Institutional Review Boards as Arbiters of Expanded Access to Unapproved Drugs

Patients facing death or serious illness who have exhausted all approved treatment options may seek access to unapproved interventions, such as drugs in development. Patients unable or unwilling to enroll in a clinical trial may try to gain access to unapproved interventions through their treating physician. In the United States, preapproval access, sometimes called compassionate use or expanded access, is governed by the Food and Drug Administration (FDA). Under the FDA’s expanded access policy, patient access to investigational interventions is dependent on the willingness of the company developing the drug (the sponsor) to make the investigational product available. ¹ If the sponsor is willing to make the drug available to an individual or group of patients with a life-threatening or serious disease for which no approved options are available, then the request must be approved by the FDA and by the treating physician’s institutional review board (IRB). Recently, this subsystem of drug distribution has experienced much scrutiny and criticism, especially by advocates of the “Right to Try” movement, who claim that FDA approval of such requests is inappropriate and unneeded. ² Little attention, however, has been paid to the role of IRBs. The role of IRBs with regard to compassionate use has been overlooked. While research is needed to reveal how IRBs understand and operationalize their role with regard to preapproval access, it is time to reconsider whether IRBs ought to be involved in any way with evaluating and approving compassionate use requests.

An IRB’s purview and mission is the ethical review of proposed research protocols. However, compassionate use is different from clinical research. Compassionate use involves experimentation, yet because the intent is therapeutic, it is primarily a form of clinical care. Clinical care is traditionally outside the domain of IRBs, which the FDA recognizes. Underscoring the fact that compassionate use is not research, the FDA allows compassionate use interventions of a time-sensitive nature to be done without prior IRB approval provided that the sponsor informs the IRB within 5 working days.

Despite the evident lack of fit between IRBs and compassionate use, this situation has, to our knowledge, received neither comment nor criticism. No data describing how IRBs deal with such requests could be located, leaving unanswered questions about the number of requests received and the percentage approved, the time frame in which they are processed, the conditions placed on investigators in order to protect patients, and other elements of IRB review and follow-up.

Since some compassionate use requests involve investigational agents in phase 1 trials (or possibly even earlier), data regarding risks and benefits may be all but nonexistent, impairing the IRB’s ability to make data-based judgments. Patients requesting compassionate use are often vulnerable, with few options, and their perception of risks is sometimes affected by hype or by early publications offering positive results and outcome reporting bias in reports of drug trials. ³ As such, IRB review could play an important oversight role in independently assessing whatever data are available. The IRB could help ensure as detailed and informed a consent process as possible including requiring discussion of all other options, such as palliative care. IRBs could assist in establishing stopping rules and in managing sponsor conflicts of interest.

Nevertheless, at least 4 unresolved ethical issues directly impact IRBs’ capability to usefully review compassionate use requests. The first involves a poor basis for approving requests when there is little safety or efficacy data. Anecdotal evidence indicates that IRBs approve a very high percentage of such requests, but what criteria these bodies use in making such judgments is unknown. Could it be that IRB members are responding to a perceived “duty to rescue,” even when the odds are slim that the experimental intervention will help the patient?

A second ethical question involves balancing compassionate use with the need to sustain the clinical studies necessary to prove new drugs safe and effective. Despite a seeming compulsion among Americans to rescue those in need, this is not a duty and is not strong enough to hold a company or, more importantly, a line of scientific inquiry hostage to the needs of those few patients seeking rescue. ⁴ Questions remain about how to balance the competing needs of patients seeking direct access to unapproved interventions and the public’s need for the clinical trials required in order for new drugs to be tested and approved. If there is a limited supply of an investigational agent or a company’s resources are too limited to run both clinical research and expanded access programs, must companies offer expanded access? The requests IRBs see will reflect varying priorities made by companies/sponsors and will not be congruent from request to request.

Third, while in some cases preapproval access is the only hope for a therapeutic option, patients are often presented with a binary choice of preapproval access or clinical trial. Such presentation leaves unmentioned other possible options such as the off-label use of approved drugs. Such options should be discussed in the informed consent process as alternatives to a trial but rarely are. ⁵ Such incomplete, nontransparent disclosure of alternatives to a clinical trial can be misleading and are likely to fuel distrust and overreliance on preapproval access. Do IRBs have the ability to ensure that the consent process covers all possible options and that neither the patients nor their physicians are swayed by therapeutic misconception with regard to a desired experimental intervention?

A fourth issue concerns the generation of data from compassionate use interventions. The FDA requires notification, to both itself and the sponsor, of serious adverse events that occur during compassionate use, but IRBs typically do not track compassionate use outcomes. Is this an abdication of the IRB’s oversight role, and does it deprive the IRB from an opportunity to learn more about the possible benefits and/or harms associated with an experimental intervention, which for which they may see another proposal in the future?

Given their composition, mandate, and training, are IRBs the correct body to review compassionate use requests? Is institutional review even necessary, given that the FDA and the sponsor have reviewed and approved the request? It is certainly not clear that the time lost to IRB review is justified by an increased level of subject protection. Nor is it clear that there is a true “value added” of IRB review for compassionate use, given that IRB review and approval can be dispensed with in emergency cases.

Those seeking compassionate use appear to constitute the sort of vulnerable population that IRBs are particularly charged with protecting from the risks and burdens of research. However, given the therapeutic intent of compassionate use, it may be inappropriate to subject compassionate use requests to review by boards intended to review research. In seeking ways to expedite compassionate use requests, consideration must be given to modifying, minimizing, or eliminating the role of IRBs.

If IRBs are deemed to have an important role to play with regard to compassionate use, they must receive clear guidance and be assisted to develop best practices. Periodically, specialized review committees have been established to work out ethically problematic areas, such as embryonic stem cell research oversight (ESCRO) committees or the Recombinant DNA Advisory Committee (RAC) reviewing all NIH-funded gene transfer protocols. Such an approach might be fruitful in resolving the issues that bedevil IRBs’ interactions with compassionate use.

There ought to be an ongoing reconsideration of the role of IRBs, based in part on concerns about their inconsistency in risk assessment and informed consent language and the lack of availability of their reasoning and of appeal of their decisions, against a background of underdeveloped definition of their benefits. ⁶ At present, the IRB landscape is undergoing a significant shift, away from institutional IRBs and toward central IRBs, often ones that operate for profit. The time is ripe as well for revisiting the requirement that IRBs review and approve compassionate use requests.