3

Strengthening the Regulatory Pathway

Although the regulatory pathway is designed to ensure that only drugs that are safe and efficacious enter the market, several participants noted that the current system has several critical challenges, such as the length of time to approval and to market, and the lack of clarity about regulatory decisions. Workshop participants explored current pathways intended to speed drugs to market in areas of unmet need for serious medical conditions, and examined the promises and pitfalls of adaptive clinical trials and conditional approval pathways for CNS drugs. Many participants underscored the importance of clarity of regulatory processes so industry decision makers can have confidence in making decisions about drug development programs. In addition, an overview of how third-party payers make reimbursement decisions was provided.

DRUG DEVELOPMENT: A REGULATORY VIEWPOINT

Janet Woodcock noted that nervous system disorders cause a vast amount of suffering to patients, family, and society, and there is a tremendous urgency to respond. For FDA, this means doing everything it can to stimulate development of therapeutics. To provide context for an informed discussion about possible policy changes aimed at stimulating innovation in CNS drug discovery, Woodcock described key aspects of the drug development pathway from the regulatory vantage point. Product developers typically require two elements to be in place to pursue a particular drug, she said. First, they need to see a market. In the CNS sector, the market is easy to recognize because disease prevalence and incidence are increasing, and the conditions are common. Second, product developers also need to know that a predictable path to market exists—not

necessarily an easy path, Woodcock said, as some failures are inevitable in the realm of drug development. However, predictability is key.

Predictability is heavily influenced by how much is known about the basic biology of the illness, Woodcock said. For example, to get a clear result about whether the drug works, a clinical trial must be long enough that, without intervention, the disease would have progressed. Along these lines, several participants noted that information about natural history is important because study design must take into account what’s known about disease progression relative to the margin of error in measurements and random symptom variability in the population. Some trials are too short to observe a change, even if the drug is effective.

In addition, Woodcock and others highlighted the importance of biomarkers. They can help diagnose a disease and thus identify individuals in whom a potential drug can be tested. Furthermore, biomarkers that reliably predict the amount of time until some particular symptom or event will appear can help inform how long a trial needs to be. The most important type of biomarkers, said Woodcock, are pharmacodynamic ones that reveal early during treatment whether the drug has the desired effect. Because trials need to be long enough to observe a change, surrogate endpoints—such as pharmacodynamic effects—can be useful, especially for diseases that occur over long periods of time. Ideal features, including information about the disease’s phenotype, natural history, and pathogenesis, are absent for many nervous system disorders, Woodcock said.

Existing Accelerated Pathways

FDA has several mechanisms that have been designed to encourage and accelerate development and review of drugs that address unmet medical needs in the treatment of serious or life-threatening conditions: priority review designation, accelerated approval, breakthrough therapy designation, and fast-track designation (see Box 3-1). Woodcock illustrated crucial features of these approvals through the lens of Alzheimer’s disease. Given that no effective treatments exist for this serious illness, drugs that have nonclinical or clinical data to demonstrate their potential to have a disease-modifying effect to address the unmet need could obtain fast-track designation. To get breakthrough therapy designation, a company would need clinical data that indicate a change in the trajectory of the disease. That is a high bar, but if such evidence existed, FDA would be asking how it can help make the drug, manufacture it, test it, and get it further evaluated, said Woodcock.

___________________

^ahttp://www.fda.gov/forpatients/approvals/fast/default.htm (accessed April 22, 2015).

^bhttp://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm358301.pdf (accessed April 22, 2015).

For priority review, the data in the application would show there is a significant advance in safety or effectiveness over existing therapies. Finally, for accelerated approval, FDA would have to see the effect of the drug on a surrogate or on an intermediate clinical endpoint. For example, FDA has issued draft guidance for developing drugs for the treatment of early-stage AD that says it would use accelerated approval for a drug that improves neurocognitive testing in pre-symptomatic patients (FDA, 2013). If a battery of tests established delay in decline of neurocognitive testing, FDA would consider accelerated approval for that application, said Woodcock.

In the context of this workshop, said Woodcock, the question is what else might be done beyond the existing pathways. She emphasized that there are trade-offs between lowering standards to get more drugs to people and maintaining high standards, but getting fewer treatments on the market.

REGULATORY CHALLENGES SPECIFIC TO CNS DISORDERS

The investment community views drug development for CNS disorders as “a green field, but with tons of boulders,” said Jeffrey Jonas, chief executive officer at SAGE Therapeutics. He pointed out that people are very aware of the large and public failures of CNS trials and how much those failures cost. The feeling is that CNS companies have overstated the efficacy of their molecules, said Jonas. Furthermore, there is a perception that the CNS regulatory environment has much more uncertainty than other areas, said Jonas, and that it is especially difficult to obtain approval for drugs that target neurological and psychological ailments.

Aaron Kesselheim, associate professor of medicine at Harvard Medical School and Brigham and Women’s Hospital, agreed with Jonas that studying new drugs for certain CNS diseases—particularly chronic ones—can be very difficult because disease courses can be highly variable and there are no validated biomarkers or other surrogate endpoints available. When investigators cannot easily predict which drugs will or will not work, obtaining regulatory approval for those drugs could be

marked with some uncertainty, because current FDA standards require demonstration of substantial evidence of efficacy based on adequate and well-controlled investigations. Kesselheim pointed out that if drug companies did not need to demonstrate efficacy and safety to FDA, the market would be deluged with ineffective or unsafe therapies, as it was in the early part of the twentieth century, before FDA’s current standards existed. In addition, he noted that current FDA standards are highly flexible with regard to medications that address unmet medical needs such as those that are innovative or that treat life-threatening conditions. In those cases, FDA often accepts earlier-stage trials and less certainty about safety or efficacy in approving new drugs.

As discussed in Chapter 1, FDA approval times for CNS disorder drugs exceed those in other disease sectors. Approval rates fall only slightly short of those for other medical areas, but this minor difference combined with the other difficulties associated with drug development in the neurological realm makes the area relatively unappealing, said Dennis Choi.

Although the pathways that Woodcock described are available for drug development across all realms, said Jonas and Steven Hyman, barriers exist to using them in the world of CNS disorders. The current state of knowledge limits use of existing mechanisms intended to facilitate drug assessment in the CNS arena, said Hyman. For example, the accelerated approval pathway depends on the existence of a suitable biomarker, and such surrogates for clinical endpoints are not available for many nervous system disorders, he added. For some conditions, delaying time to progression might provide a measurable endpoint, said Gail Maderis, but in the area of neurological diseases, that endpoint usually requires long and large trials.

Currently 88 biologics and drugs have breakthrough therapy designations (FDA, 2015). Only four of these compounds are in CNS, said Jonas. Furthermore, available evidence suggests that there is a bias away from small innovators. Of the 26 public disclosures from November 2013, 16 of the drugs that were given breakthrough designations for 2013 were from large pharmaceutical companies (Mullard, 2013). Some of the features that make a condition/drug combination amenable to study are not accessible to small companies. Large entities sometimes have data from other projects in their portfolio that can provide knowledge that helps them explore new ideas and/or design strong trials, said Jonas. Furthermore, they are more likely to have the financial resources to gather other valuable information, such as performing a natural history study.

Such capability can open up exploration of a rare and understudied disorder or a disorder for which registries do not exist or give substandard quality data. Jonas concluded that some of the regulatory initiatives may not benefit small-company innovators, especially in CNS, where datasets are often incomplete.

Jonas provided an account of an individual with Lennox-Gastaut syndrome¹ who was in an intensive care unit in super-refractory status epilepticus.² No conventional interventions were helping. With no approved treatments, the family and treatment team were faced with withdrawal of care. This individual then had a dramatic response to an experimental drug, getting better and going home. Jonas asked workshop participants to think about what happens in this context, where innovation exists, but an approved therapy does not. It is not easy to see how a conventional clinical trial could be conducted. Furthermore, statistical analysis in such situations poses a challenge because it is unclear how many patient responses constitute statistical significance. The ability to get accelerated approval on findings that are biologically plausible should trump statistical purity in such disease settings, said Jonas. Woodcock said FDA does accept compelling series in which no one expected people to rise from their bed and walk, yet they did.

FDA needs to maintain flexibility, asserted Timothy Coetzee, chief advocacy, services, and research officer at the National Multiple Sclerosis Society. Twelve disease-modifying therapies have been approved for multiple sclerosis (MS), and this situation has started “locking in the regulatory process and expectations,” said Coetzee (NMSS, 2015). When a new approach comes along, one must prove that the alternate strategy should be accepted, he said. The MS community is currently grappling with this issue in the area of progressive stage of the disease, as clinical trials that test drugs for this condition will be substantially different from those for the relapsing–remitting form. Regulators should think about new outcome measures “that perhaps haven’t been qualified,” he said. “This is where I think the advocacy community and others are going to have to come in and say, ‘this is where patient risk/benefit decision making needs to be weighed.’”

_______________

¹A severe form of epilepsy.

²Persisting and potentially life-threatening seizure activity despite anticonvulsant treatment.

POTENTIAL MECHANISMS TO ADDRESS REGULATORY CHALLENGES

Adaptive Trial Design

Given the complexity of drug discovery and development for CNS drugs, measures are being proposed to adjust the approval process to better address the realities of these illnesses, according to several participants. For example, a portion of the 21st Century Cures Act, which is currently under discussion, has grown out of acknowledgment that existing trial design and analysis might not be a good match for all disorders (USHOR, 2015a,b). A few participants noted that drug sponsors might want to test products in early disease stages, before well-established clinical endpoints have appeared. In this situation, which applies to many CNS disorders, several participants noted that adaptive trial designs might be appropriate, as they allow companies to modify ongoing studies. FDA, however, has not incorporated these types of trials into its routine approval process. The 21st Century Cures Act proposes to permit adaptive trial design and associated analytic methods, a strategy that might better fit some CNS trials, according to a few participants.

Adaptive clinical trials allow changes in study design in response to analysis of data at pre-specified points, noted George Vrandenburg. They leverage early data to guide decisions that can accelerate time lines and reduce costs by focusing the trial on the most promising doses, disease indications, and patient populations (FDA, 2010). Such a mechanism would afford companies the flexibility to adjust trials if early results reveal that a particular subpopulation of patients responds differently to a drug, for instance. A few participants noted that this pathway could save money and speed the discovery of significant clinical findings. The caveat is that such trials might produce data that confound interpretation.

Conditional Approvals

The topic of conditional approvals arose in several contexts during the workshop.

Through this proposed mechanism, a drug would be approved on the basis of information that suggests it is reasonably safe. Clinicians would then test the drug in different patients and attempt to identify what characteristics best predict a positive response, said Michael Rogawski, professor of neurology and member of the Center for Neuroscience at the

University of California, Davis. With this information, companies could design and perform properly powered and controlled clinical trials—and they could identify subsets of patients for whom the drugs will be most helpful.

Woodcock said that the main conceptual problems with that strategy is that it is not always easy to figure out why certain people respond, and identifying the predictive characteristics can be extremely difficult. Also, approving a drug based on safety data gleaned from Phase I and Phase II studies would only guarantee that the medicine does not cause a large number of people to have a “dramatic event,” she said. Detecting subtle increases in dangerous side effects requires large study populations. In addition, Woodcock stated that at this time FDA does not have statutory approval to issue conditional approvals for human drugs.

Kesselheim said that identifying subsets of people who respond to a given drug would be beneficial, but given the current state of knowledge about how to conduct rigorous studies in pharmacoepidemiology, can only be done in a reliable way in a clinically controlled environment. If results from a trial comprised of a large, relatively nonselected group turns out negative, but a responsive subpopulation emerges, such post-hoc subgroup analyses should lead to subsequent prospective trials testing that specific subgroup. Kesselheim pointed out that all biostatisticians would agree that post-hoc subgroup analyses are simply hypothesis-generating exercises. He therefore cautioned against approving drugs shown not to work in their clinical trials merely on the basis of supposedly positive effects in post-hoc subgroup analyses before those hypotheses can be further evaluated. Kesselheim added that FDA approval of a medicine for which efficacy has not been demonstrated would be problematic for many other reasons. For example, FDA’s stamp of approval is psychologically powerful, and will give physicians and patients the sense that it will work before there were robust data about the agent’s efficacy. It also complicates the conduct of subsequent clinical trials because a patient would have the choice of receiving a drug FDA-approved for his or her condition through normal channels or being put in a trial with the potential prospect of not receiving the drug; predictably, fewer patients would choose the latter option. A few participants noted that the proposed conditional approval system would rely on the fact that the therapy is not yet well validated, and therefore, has not received full endorsement. The thought is to have a mechanism in which a drug is brought forward in a conditional, preliminary fashion—with appropriate restrictions. This mechanism in turn might increase FDA and

payer leverages on reaching a conventional standard of approval, countered Choi. Woodcock asked how FDA would restrict distribution of the drug and determine who should have access to it. What are the implications if the drug is later rescinded?

Reducing Time to Market

Choi proposed one version of a conditional approval pathway for CNS drugs that might stimulate development in this area by reducing time to market (Choi et al., 2014; Eichler et al., 2012). In this system, highly selected drugs would be marketed in a limited fashion, based on biomarker or intermediate clinical data; a full period of market protection would become available after the drug sponsor obtains conventional data and clears the usual regulatory hurdles. So far, such a pathway has applied primarily to HIV/AIDS and cancer drugs, but it could be expanded to include a broader range of highly selected CNS therapies, said Choi. This mechanism entails added risk to patients, as it would reduce time in clinical trials before approval, and thus, less safety and efficacy data would be collected before approval. Given that the quality of surrogate markers in neuroscience is generally not as strong as those in HIV/AIDS and oncology, patient involvement would be crucial for making decisions about the appropriate risk/benefit spectrum for any particular drug combination, he added. In addition, several participants noted that input would be needed about providing acceptable cost to patients and payers.

Post-Approval Data Collection and Retroactive Approval

This and other ideas engendered more discussion on the topic of post-approval data collection, retroactive approval, and associated issues. Robert Armitage expressed concern about using retroactive approval measurements to determine periods of market protection, compared to a conditional approach conducted earlier on in drug development based on biomarkers, because doing so would increase uncertainty about the patent litigation. Companies would not know whether they would receive extra regulatory protection at the time when they have to decide whether to pursue any given project. The decision would be made after the product is on the market, by which time the development work is done. Furthermore, Armitage added, the political and economic environment might differ from when the decision was made, which would exacerbate the uncertainty. Hyman reinforced this idea; part of the challenge with

such a system, he said, is providing predictability. It forces companies to take risks before FDA approval, knowing they will have to demonstrate utility after a drug hits the market.

Adding to skepticism about the feasibility of this approach, Kesselheim said that FDA has limited authority to require post-approval testing and then to withdraw a drug if it does not meet the specified goals. Numerous studies show that manufacturers’ post-approval trial commitments to the FDA can be delayed or not completed. Bevacizumab (Avastin), for example, was approved based on a surrogate endpoint for metastatic breast cancer. When more convincing clinical studies showed that the drug actually did not improve survival for this condition and was associated with important safety issues, it took nearly 1 year of appeals and hearing to remove the indication. A few participants, however, endorsed some form of ex post facto approval. William Fisher, for instance, discussed the idea of using post-marketing data to determine whether a particular drug should receive “breakthrough” status.

Ethical Considerations

Alex John London, professor of philosophy and director of the Center for Ethics and Policy at Carnegie Mellon University, discussed ethical issues associated with reducing time to market. Given that companies derive great benefit from passing this milestone quickly, and many patients with serious illnesses who do not currently have useful drugs are willing to try therapies that have not been fully evaluated in clinical trials, stakeholder interests seem to align. There are, however, “trade-offs and pitfalls that come from trying to compress this time line,” he said. Early-phase studies provide crucial information about how best to use the drug—the clinical window for intervention, appropriate indications and dosage, and under what conditions toxicities appear. To shorten the time period of pre-market testing, it is likely that clinical trials would accept a more idealized and narrow population for study; as a result, some of this information would not be forthcoming before approval.

In principle, these data could be gathered after market entry, but the clinical environment is “noisier” than the trial environment, said London. Small effects are harder to detect, and doing so takes longer. Patients and third-party payers will bear this burden. “You have a trade-off,” he said. “Should you learn some of this information in smaller populations in controlled studies early on or if you rush to market, are you going to

have to learn that information later in less controlled settings where you expose larger numbers of patients to burdens?”

In addition, post-marketing studies suffer from another hazard, said London. People might hesitate to participate in studies because they could receive a placebo or the standard drug rather than the new one. “If you increase market access early, are you going to deplete the population of people who would be willing to participate in the studies that we need in order to ascertain whether these things have genuine efficacy?” he asked. “If you do that, then you have effectively closed the door” on attempts to gather crucial clinical data.

What might be more efficient for the developer could be less efficient for patients as a whole or for third-party payers. “There are issues about fairness and shifting these costs, offloading these costs onto other parties with the explicit goal of trying to increase the time to profit for one of the most profitable industries in the world,” said London. Furthermore, although patients might be willing to accept increased risk, serious adverse events can have a chilling effect on development. They also can sow distrust in innovative companies and regulators. Coetzee underscored the idea that workshop participants held widely disparate views on whether and how to improve the regulatory process. He noted a “mixed kind of view” about whether such changes are needed and highlighted the strong “differences of perspective.”

CLARITY OF REGULATORY PROCESSES AND DECISIONS

Woodcock emphasized the importance of predictability, as previously discussed, and described the agency’s efforts to communicate and work with relevant constituencies on the use of the various existing regulatory pathways. Regulatory transparency is important for companies to have a clear understanding of the development process and what is required, said many workshop participants. For example, many participants said it is crucial that terms and standards used by FDA are clearly defined and that better communication between FDA and investigators is needed regarding available pathways that are underutilized by the CNS community.

Definitions of terms such as “unmet medical need” and “breakthrough” are not explicit, and this situation creates problems, said Engelberg. Marc Boutin emphasized that the MODDERN Cures Act defines “unmet

medical need” based on what FDA uses for accelerated approval. Lengthy discussions were held among the group that crafted the legislation and people at NIH and FDA to devise the “right definition,” he said. The resulting language has been used by FDA for decades; the agency understands how to use it to good effect. Boutin gave some examples: A therapy for a condition without a treatment would qualify, as would a product that improves on an existing product with regard to measurable health outcomes and benefits. “There is a lot that goes into it, but FDA is skilled and practiced at defining what this is,” he said. The common impression that “every product would qualify for FDA’s definition of an unmet medical need” is misguided, he said. In the MODDERN Cures Act, “unmet medical need” is determined and designated not at the time a drug is approved, but at the time a drug sponsor files a clinical plan, said Engelberg. This can be problematic because companies tend to be overly optimistic; they often think that agents that are the second or third in a class will meet some unmet medical need. Perhaps those medications will work for longer than existing ones or eliminate a side effect, but those types of unmet medical needs lie on a different plane than a therapy for a serious disease that is currently untreatable. Engelberg said the only time anyone can reasonably assess whether a drug meets an unmet medical need is when it is approved; even then, that designation should be given only if the approval process contains a standardized way to compare effectiveness with existing treatments.

Engelberg also said a much stronger definition of “breakthrough” is needed to ensure that only the real value-added drugs are eligible. Fisher discussed possible ways to determine whether a drug should be categorized as “breakthrough”:

• Assembly of expert panels to predict whether a particular drug will be a breakthrough. This strategy likely would not appeal to innovators because they would have to expose themselves to such testimony, noted Fisher.
• Pre-approval (ex-ante) demonstration of increased efficacy through clinical trials. This scheme would be more empirically grounded, but it would be expensive and difficult.
• Post-approval (ex-post) demonstration of increased efficacy or social benefit in terms of saved DALYs. This determination would be made years after drug approval, when it is clear how well the drug is performing. At that point, a decision would be made about whether it has demonstrated sufficient breakthrough

“With respect to what kinds of drugs to incentivize, my suggestion is do not rely on expert panels,” said Fisher. Instead, he proposed that such decisions should depend on demonstration of health benefits through clinical trials.

In addition to standardizing and communicating key definitions, other opportunities for improving regulatory transparency exist. Jonas said there is a perception that various FDA divisions take different approaches to novel therapies. This perceived lack of consistency undermines companies’ efforts to prepare properly and efficiently for review. Part of the challenge, said Coetzee, is not just to determine what members of the FDA leadership think about the various ideas discussed in the workshop, but also how guidelines are implemented on a day-to-day basis. It will be important to work out details about how the sponsors and agency interact, and about the guidance that FDA provides when sponsors are launching clinical trials and doing other work required for drug development. “I think implementation of these issues is actually just as critical as having a high-level discussion around it,” said Coetzee. See Box 3-2 for global development regulatory opportunities.

REIMBURSEMENT: THE DECISION-MAKING PROCESS

In addition to transparency about regulatory processes and definitions, several participants sought clarity on reimbursement decisions, for example, knowledge about the size of the eventual marketplace. That number derives in part from the prevalence of conditions that a drug might treat, which directly affects reimbursement decisions. Rhonda Robinson Beale, senior vice president and chief medical officer at Blue Cross of Idaho, discussed how health plans decide what drugs to cover. The amount of money that goes into specialty drugs recently has increased dramatically. In 2012, these agents cost $87 billion, which was 25 percent of the total amount of money spent on drugs and 3.1 percent of the total national health care costs (UnitedHealth Group, 2014). In 2010, specialty drug expenditures amounted to 20 percent of total drug costs (UnitedHealth Group, 2014). Providers are part of the equation, because they are now becoming payers as they create their own insurance plans, said Roger Longman. In these situations, doctors and other health care workers are incentivized to understand the relative benefits of different therapies, including the economic ones. Doctors are no longer thinking only of therapeutic value; they are also thinking of economic value, said Longman.

Coverage choices are made in large part based on rebates: Pharmaceutical companies pay health plans (or whoever is buying the drug) a certain amount, based on usage. In the absence of dramatic, provable

differentiation in the clinic among medicines with the same indication, the importance of finances increases, said Longman. More expensive drugs must, in a cost-constrained economy, prove that they are worth more, noted Longman; otherwise, the cheaper drugs start to look more valuable.

Insurance companies describe the scope of their covered services with the phrase “medical necessity,” which relies on demonstrations of comparative effectiveness and generally accepted medical practice, said Robinson Beale. Medically necessary treatments are backed by evidence that is reproducible with fidelity, and the treatment is important for preserving life at a reasonable level of functionality. In other words, every physician who is delivering such services is getting similar outcomes, and the service is not more costly than an alternative service or drug that is equally effective. That kind of information is not easy to get, as clinical trials do not always use the same endpoints or outcome measurements, said Robinson Beale.

Health plans have technology reviews or pharmacy and therapeutic reviews that examine the research. They examine FDA decisions and all the evidence they can find that relates to the effectiveness of pharmaceutical and other technologies, with the goal of determining whether a particular treatment is clearly effective. They also look at explicit practice guidelines, which generally come from subspecialty organizations. Because a typical practice guideline costs more than $350,000, practice guidelines are generally outdated and do not include the latest research, said Robinson Beale. Insurance companies seek guidance from FDA about which patient populations benefit from new treatments, but such an approach does not take into account information gathered by practicing physicians, who commonly use drugs off label as they attempt to improve outcomes for their patients. In addition, Robinson Beale noted that clinical trials do not always compare the test treatment with a standard one, so when a new drug comes to market, insurance companies do not know how it compares with available treatments. She added that long-term and post-clinical trial information is sporadically available, which means that no reliable mechanism exists by which to understand long-term effects of drugs.

From Robinson Beale’s point of view, research evidence should clearly identify the specific affected populations based on neurocircuitry aberrancies or biomarkers, not just diagnostic classifications from the Diagnostic and Statistical Manual of Mental Disorders. Furthermore, insurance companies need to know what populations have not been test-

ed so they do not pay for an expensive drug for a one-person experiment. Robinson Beale added that she would like to see clear delineation of drug effects, duration of effects, and long-term outcomes, some of which might be established after market with a patient registry process for high-cost drugs or rush-to-market drugs. This way, the payer can continue to learn about how those drugs affect the population. Lastly, evidence reviews from FDA should be practice-guideline ready, she said, meaning they should be explicit enough that a doctor has clear directives about the conditions for which it has been approved and is effective.

OPPORTUNITIES TO INCENTIVIZE R&D BY STRENGTHENING REGULATORY PATHWAYS

• Amid the broader discussion, several participants also made a specific proposal to develop a regulatory pathway that aggressively hastens conditional approval of drugs that help patients with serious CNS diseases for which no therapies exist. Such a pathway would empower FDA to extend the existing accelerated approval pathway to CNS drugs that demonstrate substantial improvements early during clinical development based on biomarker or intermediate clinical data, with the understanding that such surrogates are not as available or strong as they are for other medical arenas in which this type of pathway has previously been deployed (HIV/AIDS and cancer). This pathway would:
  1. Harness the existing accelerated approval pathway to allow FDA to accept surrogate markers of efficacy for drug approval;
  2. Attach to the accelerated approval a set of appropriate, but stringent, restrictions on pricing and use, trying to strike a balance between those items and appropriate protection and financial return for industry sponsors.
  3. Once the drug is released onto the market, drug makers would conduct studies to confirm, refine, and/or adjust knowledge about its clinical utility; after this data collection step—using conventional clinical endpoints—the drug would receive consideration for full approval.
• Given the added risk to patients due to reduced time for accrual of safety and efficacy data before conditional approval, patient

This mechanism is intended to increase market protection for therapies that address medical problems that currently have no effective treatments, and its initial use would be targeted toward a few drugs that show unusual promise in terms of effect on a serious unmet medical need, underlying scientific support, and availability of an appropriate surrogate marker.

Woodcock raised several challenges about this proposed mechanism, stating that there are a number of legal and ethical considerations, including determining which patients should have access to the drug, restricting the distribution of the drug to the limited population, and facing the implications if the drug has to be rescinded (Choi, Engelberg, Maderis, O’Donovan,³ Reddy, Rogawski, and others).

• Promote adaptive trial designs, which could facilitate companies’ ability to focus on the most clinically relevant uses for a given drug or combination of drugs in an efficient and cost-effective manner (Rogawski and Vradenburg).
• Encourage FDA to issue guidelines that improve clarity about numerous aspects of the approval process, such as how drug sponsors can demonstrate eligibility for dormant therapy designation; improve access to FDA staff by allowing drug sponsors of dormant therapy drugs to request meetings that will inform the development of clinical plans that would support drug approval (Zorn).

_______________

³Mary O’Donovan, executive director of regulatory affairs at BioMarin Pharmaceutical Inc.

This page intentionally left blank.