Modern Approaches to Clinical Trials Using SAS: Classical, Adaptive, and Bayesian Methods

9.1 Introduction

Randomized clinical trials are the gold standard to differentiate effects of different treatments. To ensure an objective comparison of different treatments, patients' allocation should be random or unpredictable so that both known and unknown prognostic factors and potential confounders can be balanced out. In practice, selection bias should be prevented as much as possible by using randomization and blinding, including single- or double-blind schemes. A single-blind study only blocks the identity of the treatments to patients, while a double-blind study blocks both patients and clinicians. Most clinical trials attempt to balance the treatment assignment by restricting the allocation probability equally across arms, but it may ...