Learn and Confirm: A Push for Adaptive Clinical Trial Design
Adaptive designs help get new drugs to market faster and more economically.
By Tom Branna
Editorial Director
Don’t only blame the compound for the dismal success rate of new drug applications—also think about the clinical trial process, insist a growing number of pharmaceutical researchers. To help remedy the recent woeful track record in drug development, a cadre of pharmaceutical companies say there is a better way to conduct trials and they’re urging regulators to re-examine the approval process to help get better drugs to market faster and more economically.
Clearly, there’s a need for better trial methods. According to industry sources, while R&D expenses have more than doubled to nearly $40 billion during the past decade, the number of new molecular entities has plunged from more than 70 approvals in 1997 to about 20 in 2005.
Adaptive clinical trial design uses accumulating data to determine how to modify aspects of the study as it continues—without undermining the validity and integrity of the trial. According to proponents of the method, the goal of adaptive clinical trial designs is to learn from the accumulating data and apply what is learned as quickly as possible.
“There is a tremendous amount of pre-work and planning that goes into adaptive clinical trial design,” noted Michael Krams, assistant vice president, adaptive trials, clinical development, Wyeth Research. “Adaptive clinical trial design is a pre-designed process to adapt to new information.”
For example, according to Dr. Krams, many acute stroke clinical trials fail. Yet, in the preclinical trials, the compounds under investigation worked.
“One inference would be that the pre-clinical guys got it wrong, but maybe the clinical trial setup was wrong and could not show what the drug could actually do,” he explained. “Thinking about adaptive trials forces us to investigate the basics around clinical drug development even harder: Does the clinical endpoint make sense? Is it properly translating the preclinical observation? Can we apply the rigor of the preclinical experiment in a clinical trial setting? Do we have appropriate quality controls built into the clinical trial? Adaptive trials are a platform facilitating the integration of knowledge, and translational medicine and modeling based drug development are key enablers.”
Adaptive clinical trial design would have the biggest impact on three key areas designs have the potential to improve the information value of clinical drug development. In a first step, modeling based adaptive dose-response finding studies can improve our understanding of the dose-response and help us choose the correct doses to be taken into phase III.
According to Dr. Krams, the inefficiencies of traditional parallel-group dose-response finding studies have led to studying only a small number of doses, before making the decision on the dose to carry into phase III. An alternative approach may initially take a larger number of doses in “Learn” into an efficient modeling based adaptive-dose response finding study, but quickly eliminate non-viable treatment arms, either because they are unsafe or not efficacious. Once there is sufficient understanding of the viable dose-range, a small number of doses could be taken into a confirmatory phase III trial.
In a seamless phase II & III design, researchers would select the two or three doses that performed best in earlier studies and continue on in the trial. Using this tool, researchers would use a single trial to achieve the same results that are normally achieved through separate trials in phases IIb and III. Dr. Krams cautioned that that for such a trial to be considered a “confirmatory trial.”
"This method requires a case by case review with regulators, so the challenge is to find the right process that allows for interaction with heath authorities,” observed Dr. Krams.
In fact, PhRMA recently held a two-day conference to take a closer look at seamless phase II & III design.
“The consensus was that we must take any proposal on a case-by-case basis,” he noted. “There is no routine application for seamless phase II & III design.”
Sample Size Reestimation in Phase III
The number of patients enrolled in a clinical trial is determined based on the knowledge available and assumptions made at the time of designing the trial. According to Dr. Krams, sample size re-estimation ensures that assumptions made when a compound first goes to trial are correct.
“For example, what about data variability?” conjectured Dr. Krams. “What if our assumptions about variability are incorrect at the time the trial started? We should be able to adjust sample size accordingly.”
He cautioned, however, that adaptive clinical trial design, does not necessarily result in time savings. In fact, he predicted that in as many as a third of trials, adaptive clinical trial design may actually increase the time needed to develop a drug.
“It is really about improving information value,” he insisted. “It’s not about reducing the number of patients. It’s about putting the right number of patients on the right treatment.”
Moving Forward
Wyeth has been among the industry leaders in adaptive clinical trial design and is conducting and planning such studies in a number of different therapeutic areas.
“It’s not rocket science,” insisted Dr. Krams. “People have conducted interim analyses for years, and we are now taking this principle to the next level.”
While Wyeth is at the forefront of adaptive clinical trial design, other pharmaceutical companies are heavily involved in the issue as well. As Dr. Krams sees it, the industry is divided among three camps:
• Those who are very interested and have the courage to engage in adaptive clinical trial design;
• Those who will never touch it; and
• Those who are taking a wait-and-see approach to adaptive clinical trial design.
He predicted that during the next two to five years, more pharmaceutical companies will embrace the concept of adaptive clinical trial design, but cautioned that much depends on early success.
“The early protagonists must do their job well,” he insisted. “If we make a strong case for adaptive clinical trial design, it will be well-received.”
