Pharmaceutical companies will accelerate adoption of adaptive clinical trial designs, according to the Tufts Center for Drug Development, to reduce clinical development timelines and costs while increasing program success rates.
Adaptive design varies intervention based on patient response, not to cherry pick results, but to refine the gold standard of randomized clinical trials.
These trial designs, which can alter patient sample size or dosing during the course of a trial, require new technology or working practices to implement changes in randomization in the trial and to manage the trial’s clinical drug supply chain.
Wade Wirta and Steven Schwager sat down to discuss the strengths and weaknesses of adaptive design as well as the logistical considerations to manage these complex trials. Wade is a managing director of randomization and trial supply management at Medidata Solutions, and Steven is a research fellow at Medidata and professor emeritus of statistics and biological statistics at Cornell University.
What is an adaptive trial?
Wade Wirta: An adaptive trial is the type of trial design where you utilize data that’s being collected during the course of the trial in order to make adjustments to the trial design itself. These adjustments occur based on interim analyses and can drive changes in treatment arms, randomization ratios or overall trial timelines.
Steven Schwager: An adaptive trial monitors the data as the information comes in and uses the accumulated information to change the trial design when it becomes clear that this will make the trial perform better than continuing with the original design.
What are the advantages of adaptive trials?
Wade Wirta: Adaptive trials have two primary advantages. They can shorten the timeline of the trial, and they reduce the patient population and thus the trial supply that is required. Our industry understands the need to change the clinical development paradigm to yield greater productivity.
Cost pressures on R&D dictate maximizing effectiveness in executing clinical trials, and adaptive trial designs can be a major contributor in increasing productivity.
Also, treatments or doses that are futile endeavors are dropped early in adaptive trials, so there really is an ethical advantage by exposing the patients in the trial to better treatments as quickly as possible.
Steven Schwager: An adaptive trial can be more efficient – faster to complete, less expensive, and more likely to detect a treatment effect when one exists. In addition, it can benefit trial subjects by reducing the number of them who are assigned to the less effective treatment during the trial. This can increase the success rate – for example, the rate of complete remission by day 50 in a cancer trial – among the subjects in the trial by more than 20 percent.
What are the disadvantages of adaptive trials?
Steven Schwager: A major disadvantage is the additional complexity in having to determine whether to make a change, when to make it and what kind of change to make. There are lots of ways to modify a trial, and it’s possible, if not done carefully, for changes to make the design worse rather than better.
Thinking about the contingencies, what we should do and what we shouldn’t do is much more complicated than setting up an initial design and sticking with it from beginning to end.
Wade Wirta: Some of the complexities that Steve described also lead to operational challenges in adaptive trials. One primary example is in the clinical supply chain. The supply chain needs to be much more nimble and flexible, because changes in the study dictate changes in the supply chain.
In addition, there’s a lot more external interaction outside the core environment where you’re running the study. Data monitoring committees need to be able to review the data, and it can be a challenge to extract and analyze it, derive what changes are needed and express them in a way that drug monitoring committees can digest and use to make quick decisions.
Adaptive trials must be thought of not just as statistical methodology, but really as a change in the overall process and technology used to execute the trial. This adds complexities to the conduct of the trial itself.
When should you use adaptive trials?
Wade Wirta: There’s a lot of misconception in the industry about when to apply adaptive trial designs. Clinical trials fail for a variety of reasons. For example, the compound is either not effective or safe, the statistical power of the study isn’t sufficient to reach a definitive conclusion or enrollment doesn’t reach required levels.
Adaptive trials can’t address all of these risks but are able to mitigate some of the risks associated with the statistical design of the study.
Adaptive trials help improve the statistical power throughout the course of the trial by looking at the data being collected and making adjustments. A lot of Phase III research uses adaptive design, but you can apply the same concepts to Phase I and Phase II as well.
Steven Schwager: Conceptually, it is wise to modify the design of a trial when it becomes clear that the trial we should be performing differs from the trial we are actually running, because of what we have learned in the trial’s early stages. The greater the difference between what we are doing and what we now know we should be doing, the more we can gain by changing the design to reflect our best current knowledge.
How do you use the results to steer a trial in a more efficient direction?
Wade Wirta: By their very nature, adaptive trials are intended to change certain parameters as the trial is ongoing. The intent is always to make sure you come up with the best treatments targeted to the right patient populations. You’ll consider modifications like changing randomization ratios, adding or dropping treatment arms or even stopping the trial completely as a result of futility analysis.
Steven Schwager: In addition to the changes Wade just mentioned, the accrual period can be extended if it’s necessary to bring in more patients than originally planned. For example, it may turn out partway through the trial that the dose we started with isn’t the best dose for the patients. We want to change to the best dose as soon as we become aware of this improvement.
What is involved in the actual planning of an adaptive trial?
Steven Schwager: Planning an adaptive trial requires thinking ahead as much as possible to anticipate the changes we want to make. Knowing that we can’t plan for every contingency, we want to do the best we can.
Wade Wirta: Because of the complexities that we’ve discussed, there really is a need for much more coordination from a planning perspective. Clinical operations are really about defining the protocol, defining the study and executing with precision. It’s very important for the supply chain to be a part of that process.
It’s vital to run supply chain simulations so you understand the impact to supply chain logistics during the adaptation process.
I mentioned forecasting and planning earlier. This is critical for functions like packaging and labeling. Drugs need to be packaged in a way so that you can change treatments or allocation and still utilize existing drug supply. Procedures like just-in-time labeling also help introduce the required flexibility in drug supply that these designs require.
This applies to external interactions as well. The interim analysis process from a procedural perspective needs to be clearly articulated, as well as the process for the interaction with data monitoring committees. Understanding roles and responsibilities, timelines and the technical implications for these external interactions is key to reducing the risk of disruptions once the trial is underway.
Steven Schwager: Yes, keeping everyone on the clinical trial team in the loop is an additional complexity because, as Wade just said, many people and organizations are involved in a trial. Everyone needs to know when a change is going to be made. It can’t happen that some people know it and some people don’t. Communication truly matters here.
What is the FDA guidance around adaptive trials?
Wade Wirta: The FDA issued draft guidance in 2010, which has helped greatly with regulatory concerns and implications on how to run adaptive trials. The FDA really wants to push the industry to adopt adaptive trial design on a larger scale. In 2015, theFDA issued draft guidance for adaptive trials for medical devices.
Steven Schwager: The FDA guidance is a big step forward, because nobody wants to run a trial, however brilliantly conceived and executed, and then be told by the FDA you can’t do this because we insist you do it another way.
How can sponsors minimize supply management risks?
Wade Wirta: The complexity of the supply chain increases greatly with adaptive trial designs. The supply chain needs the flexibility to be able to adjust to changes, which really puts the onus on the planning of the trial. There must be tight integration between the clinical operations team and the supply chain planners.
As described earlier, forecasting and planning, which are important in traditional trials, become imperative in an adaptive design, so any changes to the trial design parameters need to be known and assimilated quickly into the forecasting and planning process.
Usually, the drug supply overage amounts increase substantially for adaptive trials. Historically, sponsors have taken a conservative approach by producing as much drug supply as possible upfront to reduce the risk of outages. In the case of biologics, where manufacturing challenges may limit available drug product, optimizing your drug supply is especially important.
Simulations will help ensure that you reduce overages as much as possible. Overage costs, however, really pale in comparison to the savings provided by adaptive trials in reduced clinical development time and cost.
Steven Schwager: As we mentioned earlier, the need for communication is crucial, because with changes unfolding in the design, it’s really vital for everyone to know what’s going to happen. Surprises are just not acceptable in a well-run clinical trial.
This article originally appeared on our Forbes BrandVoice channel.