaceutical manufacturer unless they believed that the manufacturer's drug was related to the adverse event.

The FDA ' and pharmaceutical companies ' rely on these anecdotal reports, and analyses thereof, to generate hypotheses for further epidemiological and clinical investigation. Citing the inherent limitations of these reports, the FDA, however, has repeatedly cautioned against reliance on such reports in support of conclusions as to incidence rate. (See In re Baycol Prods. Litig., 495 F. Supp. 2d 977, 987 (D. Minn. 2007) (citing FDA guidance regarding the reliance on adverse event reports to show incidence rates). Recognizing the importance of post-marketing surveillance and the attendant dangers of such activities to pharmaceutical companies, FDA regulations include a disclaimer stating that reports submitted to the FDA are not admissions that a drug caused or contributed to an adverse effect. See 21 CFR 314.80 (k) (2007).

Unlike the FDA, the European Agency for the Evaluation of Medicinal Products ('EMEA') requires that pharmaceutical companies conduct causality assessments with respect to these adverse drug reaction reports. See 21 CFR 314.80 (2007); The Rules Governing Medicinal Products in the European Union, Volume 9A (2007) are available at http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-9/pdf/vol92007-07_upd07.pdf. Because most drugs are marketed worldwide, in complying with EMEA guidance, pharmaceutical companies generally assess causality with respect to each adverse event reported to the company, regardless of the country of origin. The 'causality assessment' terminology is unfortunate, as these assessments are not meant to assess actual causation in the legal sense. These assessments aim to determine whether there is a reasonable probability of relatedness between the drug and the adverse event for regulatory reporting and safety surveillance purposes. They are intended to act as an early warning system concerning possible complications associated with the drug.

Moreover, these assessments generally do not involve the rigorous investigation and evaluation of the relationship between the adverse event and the drug typical of the scientific method. Generally, a manufacturer will compile all adverse events purportedly related to its drugs in a database. See, e.g., In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *2 (M.D. Fla. May 2, 2007) (discussing the procedures utilized by Hoffmann-La Roche in complying with the EMEA causality assessment regulations). For each report, the manufacturer includes the reporter's assessment of whether there is a possibility the adverse event was related to its drug, regardless of whether the reporter is qualified to provide such assessment or whether the reporter's assessment is based on reliable data.

Manufacturers typically also include a company assessment of whether there is a possibility that the adverse event was related to its drug. The company's assessment, however, is not the product of an independent investigation and evaluation. For instance, in making causality assessments under the EMEA regulations for Accutane', Hoffmann-La Roche considers the reporter's assessment and a score generated by applying the facts provided by the reporter to the Naranjo algorithm. See, e.g., Id. The odds are very good that if the reporter asserts there is a possible association between the drug and the adverse event, the company's assessment will similarly reflect a possible association. In fact, the company's assessment will only indicate no possible association in instances in which the reporter asserts no possible relationship, and the score generated by the algorithm is very low. Finally, a drug safety physician employed by the manufacturer generates a brief narrative regarding each report. Such narratives, however, are based exclusively on the information provided by the reporter and, accordingly, are only as good as the information provided concerning an adverse event. Moreover, the aim of these narratives, like the causality assessments, is not to identify actual causation, but rather to determine whether there is a reasonable probability of a possible relationship between the event and the drug.

Plaintiff's Arguments Regarding Causality Assessments

Stymied by pesky hearsay rules precluding adverse drug reaction reports from being directly introduced in evidence to prove causation, plaintiffs have refocused their efforts on persuading courts to allow them to introduce or rely upon the accompanying company causality assessments. Plaintiffs have attempted to use these assessments as both direct evidence of causation (or alternatively for notice purposes) and indirectly as bases upon which their experts can provide opinions on causation.

Plaintiffs have attempted to persuade courts to admit causality assessments into evidence by arguing that such assessments are admissions that the manufacturer's drug caused the reported events, and thus, exempt from the definition of hearsay. They have also contended that such assessments are relevant to establishing that a pharmaceutical company had notice or knowledge of the potential dangers associated with its drugs, and therefore, not offered for the truth of the matter asserted. Finally, plaintiffs have attempted to bring causality assessments into the courtroom using a backdoor approach, arguing that experts should be allowed to base their testimony on these assessments.

Causality Assessments Are Inadmissible to Prove Causation

Causality assessments are not admissions of causation admissible under Federal Rule of Evidence 801(d)(2)(A). Plaintiffs have argued that under Rule 801(d)(2)(A) courts should admit causality assessments as admissions by the manufacturer that the drug caused the adverse event and, therefore, is evidence of general causation. Rather, these assessments are simply the product of the manufacturer's categorizing, based upon a simple algorithm and the subjective beliefs of third-party reporters as to the relationship between a drug and a particular ailment. The causation assessments are not the manufacturer's articulations of causes of these adverse events; at best, they merely reflect that the data and opinion expressed by the reporter raise the possibility that the drug might be related to the events. This is a far cry from a manufacturer conducting its own assessment of
the patient's medical records and acknowledging that its drug caused the patient's complications.

Courts should not admit causality assessments in evidence because: 1) they are not properly characterized as a party admission concerning causation; 2) as a consequence, they are not relevant to proving causation because they are generated for much more limited purposes; and 3) what limited evidentiary value they have is greatly outweighed by their prejudicial nature. Thus, causality assessments offered to prove causation should be excluded under Federal Rules of Evidence 402 and 403.

Evidence is only relevant if it tends to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence. While causation is certainly a matter of consequence, causality assessments are neither intended to ' nor do they in fact ' shed light on actual causation. Rather, they are made in furtherance of post-marketing surveillance and pursuant to EMEA regulations requiring the company to catalogue the reports based upon a causality rating. As a consequence, causality assessments are inadmissible under Federal Rule of Evidence 402 because they are not relevant to the question of causation.

Moreover, causality assessments are inadmissible under Federal Rule of Evidence 403. As noted above, the probative value of the causality assessments is low because, at best, they are only tangentially relevant to causation. On the other hand, the dangers of unfair prejudice, confusion of the issues, and misleading the jury associated with causality assessments are very high. Even with strongly worded limiting instructions from the court, a lay jury is likely to conclude that causality assessments are admissions of causation on the part of the pharmaceutical company.

The Multi-District Litigation court overseeing the Accutane cases recently issued a series of opinions excluding causality assessments. In those cases, plaintiffs offered Hoffmann-LaRoche's causality assessments to prove that Accutane causes inflammatory bowel disease and to establish that Hoffman-LaRoche had notice of the potential dangers of Accutane. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 (M.D. Fla. May 2, 2007). The plaintiffs contended that the causality assessments, buttressed by the testimony of Hoffmann-LaRoche's Director of Drug Safety that the assessments were 'fully investigated articulations of the adverse events,' were admissions that Accutane caused inflammatory bowel disease. As such, the plaintiffs argued that the causality assessments were not hearsay and should be admitted. The court, however, properly held that the causality assessments were not admissible to prove causation. Addressing plaintiffs' contention that the causality assessments were admissible to prove causation as admissions of a party opponent, the court concluded that the testimony of Hoffmann-LaRoche's Director of Drug Safety did not indicate that the assessments reflected a company's admission that the drug had caused an adverse event. Rather, according to the court, the testimony merely indicated that the assessments reflected a company's conclusion that there was a reasonable probability that the adverse event could possibly be related to the drug. Thus, the court held that the causality assessments were not relevant to the issue of actual causation and further noted the highly prejudicial nature of the assessments in explaining their inadmissibility.

Expert Testimony Based on Causality Assessments Is Inadmissible

Courts should also preclude plaintiffs from doing an end run around the rules of evidence by presenting expert testimony concerning causation based on causality assessments. Courts should exclude such expert testimony pursuant to Federal Rule of Evidence 702 and the holding of Daubert v. Merrell, 509 U.S. 579 (1993).

Expert testimony relying on causality assessments is inadmissible because reliance on causality assessments is not scientifically valid. Rule 702 requires courts to determine whether the expert is proposing to testify as to scientific knowledge that will assist the trier of fact. In Daubert, the U.S. Supreme Court explained that in determining whether scientific expert testimony is admissible under Rule 702, courts must determine whether the reasoning or methodology underlying the testimony is scientifically valid and whether that reasoning or methodology can properly be applied to the facts in issue. Reliance upon assessments made for the purpose of post-marketing surveillance and hypothesis generation simply is not a scientifically valid method for assessing causation. Thus, the application of Rule 702 should eliminate the need for any consideration of whether the testimony and otherwise inadmissible assessments should be admitted under the more stringent standards of Rule 703.

In the Accutane cases, one of the plaintiffs' experts based his opinions, in large part, on Hoffmann-LaRoche's causality assessments. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *4-5 (M.D. Fla. May 2, 2007); see also In re Accutane Products Liability MDL 1626,
No. 804MD2523T30TBM, 2007 WL 1752593 at *8-9 (M.D. Fla. June 15, 2007). Plaintiffs argued that both the testimony and the causality assessments supporting it should be admitted under Federal Rules of Evidence 702 and 703. The court, however, properly held that the testimony was inadmissible because the causality assessments do not contain data that are reliable and upon which an expert opinion of causation can be based. The court asserted that testimony relying upon causality assessments lacked scientific validity, not having been derived from the scientific method. On this point, the court went so far as to characterize the expert's opinions arising from his review of internal documents containing causality assessments as speculative and conclusory. Moreover, addressing the applicability of the testimony to the facts in issue, the court commented that 'at most, an unbiased review of the assessments may support a conclusion that there is an association between [a drug and an adverse event].' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *4 (M.D. Fla. May 2, 2007). In its second ruling regarding the same issue, the court spelled out the deficiency even more clearly, asserting with emphasis 'an association is not the equivalent of causation.' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1752593 at *9 (M.D. Fla. June 15, 2007).

Courts Have Not Announced Blanket Prohibitions

It should be cautioned, however, that the Accutane court's rulings do not represent a blanket prohibition against the admissibility of causation assessments or the use of adverse drug reports by expert witnesses. The Accutane court noted that causality assessments themselves may also be admissible in circumstances factually distinguishable from those presented to the Accutane court ' most notably to establish notice. The court refused to allow the plaintiff to present the causality reports (or adverse drug reports) for notice purposes because the defendants admitted that they had received notice of issues possibly associated with their drug. Accordingly, the court held that such evidence was unnecessary to prove notice. More importantly, the court also reasoned that even if notice were disputed, admitting the causality assessments for notice purposes would be substantially more prejudicial than probative because while the court had gained an understanding of the purpose of causality assessments, lay jurors would likely conclude, as the plaintiffs argued, that the assessments were admissions of causation.

The court noted, however, that to the extent that notice or knowledge is disputed, a court's assessment with respect to the balance between the evidence's probative value and the dangers of unfair prejudice may shift. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *5 (M.D. Fla. May 2, 2007); see also Hoppe v. G.D. Searle & Co., 779 F. Supp. 1413, 1425 (S.D.N.Y.). In fact, the court explicitly acknowledged this, noting that if the defendants did not admit they had received notice of possible adverse events during the trial, the court might revisit the issue.

Similarly, the Accutane court did not hold that an expert witness can never rely upon adverse drug reaction reports in developing a causation opinion. There is no doubt that adverse drug reaction reports are inherently unreliable. The Accutane court characterized such reports as simply 'uncontrolled anecdotal information' and concluded that as such they 'offer[ ] one of the least reliable sources to justify opinions about both general and individual causation.' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1752593 at *9 (M.D. Fla. June 15, 2007). Because of this, courts routinely hold that adverse drug reaction reports are not sufficiently competent and reliable data upon which an expert can base a causation opinion. See In re Baycol Prods. Litig., 495 F. Supp. 2d 977, 987-988 (D. Minn. 2007) (collecting cases which have excluded expert testimony based upon adverse event reports); Dunn v. Sandoz Pharms. Corp., 275 F. Supp. 2d 672, 682-683 (D.N.C. 2003) (noting that '[c]ase reports are not controlled studies, and they cannot be verified through peer review'). The Accutane court, however, left open the door to expert testimony based in part on adverse event reports provided that the expert performed an extensive, rigorous review of these reports. Defendants, therefore, should not expect that a court will automatically disqualify an expert simply because the expert relies in part on adverse drug reaction reports. Defendants should develop the appropriate record to demonstrate that an expert placed too much reliance on such reports and did not engage in sufficient scientific investigation to justify a causation opinion.

Conclusion

The increasing reliance of plaintiffs on data collected in the course of post-marketing surveillance and assessments made thereupon is troubling, in that it creates perverse incentives for pharmaceutical companies. Both the pharmaceutical companies and the public, which relies upon those companies to produce safe and effective medication, have an interest in ensuring that manufacturers continue to maintain accurate 'early warning systems' for their drugs. Fortunately, courts have recognized the strong legal arguments for excluding the assessments. It is important, however, to recognize that these cases do not create blanket rules against the use of such data either as evidence or as the basis for expert opinion.

Andrea E.K. Thomas and Joseph K. Scully are members of Day Pitney LLP's commercial litigation group in the Hartford, CT office. Their practices focus on product liability and insurance coverage disputes.

In the present climate of constant news reports of product recalls and runaway verdicts, pharmaceutical manufacturers are concerned about certain information that can infiltrate and unfairly influence product liability trials. Manufacturers should not have to worry that a court might permit a plaintiff to utilize the manufacturer's post-marketing surveillance data ' which is collected pursuant to governmental regulations and for the purpose of keeping manufacturers apprised of the possibility of an adverse effect associated with their product ' to prove actual causation. In addition, consumers should not have to worry about such legal disincentives impacting industry pharmacovigilance efforts. Post-marketing surveillance is absolutely critical to balancing the competing public policy concerns that favor speeding up the initial approval of useful drugs and assuring that only safe drugs reach the public.

Unfortunately, adverse drug reaction reports collected and causality assessments made in the course of post-marketing surveillance have increasingly become fodder for plaintiffs' attorneys attempting to prove causation. Courts, however, properly have precluded plaintiffs from presenting post-marketing surveillance materials, most recently refusing to allow plaintiffs to introduce company causality assessments based on adverse drug reaction reports as evidence of causation and from using these reports and assessments as bases for expert opinions on causation.

Adverse Drug Reaction Reports and Causality Assessments

As with many product manufacturers, pharmaceutical companies monitor adverse events allegedly associated with their products. Pharmaceutical companies are subject to extensive governmental regulations both in the United States and abroad. Under these regulations, pharmaceutical companies must compile and submit adverse drug reaction reports to the U.S. Food and Drug Administration ('FDA') and foreign regulatory bodies on an ongoing basis. See 21 CFR 314.80 (2007); The Rules Governing Medicinal Products in the European Union, Volume 9A (2007) available at http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-9/pdf/vol9_2007-07_upd07.pdf. Adverse drug reaction reports are essentially voluntarily submitted anecdotal reports of a complication or event that the reporter may believe to be associated with a drug. The source of information regarding the underlying event may be a health care provider, but is just as likely to be a patient, family member, news report, or plaintiff's attorney. Accordingly, the reporters are self-selecting and, to a certain extent, are skewed toward an assessment that the drug was somehow associated with the adverse event. After all, it is unlikely that someone would contact a pharmaceutical manufacturer unless they believed that the manufacturer's drug was related to the adverse event.

The FDA ' and pharmaceutical companies ' rely on these anecdotal reports, and analyses thereof, to generate hypotheses for further epidemiological and clinical investigation. Citing the inherent limitations of these reports, the FDA, however, has repeatedly cautioned against reliance on such reports in support of conclusions as to incidence rate. (See In re Baycol Prods. Litig., 495 F. Supp. 2d 977, 987 (D. Minn. 2007) (citing FDA guidance regarding the reliance on adverse event reports to show incidence rates). Recognizing the importance of post-marketing surveillance and the attendant dangers of such activities to pharmaceutical companies, FDA regulations include a disclaimer stating that reports submitted to the FDA are not admissions that a drug caused or contributed to an adverse effect. See 21 CFR 314.80 (k) (2007).

Unlike the FDA, the European Agency for the Evaluation of Medicinal Products ('EMEA') requires that pharmaceutical companies conduct causality assessments with respect to these adverse drug reaction reports. See 21 CFR 314.80 (2007); The Rules Governing Medicinal Products in the European Union, Volume 9A (2007) are available at http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-9/pdf/vol92007-07_upd07.pdf. Because most drugs are marketed worldwide, in complying with EMEA guidance, pharmaceutical companies generally assess causality with respect to each adverse event reported to the company, regardless of the country of origin. The 'causality assessment' terminology is unfortunate, as these assessments are not meant to assess actual causation in the legal sense. These assessments aim to determine whether there is a reasonable probability of relatedness between the drug and the adverse event for regulatory reporting and safety surveillance purposes. They are intended to act as an early warning system concerning possible complications associated with the drug.

Moreover, these assessments generally do not involve the rigorous investigation and evaluation of the relationship between the adverse event and the drug typical of the scientific method. Generally, a manufacturer will compile all adverse events purportedly related to its drugs in a database. See, e.g., In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *2 (M.D. Fla. May 2, 2007) (discussing the procedures utilized by Hoffmann-La Roche in complying with the EMEA causality assessment regulations). For each report, the manufacturer includes the reporter's assessment of whether there is a possibility the adverse event was related to its drug, regardless of whether the reporter is qualified to provide such assessment or whether the reporter's assessment is based on reliable data.

Manufacturers typically also include a company assessment of whether there is a possibility that the adverse event was related to its drug. The company's assessment, however, is not the product of an independent investigation and evaluation. For instance, in making causality assessments under the EMEA regulations for Accutane', Hoffmann-La Roche considers the reporter's assessment and a score generated by applying the facts provided by the reporter to the Naranjo algorithm. See, e.g., Id. The odds are very good that if the reporter asserts there is a possible association between the drug and the adverse event, the company's assessment will similarly reflect a possible association. In fact, the company's assessment will only indicate no possible association in instances in which the reporter asserts no possible relationship, and the score generated by the algorithm is very low. Finally, a drug safety physician employed by the manufacturer generates a brief narrative regarding each report. Such narratives, however, are based exclusively on the information provided by the reporter and, accordingly, are only as good as the information provided concerning an adverse event. Moreover, the aim of these narratives, like the causality assessments, is not to identify actual causation, but rather to determine whether there is a reasonable probability of a possible relationship between the event and the drug.

Plaintiff's Arguments Regarding Causality Assessments

Stymied by pesky hearsay rules precluding adverse drug reaction reports from being directly introduced in evidence to prove causation, plaintiffs have refocused their efforts on persuading courts to allow them to introduce or rely upon the accompanying company causality assessments. Plaintiffs have attempted to use these assessments as both direct evidence of causation (or alternatively for notice purposes) and indirectly as bases upon which their experts can provide opinions on causation.

Plaintiffs have attempted to persuade courts to admit causality assessments into evidence by arguing that such assessments are admissions that the manufacturer's drug caused the reported events, and thus, exempt from the definition of hearsay. They have also contended that such assessments are relevant to establishing that a pharmaceutical company had notice or knowledge of the potential dangers associated with its drugs, and therefore, not offered for the truth of the matter asserted. Finally, plaintiffs have attempted to bring causality assessments into the courtroom using a backdoor approach, arguing that experts should be allowed to base their testimony on these assessments.

Causality Assessments Are Inadmissible to Prove Causation

Causality assessments are not admissions of causation admissible under Federal Rule of Evidence 801(d)(2)(A). Plaintiffs have argued that under Rule 801(d)(2)(A) courts should admit causality assessments as admissions by the manufacturer that the drug caused the adverse event and, therefore, is evidence of general causation. Rather, these assessments are simply the product of the manufacturer's categorizing, based upon a simple algorithm and the subjective beliefs of third-party reporters as to the relationship between a drug and a particular ailment. The causation assessments are not the manufacturer's articulations of causes of these adverse events; at best, they merely reflect that the data and opinion expressed by the reporter raise the possibility that the drug might be related to the events. This is a far cry from a manufacturer conducting its own assessment of
the patient's medical records and acknowledging that its drug caused the patient's complications.

Courts should not admit causality assessments in evidence because: 1) they are not properly characterized as a party admission concerning causation; 2) as a consequence, they are not relevant to proving causation because they are generated for much more limited purposes; and 3) what limited evidentiary value they have is greatly outweighed by their prejudicial nature. Thus, causality assessments offered to prove causation should be excluded under Federal Rules of Evidence 402 and 403.

Evidence is only relevant if it tends to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence. While causation is certainly a matter of consequence, causality assessments are neither intended to ' nor do they in fact ' shed light on actual causation. Rather, they are made in furtherance of post-marketing surveillance and pursuant to EMEA regulations requiring the company to catalogue the reports based upon a causality rating. As a consequence, causality assessments are inadmissible under Federal Rule of Evidence 402 because they are not relevant to the question of causation.

Moreover, causality assessments are inadmissible under Federal Rule of Evidence 403. As noted above, the probative value of the causality assessments is low because, at best, they are only tangentially relevant to causation. On the other hand, the dangers of unfair prejudice, confusion of the issues, and misleading the jury associated with causality assessments are very high. Even with strongly worded limiting instructions from the court, a lay jury is likely to conclude that causality assessments are admissions of causation on the part of the pharmaceutical company.

The Multi-District Litigation court overseeing the Accutane cases recently issued a series of opinions excluding causality assessments. In those cases, plaintiffs offered Hoffmann-LaRoche's causality assessments to prove that Accutane causes inflammatory bowel disease and to establish that Hoffman-LaRoche had notice of the potential dangers of Accutane. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 (M.D. Fla. May 2, 2007). The plaintiffs contended that the causality assessments, buttressed by the testimony of Hoffmann-LaRoche's Director of Drug Safety that the assessments were 'fully investigated articulations of the adverse events,' were admissions that Accutane caused inflammatory bowel disease. As such, the plaintiffs argued that the causality assessments were not hearsay and should be admitted. The court, however, properly held that the causality assessments were not admissible to prove causation. Addressing plaintiffs' contention that the causality assessments were admissible to prove causation as admissions of a party opponent, the court concluded that the testimony of Hoffmann-LaRoche's Director of Drug Safety did not indicate that the assessments reflected a company's admission that the drug had caused an adverse event. Rather, according to the court, the testimony merely indicated that the assessments reflected a company's conclusion that there was a reasonable probability that the adverse event could possibly be related to the drug. Thus, the court held that the causality assessments were not relevant to the issue of actual causation and further noted the highly prejudicial nature of the assessments in explaining their inadmissibility.

Expert Testimony Based on Causality Assessments Is Inadmissible

Courts should also preclude plaintiffs from doing an end run around the rules of evidence by presenting expert testimony concerning causation based on causality assessments. Courts should exclude such expert testimony pursuant to Federal Rule of Evidence 702 and the holding of Daubert v. Merrell , 509 U.S. 579 (1993).

Expert testimony relying on causality assessments is inadmissible because reliance on causality assessments is not scientifically valid. Rule 702 requires courts to determine whether the expert is proposing to testify as to scientific knowledge that will assist the trier of fact. In Daubert, the U.S. Supreme Court explained that in determining whether scientific expert testimony is admissible under Rule 702, courts must determine whether the reasoning or methodology underlying the testimony is scientifically valid and whether that reasoning or methodology can properly be applied to the facts in issue. Reliance upon assessments made for the purpose of post-marketing surveillance and hypothesis generation simply is not a scientifically valid method for assessing causation. Thus, the application of Rule 702 should eliminate the need for any consideration of whether the testimony and otherwise inadmissible assessments should be admitted under the more stringent standards of Rule 703.

In the Accutane cases, one of the plaintiffs' experts based his opinions, in large part, on Hoffmann-LaRoche's causality assessments. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *4-5 (M.D. Fla. May 2, 2007); see also In re Accutane Products Liability MDL 1626,
No. 804MD2523T30TBM, 2007 WL 1752593 at *8-9 (M.D. Fla. June 15, 2007). Plaintiffs argued that both the testimony and the causality assessments supporting it should be admitted under Federal Rules of Evidence 702 and 703. The court, however, properly held that the testimony was inadmissible because the causality assessments do not contain data that are reliable and upon which an expert opinion of causation can be based. The court asserted that testimony relying upon causality assessments lacked scientific validity, not having been derived from the scientific method. On this point, the court went so far as to characterize the expert's opinions arising from his review of internal documents containing causality assessments as speculative and conclusory. Moreover, addressing the applicability of the testimony to the facts in issue, the court commented that 'at most, an unbiased review of the assessments may support a conclusion that there is an association between [a drug and an adverse event].' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *4 (M.D. Fla. May 2, 2007). In its second ruling regarding the same issue, the court spelled out the deficiency even more clearly, asserting with emphasis 'an association is not the equivalent of causation.' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1752593 at *9 (M.D. Fla. June 15, 2007).

Courts Have Not Announced Blanket Prohibitions

It should be cautioned, however, that the Accutane court's rulings do not represent a blanket prohibition against the admissibility of causation assessments or the use of adverse drug reports by expert witnesses. The Accutane court noted that causality assessments themselves may also be admissible in circumstances factually distinguishable from those presented to the Accutane court ' most notably to establish notice. The court refused to allow the plaintiff to present the causality reports (or adverse drug reports) for notice purposes because the defendants admitted that they had received notice of issues possibly associated with their drug. Accordingly, the court held that such evidence was unnecessary to prove notice. More importantly, the court also reasoned that even if notice were disputed, admitting the causality assessments for notice purposes would be substantially more prejudicial than probative because while the court had gained an understanding of the purpose of causality assessments, lay jurors would likely conclude, as the plaintiffs argued, that the assessments were admissions of causation.

The court noted, however, that to the extent that notice or knowledge is disputed, a court's assessment with respect to the balance between the evidence's probative value and the dangers of unfair prejudice may shift. In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1288354 at *5 (M.D. Fla. May 2, 2007); see also Hoppe v. G.D. Searle & Co. , 779 F. Supp. 1413, 1425 (S.D.N.Y.). In fact, the court explicitly acknowledged this, noting that if the defendants did not admit they had received notice of possible adverse events during the trial, the court might revisit the issue.

Similarly, the Accutane court did not hold that an expert witness can never rely upon adverse drug reaction reports in developing a causation opinion. There is no doubt that adverse drug reaction reports are inherently unreliable. The Accutane court characterized such reports as simply 'uncontrolled anecdotal information' and concluded that as such they 'offer[ ] one of the least reliable sources to justify opinions about both general and individual causation.' In re Accutane Products Liability MDL 1626, No. 804MD2523T30TBM, 2007 WL 1752593 at *9 (M.D. Fla. June 15, 2007). Because of this, courts routinely hold that adverse drug reaction reports are not sufficiently competent and reliable data upon which an expert can base a causation opinion. See In re Baycol Prods. Litig., 495 F. Supp. 2d 977, 987-988 (D. Minn. 2007) (collecting cases which have excluded expert testimony based upon adverse event reports); Dunn v. Sandoz Pharms. Corp. , 275 F. Supp. 2d 672, 682-683 (D.N.C. 2003) (noting that '[c]ase reports are not controlled studies, and they cannot be verified through peer review'). The Accutane court, however, left open the door to expert testimony based in part on adverse event reports provided that the expert performed an extensive, rigorous review of these reports. Defendants, therefore, should not expect that a court will automatically disqualify an expert simply because the expert relies in part on adverse drug reaction reports. Defendants should develop the appropriate record to demonstrate that an expert placed too much reliance on such reports and did not engage in sufficient scientific investigation to justify a causation opinion.

Conclusion

The increasing reliance of plaintiffs on data collected in the course of post-marketing surveillance and assessments made thereupon is troubling, in that it creates perverse incentives for pharmaceutical companies. Both the pharmaceutical companies and the public, which relies upon those companies to produce safe and effective medication, have an interest in ensuring that manufacturers continue to maintain accurate 'early warning systems' for their drugs. Fortunately, courts have recognized the strong legal arguments for excluding the assessments. It is important, however, to recognize that these cases do not create blanket rules against the use of such data either as evidence or as the basis for expert opinion.

Andrea E.K. Thomas and Joseph K. Scully are members of Day Pitney LLP's commercial litigation group in the Hartford, CT office. Their practices focus on product liability and insurance coverage disputes.