In the world of running, the ultramarathon is anything longer than a 26-mile course. Some people run from one rim of the Grand Canyon, down to the bottom, and up to the other rim. And back. In a day. They call that an “R2R2R.”

Adaptive designs are the ultramarathons of clinical trials. Adaptive designs will test even elite athletes.

We’ve just returned from a superb ExL Pharma meeting on clinical trial design. The conference was focused on adaptive approaches. It’s clear that adaptive trials aren’t coming. They aren’t on the horizon. They’re here. The statisticians are on board. The FDA is on board. 

Wyeth is talking about considering adaptive designs for most trials. Genentech described an adaptive design that combined Phase I, II and III. Merck and Novartis aren’t far behind. For those who can handle adaptive trials, the payoff could be huge: besting a competitor to market, or reducing a trial budget by thirty percent.

At other companies, it’s not clear whether the operational agility of a sponsor is at the necessary level to execute an adaptive design. Execution will be everything. This is where even a few seasoned marathon runners may crumple.

To help companies determine if they’re working for a marathon company or an ultramarathon company, we have assembled 22 questions to provoke soul-searching and self-assessment of existing infrastructure and processes. 

Our questions were inspired by the recent ExL Pharma meeting and the presentations we heard there. We’ll be writing up a few stories about the meeting over the coming weeks.

1 Can you access selected types of clinical trial data? Electronically? Via an interactive voice response system (IVRS)? The web? Before the trial ends?

2 Is your IVRS available over the web as well? Can your IVRS and your electronic data capture (EDC) system share data seamlessly? Have your IVRS and EDC vendors ever worked together before? Is the integration of their systems a custom project? Or is it a bona fide product? Has that product been validated?

3 Is the data in your trials stored in XML format? Or is it in a proprietary format? If your data is not in a CDISC or HL7 format, what is your plan to have multiple systems use the same data?

4 How many adaptive trials has your contract research organization finished? Is their biostatistical expertise on staff?

5 If your drug works, will the indication respond quickly to your intervention? Acute pain might be a great indication for an adaptive design. Early-onset Alzheimer’s disease might not be.

6 Are your statisticians comfortable simulating clinical trials? Are your computational resources adequate to your needs? Can the same algorithm used in the simulation drive the randomization of the actual trial?

7  Can your simulations help plan drug supply requirements? Do your estimates for drug supply involve guesswork? Or have you retained a company specializing in the area?

8  Can your systems for drug supply simulation and management be connected to the IVRS and/or randomization engine? As a trial unfolds, and the randomization changes, can your supply depots receive drug and placebo according to a newly dropped or added arm of the trial? If an arm of the trial is dropped or added, what are the effects on your drug packaging and labeling?

9  How granular is the security around your clinical trial data? Can uncleaned data about blood pressure, but nothing else, be selectively and easily exported? Will that break the blind? Can that data be used by other systems?

10 How easily can interim data be exported from your clinical data systems? Is it a batch load procedure? Is a programmer or database administrator involved? Can any ordinary operational person do it?

11 How frequently will the data monitoring committee receive information about the study progress? What information will they see? What information won’t they see?

12 Who will be aware of a change to a randomization scheme? Could blinded investigators learn that an arm of a trial has failed, and telegraph that to patients? Is your company large enough that Wall Street or medical websites could publish information about an arm of the study being dropped? If that news made CNN, would it affect patient recruitment or introduce bias into the mix?

13 Can your data elements and case report forms be standardized company-wide to expedite the design of databases and work flows?

14 Can the table structure in your EDC (electronic data capture) system handle your randomization lists natively? If not, changing randomization lists manually, on a frequent basis, could become laborious or error-prone.

15 How many places do your clinical data live? Two? Five? Having all the data in one place will facilitate faster and more informed decisions than if different sets of facts reside in different systems.

16 How frequently will your randomization scheme be updated? Weekly? Monthly? Continuously? Will there be a delay between a change to the randomization scheme and those changes flowing through to your labeling, packaging and shipping systems? Are your projections for quantities of placebo and comparator as thorough as your estimates for supplies of drug?

17 Could a change in the general clinical standard of care during your trial affect your randomization? Should your trial anticipate a change in the standard of care?

18 How carefully have you selected your dosage? What statistical approach for dose-finding will be best? Could you need external consultants to answer FDA objections to your statistical methodology? Have you ever heard of Don Berry? If he’s busy, who will you use?

19 If you’re combining trials from different phases of research, such as IIb and III, are all of your sites equally adept at finding patients for both types of trials?

20 With what level of confidence can you guarantee that there will be drug for every patient who enrolls? How much overage in your drug supply can you afford?

21 How fluidly do your internal teams work together? Do they typically interact once or twice during the course of a trial?

22 Does senior management understand that adaptive trials may cost more money initially? That there will be pain and suffering until the kinks are ironed out? Is senior management truly behind adaptive approaches? Or are they just throwing around buzzwords without allocating the necessary resources?

Here’s an earlier ClinPage story on adaptive designs and a speech mentioning the topic by ex-FDA official Scott Gottlieb. If you have a better question than the ones listed here, don’t be shy. .

d9A2t49mkex