Related Experiment Videos
A comparison of methods for estimating the benchmark dose based on overdispersed data from developmental toxicity
1Department of Mathematics and Statistics, University of Windsor, Ontario, Canada.
Summary
Joint modeling of prenatal toxicity endpoints, including malformations and prenatal death, provides more accurate risk assessment. This approach yields a lower effective dose (ED05) for overall toxicity compared to separate analyses.
Area of Science:
- Toxicology
- Risk Assessment
- Developmental Biology
Background:
- Prenatal toxicity can cause offspring malformations and prenatal death.
- Traditionally, these outcomes are assessed separately in risk evaluations.
- Multivariate methods offer a new approach to risk characterization.
Purpose of the Study:
- To evaluate the accuracy and precision of effective dose (ED05) and benchmark dose (BMD05) estimates in developmental toxicity risk assessment.
- To compare joint versus separate modeling approaches for analyzing multiple toxicity endpoints.
- To assess the impact of different variance structures and the Rao-Scott transformation.
Main Methods:
- Computer simulations were used to estimate ED05 and BMD05.
- Joint dose-response models were fitted using generalized estimating equations.
- Different variance structures (Dirichlet-trinomial, generalized linear model) were examined.
- The Rao-Scott transformation was applied to address overdispersion.
Main Results:
- Different variance structures yielded comparable results for joint modeling.
- The Rao-Scott transformation reduced bias and mean squared error for ED05 estimates.
- Joint modeling consistently produced lower ED05 values for overall toxicity than separate modeling.
Conclusions:
- Joint modeling of malformations and prenatal death is preferred for developmental toxicity risk assessment.
- This integrated approach ensures a more conservative and accurate estimation of risk.
- The Rao-Scott transformation is a valuable tool for improving ED05 estimates by reducing overdispersion.