The topic of adaptive clinical trial designs is the most powerful and far-reaching issue now percolating through the community. Adaptive designs promise something for everyone.

First and foremost, as a keystone element of the agency’s unfunded Critical Path agenda, adaptive trials could burnish the lagging reputation of the FDA as a scientific innovator and guardian of public safety. Beyond that, in industry, there is ample reason to think adaptive designs could reverse a long-term decline in research productivity. Fewer patients might be needed. Trials could recruit and conclude more quickly. Shareholders at drug companies could see lower research costs. Everybody wins. Some of the big-picture themes can be found in our June, 2006 Bio-IT World feature story on the topic.

So the presentations from a recent conference on adaptive trials may be of interest. The speakers were an all-star roster of biostatisticians from government and industry.

Some of the luminaries included: Bob O’Neill (FDA), Bob Powell (FDA), Walt Offen (Lilly), Declan Doogan (Pfizer), Michael Krams (Wyeth), Robert Ruffalo (Wyeth), Marc Walton (FDA), Gordon Lan (Johnson & Johnson), Jeff Maca (Novartis), Paul Gallo (FDA), Bram Zuckerman (FDA), Jerry Schindler (Cytel) ... the list goes on. You can find all the relevant PDFs here, on a corner of the PhRMA website that has been rechristened “innovation.org.” The industry’s thinking on adaptive designs is clearly advancing rapidly. FDA comfort with the technique is high despite the absence of any formal regulatory guidance on the topic.

The one common theme to the side decks we’ve reviewed is simple. Adaptive designs will have to be charted well in advance. And followed strictly. Which is to say that once an adaptive trial starts and the key statistical triggers and amendments (a new dose, the closing of an arm of the trial) are in place, they will have to run their course. No last-minute human tinkering. Judging by the slides, everyone at the meeting felt that such an approach would be statistically and scientifically kosher.

It’s too bad such a great meeting could not have been accessible to a larger audience.

As Mark Senak’s fine Eye On FDA website chronicled throughout the fall, the FDA and the Pharmaceutical Research and Manufacturers Association (PhRMA) wanted the November 13-14, 2006 meeting to be private. Or secret. What was so hush-hush? Were the Bayesian statisticians worried about Frequentists lurking in the back row of the auditorium? The Drug Information Association could have handled the PhRMA-FDA event perfectly well, and given FDA speakers what they sometimes insist upon (a noncommercial podium).

Though the FDA-PhRMA adaptive designs conference was not advertised or publicized in the usual manner, all sorts of people knew about it. They wanted to attend. In the end, we’re assured, there was so much demand for seats that the doors of the sessions at the Marriott Bethesda North were indeed opened to the great unwashed multitudes. In the mainstream of the scientific world, scientific questions are discussed in the clear light of day. Adaptive designs, for now, appear to be outside that mainstream.

A more substantive question behind adaptive designs is technological. Does the industry have the transactional clinical data systems in place to handle adaptive designs? Will the industry have such systems in five years? Ten?

The vast majority of the data in the pharmaceutical industry is collected on paper. Even with the most advanced systems, the data on that paper are laboriously transcribed into computers. This transcription, by conventional estimates, can take weeks or months. We are not talking about transactional systems that are found in every other industry: retail, transportation, banking.

We’re talking about 1950s-era processes that would be an embarassment in any other industry. In such an environment, the pharmaceutical industry considering adaptive designs is something like Wyoming considering a navy. It might indeed be a good idea for Wyoming to have a navy. Who can say? Perhaps its coal and grain will need special protection. But there is considerable work to be done in Wyoming first.