Scientific Models of Human Health Risk Analysis in Legal and Policy Decisions

Abstract

DOUGLAS CRAWFORD-BROWN [*] I INTRODUCTION Scientists working in the fields of law or policy faced with risk-based decisions find themselves in very different positions epistemologically than those in the routine practice of science. First, risk assessment is an interdisciplinary activity, requiring the piecing together of expertise from many fields. Only a community of scientists could possess so much expertise. It is problematic when a single scientist, or a small group of them, is called upon to speak on behalf of the entire community. Second, science is by nature forward-looking. It is a process by which truth is progressively refined and approached, rather than an end point. In court, or in the policy arena, however, scientists are called on to freeze this process in a snapshot that will be regarded as final. Scientists may have little or no experience in this activity. The quality of scientific predictions of risk in the courtroom and policy arena rests in large measure on how these two differences between the normal practice and the legal/po licy practice of science are reconciled. This article considers a variety of issues that arise in reconciling these differences, and the problems that remain with scientific estimates of risk when these are used in decisions. II DEFINITION OF RISK Before turning to the issues arising from risk predictions, a definition of risk is needed. Three schools of thought are prevalent in the field of risk assessment. [1] (1) Objective risk. This school treats risk as an objective property of the world, much like any other scientific concept. Risk is identified with a combination of the probability and the severity of an outcome. [2] Often, the combination is the mathematical product of the probability and the severity, but this operation is not scientifically or logically necessary, as is demonstrated by the fact that there are competing operations in the scientific community such as Quality Adjusted Life Years (QALY). [3] (2) Subjective risk. This school treats risk as a subjective property of individuals in a society. Risk is identified as the degree of concern, or dread, expressed by an individual over some situation. It is measured using polls based on expressed or revealed opinions. [4] (3) Psychologistic risk. This school treats risk as a property of epistemological reflection, or analysis of existing evidence. Risk is identified as the degree of evidentiary support for the belief that an adverse effect will occur under certain circumstances. This school shares features with the objective school, such as the recognition that risks are related to probabilities. It is closer, however, to the Bayesian conception of probability because the objective school employs a long-term frequency concept of probability. [5] While some social scientists adhere to the subjective concept of risk, it generally is not the concept employed in the practice of natural science or in legal or policy decisionmaking. The difference between the objective and psychologistic conceptions of risk, however, is not so apparent in practical terms as might otherwise appear. For example, adherents of both would agree with the following statement by the National Academy of Science's National Research Council: Human-health risk assessment entails the evaluation of scientific information on the hazardous properties of environmental agents and on the extent of human exposure to those agents. The product of the evaluation is a statement regarding the probability that populations so exposed will be harmed, and to what degree. The probability may be expressed quantitatively or in relatively qualitative ways. Furthermore, both agree that risk involves probability and severity. Both agree that probability and severity are objective properties of the world. The objective school, however, equates risk solely with probability and severity, while the psychologistic school adds a third component to risk: rational confidence in the estimates of probability and severity. …