Novel adaptive designs: aligning drug development and patient incentives

Abstract

In the November 2014 issue of the TIRS journal Stephen Spielberg [1] wrote an editorial motivating the future direction of the journal. He remarked: ‘We are making progress in drug development, with faster, smarter and more effective ways of evaluating the efficacy and safety of new products, all of which is beginning to transform into better regulatory practices with alignment in high-quality, validated science.’ Also, ‘We need to share together what works and what does not, all in the interest of patients in need and for all who strive to develop important new products.’ The aim of this piece is to begin a discussion to focus on sharing what works and what we could be doing to work more effectively when the objective is to consider the implementation of the Novel Adaptive Design (NAD) clinical trial as a pivotal Phase III trial in a regulatory setting. The NAD is continuing to gain favor as an alternative strategy for implementation of a pivotal trial in a regulatory setting. By ‘novel,’ we mean those adaptive designs that provide for an unblinded interim analysis in order to implement a predefined adaptive design decision. Since the issuance of the draft FDA adaptive design guidance [2] in 2010 we have come to understand that methods that provide for an unblinded interim analysis to estimate the treatment effect distinguishes the less-well understood Novel Adaptive Design (NAD) from the well-understood Adaptive Design [2,3] (AD). To clarify further, ‘less-well understood’ designs tend to entail using the unblinded treatment effect to potentially modify some aspect of the design while a trial continues. We can distinguish this from a conventional Group Sequential Design (GSD), which is characterized as ‘well understood,’ where the actions are built into the study design and typically only involve whether the study terminates or continues unchanged. NAD is being advocated in R&D programs as tools for optimizing clinical trial portfolio management [4]. The potential for increasing the probability of success, accelerated timing for product approval and reduced trial costs represent the optimistic advocacy on behalf of such designs. Kenneth Getz writes: ‘Adaptive trial designs hold promise in optimizing study design. Early study terminations due to futility and sample size reestimation could save up to a hundred million dollars (USD) in direct and indirect costs annually per pharmaceutical company depending on clinical trial scope, when the trial is actually terminated, and on the sponsor’s overall implementation of this adaptive approach across the development portfolio [5].’ If NAD for pivotal clinical trials hold the promise that is being advertised why are we not seeing more of them being introduced as the implementation method of choice?