Clinical Trial Enrichment: Allowable Cheat Sheet or Stolen Playbook?

Today, Bob Temple, M.D. the Deputy Director for Clinical Science in the US FDA's Center for Drug Evaluation and Research published a blog entry at the FDAVoice entitled "Strategies for More Successful Drug Trials."

Drug companies meet with FDA officials for guidance on things like clinical trial design all the time—guidance that the FDA intends to use as a means of communicating to the drug company what strategies they should adopt if they want to see their drug get approved. There's nothing underhanded about this whatsoever, and the shocking thing is that sometimes the company ignores the FDA's recommendations. That's like asking a history teacher what you should study for an upcoming test, and then using your study time to work on paper due for an English class.

Case in point, Cell Therapeutics and their embarrassing and shameful attempt to gain market approval for pixantrone. When the FDA says you need to recruit 200 patients and you instead recruit 100, things aren't going to go well for you.

Enrichment Strategies

Dr. Temple mentions a draft guidance now published by the FDA that,

"spells out how drug developers can use such strategies, known as clinical trial enrichment, to greatly increase the likelihood that data collected during a clinical trial will demonstrate that an effective drug is effective. These are potentially powerful strategies for the pharmaceutical industry because appropriate use of enrichment could result in smaller studies, shortened drug development times, and lower development costs."

In effect, this is something of a legal, pharmaceutical cheat sheet for drug submissions. Here we have the agency itself clarifying its own methodology and attempting to communicate this to drug companies. Some of the strategies he discusses include:

"Noise" Reduction

Reducing 'noise' among a patient population amounts to not enrolling patients who it is somehow clear will either not take the drug under investigation as directed or who won't complete the study.

Prognostic Enrichment

In prognostic enrichment, drug sponsors choose patients for their study who will "have the disease manifestations the drug is intended to prevent." Obviously in doing so, it will be easier to see whether or not a drug is having the intended effect if used in these patients.

Predictive Enrichment

Predictive enrichment, which Dr. Temple calls "particularly exciting" in a way that only a bureaucrat could, involves identifying patients who might respond to treatment by making use of an aspect of their "physiology, genetics or past responses."

Codified Selection Bias

What Temple calls 'Enrichment Strategies' the actual FDA draft guidance document calls 'Selection Strategies'. And it is indeed hard not to read the draft guidance published by the FDA as well as Temple's description of it and not be wondering "isn't this just fancy-talk for selection bias"?

The draft guidance document asserts that

"Any use of an enrichment design should be explicit in the protocol and study report and should fully detail the rationale, specific enrichment maneuvers, and their impact on interpretation of results. The importance of such descriptions is obvious for trials in which high-risk patients (prognostic enrichment) and probable responders (predictive enrichment) have been selected, where the description is critical to knowing to which patients the results apply … we strongly recommend early discussions with the Agency on plans to use them."

Considering the untold number of underhanded games drug sponsors are known to play with clinical trials in order to get their drug approved, it strikes me as especially careless to leave such self-reporting in their hands.

Temple claims that "enrichment won’t save a drug that doesn’t work [but] it will help find one that will."

Unfortunately, the draft guidance document—which is currently a work in progress, not policy—can too easily be read in many ways depending on your goal. And if your goal is to get a drug pushed to market as soon as possible by spending the least amount of time or money, then the document reads not like an allowable cheat sheet, but like a stolen playbook.