As it’s name implies, the goal of adaptive design is flexible trial design that can be altered in response to the accumulation of real-time data acquired during the study. The guidance applies to all sorts of clinical trials, from simple feasibility studies to crucial pivotal trials for FDA approval.
The potential acceptance of smaller sample sizes than originally planned is one of the chief benefits of adaptive clinical trial design. The FDA said a previous trial of a bone graft achieved this by planning for an interim analysis once the trial reached 340 patients. Since the data was positive, the FDA approved the device, even though a sample of up to 500 patients was planned.
In addition, the sample size can be determined based on a predefined metric, the FDA says. For example, a trial can be designed so that enrollment ends once a specified standard deviation (or variance) in the primary endpoint is achieved, the guidance explained.
Another example of adaptive design is a change in the claim or hypothesis of the study. In that regard, “two different strategies may be used: one is to plan the study as a superiority trial and have a fallback hypothesis of non-inferiority; the other is to plan (and size) the study originally as non-inferiority but allow for an investigation of superiority,” the FDA said.
The document stressed that the use of alternative sample size determinations and other design adaptive components should be determined in consultation with the FDA and prior to the initiation of the trial. The FDA said unplanned changes are also allowed on a case-by-case basis, but they will face greater scrutiny from the agency.
Adaptive design should be carried out in a manner that minimizes the chance of erroneous conclusions and operational bias, the agency says. One risk of a midstream change, such as increase in the sample size, is that “investigators, study subjects and/or third-party evaluators may behave differently, either unconsciously or subconsciously, if the existence or size of the increase, or the reason for the increase, becomes known to them,” the FDA warns.
The guidance contains extensive discussion about the types of a trial for which adaptive design is appropriate (such as those with shorter endpoints and longer recruitment times), and the special challenges faced when making adaptive changes using unblinded data, since that carries a greater risk of inducing bias among a trial’s investigators and/or patients. The use of computer simulations and the role they can play in making adaptive changes is also reviewed in the guidance.
The plethora of recent guidance documents about topics related to the clinical trial paradigm, like the use of pre- and postmarket data, acceptance of data from outside the U.S, and extrapolation of existing clinical data, means that industry observers should pay close attention for signs that the new clinical trial tools and tricks they enable are actually being accepted by the agency and implemented in companies’ clinical trials.
REFERENCE: Fierce Medical Devices; 22 MAY 2015; Varun Saxena