EFSA Platform for Bayesian Benchmark Dose Analysis
Meta data
Disclaimer: The present document has been produced and adopted by the bodies identified above as author(s). In accordance with Article 36 of Regulation (EC) No 178/2002, this task has been carried out exclusively by the author(s) in the context of a grant agreement between the European Food Safety Authority and the author(s). The present document is published complying with the transparency principle to which the Authority is subject. It cannot be considered as an output adopted by the Authority. The European Food Safety Authority reserves its rights, view and position as regards the issues addressed and the conclusions reached in the present document, without prejudice to the rights of the authors.
Abstract
The main goal of this grant is to contribute to the harmonization of BMD approaches across international organizations and to contribute to the development of an EFSA online platform that performs Bayesian benchmark dose (BMD) analysis. The objectives include (i) to further develop the prototype Bayesian Model Averaging for continuous data (exploring the inclusion of covariates, as well as exploring models including litter effects) using the new family of models, (ii) to develop Bayesian Model Averaging for quantal data (exploring the inclusion of covariates, as well as exploring models including litter effects) using the new family of models, and (iii) updating the R4EU web platform to include Bayesian model fitting. In this scientific report, we extensively describe the developed methodology. Bayesian Model Averaging was developed for continuous and quantal data using 8 candidate models. Additionally for continuous data, the normal as well as the lognormal distribution are considered, resulting in 16 candidate models. The use of informative prior distributions was implemented for both types of data. For both types of data, the methodology was extended to the case of clustered data, and a first exploration of the inclusion of covariate effects was done.