5

Situating CAN Within the Drug Development Ecosystem

CAN is part of a much larger system engaged in or supporting the development of new cures. Throughout the workshop, participants referred to this system as an “ecosystem” in which each part is shaped by, and dependent on, other parts. Even though CAN is currently small, it can influence this much larger system if it is both catalytic and strategic. And if it is successful, it can be expected to grow.

The penultimate session of the workshop featured presentations and panel discussions among participants from a variety of organizations commenting on the role of CAN within the drug development ecosystem. The session was chaired by Margaret Anderson, FasterCures, who also moderated the first panel, which explored regulatory science priorities that are important for drug development. Myrl Weinberg, President, National Health Council, moderated the second panel, which considered the role of CAN in advancing cross-sector and other collaborative translational science activities. This chapter provides an integrated summary of the presentations and panel discussions, organizing the remarks by sector to offer a multifaceted perspective on CAN’s role in the drug development ecosystem and future.

FDA

FDA Regulatory Science

CAN’s authorizing legislation provides that one of its functions is to “facilitate review in the Food and Drug Administration for the high need cures funded by CAN” (see Appendix B). CAN is well positioned to support such collaboration, said Jesse Goodman, Chief Scientist, FDA. It could, as an explicit part of its work, address gaps in regulatory science that, if they were filled, could greatly improve product development. In the process, CAN could help develop a regulatory science toolbox that could create a more efficient pathway to develop and evaluate products. For example, Goodman noted, there is a need for end-to-end project management and support. Innovators often do not have experience managing projects or running businesses. He suggested that it could be very helpful if NIH and CAN could provide these kinds of services or connections.

In 2011, FDA issued a regulatory science plan that also addresses gaps and opportunities in the science and in product development (FDA, 2011). Though it is a high-level plan, according to Goodman, it also lists specific areas where targeted work could accelerate the development of needed products. The report lists several FDA priority areas that are relevant to CAN:

• Modernize toxicology to enhance safety.
• Stimulate innovation in clinical evaluation and personalized medicine.
• Support new approaches to improve product manufacturing and quality.
• Ensure readiness to evaluate emerging technologies.
• Harness diverse data through information sciences to improve health outcomes.
• Facilitate development of medical countermeasures to protect U.S. and global health and security.
• Strengthen social and behavioral science to help consumers and professionals make informed decisions.

Basic science will continue to be important to develop cures for diseases, said Goodman, but so will such steps as building precompetitive cross-cutting consortia, developing better evaluative tools and measures, and supporting relevant data-gathering and data-sharing initiatives. CAN could stimulate such opportunities through project design and evaluation criteria in areas such as toxicology or Alzheimer’s disease.

The Big Picture

Goodman also addressed some of the big picture issues that CAN faces. Can the program find a “sweet spot” where industry is not investing, but promising opportunities exist, thus catalyzing other interests? Perhaps such spots are characterized by higher risks, occur earlier in the development process, require new collaborations, or would benefit multiple diseases rather than a specific disease. Goodman encouraged thinking about how incentives could be created that might be missing in the commercial market to drive the timely development of needed products, including consideration of whether CAN could pick where success would have follow-on benefits in a much broader domain.

These and other projects supported by CAN require not just a scientific motivation but a business and management plan with timelines and deliverables, according to Goodman. For that reason, he proposed broadening the evaluation criteria and evaluators beyond the typical study section model. Scientific excellence must be ensured, but evaluators also could include clinicians, business people, and patients.

Management teams should have multidisciplinary representation, and projects should undergo periodic independent review by people who are not deeply invested in the project and have the authority to advise that a project be terminated or changed, Goodman said. “A lot of this is about focusing not just on what the grantees will do but how they plan to do it.”

The investment made in CAN could be uniquely valuable, Goodman concluded. New values, management approaches, and team organization, combined with new tools, methods, and public knowledge, could produce tremendous benefits. Goodman noted the importance of FDA’s commitment to working with CAN and its partners in the broader ecosystem.

Specific Proposals

Goodman suggested several specific potential projects for CAN related to regulatory science:

• Develop new approaches to clinical studies. An emphasis on relevant populations, comparator arms, and clinically meaningful outcomes, such as survival, quality of life, and patient-reported outcomes, could both help products get approved and educate physicians and patients about those products. A stronger clinical trials infrastructure and support for the development of relevant biomarkers also could enhance clinical trials, perhaps through systematic leveraging of the CTSAs.
• Data creation and sharing. Strong natural history data are often lacking today and could greatly help in the design of studies. Data from related products and studies could be leveraged. And some data generated in the process of product development could contribute to the field and help patients even if the product fails.
• Early, continuous engagement of industry with FDA. Product development pathways need to specify the indication and the anticipated risks. Scientific uncertainties can throw off a project for 2 to 3 years, and energy devoted up front to anticipating those uncertainties can be a valuable investment. Engaging with industry is resource intensive for FDA. Perhaps FDA and NIH staff or fellows could work together with grantees to extend FDA’s ability to do this work, Goodman said.

Drug Development Needs

ShaAvhrée Buckman-Garner, Director, Office of Translational Sciences, Center for Drug Evaluation and Research (CDER), FDA, addressed CDER’s efforts to support drug development. In 2011, CDER approved 30 new molecular entities, the highest total of new molecular entities approved in 7 years. Of those 30 new molecular entities, 12 were first in class, 11 were orphan drugs, and 19 were approved first in the United States.

The key issues have become more apparent. The development of evaluative tools is a tremendously neglected area. Better science is needed both to predict and to assess the safety and efficacy of investigational products. The major causes of failure in Phase 3 clinical development are a lack of effectiveness compared with a placebo or active control, unexpected drug toxicities, and commercial nonviability because a new therapy is no better than an existing therapy. While a large amount of biochemical and molecular knowledge may exist, there are few ways to assess the state of a whole organism or the impact of interventions at an organismal level. In addition, most assessment tools are not standardized, so the ability to compare one experiment against another is limited. The sources of variability in treatment response are largely unknown, even with current therapies. As a result, most clinical development programs are brute force empirical efforts that are extremely costly and time-consuming.

To predict, measure, and improve efficacy, major advances are needed, including new endpoints, new trial designs, better biomarkers to divide diseases into subsets according to prognostic or response predictors, patient-reported outcomes that have credibility, and natural history studies to understand disease course, particularly for rare diseases. These are “great concepts,” said Buckman-Garner, but no one organization has the job of developing these ideas.

CDER has been engaging in a wide variety of collaborative efforts to help translate these concepts into action. The goal is to pull together key stakeholders, whether in academia, industry, or government, with strong project management and specific goals to ascertain the key questions and potential approaches to resolve the problems.

Medical Devices

Elizabeth Mansfield, Director of Personalized Medicine, Office of In Vitro Diagnostic Device Evaluation and Safety, Center for Devices and Radiological Health (CDRH), FDA, spoke to the relevance of CAN for medical device development. Diagnostic devices, in particular, feed drug development, approval, and use, but this area often gets short shrift in terms of funding and attention. Many developers of these devices are small companies run by people with little regulatory experience.

CDRH is increasingly working with the centers for drugs and for biologics on companion diagnostics, and “we want to be able to feed that as fast as it can go.” Some drugs and biologics work well in a subset of the population, but without available diagnostics, there is no way of identifying this subset.

A Role for CAN

Given FDA’s interest in trial designs, drug development tools, and modeling, Buckman-Garner pointed to four specific areas where CAN could play a very helpful role: data standards, data sharing, training, and clinical trial networks.

• Data standards. FDA serves as a huge repository of data, but it is not in a standardized format. Greater uniformity in the representation of data, both in case report forms and disease-specific domains, higher levels of quality, increased interoperability, and more collaboration on international standards all require a much more focused effort.
• Data sharing. The consortia with which FDA is involved have been seeking to facilitate precompetitive sharing of data and reduce barriers to access relevant information. As an example, Buckman-Garner cited a collaborative database of tens of thousands of electrocardiograms. However, because FDA does not own the data, only a portion of the warehouse is open for research purposes, and sponsors have to be asked before using other data. “This is a big challenge [and] just one example of the challenges that we deal with.”
• Training. Creating an integrated workforce for translational science requires professionals skilled in clinical investigation, drug development, regulatory science, medical informatics and computer science, statistics, and other fields. Today, unmet needs exist in many of these areas, said Buckman-Garner.
• Development of robust clinical trial networks. Buckman-Garner noted that this effort requires the establishment of hubs for clinical trial networks that incorporate medical practitioners and also have the capacity for integration of sophisticated bench science.
• Devices. Mansfield added that CAN could support the development of devices by helping to build a fundamental understanding of how to translate a good idea into a product that is well understood. Many wonderful ideas emerging from NIH research, such as next-generation sequencing, will not automatically leap the chasm to become useful products unless more people know how to convert this knowledge into the needed tools. One challenge with devices is that their life cycles are extremely short compared to drugs. A device can evolve into the next-generation device within just a few years, making it difficult for FDA to predict the questions that will need to be answered. In this regard, CAN could act as a sentinel, said Mansfield, preparing the expertise and knowledge needed to expedite this process. CAN could help develop the knowledge of

To drive CAN’s investments, Buckman-Garner suggested identifying key decision points in the drug development process and prioritizing related knowledge gaps to determine

• the low-hanging fruit,
• areas of unmet needs that other consortia are not currently focusing on,
• ways to partner, and
• ways to leverage current efforts for further success.

Role of the CAN Board

Douglas Throckmorton, Deputy Director, Regulatory Programs, CDER, noted that CAN’s relationship with FDA will be pivotal given FDA’s prominence in CAN’s authorizing legislation. Furthermore, their goals are aligned, since both organizations emphasize the development of products that are efficiently used and effective in delivering benefits.

Throckmorton also remarked on the number of disease advocates that will be on the CAN Board—as many as one-third. This is “spectacularly the right thing to do,” he said. These are committed and sophisticated groups that have gone beyond getting money and providing grants to engage in careful management of ongoing efforts in an area. Patient advocacy groups straddle the public and private sectors in a productive way and, through translation or communication between these sectors, can make the CAN Board very effective.

The CAN Board can do prioritization and management, he said. It also can decide more specifically what the goals of CAN should be. It can foster communication among sectors, since each sector has less than a full understanding of the others. Finally, it can resolve misunderstandings among groups.

THE OFFICE OF SCIENCE AND TECHNOLOGY POLICY¹

Thomas Kalil, Deputy Director for Policy, Office of Science and Technology Policy, Executive Office of the President, said that, according to an analysis by Warburg Pincus, venture capital returns on investments in the life sciences are only around 1 percent. He commented that “limited

____________

¹ This section is based on the presentation by Thomas Kalil, Deputy Director for Policy, Office of Science and Technology Policy, Executive Office of the President.

partners are not going to be lined up around the block to invest in something that is generating a 1 percent return.”

DARPA has been a model for CAN, as demonstrated by the legislative direction to use OTA. But OTA is only a small piece of what makes DARPA successful, according to Kalil. Another major element is DARPA’s ability to attract world-class program managers who are the peers of the best researchers and innovators in the field. These managers are “willing to swing for the fences,” said Kalil, whether that means developing a Mach 20 aircraft or prosthetics that would allow a veteran to play the piano again. Such program managers focus on specific goals rather than simply funding projects with the highest-priority scores up to the pay line. They integrate across disciplines rather than following the lead of study sections organized around disciplinary lines. They actively manage programs and are willing to support technology development even if it is not hypothesis driven.

NCATS, CAN, and NIH as a whole should invest at least a fraction of their resources according to such a model, said Kalil. DoD invests about $12 billion in R&D, and of that, $3 billion is invested in the DARPA model. “I am not suggesting that NIH invest a quarter of its resources in the DARPA approach, but I think it should be some fraction where that approach makes sense,” Kalil said.

CAN also needs to invest in things that are not high-risk and high-return but that build a strong infrastructure. For example, figuring out an XML schema for case-reporting formats does not require a DARPA approach, but the argument can be made that it is important.

Kalil agreed that CAN needs to have some early wins to build congressional support. In particular, moving the needle on costs or success rates would be “the strongest argument for continuing investment.” However, CAN will need to grow from the current funding levels of $10 million before it could be expected to have a major impact.

THE PHARMACEUTICAL INDUSTRY

Garry Neil, Corporate Vice President, Science and Technology, Johnson & Johnson, addressed trouble in the pharmaceutical industry. With the cost of getting a single new drug to market approaching $4 billion by some estimates, industry is highly motivated to do what it can to boost productivity in the drug development ecosystem. It knows that many stakeholders are depending on industry and counting on companies to succeed not just in getting new products into the market but in saving lives, improving quality of life, and providing cost-effective access to drugs to the maximum number of people.

Neil had the following specific suggestions for what CAN could contribute to the drug development ecosystem:

• Help define and clarify clinical endpoints. In the past, researchers have largely tried to adapt clinical measures of diagnosis to treatment effect, and, not surprisingly, many of these measures lack sensitivity, specificity, or validity. In addition, an imperfect disease taxonomy is driven by the same clinical markers, the population is heterogeneous, and the clinical measures of diagnosis occur late in the process because the disease is driven by clinical manifestations.
• Contribute to target identification and validation. Good targets linked to well-established clinical data remain scarce, which limits progress.
• Outreach to patients and patient communities for participation in clinical trials. Only a few percent of cancer patients who are eligible enroll in clinical trials.
• Help advance regulatory science. Drugs need to get into the hands of physicians and patients faster while not exposing people to unnecessary risk. At the same time, potential risks and potential benefits need to be balanced. For example, it could be helpful to reframe the question of benefit—risk to ask: what are the risks of not treating an individual or a group that could benefit from a new therapy?
• Industry—FDA engagement. Could large meetings be replaced with one-on-one engagement? Could a help desk provide information with less strain on resources? If more meetings add value, then FDA needs more resources through extension of the Prescription Drug User Fee Act or some other means.
• Postmarketing surveillance. It is critical to be able to adequately monitor in real time what is going on after a drug is approved and is out into the marketplace. Physicians tend to use some products quite differently than expected once they are available. Various projects are under way to monitor how products are used, but bigger investments in these areas are needed.
• Precompetitive research. In many areas, industry can collaborate on precompetitive research because there is no competitive advantage in owning this knowledge.

A Master Plan for Cures Acceleration

Freda Lewis-Hall, Chief Medical Officer, Pfizer Inc., also commented on CAN’s role from an industry perspective. She suggested that CAN contribute to the development of a “master plan” for the drug development ecosystem that would establish the vision for the system, science, and tools that are needed. Through its master plan, CAN could systemati-

cally identify the highest-priority needs and barriers and then establish a strategy to tackle them in a coordinated way.

Lewis-Hall suggested thinking about four specific areas of opportunity that CAN could advance, as follows:

• CAN could serve as a project manager to conduct the various players, through encouragement, dissemination of best practices and worked examples, and education about opportunities to advance the field.
• CAN has the opportunity to facilitate harmonization of standards and best practices, for example, in the area of data management.
• CAN could help connect people who are working in the same area, by facilitating networking, matchmaking, communication, and sharing.
• CAN’s limited funds preclude its being able to tackle all of the problems confronting the drug development ecosystem. However, CAN has the opportunity to catalyze work through provision of seed funding and following on by, for example, helping to create a foundation or other group that would be resourced to extend the work.

PUBLIC—PRIVATE PARTNERSHIPS

Public—private partnerships need to serve patients, said Ellen Sigal, Chairperson and Founder, Friends of Cancer Research. Patients want treatments that work for them and that are safe and effective. A great advantage of CAN, she said, is that it is differentiated from the other entities in the ecosystem, which provides it with unique opportunities.

Entities that have complementary missions to CAN include FNIH, the Reagan-Udall Foundation for the FDA, and the Patient-Centered Outcomes Research Institute (PCORI). FNIH has evolved over the past decade and a half. It started with small important projects and now is doing very large important projects. Reagan-Udall shares the same mission as FNIH, but for FDA. Initial funding difficulties for the organization are being resolved. PCORI is the newest entity and has substantial funding, but it will only be successful, said Sigal, if it answers questions that are important to patients.

Sigal briefly described several lessons learned from her experience. Public—private partnerships work well when they answer important questions, but they also go to the same sources for support over and over. The pharmaceutical industry has been generous, but other resources are needed. Companies’ resources are limited, said Sigal, and “they have to work on the things that are most interesting to them.”

Public—private partnerships are nimble and can do things that government cannot. They are not constrained by personnel or contractual issues to the same extent. They need to retain this ability to do things quickly and get nontraditional groups together and not start to act more like government.

Public—private partnerships need to have a training component for new partners, Sigal observed. For example, training programs can help patient representatives be more effective and also have the effect of bringing in new people rather than using the same people repeatedly.

CAN could do things that others are not doing by working with existing foundations, partnerships, companies, agencies, and other parts of the drug development ecosystem.

The Critical Path Institute

The Critical Path Institute (C-Path) is a public—private partnership that works with FDA to accelerate the development and review of medical products. Over the past 6 years, it has built 6 global consortia with 41 biomedical companies and more than 1,000 scientists to create tools for drug development and advance regulatory science. But it has not been easy, according to Carolyn Compton, President and Chief Executive Officer, C-Path. The collaborations require extensive multidisciplinary teams of engineers, molecular biologists, health care providers, information technology experts, project managers, and many others. Among those who participate on multiple consortia, “consortium fatigue” can be a problem.

With its partners, C-Path has brought some of the first biomarkers through the qualification process with FDA. This process of qualification improves the conversation among industry, academia, and FDA and could be supported by CAN, Compton suggested. Like Goodman, Compton also suggested that CAN contribute to the development of standards in such areas as the qualification of biomarkers, patient-recorded outcomes, and data sharing.

Compton referenced Sematech, which brought together 14 semiconductor companies to do precompetitive research in response to the perceived threat to the U.S. semiconductor industry. The first thing Sematech did, she said, was to gather metrics from the participating companies to form standards. As applied to the case of drug development, standards could be formed with regulatory decisions in mind—for example, by requiring their use in the CTSAs. Standards also could improve the end product by combining data to create a better product. And because FDA would be part of the development process, industry could be confident that the data collected for regulatory submissions could meet FDA’s needs.

“I would submit that the real cool tools are the fundamental tools that support standards,” said Compton. “This is an opportunity to change the culture.”

The Biomarkers Consortium

The Biomarkers Consortium within FNIH identifies, develops, funds, and executes projects designed to develop and qualify biomarkers that can improve either drug development or clinical care. It is supported by the private sector, nonprofits, and NIH. It has launched 14 projects and has completed 3 of them. As an example of a recent success, David Wholley, Manager, The Biomarkers Consortium, FNIH, mentioned the establishment of new endpoints for clinical trials of community-acquired bacterial pneumonia and acute skin infections so that FDA can continue to approve new anti-infectives in these diseases.

Wholley suggested that The Biomarkers Consortium could be a model for NCATS and for CAN because of its diverse representation from all sectors of the drug development ecosystem. Steering committees and an executive committee design and then oversee the management of projects. Furthermore, all stakeholders are represented in all levels of this infrastructure. “We are able to bring the right expertise at the right time to the projects,” said Wholley. “We believe you can put the 15 to 20 smartest people in a given disease area in the room, provided you have this representation, and they can come up with good ideas.”

The projects take 6 to 18 months to develop and run for anywhere from 6 months to 5 years. The consortium raises funds from industry, and NIH can make parallel investments. The consortium can contribute funds directly to an NIH institute that then manages the project, but the majority of the portfolio is managed directly within FNIH.

Wholley also offered some lessons learned from his experience. He offered the opinion that Sematech is not a good model for this work. The semiconductor industry has a vertically integrated R&D system and is not a regulated industry. The challenges in drug development are “much deeper,” he said.

Interactions of scientists from industry, NIH, and FDA can produce cultural change. Even small interactions have a ripple effect and can make a difference.

Wholley also mentioned possible conflicts. NIH asked that The Biomarkers Consortium be precompetitive. CAN is allowed to do some things that are in the competitive product development space, but that will raise such issues as conflict of interest, antitrust, confidentiality, data access, publication, and intellectual property. Policies at NCATS and CAN will be needed to deal with these issues.

Another issue is global harmonization. Pharmaceutical companies are global entities. The regulatory environment includes not only FDA but agencies in other countries and regions, and working with these agencies introduces an additional complication in drug development.

Finally, given the tight limits on resources, CAN will need a clear focus, and all sectors will need to be represented from the beginning. He said there is a role for third parties to work with CAN, which could extend the funding available. Public—private partnerships could leverage not just the money but the resources from across sectors.

The Reagan-Udall Foundation for the FDA

Jane Reese-Coulbourne, Executive Director, Reagan-Udall Foundation for the FDA, described projects that the Reagan-Udall Foundation has up and running and others in the works. Examples include an evaluation of multidrug tuberculosis regimens, cardiotoxicity in cancer drugs, and the reformulation of pediatric drugs.

Public—private partnerships are essential, said Reese-Coulbourne. The question is not whether to do them but how. But the members of a partnership often speak different languages, whether because they represent different scientific disciplines or different parts of the ecosystem. They also have different reward systems, which affect such issues as the sharing of data. Cultural barriers do not necessarily need to be broken down but they do need to be worked through. One way to break down these barriers is through cross-sectoral training, she said. For example, the National Breast Cancer Coalition trains not just advocates slated to serve on panels, but also scientists to understand why the advocates are there, so that information flows in both directions. Also, scientists teach some of the classes to advocates, partly to convey an understanding of scientific terminology and also to discuss and model effective behaviors in advisory roles.

CAN is designed to produce revolutionary advances, Reese-Coulbourne noted. But many existing public—private partnerships rely on old systems, not revolutionary systems. They try to do faster and harder what people have already been doing, rather than looking at the system as a whole. CAN has the opportunity not only to solve specific problems but, also to address systemic issues that are at the root of problems. It can map out the system and its problems and figure out how the parts of the system can work together to meet the needs that exist.

TOOLS AND CURES

Panelists and workshop participants discussed the goals of CAN. Sudip Parikh, Battelle Memorial Institute, expressed his opinion that Con-

gress will not take the view that development of “cool tools” should be seen as a metric of success for CAN. It would be a success for NCATS, but for CAN, he said, “the success is cures, treatments, devices, therapies, behavioral interventions.” Goodman noted that he sees CAN’s responsibilities as including an examination of the process for product development so that CAN advances the field and not just an individual project, thus serving the public good. He noted that industry already has incentives in regard to development of cures, and it is in the scope of a federal program such as CAN to do more. Parikh added that there are areas where treatments are needed but are not being developed. Vertex Pharmaceuticals is an example of a company driven by patient advocates where there is a high need but not much interest in the pharmaceutical industry. Kristin Schneeman, Program Director, FasterCures, offered a potential semantic clarification. She suggested that CAN, as implied by its name, is accelerating cures, not necessarily producing them, which could be helpful in framing CAN’s metrics for success.

The discussion also touched on the role that CAN could play in helping partnerships of patient advocates, small companies, and academic researchers get high need treatments through the approval process. Goodman noted that CAN could require the developer of a product to have a plan for project management and engagement with FDA. NIH could partner with another group that could provide product developers with assistance in these areas.

Buckman-Garner added that, in this regard, CDER has been involved in an exploratory data submission program, which is designed to encourage early conversations about what is needed in product development. This program has led to meetings with groups from academia, industry, and NIH to have discussions about how to get through the regulatory process.

Neil said he could think of at least half a dozen cases of high need populations with no treatment alternatives where the problems are potentially tractable based on current scientific understanding. These cases may not be commercially attractive projects, but they offer excellent case studies of whether the drug development ecosystem could work in a different way to come up with an effective, safe, small-molecule treatment for a high need population.