How to add flexibility to clinical trial designs

 Introduction

Adaptive clinical trialdesigns allow for changes to a study's design in response to accumulating data, making trials more adaptable, ethical, and efficient. These advantages are obtained while maintaining the trial's integrity and validity, thanks to pre-specification and suitable adjustments for anticipated charges throughout the trial. Unfortunately, adaptive designs have been reluctant to catch on in clinical research, despite significant statistical literature demonstrating their potential advantages over typically fixed designs. One important reason for this is that many clinical community members are still unaware of various trial design adjustments and their benefits and drawbacks. This blog aims to explain when adaptable designs can be utilized to answer certain scientific issues.

Considerations for Adaptive Designs

The 2019 guidance identifies four key principles to consider when designing an adaptive design trial:

1.     Controlling the chance of erroneous conclusions

2.     Estimating treatment effects

3.     Trial planning

4.     Maintaining trial conduct and integrity

Controlling the Chance of Erroneous Conclusions

One adaptive design method is to plan an unblinded preliminary test midway through the planned trial to see if an efficacy endpoint has been fulfilled. Early completion of an endpoint can cut down on the amount of time and resources needed for the experiment. However, if the endpoint is not met, the trial will proceed with a follow-up test after it is finished. In the second situation, increasing the number of tests will raise the final analysis' error probability. As a result, any consequences on the statistical validity of the final analysis should be taken into account during the design stage.

The statistical theory has been used in the past to ensure that type I and II errors are effectively handled in non-adaptive trials. It usually entails utilizing a pre-determined significance level, such as 5%. This strategy, however, is not practicable for designs that incorporate multiple parts. Clinical trial simulations may be a valuable tool to aid with adaptive trial design in such instances. Simulating hypothetical clinical trials under a set of assumptions can yield an estimate of error under those assumptions.

Estimating Treatment Effects

Changes in the data type in the primary analysis (e.g., endpoints, populations) could be one source of bias, making interpretations of treatment impact challenges. If available, methods for altering estimates to eliminate bias should be planned and implemented for reporting results. When such approaches are not accessible, the level of bias should be assessed at the very least, and treatment effect estimates should be given and interpreted with caution.

 Trial Planning

The quantity and timing of intermediate analyses, the type of adaptation(s), the statistical inferential methods to be utilized, and the precise algorithm driving the adaptation decision should all be included in the prospective planning. A thorough analysis plan created before starting the experiment enhances confidence that adaptive decisions were not taken based on haphazardly collected knowledge.

 Maintaining Trial Conduct and Integrity

The discovery of accumulating data in a trial may have an impact on the trial's direction and conduct, as well as the Sponsor's actions. As a result, it is strongly advised that access to comparative interim results be restricted to those not involved in the trial's conduct or management.

When planning for an adaptive trial, possible sources and consequences of trial conduct issues must be identified. Plans must be in place to avoid these issues, including processes intended to control blinding and document access throughout the trial. These and similar problems are often impossible to adjust after the data have been collected.

 Potential Challenges of Adaptive Design

Besides to the issues mentioned above, there are also potential drawbacks to consider when selecting an adaptive design. While an adaptive design may reduce the number of trials, crucial insights that more thorough analyses may have obtained following exploratory research may be lost during a hasty interim analysis.

It could lead to a failure to recognize safety concerns or other essential information about treatment response, interactions with concomitant medicines, or other factors. Such omissions can be costly and cause overall development times to be extended. In the end, adaptive design may not be the most appropriate approach for all clinical studies. For example, short studies (e.g., 2-8 weeks) in groups that can be recruited rapidly are included (i.e., less than 3-6 months) because recruitment must come to a standstill until interim analyses are carried out. An adaptive design, on the other hand, maybe well suited for lengthier research in which provisional data from a short-term endpoint (e.g., at six weeks) is used to anticipate a long-term endpoint (e.g., at 6-12 months), as pausing patient recruitment is unnecessary in this scenario.

When only a few concerns (e.g., dose, demographic subsets, and endpoints) need to be investigated, adaptive designs perform best and with the least risk. Running an exploratory trial before planning the "adequately controlled" trial may provide insights into some of these factors for projects with high ambiguity around several parameters. Thus, it can reduce uncertainty and make the approach more efficient and informative.

 Conclusions

Clinical trials with adaptive design may have several advantages over trials with traditional designs, including making prospectively planned changes to particular aspects of the study design and obtaining more informative and efficient research outcomes. Adaptive design, on the other hand, is not without its drawbacks. As a result, all research design decisions must be thoroughly evaluated and implemented.