This note explains some of the concepts swirling around in the media right now, relating to medicine approval. Much of what follows appears (or will appear) in an article on the U.S. “right to try” law, which I recently wrote with a colleague at the University of Bourgogne in Dijon, France. Some of the background discussion will be useful here.
A new medicine must be approved by FDA before it can be shipped in interstate commerce (effectively, before it can be sold commercially for use by patients). There are two pathways to market in the United States: a biologics license application for a biological drug and a new drug application for any other type of drug. FDA requires proof of safety and effectiveness, which takes the form of data from laboratory and animal (“preclinical” or “nonclinical”) testing as well as human (“clinical”) trials.
A variety of legal, scientific, and ethical considerations mean that developing the safety and effectiveness data for premarket approval of a medicine is iterative. That is, after trials in relevant animals show that it would be safe to begin testing in humans, the applicant begins with small safety tests (often in healthy volunteers) and moves gradually to larger and larger trials. During this time the medicine is considered “investigational” or “experimental.” And it can’t be introduced into interstate commerce.
The traditional approach involves three phases of testing.
- Phase 1 trials entail the initial introduction of the investigational medicine in humans and focus on how the body reacts to the medicine — questions of absorption, distribution, metabolism, excretion, and side effects of increasing dose. These trials sometimes also generate early evidence of effectiveness, if the subjects are patients rather than healthy volunteers.
- Phase 2 trials are usually controlled and assess the effectiveness of the medicine in patients, as well as common short-term side effects and risks. In a “controlled” trial, each subject is randomly assigned to one of two groups — one group receives the experimental medicine, and the other receives a “control” (typically either a placebo or a medicine that treats the condition) for comparison.
- The pivotal trials proving statistically robust proof of effectiveness — phase 3 trials — often involve thousands of patients and clinical trial sites around the country or world.
The gold standard for proof of effectiveness is a prospective, randomized, controlled, double-blinded trial. Although randomized, controlled trials aren’t perfect, the design minimizes the potential for bias (which might happen, for example, if the patient or doctor knows who is receiving the test drug and who is receiving the control) as well as the problem of confounding variables. This form of evidence is superior to real world evidence and superior to personal experience and anecdotal information.
It’s possible for a trial to combine elements of phase 1 and 2, or phase 2 and 3 … there’s no law requiring companies to proceed through three discrete and sequential sets of trials. But the basic principle applies: you start small, you generate an adequate database in humans to characterize the drug’s safety profile, and you complete trials designed to support a statistically robust conclusion that the drug is effective (causes the result you observe).
The Tradeoff Involved
In theory, a new medicine must be shown safe and effective before it may be marketed. In reality, no medicine is perfectly safe or always effective in the patients for whom it is labeled. Patients are heterogeneous, and clinical responses vary. Side effects are inevitable; medicines are biologically active, and the relationship between a patient’s body and the chemical can be complex. As a result, when approving a new medicine for the market, the most a regulator can ask for is proof that the medicine’s benefits outweigh its risks for most people most of the time.
The challenge is that it is impossible to be certain about this. No premarket research and development program can generate complete information about a medicine’s clinical profile. Approval really means only that the medicine’s benefits outweigh its risks based on the data generated to date. Requiring premarket approval therefore always entails deciding when to make the call on benefit and risk — how much data must be generated before FDA makes this call.
This creates a tradeoff. On the one hand, while it is impossible to eliminate all uncertainty about the clinical profile of a proposed new medicine, more testing will always provide more certainty. On the other hand, testing delays the regulatory decision (and thus market entry), and requiring more testing delays the decision even longer. If the regulator still approves the medicine, the testing delayed access to a medicine that had a positive benefit-risk ratio the whole time. Patients who could have benefitted from the medicine had to wait. And if the medicine treated a serious or life-threatening disease, some patients — hundreds, thousands, or more — may have never had the chance to use the medicine. They may have died waiting.
How much information is enough for a decision on the risk-benefit profile of a new medicine depends on the relative weight given to two values — earlier (rather than later) release of new medicines to patients, on the one hand, and reduction of uncertainty about the effects of those medicines, on the other hand. Doctors, patients, regulators, and policymakers may disagree about the relative importance of avoiding errors (rejecting good medicines and approving bad medicines) and the cost of delay, just as they may disagree about the weight to be placed on particular benefits and particular risks.
Our legal framework give these calls to a regulator. A medicine may not be sold for use by a patient — even if the benefits exceed the risks from the perspective of the patient, even if avoiding delay is more important to the patient than knowing more about the drug — until a regulator agrees, based on its own assessment of benefits and risks. And it considers these at the population level, not the individual level.
Evolution in the Gatekeeping Model
Over the last half century, however, the regulatory gatekeeping model has evolved, as the broader relationship between the individual and state on matters of personal health has evolved. Our article explores this evolution in France and the United States, considering scientific developments, sociocultural changes, and broader legal pressures that contributed to the evolution. And we describe two innovations in the gatekeeping model: early decision mechanisms and early access mechanisms.
Earlier Decisions … e.g., Fast Track and Accelerated Approval
An early decision mechanism shifts the timing of the regulator’s benefit-risk judgment call to an earlier moment in time (and possibly an earlier moment in the process of generating evidence). Some early decision mechanisms are designed to accelerate premarket research and development, or regulatory review, with no change in the standard for approval. For example, in the United States, “fast track” designation entitles a company to more frequent meetings with FDA to discuss trial design and data needed for approval. It is available for a drug intended to treat a serious or life-threatening disease or condition, if it demonstrates the potential to address unmet medical needs for that disease or condition. Other early decision mechanisms permit approval on a different evidentiary basis. For example, U.S. law permits accelerated approval of a medicine intended for treatment of serious or life-threatening illness, based on data that do not show clinical benefit but rather are thought to predict it.
Earlier Access … Expanded Access
Early access mechanisms emerged during the worst years of the AIDS crisis and responded to the fact that better informed and newly empowered patients were willing to take greater risks in exchange for earlier access to new medicines. (But, to be clear, even before the AIDS crisis, FDA had permitted seriously ill patients access to experimental drugs.) Our article explores this history — in France and the United States — but I am just going to summarize current law here. Current law is reflected in the statutory provisions and regulation governing expanded access to investigational drugs and devices.
With expanded access, FDA still plays a gatekeeping role. Expanded access requires a showing that (1) the patient has a serious or immediately life-threatening disease or condition for which there is no comparable or satisfactory alternative therapy, (2) the potential benefit for the patient or patients justifies the potential risks, and the potential risks are not unreasonable in the context of the disease being treated, and (3) providing the drug will not interfere with clinical trials that could support marketing approval.
FDA will permit expanded access for individual patients as well as for groups of patients.
For an individual patient, the general criteria for expanded access must be satisfied, and (1) the treating doctor must determine that the probable risk to the patient from the drug is not greater than the probable risk from the disease, and (2) FDA must determine that the patient cannot obtain the drug any other way (for instance, by enrolling in a clinical trial). The agency ordinarily looks for completed phase 1 trials at doses similar to those proposed for the patient, together with preliminary evidence suggesting effectiveness. In some cases, however, FDA will permit a single patient access based on preclinical (animal) data or even mechanism of action.
In addition, the U.S. permits expanded access for groups of patients.
First, it permits widespread use. The ordinary standards for expanded access apply. If the medicine is intended to treat a serious disease or condition, FDA will look for data from phase 3 trials showing safety and effectiveness, but in some cases it will accept compelling data from phase 2 trials. If the medicine is intended to treat an immediately life-threatening disease, FDA will consider whether “the available scientific evidence, taken as a whole, provides a reasonable basis to conclude that the investigational drug may be effective for the expanded access use and would not expose patients to an unreasonable and significant risk of illness or injury.” This will “ordinarily consist of clinical data from phase 3 or phase 2 trials,” but it could comprise “more preliminary clinical evidence.”
Second, it permits use by “intermediate-size” groups. This is appropriate if patients cannot participate in the ongoing trials — because they do not meet enrollment criteria, because enrollment has ended, or even because the trial site is not geographically accessible. The agency’s regulations also describe use of this arrangement when a drug is not under development at all — for instance, because it is so rare that the sponsor cannot recruit patients for trials. For intermediate-size groups to enjoy early access, the ordinary standards for expanded access must be met. In addition, there must be (1) enough evidence of safety to justify a clinical trial at the same dose and duration in the same number of people, and (2) preliminary clinical evidence of effectiveness, or of a plausible pharmacologic effect, sufficient to make use a reasonable therapeutic option for the patients in question.
Right to Try
In 2018, the U.S. Congress passed a law that takes a different approach to the challenge of balancing the desire for robust data about new medicines, on the one hand, and the desire to allow individual patients access to incompletely tested but potentially beneficial compounds at their own discretion, on the other.
The new right-to-try law permits early access without a regulator’s involvement. Congress added a new section to the U.S. drug statute, which — when its terms are met — exempts certain drugs provided to certain patients from the gatekeeping provisions of that statute and from FDA’s regulations implementing those gatekeeping provisions. The patient in question must be diagnosed with a life-threatening disease or condition — generally meaning the likelihood of death is high unless the course of disease is interrupted. And the patient must have exhausted approved treatment options and must be unable to participate in a clinical trial involving the drug. The drug itself must be the subject of a pending marketing application or a clinical trial intended to form the primary basis of a claim of effectiveness in support of marketing approval, and it must have completed phase 1 trials. If these things are true, the drug may be provided to the patient.
The federal government does not play a role in determining whether these things are true. Neither the company nor the doctor seeks permission from FDA. If anyone plays a gatekeeping role, it is state-licensed doctors. Before the drug can be provided to the patient, a physician in good standing with the appropriate licensing board must determine that the patient has exhausted approved treatment options and cannot participate in a clinical trial. The right-to-try law specifies no actor to enforce the other two threshold eligibility requirements — that the patient’s disease is life-threatening and that the patient provided informed consent.
FDA’s role here is, at best, after the fact. The agency would have to learn of the procedure in the first instance and then, believing that the patient had not provided informed consent or did not suffer from a life-threatening disease, claim that the patient had not been eligible for right-to-try access. If true, the drug was not exempt from FDA’s gatekeeping authorities, and FDA could take enforcement action. But the agency will not learn about right-to-try treatments until the company’s annual summary of right-to-try uses, and the statute does not require identification of the investigators or patients. So these limitations may turn out to be a sham. To be fair, though, state law will usually impose its own informed consent obligation on treating doctors. And it may require that access proceed through the same kind of ethics review as FDA would have required.
Although the primary feature of the right-to-try law is its removal of FDA as a gatekeeper, the law also addresses two grounds on which companies supposedly decline to provide their medicines before approval. First, the law limits their liability exposure. A company faces no liability arising out of any act or omission with respect to medicine provided to patients under right-to-try. Second, the law limits FDA’s use of the data arising out of the patient’s use of the medicine. The agency cannot use a clinical outcome from right-to-try to delay approval of the medicine unless the sponsor requests that use or the agency finds that using the clinical outcome is critical to determining the medicine’s safety. (But it is not clear these are the real reasons companies decline to provide expanded access. Recent scholarship suggests that concerns about adverse regulatory outcomes and liability exposure would not have been well-founded.)
Comparing Expanded Access and Right-To-Try
The essence of the right-to-try law is elimination of the regulator’s role. Compassionate use under the right-to-try law is a matter for the company, doctor, and patient. FDA receives no information, has no review authority, and as a practical matter has no role. The right-to-try law also strips the agency of the authority to impose conditions on access. Ordinarily, even in expanded access situations, the sponsor of the trial (usually the drug company) notifies FDA within 15 days of any serious and unexpected adverse reaction. It also notifies investigators working with the drug. These rules do not apply. Nor do FDA’s rules relating to maintaining control of the investigational medicine or recordkeeping. And the agency has no power to call a halt to the process when patients are subject to unreasonable risk of injury or when the doctors lack the training and experience necessary to administer the drug. Only three FDA regulations relating to investigational medicines will apply: a regulation governing labeling, a regulation prohibiting promotion, and the regulation limiting how much the company can charge (only the direct costs of making the medicine available). And the agency will have to enforce these rules after the fact, when it receives the company’s annual summary.