Stratified Ordinal Regression: A Tool for Combining Information from Disparate Toxicological Studies (1994)

Abstract:

This article focuses on risk assessment for acute inhalation exposure to perchloroethylene (PERC), and develops techniques for combining information from the available studies. The use of stratifi­cation is discussed as a way to address systematic aspects of data heterogeneity. It is demonstrated that failing to stratify on key variables can result in a misleading analysis. Humans appear to be much more sensitive to PERC than rats or mice given the same ambient air concentration. In fact, for the PERC data, the common practice of dividing the animal effective dose by ten seems close to the mark. An advantage of the analysis presented here is that it provides a way to scientifically investigate aspects of risk assessments such as species extrapolations. The same kinds of methods can be used for other chemicals for which sufficient data are available to estimate the effects of interest.

Author: 
Raymond J. CarrollDouglas G. SimpsonHaibo Zhou
Publication Date: 
Thursday, September 1, 1994
File Attachment: 
PDF icon tr26.pdf
Report Number: 
26