Adaptive design (AD) is a clinical trial design that uses accumulating data to decide how to modify aspects of the study as it continues, without undermining the validity and integrity of the trial. AD appears to be growing at a significant pace.
At a recent symposium, there were at least 14 individual presentations on this topic and, according to FDA’s Boguang Zhen, approximately 10% of the INDs and IDEs submitted for review in the years 2008 through 2011 contained adaptive designs.
Why are we interested in adaptive designs? Personalized Medicine has so many possible ways to “personalize” it that it is virtually impossible to conduct full formal studies on them all. We rely on biological and medical knowledge to limit the field to the most promising of these and then devise studies that are intended to be efficient and conclusive. There has always been a problem in optimizing the dose (regimen). Many products have been studied to approval at one dose level and when unexpected adverse events show up, there is not enough safety and efficacy information available at other doses to allow the product to stay on the market. Part of the AD effort seems to be to carry more reasonably promising doses (regimens) longer and to study the relationships of effectiveness and safety to dose more thoroughly.
The FDA Guidance
So does the use of AD solve this problem? Is it all good? Not so fast. In February 2010 the FDA issued a draft guidance “Adaptive Designs for Clinical Trials of Drugs and Biologics” that separates ADs that the FDA considers well understood from those for which there is insufficient information. It also describes the nature of an acceptable study that uses AD. A major issue is that potential adaptations must be planned in advance (including all possible outcome branches of the decision tree) and very well documented. It is a good idea to get agreement with FDA prior to initiating any study that uses an AD.
In an AD, it is expected that there will be little time lost in the study at a decision point as it moves on under the (possibly) modified design. It is also expected that the data before and after the modification will be used in the final analysis, and that all possible outcomes and the resultant change in direction, if any, must be considered.
Consequences of using AD
AD is intended to bring efficacious and safe products to market faster and more efficiently (cheaper). However, its use in conjunction with biomarker development means that the population identified as finding the product to be useful will be only a subset of the diseased population. It also means that the safety profile of the product will be substantially less than we are accustomed to having for new products. The consequences both to this subset population and to the unserved subpopulations are not yet clear.The author is a consultant in The CECON Group network and has extensive experience in regulatory statistics in both clinical and preclinical areas. For 25 years, he was a statistical reviewer, supervisor, and manager in the Center for Drug Evaluation and Research (CDER) of the FDA. His expertise includes drug quality, stability, clinical trials, and data integrity. Click here to read more about this regulatory statistician consultant and expert witness.