From the Editor: No Substitutes

Don't use surrogate measures in clinical trials, says a major industry critic. OK, but don't the critics base a lot of their points on surrogate measures, too?
Oct 01, 2005

In an opinion piece in the September 8 New England Journal of Medicine, Harvard medical professor and long-time industry critic Jerry Avorn takes a whack at FDA, accusing the agency of practicing a level of science that wouldn’t pass muster anywhere else in research—science that’s only "good enough for government work."

Patrick Clinton
He qualifies that a bit: FDA isn't sloppy about execution, but it pursues the wrong goals. "Like a patient with obsessive–compulsive disorder," he writes, "the agency is single-mindedly preoccupied with demanding the meticulous performance of a series of relatively simple acts—proving that a new medication is superior to a usually irrelevant comparison treatment (such as placebo) in achieving a potentially irrelevant outcome (such as a surrogate measure)."

What the argument comes down to is pretty simple: Avorn thinks that it's not good enough for a drug to prove itself more effective than a placebo—that is, simply better than nothing. It should have to be tested against other therapies, and the agency should assess relative benefit against relative risk across the board.

It may sound like Avorn just wants to raise regulatory standards a bit. But going beyond the traditional standard of safety and efficacy is actually much more. He seems to want FDA to make comparisons that traditionally have been left to physicians, patients, managed care organizations, and others. It's not clear how such a policy could be carried out. (For instance, if you can't gain marketing approval unless your drug presents a better profile than others, do you lose it when others have passed you by?) More important, it's not clear that patients will benefit if FDA starts limiting physician choice to preferred alternatives. As Peter Pitts points out on the excellent DrugWonks weblog (, "Patients (otherwise known as 'people') respond differently to different drugs—even drugs within the same therapeutic category. Asking FDA to recalibrate 'safe and effective' to 'as safe mostly but not as effective in some circumstances but somewhat more effective in others' is a poor public health course.... There are a lot of places that FDA belongs—but subjective intermediary between patient and doctor is not one of them."

Avorn's essay isn't stupid, but it's extremely frustrating, partly because it doesn't follow its own suggestions. He wants longer trials, direct comparison of products and treatments in the real world, and he wants trials to stop using surrogate endpoints. But his article doesn't take a long view. It doesn't compare real-world alternatives, instead putting his proposal up against a straw-man parody of how things work today—a sort of policy placebo, if you will. Most important, like so many drug industry critics, Avorn is caught up in numbers that look a lot like one of his bugaboos: surrogate endpoints.

You see this all the time: Critics argue that FDA is doing a bad job, or healthcare is going off-course, and their proof is that prices are rising, or that prescription numbers are up, or that pharma spends X number of dollars on TV advertising. It's obvious why critics, including Avorn, use these statistics. They probably should look at them. But they're surrogates, with all the problems that surrogates present.

In a way, Avorn's got the right advice. He's just aiming it in the wrong direction. Here's a clinical trial I'd like to see: In one arm, we make FDA get tough, reduce drug approvals, and clamp down on marketing. In the other, we make the debate on drugs actually follow Avorn's rules of evidence, and go wherever the data lead. I've got a hunch which one's going to reach the endpoint we all want.

Patrick Clinton is Pharmaceutical Executive's editor-in-chief and can be reached at

lorem ipsum