ICSA Webinar - Preclinical Safety Statistics: Introduction, Perspectives, and New Developments

Preclinical safety studies are central to pharmaceutical development, informing decisions from first‑in‑human dosing through late‑stage clinical trials. These studies—including general toxicology, safety pharmacology, and reproductive and developmental toxicology—pose statistical challenges driven by complex designs, heterogeneous endpoints, limited sample sizes, and regulatory and animal‑welfare constraints. This webinar reviews core statistical design and analysis principles for preclinical safety assessment, including dose‑response evaluation, trend‑based testing, handling monotonic and non‑monotonic responses, baseline adjustment, sex and time effects, and the use of historical control data. Emphasis is placed on how these choices affect power, Type I error control, and interpretability in highly automated analysis environments. The session then focuses on emerging developments that reshape study efficiency and decision‑making. Virtual Control Groups (VCG) are discussed as a design and analysis strategy that leverages external and historical data to reduce reliance on concurrent controls while preserving operating characteristics. Bayesian methods are presented as a complementary framework for uncertainty quantification and information borrowing, enabling transparent evaluation of trade‑offs through operating‑characteristic assessment. Examples illustrate how these approaches can be integrated to support robust, ethical, and future‑ready preclinical safety studies.