Patients dying because of FDA inflexibility

On May 26, the Food and Drug Administration will decide whether to approve eteplirsen, a therapy for Duchenne multiple dystrophy (DMD), on a fast-track basis. The drug, made by Sarepta Therapeutics, is the first for this brutal disease, which strikes about 500 boys annually in the U.S. It shows up before the age of five, putting kids in wheelchairs by their teens and ending tragically in premature death. An advisory panel in April voted against approval of the medicine, 7-6. At the hearing, parents of boys afflicted by DMD became hysterical. They pleaded for the majority to reverse course and let their children try the medication. 

They are not stupid people. They are up against some bad biology. But, almost worse, they are now confronted by some bad regulatory policy DNA. The FDA advisory board experts see this. A professor at the Medical University of South Carolina, who voted no, was even apologetic: “Based on all I heard, the drug definitely works, but the question was framed differently.” 

{mosads}That almost perfectly captures the deep public policy failure that now reinforces this horrifying disease — and countless others. The anguished parents frame the issue properly: comparing life with and without the medicine. The government does not doubt that the therapy is safe, and it has been shown to be effective in a study involving 12 participants. But the standard that the FDA uses may require more proof. This would involve additional trials to demonstrate success in specified biomarkers. While data are generated and theories resolved, people will die. 

That is perverse. The humane and smart policy is for the FDA to set aside its advisory committee recommendation and approve access to this potentially life-saving drug. Informed consent by patients and careful data collection thereafter will not only increase their chances, it could pave the way for scientific breakthroughs beyond. This approach has been endorsed for eteplirsen by 24 U.S. senators who span the political spectrum, from Elizabeth Warren (D-Mass.) to Jeff Sessions (R-Ala.). As the senators told the FDA in April, “the risk of doing nothing for a patient with DMD is their certain death.” 

Historically, the FDA mandates clinical experiments with large numbers of human sufferers. Some get a new drug, others a placebo. The differences in outcomes measure the innovation’s effectiveness. While called the “gold standard” for research, the approach is slow, expensive and imposes huge burdens on “small” diseases like DMD, where large samples are unavailable. It also ignores vast terabytes of data about the genetic differences among the patients, on the one hand, and the specific diseases, on the other. 

Indeed, the best example of fast, life-saving biotech discovery — new drug cocktails to save millions of HIV patients from death via AIDS — came when the FDA abandoned its “gold standard.” It allowed patients and doctors to try risky medicines, not yet proven effective, and observe how they worked. As Peter Huber, in his compelling 2013 book “The Cure in the Code,” describes: “Many Americans are alive today” because regulators in the 1980s and 1990s wrote “new rules — anti-rules, actually, that spelled out when big chunks of the old rules wouldn’t be fully enforced.” 

Sham remedies were a problem a century ago. Government reforms brought higher standards, guarding against medicines coming to customers without proper vetting. They drive an FDA official today to defend the scrutiny given Sarepta’s eteplirsen, announcing last month that “anecdote and emotion do not change the data with which we are confronted.”

Yet that view is itself susceptible to quack bromides. The emotion of DMD parents stems from the dire consequences of delay, and that forms a crucial baseline for properly evaluating new drugs. And case studies, systematically analyzed, can yield rich results: “the plural of anecdote is data.” Given that human beings differ genetically in complex ways, the old statistical approach — focusing on crude mean differences — fails to incorporate vital information. In contrast, doctors prescribing medications “off label” are now pinpointing subtle differences and responses, often improving on placebo studies. That is why the Congress in 2012 passed legislation, signed by President Obama, that granted the FDA “broad flexibility” to “take into account the views and experiences of patients” in drug reviews, as the senators’ letter summarized.

Fly-by-night mountebanks of the 19th century motivated a change in law. But there are new threats and opportunities today. We ought to adjust our models of medicine accordingly. When patients, their loving parents and even their U.S. senators understand what the relevant risk-reward choices are, the humane and science-friendly choice is to avoid making the perfect the enemy of the good. 

Hazlett is H.H. Macaulay Endowed Professor of Economics at Clemson University, where he also directs the Information Economy Project.