ABSTRACT
From a statistical perspective, developmental and reproductive toxicology (DART) studies offer unique issues not found in other biological research. The objective of DART studies is to evaluate the effects of test article administration not only on a first generation of animals (F0), but also on a second (F1), and sometimes a third (F2). It is the information collected on the F1 (and later generation) animals that present unique statistical issues. In particular, two aspects of the study designs and objectives of such studies set them apart from other nonclinical, and even clinical, studies from a statistical perspective: F1 animals (i) are not all independent of one another, and (ii) in most cases, are not directly exposed to the test article. The objective of this chapter is to develop concepts to aid in
the understanding of statistics as it applies to DART. First, the DART study design will be reviewed from a statistical perspective, followed by some basic statistical concepts presented in a toxicological context. Next, statistical methodologies will be presented for data collected on the F0 generation. These same methodologies will then be adapted to the unique issues presented by the F1 generation. Finally, related issues are considered, such as specific details of the statistical aspects. This chapter is not intended to provide cookbook recipes for
applying statistical methods to data collected in DART studies. As such, formulas and calculations are kept to a minimum unless used to illustrate a concept. Rather, this chapter is intended to provide a basic foundation of statistical principles that will enable the reader to critically appraise the design, conduct, statistical methodology, and interpretation of DART studies in the literature and in a regulatory environment. The example data used to illustrate concepts have been
constructed to reflect real data in terms of magnitude and variation.
