There is one certainty about the current regulatory process for drug approval in the United States and Europe: No one likes it.
Manufacturers are frustrated by the need for large, complex, and lengthy clinical-development programs that often hinge on meeting a single endpoint in one pivotal clinical trial. As a result, the cost to bring a drug to market has been estimated to be well over $1 billion — and it may be much higher. Patients and providers are disturbed by lack of timely access to medicines that show early promise in addressing significant unmet needs. Even regulators, who are responsible for enforcing the current structure, chafe at what manufacturers typically present to them: successful trial results in patients who are carefully selected to show the drug offers benefits but who are not very representative of the broader population likely to receive it. Payers then have a mess on their hands: pressure to pay for premium-priced medications that, when broadly employed, don’t offer much therapeutic benefit over existing alternatives.
Can this process be altered? Yes, but it will require a true reinvention of the regulatory drug-review process that addresses all of the flawed assumptions that exist with the current framework. One potential solution is adaptive licensing (also called “staggered approval” and “progressive licensing”), a novel approach that is currently being discussed by manufacturers, regulators, and other stakeholders.
Under adaptive licensing, the clinical-development program is restructured to allow for early approval of a new compound for a limited, typically high-risk population based on valid clinical measures from smaller human studies. Approval would be revisited at several points along the clinical-development pathway as candidate populations are broadened, longer-term outcomes are evaluated, and risks of treatment are better understood. Indications for treatment would be broadened (or restricted) accordingly at each step.
Initial Reforms Don’t Suffice
I would argue that some of the attempts to streamline the current review process are only minor modifications and fall short of what is needed. For example, the “accelerated approval” pathways used by the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for certain drugs have generally been limited to small reductions in the time allotted for evaluation of the traditional submission package or use of “surrogate” endpoints (e.g., reductions in blood levels of a cardiac biomarker rather than reductions in risk of stroke or MI). This latter approach has been controversial, because many proposed surrogates have proven themselves merely to be correlates of disease rather than important components of the causal pathway.
Advantages of Adaptive Licensing
The basic concept at the heart of the adaptive licensing approach is that the review process should be a “learning system.” Breakthrough therapies would be made available to patients much earlier in the regulatory process. Presentation of clinical data to regulators in a stepwise fashion throughout drug development should allow the drug company to better target those patients for whom the benefits of the therapy outweigh the risks and should result in more informed decisions on withdrawing the drug from the market if the risks are simply too great for anyone.
(In the current paradigm, safety concerns are often not raised until well after regulatory approval. The reasons: regulators don’t get any hint of safety issues until they are presented with the submission package, and given the significant investment required to bring the drug to regulatory attention in the first place, manufacturers have no incentive to undertake a full exploration of safety.)
Unanswered Questions about Adaptive Licensing
One is whether current regulatory controls are sufficient to prevent a new medicine from being used for purposes that go well beyond the initial, limited indication. Do manufacturers need to commit to adding their own controls? It also may be the case that manufacturers will need to augment traditional efficacy studies with long-term observational studies that start earlier in clinical development in order to detect potential safety signals much sooner.
It is also clear that little progress on implementing adaptive licensing will be made without explicit cooperation from all stakeholders. All parties must feel that they can participate in confidential discussions that involve full disclosure. An ambitious program created by MIT’s Center for Biomedical Innovation, known as NEW Drug Development ParaDIGmS (NEWDIGS), has been created to serve as this “safe haven.” The program has made substantial progress in convening stakeholder workshops to (a) discuss the promise and pitfalls of an adaptive licensing approach, (b) quantify the potential benefits of adaptive licensing through historical case studies, and (c) develop standards for determining candidate compounds for adaptive licensing.
An Obstacle: The Existing Mind-Set
But the biggest unanswered question is whether we can collectively shift the paradigm enough. I was privileged to be invited to a recent NEWDIGS workshop to help represent the payer community. During the workshop, we discussed an actual compound that had been developed for a rare condition that represented a significant unmet need. The manufacturer had clearly done its homework on adaptive licensing and presented a thoughtful approach to clinical development that featured early authorization based on a clinically-validated endpoint, followed by longer-term studies evaluating both drug safety and “harder” outcomes (in this case, survival).
The initial pushback from regulator and payer representatives (including me) was all-too traditional: “How do we know this endpoint is associated with improved health status for the patient?” “How much bigger and longer would the first study need to be to show impact on survival?”
The frustration was evident on the manufacturer representatives’ faces. It took us until well into the afternoon to realize that we were asking them to chop a traditional clinical-development program into bite-size chunks rather than engaging in a true rethinking of the approval process. They were presenting us with their best approximation of a clinical-development program that would reduce initial investment, provide useful early clinical information, and mitigate patient risk, and we threw the old process back at them. We did eventually arrive at a good place (i.e., discussions will continue based on the adaptive program as presented) but not without a lot of pain.
And so it is clear that yes, adaptive licensing requires cooperation and information-s
haring among all stakeholders as well as candidate compounds that meet the criteria for such an approach. But the success of an adaptive approach really only hinges on whether we are willing to shift the current regulatory paradigm — a radical change in thinking about what we as a society want the products of a drug-development program to be, when we want to see results, and how much risk we are willing to bear at each step in the process.