The Use and Misuse of Bradford Hill in American Tort Law

摘要

More than fifty years ago, Sir Austin Bradford Hill's Presidential Address to the Royal Society of Medicine-The Environment and Disease: Association or Causation?-identified some examinations that could be conducted upon features of epidemiological data that researchers could undertake to assist inferences about the ability of environmental agents to cause disease. We conducted a Westlaw search of largely federal district and appellate court decisions citing "Hill's aspects," "Hill's considerations," or "Hill's criteria" that identified numerous illustrative instances in which Hill's address has been invoked by judges. We will analyze both incorrect and correct uses of the results of the examinations that Hill proposed. Among the legal misinterpretations of Hill that we have found are: (1) Requiring that all of Hill's recommended examinations be carried out (2) Treating the examinations as criteria with yeso answers as if they were like Koch's Postulates and with symmetric impact of supportive and un-supportive results (3)Requiring statistical significance of associations in each considered study (4) demanding that Rate Ratios be greater than 2.0 (5) Demanding that the body of evidence include epidemiology (6) demanding that the body of evidence include toxicology. All of these demands are contrary to Hill's recommendations. In addition, we argue that understanding the probabilistic foundations of epidemiology shows that Hill is well justified in his recommended examinations of epidemiological evidence and his recommendations for interpreting the results of those examinations to warrant causal inference. We provide numeric examples to show how many of Hill's examinations are like rule-in/ but not rule-out laboratory examinations and that the prior plausibility of an association, ought, indeed to influence our degree of causal certification. Understanding the probabilistic roots of Hill's recommendations provides a deeper justification for them.