In response to:
FDA: This Agency Can Be Dangerous from the September 30, 2010 issue
FDA: This Agency Can Be Dangerous from the September 30, 2010 issue
To the Editors:
I agree with Marcia Angell’s judgment that the Food and Drug Administration needs further independence from drug companies [“FDA: This Agency Can Be Dangerous,” NYR, September 30]. Yet her entitlement to this valuable perspective cannot serve as privilege to synthesize new facts and cartoonish institutional portraits. Her essay profoundly mischaracterizes the FDA’s history and behavior, and hence undermines our common goal of its science- and safety-based reform.
Angell scarcely touches upon my book Reputation and Power; it is rare to read a book review in which not a single word of the reviewed volume is quoted. Her claim that I ignore the FDA’s failures is easily refuted. In a book about reputation, I show in the introduction and twelfth chapter of the book that the FDA has lost public trust. And far from demonstrating “little apparent concern” for the effect of user fees, the book’s reputation-based account explains that the Prescription Drug User Fee Act embedded industry as an audience as if inside the agency’s head (Chapters 6 and 12). The book and related research also illustrate how the PDUFA’s deadlines distorted agency decision-making. Nor do I ignore “ads aimed directly at consumers,” as I examine these squarely in Chapter 12 (p. 733).
More worrisome, Angell’s essay drips with factual errors and profound historical mischaracterizations. I provide a longer list at my website.1 Several are glaring:
• Angell offers a stunningly inept characterization of FDA evidence requirements. Drug companies lobbied hard against the placebo trial requirement (they often still do), and the FDA has for decades hardened its approval criteria when the drug is a me-too therapy (Chapter 7). The agency increasingly requires an active control trial, and it recently negated the role of non- inferiority trials in whole drug classes.
• A 2007 law already requires what Angell proposes: registration of clinical trials. Contrary to her claim, the same law also directs $375 million toward postmarket surveillance.
• The agency has diligently policed conflict of interest on advisory committees in the last two years; conflict-of-interest waivers have sunk to historic lows, as just 2 percent of them were granted last year.
• New FDA leadership, far from being passive, has dismissed agency medical device regulators and reopened old device approval cases; meanwhile consumer advertisement warning letters to drug companies have more than doubled since 2008.
These errors and mischaracterizations, among many others, demonstrate the risk of converting a book review into a personal soapbox to rehearse talking points the reviewer has been making elsewhere for a decade. In that time, science, the FDA, and scholarship itself have been passing by Angell’s hollow and specious portraiture.
Allie S. Freed Professor of Government
Director, Center for American Political Studies
In his waspish letter (how helpful are words like “cartoonish,” “hollow,” and “specious”?), Daniel Carpenter objects to my criticisms of his book and of the FDA itself.
My major criticism of the book was that although it touches on nearly everything about the FDA’s Center for Drug Evaluation and Research (CDER), it hardly considers whether the agency is actually doing its essential job of ensuring that prescription drugs are safe and effective.
For example, Carpenter claims in his letter that he examined “squarely” direct-to-consumer ads on page 733 of the book, but that examination consists of just three sentences describing the loosening of regulations in 1997. He says nothing about whether ads are adequately regulated, or whether they contribute to the overprescribing of drugs, despite the fact that these are contentious issues and such ads are banned in nearly every other advanced country. Similarly, while he describes a “loss of trust” in the FDA, he does not examine whether this is justified. And while he notes the effects of the Prescription Drug User Fee Act on the timing of drug approval, he does not discuss its many other effects on how CDER does its job, nor does he consider whether it should be repealed.
In defending the FDA, Carpenter contends that my review “drips with factual errors and profound historical mischaracterizations” (“drips”?). He gives four examples, which I will address in order:
• Drug companies may have opposed the requirement for placebo controls some thirty years ago, as Carpenter claims, but this is because they objected to any regulation of trials. In fact, they preferred no concurrent controls at all, or active controls chosen by them, without oversight. In recent years, however, the industry has clung to placebo controls as the lesser of evils, and lobbied against proposals that me-too drugs be compared with existing ones at equivalent doses. The FDA has considerable discretion, but generally requires active controls only when it is grossly unethical to use placebos (as when there is an existing treatment for a life-threatening disease).
• I proposed that “all clinical trials, without exception, should be registered at inception in a public database and the results shown when the research is completed.” Contrary to Carpenter, the 2007 law, while a step in the right direction, does not require that. For example, it excludes phase 1 trials and those that do not lead to approval, and there are limitations on access to results. (See Alastair J.J. Wood, “Progress and Deficiencies in the Registration of Clinical Trials,” The New England Journal of Medicine, February 19, 2009.)
• Advisory panel members may receive income from drug companies, apparently without limit, although they are banned from attending any meeting in which they have income from companies affected by the outcome of that meeting. However, panel members may receive waivers to participate anyway, according to the August 2008 guidelines, if “it is necessary to afford the committee essential expertise,” as long as waivers are not given to more than 13 percent of advisers (this year’s legal limit). If an adviser’s income from relevant companies is above $50,000 per year, then the adviser would “generally” not receive a waiver. It is true that the new FDA commissioner is publicly disclosing conflicts of interest and holding waivers below the limits. Still, whether the rules are diligently policed is not the main issue, but rather whether they are too lax, which I believe they are.
• It is absurd to imply that direct-to-consumer advertising is being adequately regulated because warning letters to drug companies have more than doubled. The FDA has only fifty-one employees to review tens of thousands of promotional campaigns, an impossible job no matter how committed they are. Under the George W. Bush administration, there was a deliberate slowdown in the already small number of warning letters sent to drug companies because of false or misleading ads. That the number has increased under the new administration is hardly surprising. But because there are no meaningful penalties, the letters have little effect, as any knowledgeable viewer of TV drug ads can attest.
Carpenter and I share a belief in the work of the FDA, but he seems unable or unwilling to admit that the agency is not doing its job because of industry influence, and sees marginal improvements as evidence that all is well. The FDA needs constructive criticism, not an apologist.
Update: In my review, I said that the FDA had not yet decided whether to pull from the market the diabetes drug Avandia because of heart risks. On September 23, in a joint announcement, the European Medicines Agency and the FDA announced that sales of Avandia would be banned in Europe and sharply restricted in the US. However, the FDA decided that US patients currently taking Avandia may continue to do so without restrictions. Thus, some 600,000 current users could go on being exposed to cardiovascular risks deemed too great for everyone else. This split-the-baby decision reflected the internal struggle at the FDA between the drug safety office (OSE), which favored pulling the drug from the market altogether, and the drug approval office (OND), which favored leaving it on the market without restrictions.