Adaptive design: Points to consider
Informed consent, therapeutic misconception
Adaptive design clinical trials raise many of the same issues in IRB review as do unplanned modifications to existing trials, says Marjorie Speers, PhD, executive director of the Association for the Accreditation of Human Research Protection Programs (AAHRPP) in Washington DC.
"IRBs are faced every day with adaptive designs, because every day, they approve research studies and something will change in that study," she says.
But adaptive design does add a few wrinkles to some of the usual IRB issues:
Informed consent: Researchers and IRBs must decide how much to tell subjects about the potential changes that might occur during the study.
The informed consent for adaptive studies should explain to participants that there will be examinations of data at set points in the study, approximately when those points will occur and what kinds of changes might result, says Marc K. Walton, MD, PhD, associate director of the Office of Translational Science of the FDA's Center for Drug Evaluation and Research in Silver Spring, MD.
"For example, (the informed consent) might state that two of the four dose groups will no longer have patients randomized to them, and only two of the dose levels will be continued for the assignment of new patients," he says.
Participants would need to be told if they were being enrolled after a decision point in an adaptive trial.
The complexity of some versions of adaptive design can lead to increased challenges in making consent understandable, says Daryl Pullman, PhD, an associate professor of medical ethics at Memorial University of Newfoundland in St. John's.
"Consent documents tend to be very long and technical already," Pullman says. "And I think they could become even longer and more technical if you're trying to explain all the intricacies of an adaptive design to potential research subjects. More description in those situations can actually undermine the informational content of it."
He says he doesn't believe it's necessary to reconsent subjects once the study is under way every time there's an adaptation.
Therapeutic misconception: Pullman also worries that the nature of an adaptive design changing interventions as researchers know more about what seems to work will make it more difficult for subjects to grasp the difference between treatment and research.
"Adaptive design tends to cloud that," he says. "I think it's going to become more difficult for clinicians to make that distinction."
The distinction might be especially difficult if the adaptive design calls for particular groups of subjects to be moved from one arm of a study to another.
Walton says that while it is possible to design an adaptive trial that could result in changing an already enrolled participant's treatment increasing or decreasing a dose, discontinuing treatment to one arm of a trial, etc. most designs he sees use interim data to change enrollment patterns for future patients.
"The adaptations that have been focused on in the (FDA) guidance document are on what happens to patients who are enrolled after the point of interim analysis, not changes to the management of patients already enrolled in the study," he says.
Computer simulations: One new tool that IRBs may find helpful in reviewing a proposed adaptive design is a computer-generated trial simulation, which runs through potential scenarios for adaptations proposed during the trial to see how they might play out. The FDA draft guidance on adaptive designs says such simulations can help evaluate various design options ahead of the trial.
"A report of the study simulations that were conducted will be extremely helpful to FDA in understanding what may occur during the study," Walton says. "And I think that IRBs will find that also to be very valuable to look at. It's a new kind of information that they are not currently familiar with seeing."
He says FDA hopes that simulation reports will be sufficiently understandable to people who are not statistical experts, but IRBs still might want to seek outside help when evaluating some of the more complicated studies.
"It might be that in some of these really complex ones, they'll want somebody who has some specialized statistical expertise to help advise them on whether or not the simulations have adequately assessed what might happen," Walton says. "We are hoping that the simulation report itself will be understandable to people without deep statistical expertise."