Trial Design

Accelerating Progress

For various reasons, touched on in more detail in Bayesian analyses. Dr. Dave Alberts the Arizona Cancer Center has also done some excellent work in this area, Dr. Bruce Chabner contributes some thought-provoking work, and Dr. Mark Ratain at the University of Chicago has published several good articles on the subject as well.  We believe those pieces form a reasonable starting point for any discussion about improving oncology trials.

For example, Dr. Gottlieb’s speech before the 2006 Conference on Adaptive Trial Design, given while he was Deputy Commissioner for Medical and Scientific Affairs at FDA, provides important historical context and direction for future emphasis.  That speech can be found here.

Dr. Berry authored a review of progress in use of Bayesian statistical methods within clinical trials and the drug development process in the January 2006 issue of Nature Reviews: Drug Discovery.  His article can be found here

A very good PowerPoint presentation (in PDF format) by Dr. Dave Alberts from the Arizona Cancer Center can be found here. Even though it’s in presentation form as opposed to scholarly article, Dr.Alberts’ points and support are clear and effective.  We hope to bring more of Dr. Alberts’ work to Accelerate Progress in the future as we look at his work as truly pushing clinical trial designs in the right directions for patients and for systemic improvement.

A high-level piece on the subject ran as the cover story in Bio-IT World’s June 2008 issue, found here. There is a brief discussion in this piece about the differences between EMEA thinking on Adaptive Clinical Trials (ACTs) and FDA approaches as well as a note about the lack of expertise at FDA which results in a lack of guidance for industry on the issue.  The net result, then, is little use of ACTs in U.S. designed trials, to the detriment of patients, the development process, and actual learning about a new therapeutic. 

Drs. Freidlin and Simon at the National Cancer Institute (NCI) published a piece in 2005 entitled “Adaptive Signature Design: An Adaptive Clinical Trial Design for Generating and Prospectively Testing A Gene Expression Signature for Sensitive Patients” which can be found via the following link: 

The piece provides background and simulation data on using an adaptive design created to be able to incorporate gene expression (or other biomarker type data) signatures that aren’t discovered until after the trial has begun.  Given the frequency with which such learning happens within a trial, this could be a good way for companies to design trials where the subset of patients most likely to benefit is not prospectively identified but is detectable once the trial has reached an initial enrollment level.

The prior piece purports to build on the concept of randomized discontinuation trials (RDT), well-discussed in this article by Rosner, Stadler, and Ratain.

The RDT design was implemented and discussed by Ratain, et al, in the following paper:

A Powerpoint primer on the benefits of Bayesian analysis and adaptive designs (including, but not limited to, RDTs) by Jane Perlmutter is here.

Uptake and use of such advances in trial design remains slow in real terms, but appears to be accelerating somewhat and, as Critical Path initiatives receive increased attention and funding, Accelerate Progress will continue to push for newer, better trial designs to accelerate our progress against these difficult and terrible diseases.  Acting as an “Acceleration Engine”, we will help other organizations and researchers funding disease research create better trial designs, incorporate more use of adaptive elements, increase use of Bayesian analysis within statistical protocols, and further help accelerate the learning loop embedded within the research and trial process.  This will allow researchers to learn more, more quickly, accelerating safe and effective new therapies to patients and more quickly discarding less effective or more toxic options, saving lives, dollars, and time, further improving the efficiency of the overall system. 

To paraphrase an old ad line, “We don’t do most of the research you fund.  We make most of the research you fund better.” Better meaning more effective in terms of providing actionable knowledge faster, accelerating returns on investment, accelerating learning, speeding progress, and helping more patients more quickly and efficiently.