Print

Regulatory Science Rules Governing Risk Assessment

Despite the abundance of information including peer reviewed articles and reports resulting from consensus of credible panels there appears to be a lack of clear rules governing various aspects of risk assessment. The following rules are derived from generally available and credible information.

Rule 1: Truth in Risk Assessment

This rule requires that the risk assessor clearly indicate the choices, assumptions, and other decisions and justify them. For example for human health risk assessment the risk assessor must indicate: why one set on animal tests was chosen and what would have happened if all animal sets would have been used. Similarly, what options were available for high-dose-to-low-dose and rodents-to-human extrapolations. Subsequently, the risk assessor must provide actual computations to compare the results. In effect the risk assessor must essentially analyze all options and indicate which one of them would be preferable and why. Applying a probabilistic approach will result in conclusions that rely on the central trend (e.g., average, median) and are unlikely to be significantly affected by the outliers.

Rule II: Honesty in Communication

The risk assessor must include the values resulting from the risk assessment in a common statistical form. The risk assessor must provide the central trend (e.g., average, median) together with the usual (e.g., 66%, 90%, 95%) confidence intervals (as specified by their lower and upper limits). Depending on the desires of risk managers, the risk assessor may also provide other statistics (such as the 25th and 75th percentile). All regulations must include the central trend (e.g., average, median) together with the usual (e.g., 66%, 90%, 95%) confidence intervals (as specified by their lower and upper limits).

Rule III: Nothing but Science:

 

 

This rule otherwise known as Shoemaker Stick to Your Shoemaking requires that risk assessors provide the risk managers with scientific information following Rules I and II. In particular, risk assessors may not include in the assessment any societal goal. The level of protectiveness in various regulations is not the domain of science but the result of the democratic process. The scientific staff of a regulatory agency must provide the regulatory system with the most accurate information, including the uncertainties related to all estimates. The risk manager is responsible to: provide the entire picture to the public during the rule-making process; and justify the choice of a specific confidence limit. These choices may be based on economics, unique populations, and other requirements.