Abstract of Meeting Paper

Society for Risk Analysis 2001 Annual Meeting

The Problem of Quantal Data and Consequences for the Benchmark Approach. W. Slob, National Institute of Public Health & Environment (RIVM)

The benchmark is gaining acceptance and application in toxicological risk assessment. Most of the applications concern quantal data. The application of the benchmark approach to continuous data is usually considered as less obvious. Various authors have proposed to deal with continuous data by transforming them into quantal data. In this way a benchmark response can be defined in terms of extra risk also for continuous data. This approach is not recommendable for various reasons. One of them is that dichotomizing continuous data is a waste of valuable information. More importantly, defining the benchmark dose in relation to a particular response rate (extra risk) is problematic: defined this way, the concept itself (apart from its estimate) is sensitive to the experimental error in the data. Therefore, the concept of the benchmark dose when relating to quantal data cannot be defined in biologically meaningful terms. In contrast, in the case of continuous data a biologically meaningful definition is possible, and the problems that quantal data suffer from do not apply here. As a result, the crucial issue of the benchmark approach (and of risk assessment), i.e. the decision of what size of a response can be considered as nonadverse for any given endpoint, is rendered into a biological issue. This does not make things more easy, but it does make them more appropriate.


Go to . . .

2001 SRA Annual Meeting Table of Contents
2001 SRA Annual Meeting Author Index
Main Abstracts Menu Page
RiskWorld Home Page