FDA Draft Guidance on Adaptive Designs for Medical Device Trials

The Food and Drug Administration (FDA) has historically supported, and often encouraged, the use of adaptive trial design in clinical trials in order to increase the success rate of a trial. Last week, the FDA released draft guidelines concepts in medical device trials, stating that, “adaptive clinical designs can be used in medical device studies as long as they are for prospectively planning modifications based on accumulating study data without undermining the trial’s integrity and validity”.

That being said, the FDA draft guidance applies specifically to medical device trials including applications for premarket approvals, 510(k)s, de novo, humanitarian device exemptions and investigational device exemptions. The FDA announcement comes at a time when Sponsors face various challenges and uncertainties in conducting medical device studies including:

  • Safety Problems
  • Unexpected treatment effects
  • High variance
  • Effects in secondary endpoints and subpopulations

One of the challenges Sponsors will face is understanding the implications of adaptive trial design and how to properly present it to the FDA.

For example, the most common parameters that are subject to adaptation include sample size calculation, treatment arms and hypotheses. Statisticians apply sequential designs based on accumulated data to reassess the sample size. This is a significant benefit according to the FDA, Because the chance of an erroneous positive conclusion is no longer well controlled, the approach of simply extending a study at the end in a manner that is not pre-specified is neither scientifically sound nor recommended. In contrast, an adaptive design can permit sample size reassessment and appropriately control the Type I error in hypothesis testing or, correspondingly for interval estimation, the confidence coefficient”.

Another method encouraged by the FDA is the Bayesian Framework in which statisticians combine prior information and new experimental data to produce the “posterior probability”. Bayesian models can be easily applied to Adaptive Designs, sharing the same idea of translating information into a decision as quickly as possible. Therefore, studies doomed for failure can be terminated earlier to reduce the number of patients being treated with a non-effective device or drug.

Implications of the FDA’s Draft Guidance on Medical Device Trials

Implementing an adaptive design approach can be beneficial to medical device Sponsors in terms of greater trial efficiency, cost savings and ethical considerations (i.e. early termination and avoiding risk to patients). However, the FDA points out, “procedures to assure the proper conduct of adaptively designed studies must be put into place so the study will provide valid scientific evidence that can be relied upon by FDA to assess the benefits and risks of the investigational medical device. Sponsors are strongly encouraged to discuss the planning of adaptive clinical study designs with the appropriate FDA review division in advance”.

Therefore, it is crucial to involve a biostatistician in the planning phase including in pre-submission briefings with the FDA.

A very detailed description of methods and adaptations have to be in the protocol along with setting up an independent Data Monitoring Committee. In order to have fast access to data and efficient data management, sponsors should consider electronic data capture, a central database, central randomization, prompt data entry and efficient data cleaning processes.

CROS NT and Adaptive Trial Design

CROS NT has a group of statisticians with 20-30 years of experience in adaptive trial design who offer consultancy on a case-by-case basis.

CROS NT can assist with the following regarding Adaptive Trial Design:

  • Work with Sponsors to improve the trial design
  • Use simulations to calculate the power of the sample size and probability of success
  • Define the decision rule for interim analyses and provide statistical justification
  • Provide Electronic Data Capture and reporting to efficiently manage data for interim analyses