The FDA released its draft guidance on “Adaptive Design Clinical Trials for Drugs and Biologics” in 2010 and in so doing provided a framework for designing more efficient and often more informative clinical trials. The inclusion of adaptive design elements in a trial can reduce the number of patients needed, decrease trial duration, and provide more informative trial results.
Adaptive design is of particular interest for studies intended to support marketing – in other words, those studies that fit the definition of “adequate and well-controlled.” However, the utility of adaptive design elements is not limited only to these types of trials, but may also be quite useful in earlier phase studies, including exploratory studies.
In exploratory phases, adaptive designs can allow initial evaluation of a broad range of design choices with the opportunity to discontinue evaluation of suboptimal choices. Because exploratory studies have less impact on regulatory approval decisions, FDA encourages sponsors to gain experience with adaptive design methods in these phases.
Adaptive Design vs. Conventional Clinical Trial Design
So, what is adaptive design? The draft guidance from the FDA defines adaptive design as “a study that includes a prospectively planned opportunity for modification of one or more specified aspects of the study design and hypotheses based on analysis of data (usually interim data) from the subjects in the study.” The key concepts for defining an adaptive design are “prospectively planned” and “analysis of data from subjects in the study.” This means that 1) any modifications or “adaptations” must be planned prior to any unblinded data analysis by anyone involved in planning or implementing the modifications and 2) that decisions to implement modifications must be based at least in part upon study-internal data.
A familiar example might be an early ascending dose study, which will likely employ prospectively planned interim reviews of pharmacokinetic and safety data by a specified committee that will determine whether the study should proceed to a new dose or be stopped. Defining and prospectively documenting who is on the committee and the criteria for which decisions to stop, repeat, or proceed to a higher dose are made establishes a prospective plan for altering the design of the study based on data internal to the study.
To further clarify what constitutes adaptive design, it might be helpful to touch briefly upon key design aspects of a conventional trial. In a conventionally designed clinical trial, critical study parameters are planned using assumptions and best estimates of things like population means and event rates, variance, dose-response effect size and location, and discontinuation rates. Such a design can work well when these estimates and assumptions are accurate, but problems can arise if they are not. To remedy this, conventional study designs often include elements to reduce risks associated with uncertainty. For example, if the study is meant to determine dose-response, the protocol may include multiple fixed-size randomized groups to ensure that an optimal dose is captured. Such design decisions are made with the understanding that several groups will likely be treated with suboptimal doses. In this way, the study design trades efficiency in exchange for reducing the risk that the optimal dose will be missed.
But what if you didn’t have to compromise efficiency to ensure, for example, that the optimal dose is captured? Going back to the example of an ascending dose study, the application of appropriately and prospectively designed modification rules may allow you to alter an otherwise rigid escalation plan in the event that drug exposures are not escalating as initially anticipated. In contrast to a conventional study design, applying an adaptive design could also allow identification of subpopulations with differing dose responses as data are being collected, leading to the discontinuation of data collection for patients in suboptimal groups, and potentially saving both time and money, without negatively impacting study outcomes.
Another example of where adaptive design might be useful is in the determination of an appropriate sample size for the study. In a conventional trial, there is a risk of underpowering a study if estimates of variance and treatment effect are overly optimistic; likewise, if these parameter estimates are too conservative, then you can end up with an excessively large study population. If a study is underpowered, then the study data will not be sufficiently robust from a statistical standpoint to support meaningful conclusions. If the study population is too large, then significant amounts of time and money can be wasted. Using an adaptive design, sample size could be adjusted based upon accumulating study data in such a way that these undesirable outcomes can be avoided.
Together, these examples highlight the central advantage of adaptive design approaches, which is, to reiterate, the ability to include prospectively planned opportunities for modification of one or more specified study design elements and hypotheses based upon analysis of data from study subjects at prospectively planned time points. Again, the keys here are that the modifications must be prospectively planned and that they must be based at least in part upon study-internal data.
A Word of Caution
Before we get too far ahead of ourselves, it is a good idea to point out what adaptive design is not. First, just because a study is modified does not mean that it is “adaptive.” For example, studies where design features are changed, but not prospectively planned, do not qualify as adaptive design studies. Indeed, unplanned changes to the study design in response to an unblinded interim analysis introduce operational bias and can raise serious concerns about the integrity of the trial data.
In addition, revisions to the study design based on information obtained entirely from sources outside of the study are also not considered adaptive designs, regardless of when they were planned. This includes, for example, changes in dose or dosing regimen that are based upon new pharmacokinetics information obtained from a separate trial. While such modifications may make sense on a practical level in certain contexts, it is not an adaptive design.
What Can Be Modified?
Now that we’ve learned a little about what adaptive design is, what it is not, and a few of the advantages that it offers, let’s take a moment to highlight some of modifications that can be considered in an adaptive design clinical trial.
Specific examples of prospectively planned study design modifications include:
- Eligibility criteria
- Randomization procedures
- Treatment regimens of future study groups
- Study sample size
- Concomitant treatments
- Schedule of evaluations
- Primary endpoint
- Selection and ordering of secondary endpoints
- Analytical methods for evaluation of endpoints
While these parameters are common candidates for adaptive design decisions, not all may be appropriate for every trial. Indeed, careful consideration should be given to which parameters to make “adaptive,” as bias may be inadvertently introduced through inappropriate choices or by making the trial too flexible.
Potential Pitfalls of Adaptive Designs
As suggested above, even designs that qualify as adaptive can pose challenges. One of the most prominent ways this happens is through the introduction of bias into the trial. One source of potential bias may arise from the existence of multiple design elements that can be changed at one or more timepoints during the study. Another potential source could be changes involving the type of data in the primary analysis (e.g., endpoints, populations), which may make interpretations of treatment effect difficult. Also, operational bias may be introduced if analysts or investigators charged with implementing modifications are unblinded.
Because of the potential for introducing bias into adaptive design trials, any modifications to the study design should be limited to the most critical and least understood parameters (e.g., dose, schedule, population, endpoint, concomitant therapies) and should be planned and implemented only by those who remain blinded to the study data. Additionally, the potential impacts to the statistical validity of the final analysis need to be considered. It is important to understand, however, that even in a well-planned adaptive design trial, uncertainties may still exist and sources of these uncertainties should be actively investigated.
Furthermore, while time efficiency is one of the advantages of adaptive design (e.g., by eliminating an exploratory study in favor of including exploratory goals within an adaptive design study), it is possible that critical insights may be missed during a rapid interim analysis that might have been captured by the more thoughtful analyses that tend to occur following an exploratory study. This may result in inadequate recognition of safety issues or other critical information related to treatment response, interactions with concomitant therapies, or other variables. Such oversights can be costly and may extend timelines for the overall development program. Therefore, allowing sufficient time for interim evaluations prior to the implementation of prospectively planned alterations is highly advisable to avoid counteracting any efficiencies to the program that might otherwise be gained.
Finally, it is important to realize that adaptive design may not be a great option for all clinical trials. Indeed, adaptive designs tend to work best and with less risk when only a few issues (e.g., dose, population subsets, endpoints) need to be examined. For programs where there is significant uncertainty around many parameters, running an exploratory trial prior to designing the “adequate and well-controlled” trial may provide additional insights regarding at least some of these parameters, thereby reducing uncertainty and making the approach more efficient and informative.
Adaptive design clinical trials can offer key advantages over conventionally designed trials, including the flexibility to make prospectively planned modifications to certain elements of the study design to achieve more informative and efficient study outcomes. However, adaptive design is not without risks and it is imperative that all study design decisions are thoughtfully made. If done properly, including adaptive design elements may speed up the overall drug development process and help get products to market quicker and at a lower cost.
Are you considering designing an adaptive clinical trial? Contact one of our senior scientists to ensure you get the most out of your study design.