Journal Article

Commentary on Hey and Kimmelman

Saxman, Scott Brian

2015

Overview

Prospective clinical trials have been the cornerstone for testing new therapies in medicine for several centuries, dating back at least to James Lind who is acknowledged for conducting the first such study for the treatment of scurvy in 1747. Randomization has also been firmly established as a means to minimize bias and enhance precision when two or more therapeutic interventions are being compared to each other. Traditionally, the randomization allocation in a particular clinical trial is fixed throughout the duration of the trial, using either a balanced approach in which the distribution of patients into each arm is equal (1:1) or an unbalanced schema in which the allotment of patients is unequal (2:1, 3:1, etc.) with the higher number of patients usually being assigned to the experimental arm. A relative newcomer to the clinical trials scene is a trial design using a randomization method referred to as outcome-adaptive randomization (OAR). In this type of comparative clinical trial, the randomization probability is not fixed, but adapts throughout the conduct of the study. Specifically, as clinical outcome data become available, the randomization scheme is modified such that more patients are enrolled in the arm for which the results seem to be more favorable. This adaption of the treatment allocation is fluid, and may go up or down throughout the duration of the trial as more clinical data on the patients who were previously enrolled become available. The ethical intent of OAR is that at the end of the trial, more patients will end up being treated with the better therapy, and thus fewer patients will receive an inferior treatment.

• Publisher: SAGE Publications,Sage Publications Ltd
• Subject: Clinical trials / Ethics / Humans / Patients / Random variables / Randomized Controlled Trials as Topic - ethics / Research Design