|
Abstract:
|
A widely used method in risk studies involving multiple chemicals is the relative potency approach, where the chemicals are assumed to be toxicologically similar – that is, they have parallel dose-response curves - and thus have constant relative potency. It has been long recognized by researchers that the similarity assumption is not always reasonable, and thus the common strategy of treating relative potency as fixed when it is not may distort conclusions and potentially mislead investigators. Despite growing awareness, risk assessors continue to apply this traditional approach even when similarity assumption is violated, mainly due to the apparent simplicity of this method and the limited availability of strategies for handling non-similar chemicals. In this presentation, we will employ the method of the relative potency function, which describes the relative potency as a function of the mean response, rather than a constant. The estimation and statistical inferences (that is, multiplicity-adjusted confidence limits) of relative potency at multiple response levels under the log-logistic model for quantal response data will be discussed.
|
