Epidemiology Can Be Your Friend: Using Epidemiology in the Courtroom

By Raphael Metzger, Esq.

Introduction

As in all personal injury cases, the toxic tort plaintiff must prove that the defendant proximately caused his injury.  All toxic tort practitioners know that proving causation is usually the most difficult part of a toxic tort case (and usually the most costly).

Proving causation is a two-step process.  First, counsel must prove that the chemical in question is capable of causing the disease which the plaintiff has.  This aspect of proof is commonly called "general causation."  Second, counsel must prove that the chemical did, in fact, cause the plaintiff to suffer the disease.  This aspect of proof is commonly called "specific causation."

Epidemiology is often considered the bane of toxic tort cases, because some courts have required supportive epidemiologic studies as proof of general causation.  The absence of any supportive epidemiologic studies is fatal to the plaintiff’s toxic tort case in those jurisdictions which require epidemiologic evidence.  However, epidemiology can also help prove a toxic tort case by providing evidence supporting both general and specific causation.

This article will show how epidemiology can be the toxic tort litigator’s friend, i.e., how it can be used to the plaintiff’s advantage, rather than being the defense’s weapon of destruction.
The effective use of epidemiologic evidence depends greatly on the nature and quality of the epidemiologic data.  This article will explain how robust epidemiologic data can help the plaintiff prove general causation, specific causation, and causative dose.  This article will also show how minimal epidemiologic data can be used to support the proof of general causation.  Lastly, this article will explain why the absence of epidemiologic studies should not preclude a finding of causation where other supportive evidence exists.

Epidemiologic Principles

Epidemiology is the medical science that evaluates the distribution and causes of disease ... by studying their patterns in populations. . . .  Epidemiology seeks to establish associations between suspected causes and effects by one of two methods: either comparing the incidence of disease ... across exposed and unexposed populations, or comparing the incidence of exposure across sick and healthy populations.  Smith v. Ortho Pharmaceutical Corp., 770 F.Supp. 1561, 1573 (N.D. Ga. 1991).

In an individual case, epidemiology cannot conclusively prove causation; at best, it can only establish a certain probability that a randomly selected case of disease was one that would not have occurred absent exposure, or the 'relative risk' of the exposed population.  Smith v. Ortho Pharmaceutical Corp., 770 F.Supp. 1561, 1573 (N.D. Ga. 1991).

Because epidemiology involves evidence on causation derived from group-based information, rather than specific conclusions regarding causation in an individual case, epidemiology will not conclusively prove or disprove that an agent or chemical causes a particular effect.  Smith v. Ortho Pharmaceutical Corp., 770 F.Supp. 1561, 1576 (N.D. Ga. 1991).

"By comparison to the clinical health sciences, which are directly concerned with diseases in particular patients, epidemiology is concerned with the statistical analysis of disease in groups of patients.  The statistical associations may become so compelling, as they did in establishing the correlation between asbestos exposure and mesothelioma, that they raise a legitimate implication of causation.  Statistical associations, however, do not necessarily imply causation....  It is important, therefore, to have some basis for deciding whether a statistical association derived from an observational study represents a cause-and-effect relationship.”  Landrigan v. Celotex Corp., 127 N.J. 404, 605 A.2d 1079 (N.J. 1992).

"In general, the stronger the association, the more likely it represents a cause-and-effect relationship.  Weak associations often turn out to be spurious and explainable by some known, or as yet unknown, confounding variable.  In order for an association to be spurious, the underlying factor that explains it must have a stronger relationship to the disease than the suspected causal factor.  When the causal factor under consideration is strongly related to the disease, it is likely, although not certain, that the underlying variable with the necessarily even stronger relationship to the disease would be recognizable."  Landrigan v. Celotex Corp., 127 N.J. 404, 605 A.2d 1079 (N.J. 1992).

Basic Epidemiologic Concepts

To use epidemiology effectively, one must understand a few basic epidemiologic concepts.  The most important epidemiologic concepts are magnitude of association, statistical significance, odds ratio, relative risk, and attributable risk.

Magnitude of association refers to the extent to which disease is increased in an exposed group compared to controls.  The greater the magnitude of the association, the greater the likelihood that the association represents a true cause-and-effect relationship.  Thus, as a general proposition, when an exposed population has a disease at 20 times the rate of an unexposed population, the association is more likely causal than when the exposed population only has the disease at twice the rate of the unexposed population.

Statistical significance refers to the probability that an association is real and not the result of chance.  Conventionally, epidemiologists consider statistical significance to be achieved when the “confidence interval” reaches the 95% mark, i.e., where the probability that the association is due to chance is less than 5%.  When the confidence interval of an association is less than 95%, most epidemiologists do not consider the association sufficiently reliable to be deemed causal.

Magnitude of the association and statistical significance are important concepts.  Although the magnitude of the association from an epidemiologic study may be high, if the confidence interval is less than 95%, epidemiologists do not deem the association causal.  Likewise, although an association may have a confidence interval of 99% (generally considered to be “highly” significant), if the magnitude of the association is low, epidemiologists do not consider the association causal, because the association may be due to a confounding variable rather than the exposure studied.

The strength of an association between exposure and disease can be stated as a relative risk (RR), an odds ratio (OR), or an attributable risk (AR).

Relative Risk (RR) is defined as the ratio of the incidence rate of disease in exposed individuals to the incidence rate in unexposed individuals.

The Odds Ratio (OR) is similar to a relative risk in that it expresses in quantitative terms the association between exposure to an agent and a disease. 

Attributable Risk (AR) represents the amount of disease among exposed individuals that can be attributed to the exposure.

Using Epidemiologic Evidence to Establish General Causation

In determining whether an association represents a cause-effect relationship, epidemiologists consider nine factors derived from Sir Austin Bradford Hill: (1) Strength of the association, (2) Consistency of the association, (3) Specificity of the association, (4) Temporality of the association, (5) Biologic gradient (dose-response relationship), (6) Biological plausibility, (7) Coherence of the association, (8) Experimental evidence, and (9) Analogy.

Epidemiologic studies may provide data pertinent to the evaluation of the first five of these factors.  Epidemiologic studies always report the strength of the association they investigate.  Where multiple epidemiologic studies explore the same association, they can be evaluated for consistency.  An association exhibits specificity if the exposure is associated only with a single disease or single category of diseases.  Epidemiologic studies often report temporal relationships between exposure and manifestation of disease and may thus provide useful information for evaluating temporality of the association.  Epidemiologic studies which compare disease outcomes between individuals with lesser exposures and those with greater exposures can be useful in evaluating biologic gradients, i.e., dose-response relationships.

Using Epidemiologic Evidence to Establish Specific Causation

The second edition of the Reference Manual on Scientific Evidence, published by the Federal Judicial Center in 2000 makes the bold statement that epidemiology has no relevance to specific causation: “Epidemiology is concerned with the incidence of disease in populations and does not address the question of the cause of an individual’s disease.”  Id. at 381, citing DeLuca v. Merrell Dow Pharmaceuticals, Inc., 911 F.2d 941, 945 & n. 6 (3d Cir. 1990) [“Epidemiological studies do not provide direct evidence that a particular plaintiff was injured by exposure to a substance.”]

Although epidemiologic evidence cannot, by itself, prove specific causation, epidemiologic data may nevertheless provide evidence that is both relevant and highly probative of specific causation.  For example, where the magnitude of an association between an exposure and a disease in an epidemiologic study is great and statistically significant, a plaintiff whose exposure is like that of the study subjects and who has the disease found in the study subjects can accurately be said to have an increased risk of the disease.  Where the study is robust enough to determine the causal dose and the plaintiff is shown to have had such dose, the study is also relevant to establish sufficiency of exposure.

Using Relative Risk to Prove a Probability of Causation

Relative Risk can be used to establish general causation and may also be helpful in proving probability of specific causation.

Relative Risk (RR) is a statistical measure of the strength of the association between exposure and disease; it indicates the probability of developing the disease in an exposed group relative to an unexposed group.

Relative Risk (RR) is calculated by the following formula:

A Relative Risk of 3 means that a person who is exposed to the chemical at issue is 3 times more likely to develop the disease in question than a person who is not so exposed.

Relative Risk can be a powerful tool in establishing causation by a simple formula that converts Relative Risk to a probability statement:

Thus, if the Relative Risk in a study is 3, the probability is 2/3, or about 67%.  What does this actually mean?  It means that if a plaintiff has the disease in question and was exposed to the toxin chemical like the subjects in the study, the probability that the exposure caused the person's disease is 67%.  That says a lot!

Any Relative Risk greater than 2 yields a probability greater than 50%.  Thus, such a Relative Risk can help satisfy the plaintiff's burden of proving general causation by a preponderance of the evidence.  Such a Relative Risk also goes a long way to proving specific causation.  It's all in the numbers!

An excellent discussion of the use of epidemiology in toxic tort cases and the legal significance of Relative Risk is found in In re Joint Eastern and Southern District Asbestos Litigation (Maiorana), 827 F. Supp. 1014 (S.D.N.Y. 1993).

Benzene-Induced Acute Myeloid Leukemia: How to Use Epidemiology to Prove Causation

A good example of how epidemiology can help prove causation is the case of benzene-induced acute myelogenous leukemia.  Several epidemiologic studies have been conducted of the association between benzene and leukemia.  These studies consistently find a causal association between benzene and acute myelogenous leukemia -- one of the four major types of leukemia.  The studies are so consistent that all experts agree that benzene is a human leukemogen -- specifically that benzene causes acute myelogenous leukemia.  The studies exhibit a clear dose-response relationship and also provide useful data regarding the latency period for benzene-induced leukemia.  Additionally, a major risk assessment has been performed which shows that the risk of developing acute myelogenous leukemia from benzene is 1.7 at a cumulative dose of 1 part per million year(exposure to a cumulative dose of benzene of 1 part per million over the course of a work year).  Rinsky, R., et al., “Benzene and Leukemia,” New Engl. J. Med. 316:1044-50 (1987).

Such robust epidemiologic data is very helpful.  First, the epidemiologic data is the cornerstone of proving general causation.  Second, where a worker has acute myelogenous leukemia and has a history of benzene exposure, the epidemiologic data can be used to prove specific causation.  An industrial hygienist or environmental scientist can model the worker’s benzene exposure and calculate a cumulative dose.  Where the worker’s cumulative benzene exposure exceeds 1 ppmy, it is more probable than not that the workers’ leukemia was caused by benzene exposure.  A clinician can then rule out other possible causes of leukemia such as ionizing radiation and certain drugs known to cause the disease, and thereby prove specific causation to a reasonable degree of medical probability.  The formula for relative risk can also be used to illustrate the probability of causation to the jury.  For example, if the worker’s cumulative benzene dose is determined to be 10 ppmy, the probability of causation is 10-1 ÷ 10 = 90%.

Conclusion

Epidemiology can be your friend, rather than your nemesis.  If you know how to use epidemiology to your advantage, it can be the sword which slays the Daubert dragon, rather than the sword upon which the injured plaintiff falls.

Raphael Metzger is the principal of the Metzger Law Group, a Professional Law Corporation.  His firm concentrates its practice on occupational cancer, lung disease, and toxic organ failure cases.  The firm’s offices are located in Long Beach, California.