ABSTRACT
This chapter introduces key statistical concepts that are commonly employed in the review and approval of regulatory submissions. These key concepts include, but are not limited to, confounding and interaction, hypotheses testing and p-value, one-sided versus two-sided hypotheses, clinical difference versus statistical difference, and reproducibility and generalizability. In addition, some innovative designs and analysis methods for pharmaceutical development and evaluation are discussed. These innovative designs include complex innovative designs such as n-of-1 trial design, adaptive designs, master protocols, and Bayesian adaptive designs. During the review and approval process of regulatory submissions, some controversial and challenging issues that are inevitably encountered are briefly described. These controversial and challenging issues include totality-of-the-evidence versus substantial evidence, (1 − α) confidence interval approach for assessment of new drugs versus (1 − 2α) confidence interval approach for evaluation for generic/biosimilar drugs, endpoint selection, non-inferiority (similarity) margin selection, consistency test in multi-regional trials, sample size requirement for rare diseases drug development, extrapolation, drug products with multiple components, and the role of Advisory Committee. In addition, recent FDA critical clinical initiatives such as precision/personalized medicine, biomarker-driven clinical trials, model-informed drug development (MIDD), big data analytics, and real-world data/real-world evidence are also discussed. Aim and scope of the book are also described in this chapter.
