Approaches to handling uncertainty when setting environmental exposure standards

Abstract

The statistical properties of the two most widely used methods for setting environmental exposure standards are explored. The traditional NOAEL approach handles uncertainty in disagreement with the precautionary principle: a smaller and less sensitive Study will tend to yield higher exposure limits. As an attractive alternative, the Benchmark dose approach estimates the exposure associated with a predefined risk increase above the background. Although advantageous in several respects, this method is sensitive to sources of uncertainty arising from measurement error and data dependent selection of the dose-response model as well the choice of critical endpoint. An improved Benchmark analysis can be conducted using Structural equation models and Bayesian model averaging.