Report on the current state of play in the adoption of adaptive trials designs by Kenneth Getz of Tufts CSDD

Adaptive trials designs Kenneth Getz reportThere’s a really good article in Applied Clinical Trials on the state of play in the industry of the use of adaptive clinical trials by Ken Getz of Tufts, click here for full article  reports that in their survey of the industry, about 20% of phase III (late stage, confirmatory) trials are adaptive; these will almost certainly be group sequential designs or versions of these that only stop early for futility. Study designs that allow early study termination for futility are reportedly easily implemented and the most common form of adaptive design (presumably in both phase II and phase III).

In line with our own thinking and experience, however, Getz reports that there are still very low rates of “sophisticated adaptive designs during the exploratory phase clinical trials” that “could have the greatest impact on improving late stage success rates”.

Again in line with our experience, Getz reports that, the main resistance to using adaptive designs come from the internal groups that would be responsible for running them. Regulators are not the main barrier, especially for early phase trials, though the regulators’ enthusiasm for them could be conveyed more strongly and consistently.

Getz emphasises that where adaptive trials have been used they have already provided substantial cost savings – through early stopping and the reduction in protocol amendments – but that these savings are small compared to the gains that adaptive trials used in the early phases could offer in terms of improving late stage success rates. And while the study found that sponsor organisations recognised this, it has not yet translated into these types of designs actually being run.

I think that something that is difficult for proponents of these more sophisticated designs, is to be able to demonstrate clearly when such designs can improve the likelihood of success of a drug development program. That the small amount of additional time, and cost, of a more sophisticated design is more than outweighed by the increased probability of success and increased value of a better characterised treatment.

Demonstrating the value of this trade-off is complex. It requires, firstly, estimating the current uncertainties in the effects the drug may have at different doses or treatment regimens, and secondly how the value of the drug is likely to vary with the efficacy and tolerability of the selected treatment and target population. We need to be able to evaluate, over the range of likely drug characteristics, how well different types of learn phase trial perform in enabling us to then design the right kind of phase III trials and how these then in turn impact on the probability of technical success and value of the drug.

However, we think this can be done using simulation. We have developed a prototype software system that can simulate the various phases of development along with some of the between phase design choices.

Tom Parke

Tom Parke

Tom Parke has been working at Tessella for over ten years. For a large part of that time he has ...

© Copyright 2018 Tessella
All rights reserved