The FDA Proposes a New Guidance About Distributing Scientific and Medical Publications on Unapproved New Uses: It Doesn't Address the Problem of Publication Bias
On Friday, February 28th, the FDA made available a new proposed Guidance on the subject of pharmaceutical companies' distribution of published materials that report on studies of their drugs for "unapproved" uses.1 The marketing of pharmaceuticals is a huge enterprise. According to a study by the Pew Charitable Trusts Health Initiative, pharmaceutical companies spent $29 billion in 2012 to promote the use of their drugs, with $15 billion of that spent on direct marketing to doctors by pharmaceutical reps, a practice known as "detailing." For some drugs, a major portion of the promotion effort is aimed at generating "off-label" use, which is the prescription of a drug for diseases or conditions for which the FDA has not received sufficient evidence to determine the drug's safety and efficacy. Until the mid-nineties, the FDA took a hard and fast position limiting the ability of pharmaceutical companies to promote off-label use of their drugs. The FDA's position was straightforward: placing a drug on the market for a use which has not been approved by the FDA is selling the drug without adequate information about its safety and efficacy for that use, which renders the drug "misbranded" in FDA parlance and the marketing illegal. If a drug maker wanted to promote a new use of an already approved drug, then the FDA's view was that the only proper way to do so was to do studies that were reviewed by the FDA and sufficient to support adding that use to the drug's label. Then, in a series of decisions, the District of Columbia Court of Appeals ruled that drug makers' dissemination of information about unapproved uses was commercial speech and merited protection under the First Amendment. That view was partly adopted by statue with the Food and Drug Administration Modernization Act of 1997. Since then, the FDA's efforts have been directed at defining the appropriate limits on off-label promotion, consistent with the First Amendment.
Doctors are free to prescribe any drug for any condition, with only their professional judgment (and sometimes the patient's insurance coverage) as a constraint. If a doctor wants to prescribe a drug that has only been approved for the treatment of major depression for a patient that has obsessive compulsive disorder, then the doctor is free to do so. Of course, the doctor is more likely to do so if a pharmaceutical sales representative has provided the doctor with a copy of a published study that concluded the drug was in fact useful for the treatment of OCD. Journal articles of this general type represent by far the biggest source of concern, and the FDA's attempt to ensure that pharmaceutical companies' dissemination of such journal articles does not mislead doctors and patients is front and center in the proposed Guidance. Why would the distribution of reprints of published articles be misleading, or worse, potentially dangerous? For several reasons. For one, such studies are often too small to allow any firm conclusion about the relative risks and benefits of the drug in a new use, or reach conclusions based on measurements of an endpoint that is only suggestive of real benefit. This can be the case even with FDA approved indications, as the recent example of Avastin's use in breast cancer illustrates. The FDA approved the use of Avastin in metastatic, HER-2 negative, breast cancer based on studies that demonstrated that Avastin's use resulted in a significant increase in progression-free survival (the length of time after treatment during which the disease does not get worse). However, further studies demonstrated that despite the progression-free survival, the patients over-all survival, which is the time from treatment until death, did not change. Avastin, a costly drug which has its own adverse effects, did not help breast cancer patients live longer, and that use was removed from the label. Because the FDA approved the use on the basis of the more preliminary finding, the maker of Avastin was required to do the further study that showed no ultimate benefit. Of course for studies of uses that are and remain off-label, there is no such requirement, and off-label uses that are stimulated by preliminary findings would likely continue in the absence of the follow-up definitive studies.
An even bigger pitfall for the dissemination of materials concerning off-label use is publication bias. The FDA's proposed Guidance deals with the problem of handing out reprints of publications by focusing on fundamental elements of reasonable science and fairness, such as the character of the journal (peer-reviewed by independent experts), the balance of the material distributed (not inconsistent with the weight of the evidence from all published studies on the issue, the absence of conflict of interest, and so on. All this certainly is necessary, but it is hardly sufficient. The problem goes beyond ensuring reasonable peer-review and disclosure of financial conflicts, to the heart of the very publication process, which is rife with publication bias, and the often-flawed data analysis and reporting in studies that are published. Vedula et al. published a stunning analysis, in the New England Journal of Medicine, of the publication bias and outcome reporting bias in the case of off-label studies for the drug gabapentin.2 My colleague, Joanna Sax, examined the problem of publication bias and deception in drug promotion at length in her article Protecting Scientific Integrity: The Commercial Speech Doctrine Applied to Industry Publications3 and, in addition to documenting and detailing the scope of the problem, suggested that the problem be considered one of deceptive marketing and proposed a statutory fix. In the case of the FDA's proposed Guidance, a statute is not needed. A reprint that reports positive results in a study of the off-label use of a drug is misleading not merely if other published studies report negative results, but is just as misleading if unpublished studies reached negative conclusions. The FDA needs to require pharmaceutical companies to disclose and distribute the results of any negative unpublished studies, in reasonable summary form, if they wish to disclose and distribute positive published studies of the same drug for the same general indication.
1 FDA, Revised Draft Guidance for Industry on Distributing Scientific and Medical Publications on Unapproved New Uses--Recommended Practices; Availability, [Docket No. FDA-2008-D-0053] (February 28, 2014).
2. S. Swaroop Vedulla, Outcome Reporting in Industry-Sponsored Trials of Gabapentin for Off-Label Use, N. Engl. J. Med. 2009; v. 361:1963-71 (Nov. 12, 2009).
3Joanna Sax, J.D., Ph.D., Protecting Scientific Integrity: The Commercial Speech Doctrine Applied to Industry Publications, 37 American Journal of Law & Medicine 203 (2011)