Risk managing your development with novel trial designs

Risk managing your development with novel trial designs
10th February 2017 Sue Carr

Novel, adaptive designs are seen by many drug developers as highly attractive as they seemingly offer more for less: more in terms of considerable flexibility in trial execution, increased chances of success and opportunities to stop early for efficacy or futility, and less in terms of cost and time.

However, realising the potential benefits of an adaptive design is not straightforward. To do so requires in-depth expertise and experience, both technically in terms of statistical simulation and also with the likes of EMA and FDA who rightly have reservations regarding the use of such designs to provide pivotal, phase III evidence.   KJC Statistics is fortunate to have considerable experience in designing adaptive trials and successfully helping Sponsors navigate the ensuing discussions with regulatory agencies.

Recent examples of such designs provided by KJC statistics include:

  • Phase III enrichment designs involving the overall and a biomarker defined subpopulation, with an interim decision point on whether to proceed with the overall population, the subpopulation or both – positive feedback was received from both the FDA and EMA and study is now going forward.
  • Novel phase III design with an early, biomarker driven interim for futility and, if not futile, a sample size re-estimation is affected. This is then followed by a second interim to deliver data that could be used to support an application for accelerated approval while the trial continues recruitment and follow-up to a final analysis on clinical outcomes.
  • Oncology phase II ‘bucket’ design where a new drug is simultaneously trialled in a number of different phenotypes, with frequent interim analyses that use Bayesian methods to combine response data across phenotypes with decision rules that allow for the early curtailment of phenotypes for either futility or activity.

If you are considering such designs, or think an adaptive design might be the right choice for your development, contact us here at KJC statistics, we will be delighted to help you.