FDA Guidance Review: Waiver of Informed Consent
Imagine a study that requires access to the health information of a few thousand patients who have received an implantable device, in order to study the effectiveness of that device in a “real world” setting. Let’s further assume the study does not require any interaction with these patients or the collection of identifying information and that the results may be submitted to the FDA.
Is prior informed consent required to conduct this study?
In the past, likely, yes. Now, most probably not. Why the swing?
In response to the dictates of the 21st Century Cures Act, the FDA has taken a significant step in harmonizing the sometimes disparate systems of mandated ethics review in the U.S., recently issuing a new guidance document.
Working Toward Harmonization
The Common Rule has long provided a means for IRBs to waive the informed consent requirement for minimal risk research, pursuant to 45 C.F.R. § 46.116(d). The FDA’s regulations for the protection of human subjects have not included this flexibility. This is perhaps due to the perception that drug and device research is generally never minimal risk, or due to concerns that allowing waivers of informed consent may result in widespread retrospective clinical trials, which would involve systematic uses of approved drugs and devices for off-label indications in possible support of broadening market approval.
The new guidance document reports that the basis for only permitting exceptions to the informed consent requirements in life threatening situations or for emergency research stemmed from such stipulation in the Food, Drug, & Cosmetic Act regarding devices; the FDA then elected to apply the statutory limitation to clinical investigations involving drugs as well, to promote consistency—at least within the FDA. Now, in light of the statutory change, the agency has decided to exercise immediate enforcement discretion regarding an IRB’s ability to waive the requirement for informed consent in FDA-regulated studies.
The long held answer to the question posed above was “likely,” due to ambiguity regarding how the FDA may interpret what it means for a clinical investigation to “involve” a test article and a human subject(s). The general impression—having been reinforced anecdotally through various FDA interactions—considered retrospective chart reviews as involving a test article (even though the drug or device would not be administered as part of the research), in particular when there was an intention to submit the data to the FDA.
The FDA’s definition of a human subject—“an individual who is or becomes a participant in research, either as a recipient of the test article or as a control; (a) subject may be either a healthy individual or a patient”—is likewise susceptible to a degree of uncertainty, namely in that it is unclear whether the individual in question must receive the test article/control pursuant to the study protocol itself or whether observational research is also included.
Further complicating matters, IRBs frequently address some permutation of the following contention: since the study will not involve the collection of identifiers or an interaction with the participant, informed consent is not necessary. This may or may not be true, depending on other factors. Unlike the Common Rule, identifiability has no bearing on the applicability of Parts 50 and 56, and unfortunately the Common Rule framework is often mistakenly superimposed on the FDA regulations. The definition of human subject does not contain an explicit limitation regarding identifiability. If the data involves a test article and is intended to be submitted to the FDA, the question of identifiability of participants, at least with regards to the FDA, is fundamentally irrelevant.
Thus, essentially, if the sponsor intends to submit the data to the FDA, Part 50 likely applies, and, up until very recently, informed consent would be required. Researchers were left to characterize any medical record access as a screening activity to identify eligible participants from which to subsequently obtain informed consent, typically with a request for a waiver of documentation of informed consent to simplify the consent process as much as possible (21 C.F.R. § 56.109(c)), or to abandon the endeavor altogether. After all, it stands to reason that data collected for and submitted to the FDA should be in compliance with—or at least in deference to—its regulations. Failure to obtain consent, even with an IRB’s blessing, may jeopardize the FDA’s willingness to include the resulting data in its determinations (21 C.F.R. § 56.103(b)).
So, why now?
While seemingly having a minimal impact on the bulk of FDA-regulated research, which still almost invariably exceeds the minimal risk threshold, sponsors and investigators have repeatedly confronted the challenge of conducting minimal risk research that may offer significant information regarding risks and efficacy of approved medications. Such scenarios only stood to multiply in frequency with the momentum building behind real-world evidence, big data, and interoperability of electronic health records—as the 21st Century Cures Act attests.
However, the optimism for greater efficiency and uniformity certainly comes with legitimate safety, privacy, and confidentiality concerns. To what extent is society comfortable with allowing access to health information without individual and particularized consent? And, further, are patients and regulators comfortable with “real-world” trials in lieu of the classic randomized clinical trial?
Looking forward, the FDA will promulgate new rules, presumably to remain generally consistent with the issued guidance document and, ultimately, the revised Common Rule. In the interim, researchers are empowered to proceed with valuable minimal risk research, under an IRB approved waiver or alteration of informed consent. Finally, the practicability of obtaining consent may be considered in the conduct of minimal risk research regulated by the FDA.