Although the U.S. Food and Drug Administration (FDA) does not regulate the practice of medicine, it is nevertheless responsible for protecting the public from unsafe or ineffective medical interventions. The primary way the FDA does this is by requiring medical drug and device manufacturers to demonstrate safety and efficacy in well-controlled clinical trials. This is to ensure that an “FDA-approved” medical intervention is supported by a robust body of scientific evidence.
Current FDA rules also generally prevent manufacturers from proactively promoting their drugs to physicians and patients for any use not specifically approved by the FDA (so-called “off-label” uses). In a recent draft guidance for industry1 and a detailed memorandum,2 the FDA detailed numerous arguments supporting the current restrictions on off-label product communications. For example, this restriction protects patients from harmful or ineffective uses of medical interventions which the FDA has not authorized based on a review of the totality of data. It also encourages manufacturers to produce robust scientific evidence necessary to support good clinical decision-making.
However, the FDA's rules relating to off-label drug promotion are being challenged. In 2014, a split decision in the federal Second Circuit Court of Appeals interpreted the First Amendment protection of commercial speech to restrict the FDA's ability to regulate off-label promotion so long as the promotional claim is “true and non-misleading.”Reference Robertson and Kesselheim3 The pharmaceutical industry favors this alternative benchmark, as does FDA commissioner Scott Gottlieb, who has long argued that restrictions on off-label promotion are over-paternal-istic toward physicians and do more harm than good.Reference Gottlieb4
But as we recently described, “true and non-misleading” is not a clear and objective standard.Reference Hey and Kesselheim5 The courts, for example, have often permitted very narrow interpretations of “truth.” Something can be “true” because it is vague and non-informative, rather than because it conveys any useful information. For instance, in May 2015, Amarin Pharmaceuticals filed suit against the FDA, seeking protection from potential criminal misbranding charges for off-label promotional statements concerning its prescription omega-3 fatty acid pill, Vascepa. Vascepa was already FDA-approved to lower “very high” triglyceride levels, which could protect against development of pancreatitis. The company wished to promote Vascepa to treat “moderate” triglyceride levels as preventative of cardiovascular disease, a much larger pool of patients. The FDA did not agree that the existing scientific evidence supported this second indication, and thus contested the company's ability to promote its product in this way, provoking the lawsuit. The U.S. district court reviewing the case cited the Second Circuit's decision (which was controlling precedent for the court) and initially ruled in Amarin's favor.6 The FDA then settled the case, permitting Amarin to make its claims so long as they were worded in a certain way — indicating that “supportive but not conclusive evidence” backed the use of Vascepa to lower cardiovascular risk in patients with moderate triglyceride levels — and accompanied by a disclaimer that the FDA had not approved the use.
This case reveals an ambiguity in whether “truth” should be understood in a scientific sense or a logical sense of the term. The contested promotional claim that “supportive but not conclusive evidence showed that [Vascepa] may reduce the risk of cardiovascular disease” was true, but only vacuously so.7 That is, the statement is logically true, but only by virtue of its linguistic construction, not because it conveys information about the state of scientific evidence. Logical truth is necessary, but not sufficient as a criterion for what should be acceptable promotional speech.8 One can make a highly dubious claim, such as “drug X is a cure for cancer,” into a vacuously true claim by simply attaching a qualifier like, “It is possible that” or “Supportive preclinical evidence suggests that.” We have previously proposed that the scientific sense of truth, perhaps best characterized as “informativeness,” is the more appropriate standard. For example, the other promotional statements that Amarin wished to make about Vascepa described specific results in their clinical trials. These statements — which the FDA was prepared to allow Amarin to make under its current rules — were true in the informative sense of the word because they communicated a summary of scientific data.
This case reveals an ambiguity in whether “truth” should be understood in a scientific sense or a logical sense of the term. The contested promotional claim that “supportive but not conclusive evidence showed that [Vascepa] may reduce the risk of cardiovascular disease” was true, but only vacuously so. That is, the statement is logically true, but only by virtue of its linguistic construction, not because it conveys information about the state of scientific evidence.
But informativeness is still not a strict enough criterion on its own to guard against harmful uses of medications. The Amarin case also shows how a statement can be true, informative, and yet still be misleading. Another of Amarin's promotional statements described the results of its clinical trial, which showed the Vascepa could modify patient triglyceride levels. This statement is true and informative. Yet, it is also lacking context and conveys an incomplete picture of the total evidence. The total body of scientific evidence does not in fact support the intended implication of this promotional statement — that triglyceride modification is a reliable predictor of cardiovascular benefit. Thus, this statement could only serve the company's interest by misleading physicians into thinking that there is evidence of the drug's clinical utility when there is not. But again, the court took a narrow view of the “misleading” standard.9 Since reporting a trial's result is neither false nor fraudulent, it was allowed to stand.
Acknowledging these difficulties for the “true and non-misleading” standard, the FDA has announced that any marketing communications referencing clinical trials must “disclose in a clear and prominent manner” any “limitations related to the study design, methodology, and results,” as well as accurately conveying information about the scientific context.10 But disclosing limitations is only one factor that can help avoid a problematic promotional statement about a prescription drug. We have identified 5 other questions relevant to whether such a communication is misleading:
What is the relevant body of scientific evidence? Promoting a drug as effective because one trial has shown a benefit, when there are four other trials, unmentioned, that showed no benefit, can be true, but is also misleading. Thus, the first step to adjudicating misleadingness is to determine what kinds of studies or evidence are relevant to the intervention's safety and efficacy profile as asserted by the promotional claim. A large body of data is typically better for supporting strong safety and efficacy claims than a small body of data. Therefore, truthful and non-misleading promotional statements should include a broad description of the relevant disease area, the presumed mechanism of action, and quantitative information about the total number of studies conducted and the number of research participants that have been exposed to the product. In more operational terms, one can think of this step as providing a physician or patient with enough information so that they could conduct a trial registry search to see for themselves the kinds of data that may already exist about the product.
How much uncertainty is there in the existing evidence? A large body of data does not necessarily imply certainty. A drug's mean effect size may look promising, but if there are wide confidence intervals around the mean, then it is important to convey this uncertainty to the physician. A truthful and non-misleading promotional statement should thus include confidence intervals (or equivalent representations of the degree of uncertainty) around the claimed effect size, along with any serious adverse event rates (and their confidence intervals) for the product.
How concordant is the evidence? Some uncertainties cannot be quantified precisely. For example, existing studies may be too heterogeneous in their design or study population to reliably pool their results into a single effect estimate. If this is the case, promotional materials should still acknowledge discordant evidence — explicitly including results that either disagree with the company's statement or, to protect against publication bias, acknowledgment of registered studies for which the results are not known.
Is the claim consistent with current well-confirmed scientific theory? Biomedical theories are important drivers of research and development activities. Although theories often turn out to be incomplete or false after rigorous testing, it is nevertheless informative to locate a scientific claim in the context of any underlying theory. In particular, identifying the relevant scientific theory can help a physician to demarcate the relevant comparison drugs or drug classes (which directly informs point 5 below).
Are there supportive or contradictory analogies from related fields of research? Reasoning by analogy is a powerful tool in science. Research strategies and inferences that have proven successful in other domains may therefore provide valid inferential support for the company's promotional claim. However, valid reasoning by analogy must also acknowledge failed or unsuccessful outcomes relevant to the promotional claim. Indeed, given the high failure rate of drug development, it is essential that a manufacturer acknowledge and discuss analogous negative instances (particularly those falling under the same biological or patho-physiological theory) and provide an argument for why their claim/product is different.
It is no coincidence that this list of topics is similar to what FDA reviewers address when they consider an application for a new drug approval or a request for a supplemental indication for an already-approved drug. Evaluating how to use medical interventions safely and effectively, in accord with the available evidence, requires putting the evidence into context and considering relevant uncertainties and counter-examples. The rigorous standards for FDA approval and restriction on off-label promotion is intended, in part, to provide physicians and patients with assurance that this work has been done.
By eliminating this assurance, the current push toward a vague and/or narrow “true and non-misleading” standard — exemplified by the Amarin ruling — thus poses a serious risk to public health. Truthful and yet uninformative statements can conceal a lack of evidence. Even informative statements can fail to provide adequate context and mislead physicians into prescribing ineffective or harmful drugs. Unfortunately, both of these kinds of promotional statements have recently been judged as “true and non-misleading” test.
A fundamental goal of the FDA restriction on promotional speech is to ensure informed physician decision-making and valid patient consent. This goal is well served by restricting promotional claims to only approved uses of medical products. In the absence of such regulation — whether due to other court decisions or political pressures on the FDA — a more useful and rigorous “true and non-misleading” standard, such as the one we have outlined would need to emerge to help protect patients and the public health.
