EcoSta 2018: Registration
View Submission - EcoSta2018
A0255
Title: Robust dose-level designs for binary responses in environmental risk assessment Authors:  Wanchunzi Yu - Bridgewater State University (United States) [presenting]
John Stufken - University of North Carolina at Greensboro (United States)
Abstract: The estimation of the lower confidence limit of a benchmark dose (BMD) is a common objective in environmental risk assessment. A risk function is involved to determine the BMD that corresponds to a given, low-level benchmark response (BMR). In a BMD study, measurements are taken at different dose levels for a pollutant of interest. These dose levels need to be selected in a controlled experiment, which is a design problem. We proposed a weighted c-efficiency criterion to obtain a design which is robust to risk function uncertainty and misspecification. Furthermore, a min-max weighted c-efficiency criterion is also developed to find a design not only robust to various risk functions, but also to different function parameters. In the simulation studies, we apply the particle swarm optimization (PSO) algorithm to search for designs. The presented methods for identifying robust designs is also demonstrated through the mammalian carcinogenicity of cumene (C9H12) example.