Design and Analysis of Randomized Clinical TrialsDesign, Analysis & Theory

Published October 2007 Updated September 2016 18 lectures
Dr. Vance Berger
Bethesda, MD, USA
Summary

Evidence-based medicine is one of the corner stones of the evaluation of all types of medical interventions, including drugs, vaccines, diets, and complementary medicine.... read more

At the very top of the evidence-based hierarchy is the randomized clinical trial, which by its very design offers a parallel control group, and often masking and allocation concealment too. Many of the biases that can compromise the integrity of other designs are eliminated by randomized clinical trials, but some of these biases are eliminated by the add-ons of masking and allocation concealment, which do not necessarily apply to all randomized trials, or even to all randomized trials that are labeled as masked and with allocation concealment.

Generally, the wrong definition of masking and allocation concealment are used, so that only the effort, but not the result, is taken to be satisfactory. Moreover, other issues also need to be considered in the design, conduct, analysis, and interpretation of randomized clinical trials, including multiplicity (multiple endpoints, multiple time points, and sometimes interim analyses), surrogate endpoints (including the often misconstrued meaning of the “validation” of surrogate endpoints by the Prentice criterion), noncompliance, missing data, and ethical considerations.