CHAPTER § 9.04 The Admissibility of Medical Causation Evidence

• Jurisdiction: United States

§ 9.04 The Admissibility of Medical Causation Evidence

In most cases, the plaintiff will attempt to prove medical causation by presenting expert testimony that the drug or device: (1) was capable of causing his or her injury (general causation), and (2) actually caused his or her injury (specific causation). This section discusses some of the issues pharmaceutical and device manufacturers should look out for when confronted with such testimony.¹⁸⁰

[1] General Causation

Courts use a two-prong approach to assess the reliability of medical-causation evidence: First, a plaintiff must demonstrate that the product in question causes the alleged injury in the general population, i.e., "general causation." Second, a plaintiff must demonstrate that the product was, in fact, the cause of his or her injury, i.e., "specific causation."¹⁸¹ If plaintiff provides sufficient expert evidence or testimony demonstrating general causation, then the plaintiff may proceed to proving specific causation.¹⁸² However, if there is insufficient evidence of general causation, the defendant may be entitled to summary judgment.¹⁸³

General causation asks whether the particular exposure was capable of causing the plaintiff's alleged injuries.¹⁸⁴ When the medical community generally recognizes that a particular product causes a condition or disease, then extensive inquiry into general causation is not necessary.¹⁸⁵ In cases where the toxicity of a product is not generally accepted, however, the court will need to undertake a more extensive analysis of the expert testimony.¹⁸⁶

Expert proof of general causation may be based on a wide variety of scientific evidence, including epidemiological studies, animal studies, case reports or adverse-event reports, chemical analogies, and other sources such as regulatory-agency findings. Not all forms of scientific evidence, however, are equally reliable. For example, case and adverse-event reports, FDA decisions, and other anecdotal evidence are generally not sufficient to demonstrate general causation.¹⁸⁷ The following sections discuss the relative weight of various forms of general causation evidence.

[a] Epidemiological Studies

Epidemiology is the "study of the distribution of disease in populations and the risk factors associated with particular diseases."¹⁸⁸ It employs surveys and studies to determine the probability of an association between a specific exposure and the occurrence of disease. A number of courts have noted that epidemiological studies are the most reliable evidence of general causation.¹⁸⁹ Courts accord greater weight to some types of epidemiological studies rather than others.¹⁹⁰ Because of the "various research designs" of epidemiological studies, they can "differ in the evidentiary weight they lend to a hypothesis that exposure to a given substance causes a condi tion."¹⁹¹

For example, two forms of epidemiological studies, cohort studies and case-control studies, have been noted to demonstrate more reliably an association between a particular substance and a condition or injury.¹⁹² In a cohort study, epidemiologists compare a group of individuals who have been exposed to a particular substance to an unexposed group, and they evaluate the two groups over time.¹⁹³ In a case-control study, epidemiologists retrospectively compare individuals who have experienced a particular injury or event to those who have not.¹⁹⁴ Epidemiologists then compare the background rate¹⁹⁵ of such injury in the unexposed group to the incidence of the injury in the exposed group. In this manner, epidemiologists are able to determine whether exposure to some particular substance is associated with a particular injury.

Epidemiologists assign a "risk ratio" based on the incidence of an injury or condition in a particular study. A risk ratio of 1.0 means the incidence of that particular condition is equal to the incidence of that condition in the unexposed population.¹⁹⁶ A risk ratio of 2.0, called a "doubling," means that the incidence of that condition is twice that of the incidence in the general population.¹⁹⁷ If the risk ratio is above 1.0, there is then a potential association between the substance and the condition at issue.¹⁹⁸

Epidemiologists utilize "confidence intervals" as a means of discussing how precise the risk estimate is by showing the boundaries of relative risk consistent with the data found in the study.¹⁹⁹ This is usually discussed in terms of whether the estimated risk is "statistically significant." The risk estimate is not statistically significant if the confidence interval includes the value 1.0, because the data are consistent with no increased or decreased risk.²⁰⁰ If the upper and lower bounds of the confidence interval are both greater than 1.0, then the risk estimate is statistically significant.²⁰¹ It then can be said that there is a 95-percent chance that the "true risk" is increased; however, that does not rule out the possibility that the elevated risk may be due to bias, error, or confounding as an explanation for the increase.²⁰² In other words, there must be a statistically significant difference between the frequency of the outcome (i) in patients exposed to the drug and (ii) in a similar population of patients who have not been exposed to the drug. If there is a statistically significant increase in outcome in the exposed population, then it can be concluded that the exposure and the outcome are associated or that the exposure increases the risk of the outcome. On the other hand, if an observed increase in frequency of the outcome in the exposed population is not statistically significant, then the possibility that it occurred by chance cannot be ruled out, and no association is established.

As a number of courts have recognized, an association between exposure to a drug and a particular outcome that is not statistically significant does not support the hypothesis that the drug caused the outcome.²⁰³

Although epidemiological studies are often viewed as some of the strongest evidence of general causation, courts will not automatically admit these studies into evi-dence.²⁰⁴ Rather, acting in their role as "gatekeeper," courts will delve into the sufficiency of the methodology employed and the sufficiency of the evidence regarding the question of medical causation in general. Some courts require a risk ratio of at least 2.0 to show that there is a causal connection between the product and the injury or condition.²⁰⁵

Regardless of whether the court requires a risk ratio of 2.0 or higher,²⁰⁶ a single study, or even a number of studies, showing such an association is not sufficient to establish general causation.²⁰⁷ Rather, other factors must be considered.²⁰⁸

In considering such "other factors," a number of courts have turned to the Bradford-Hill criteria to determine whether an epidemiological study adequately demonstrates a causal association between a drug and a particular condition.²⁰⁹ The Bradford-Hill criteria, consisting of nine separate factors, were set out in 1965 by noted epidemiologist Sir Austin Bradford Hill.²¹⁰ The criteria are:

(1) Strength: How strong is the association between the suspected risk factor and the observed outcome?
(2) Consistency: Does the association hold in different settings and among different groups?
(3) Specificity: How closely are the specific exposure factor and the specific health outcome associated, i.e., how unique is the quality or quantity of the response?
(4) Temporality: Does the hypothesized cause precede the effect?
(5) Biological Plausibility: Does the apparent association make sense biologically?
(6) Coherence: Is the association consistent with what is known of the natural history and biology of the disease?
(7) Experimental Verification: Does any experimental evidence support the hypothesis of an association?
(8) Biological Analogy: Are there examples of similar risk factors and similar outcomes?
(9) Dose-Response Relationship: Has a dose-response relationship been established, i.e., does the magnitude of the response increase as the magnitude of the dose increases?²¹¹

Although it is not necessary that all of these criteria be met for a specific epidemiological study to be considered scientifically reliable, these criteria are generally viewed as useful tools to guide courts in assessing expert testimony.²¹²

A few courts have held that the absence of epidemiological evidence is not necessarily fatal to a determination of general causation.²¹³ For example, the Eleventh Circuit noted that, "while epidemiological studies may be powerful evidence of causation, the lack thereof is not fatal to a plaintiff's case."²¹⁴ Nevertheless, at least one case has held that if the defendant proffered epidemiological evidence showing no statistically significant association between the drug and the injury, the plaintiff must present contrary epidemiological evidence supporting his or her claim.²¹⁵

[b] Animal Studies

Animal studies are another form of scientific evidence used to demonstrate a causal connection between a product and a particular condition. In general, courts view animal studies with a certain amount of skepticism and have found them to be insufficient on their own to prove general causation.²¹⁶ As one court noted, "Extrapolations from animal studies to human beings generally are not considered reliable in the absence of a credible scientific explanation of why such extrapolation is warranted."²¹⁷

Specifically, animal studies are weak indicators of causation in a human population because "'differences in absorption, metabolism, and other factors may result in inter-species variation in responses.'"²¹⁸ Further, animal studies frequently use higher doses than would ever be used in humans.²¹⁹ There are also differences between animal and human reactions to drugs. For example, a common insecticide, carbaryl, causes fetal abnormalities in dogs because of a particular enzyme that dogs possess, but does not have any effect on humans,²²...