For the all the excitement and anticipation around adaptive designs for clinical trials, there’s always been one pesky fly in the ointment. Do regulators actually bless adaptive approaches when sponsors propose them? Has hype inflated expectations for such a promising technique?

Adaptive trials—which include a statistical regimen to use the results of an ongoing trial to redirect it in midstream—have daunting operational issues in an industry burdened by paper in every process, every aspect of a trial.

But there are also non-trivial regulatory worries. Those were the topic of a webinar last month. Sponsored by electronic data capture company Medidata, the online presentations attempted to summarize (or perhaps telepathically channel) European and U.S. regulatory thinking. It’s clear that while publicly touting the merits of new trial techniques such as adaptive designs, regulators have in an unknown number of instances been forced to say “no” when assessing specific protocols.

Bursting The Bubble

So what do regulators think? “They are very cautious about adaptive designs being deployed too quickly,” said Tom Parke, head of clinical trial solutions at Tessella.  That firm is one of a small, busy circle of companies focusing on consulting, services and software for adaptive designs. (Others might include Cytel and the father-son duo lead by Don Berry,  who is most publicly affiliated with the M.D. Anderson Cancer Center. )

Tessella has helped some very high-profile customers in big pharma on both sides of the Atlantic. Parke sketched (without naming any firms) a few proposals for adaptive designs that had not been well received by regulators (who also remained unidentified.) There is a host of reasons that adaptive designs fail to impress government reviewers, Parke said. “At the top of the list was that people didn’t justify why they wanted to do the adaptive design,” he said.

Old Designs

Translation: sponsors can’t simply say, We hope to save money. You need a scientific reason. Added Parke: “Other trials have been turned down because the total evidence that would result would be inadequate or the methodology that had been proposed had not been one of the modern, sophisticated designs.”

Overall, Parke says, there is a more-traveled path to adaptive designs in the medical device arena, largely because a new device can quite plausibly use some of the knowledge acquired in a trial of a similar, previous device. “The use of prior information is very natural in that context,” he said.

Parke added that the use of Bayesian (as opposed to frequentist) approaches is well-accepted in the academic community, but only just coming into wider use in the pharmaceutical industry. “Its penetration in clinical trials lags behind other industries,” he noted. “[The Bayesian method] has a solid statistical underpinning that frequentist statistics doesn’t have.”

Different Perspectives

Parke is a proponent and an advocate for adaptive designs, but he didn’t sugarcoat the different ways that such trials may be perceived in regulatory circles. The faster, cheaper trials prized by adaptive design proponents may be viewed in a different light in government, Parke hinted. “The regulators said, and I think I remember this word for word, ‘you call it removing the white space, but we call it removing the thinking time.’”

The other main speaker on the webinar was Michael Borkowski of United Biosource, the contract research organization. He’s right up there with Merck’s Jerry Schindler in having an uncanny knack for talking about adaptive approaches rigorously without getting lost in equations.

United Biosource has been beefing up its internal biostatistical resources for adaptive projects. And Borkowski’s bottom line is that the regulators are receptive to them. Officials just want such trials done right. As such, he says the reception for properly conceived adaptive trials will be warm.

“It’s a fairly positive sentiment change. In the recent meetings,” Borkowski said, “it’s not whether or not it will be acceptable. It’s a matter of just defining what is acceptable. There are not issues with adaptive designs being done.”

Warning Shot

“No one is saying we can’t do adaptive trials,” Borkowski said. “The regulators are ready to be worked with. As companies do these interim analyses for a variety of reasons, for safety or anything else, there are very simple adaptive methodologies that can be applied that are readily acceptable.”

He continued: “As long as the trial has a good purpose and the issues are being addressed and documented, you should receive positive regulatory feedback.”

But he also seemed to be keen to not re-inflate the balloon of excitement around adaptive trials. Some of his presentation was as cautionary as Parke’s. As an example, he cited group sequential designs. “Those types of trials are under increased scrutiny,” Borkowski said.