Against the Odds: Proving Causation of Disease with Epidemiological Evidence
George W. Conk
Fordham Law School
Shepard's Expert & Scientific Evidence Quarterly, Vol. 3, pp. 103-136, Summer 1995
The consequences of the U.S. Supreme Court's decision in Daubert v. Merrell Dow Pharmaceuticals (1993) have been felt most strongly in product liability cases. Many of those cases have centered on 'toxic tort' claims - arising from exposure to minerals and chemicals, and from ingestion or injection of therapeutic toxins. Courts have shown a pattern of engaging in a mechanical and rigid application of the rudimentary four factors suggested by the Supreme Court to evaluate the reliability of the opinions asserted by experts regarding causation of disease. Courts' reductionist reliance on the 'Daubert factors' (testability, peer review, known error rates, general acceptance) are a failure of governance.
The central wisdom of Daubert is that law, not science governs, and that law must make an independent judgment regarding the sufficiency of evidence. But law can properly carry out that assessment only if it gives to science and scientists a full measure of recognition of their expertise and their methods. Scientific conclusions are reached by reasoning from the best available data, not by wooden use of the four part Daubert checklist - which itself lacks coherence. Thus, as practiced, post-Daubert, courts often replace the Frye general acceptance test, rejected as too 'austere' by Justice Blackmun, with a truncated, pidgin evaluation of scientific evidence.
Courts' awkward, formulaic vision is particularly apparent in their handling of epidemiologic evidence, of which a far too narrow and arithmetic view is taken. The essential power of modern epidemiology and toxicology is that they are practical arts, combining knowledge from many disciplines and seeking to review it in a systematic way. Brian MacMahon and Thomas F. Pugh, two authorities on epidemiologic methods put it this way:
"While epidemiologic information is at times derived from a much wider spectrum of biologic and medical disciplines, these three - clinical medicine, pathology and biostatistics - have almost universal application in epidemiology. Indeed, epidemiology may be thought of as the joint application of the three in the search for further understanding of disease etiology."
The requirement of some courts that statistical epidemiological evidence must be produced in order to prove that something causes injury or disease presents great dangers for the plaintiff. The ability to construct an effective narrative through epidemiology is much obstructed by courts' treatment of epidemiology as a branch of statistics which makes decisions about causes based on 'odds ratios' and other numerical devices.
Images of dice tumbling across the table are among the least appealing bases for judgment to jurors. Jurors are much more likely to agree with Albert Einstein that He does not play dice with the universe and to accept causal explanations derived from biologically coherent explanations of the evidence that a particular substance or condition probably plays a role in causing human disease or injury.
In trying to prove causation of disease plaintiffs' lawyers confront our conservative scientific culture which sustains a conservative judicial attitude about the kind of scientific evidence needed to show that a substance is toxic and that it actually injured the plaintiff.
Courts should adopt the approach of epidemiologist and philosopher of science Mervyn Susser: "The philosophy of causal inference reaches deep into abstractions. Applied in clinical (practice), however, it yields some practical benefits. For clinicians, inference is a constant and every day activity. In going about their business of diagnosis and treatment, clinicians are constantly making logical inferences, that is to say, drawing conclusions from a set of facts and premises...This logic (of causal inference) enables the contemporary clinician, in dealings with patients, to shift from charismatic priest-like authority to the authority of tried knowledge and of rational predictions founded on explicit models of causal relationships."
In the end, observes Susser, judiciousness matters most.
A discussion of these issues is followed by an example - application of a fuller, more flexible approach to causal inference in assessing the evidence for a causal relationship between a particular disease - kidney cancer - and exposure to asbestos.
Number of Pages in PDF File: 33
Keywords: daubert, scientific evidence, causation, epidemiology, toxic torts, asbestos, proof
JEL Classification: C10,N40,O30,C90Accepted Paper Series
Date posted: April 7, 2006
© 2014 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo3 in 1.047 seconds