To set the stage for discussion of next steps for regulatory science at FDA, workshop co-chair John Rex, of AstraZeneca, outlined the key crosscutting themes that emerged over the course of the meeting. Throughout the workshop, stakeholders and participants from all sectors emphasized that collaboration is essential for advancing the development and evaluation of MCMs. No single partner has all of the tools. Participants discussed how, in the current research environment, “you get what you reward,” and the MCM enterprise needs to find ways to reward flexibility and innovative thinking. Education and training were also highlighted as being essential to advancing MCM development (including scientific training, leadership development, and education of the public). Participants discussed the practical and ethical limitations of conducting clinical trials for MCMs (in both the general population and in at-risk groups such as children2and pregnant women), and the available alternative regulatory mechanisms to demonstrate efficacy. In this regard, there was much discussion of the Animal Rule, with a particular focus on the challenges of validation of animal models and establishing true correlates of efficacy.
Rex summed up workshop discussions regarding metrics of success
1 This subsection is based on the summary remarks provided by workshop co-chair John Rex of AstraZeneca.
2 Nelson of FDA advised participants that the agency is planning a workshop focusing on the ethical issues of pediatric MCM development for the first quarter of 2012.
in MCM development by noting that definition of metrics is challenging. He commented that there is limited opportunity to assess the true public health benefit of an MCM; success cannot simply be measured by the number of MCM approvals (as not all products will or should succeed). Participants offered a variety of suggestions for metrics, from a goal of adding a defined number of new, approved MCMs to the SNS within a defined time period; to smaller, incremental steps such as developing an assay that solves a key problem, thereby reducing time and/or cost of development;to finalizing MCM-related FDA guidance documents; to approving a product under the Animal Rule. It was repeated throughout the workshop that providing a clear regulatory pathway forward can foster innovation and enhance the quality of sponsor submissions;this, it is hoped, would lead to increased speed of review and success of applications.
From a defense perspective, U.S. troops face threats around the globe, not just traditional biothreats, but endemic diseases as well. It is important to remember that MCM development must include not only vaccines and therapeutics, but also point-of-care diagnostics (for both organism identification and drug resistance profile). An ongoing challenge for both civilian and military populations is getting “the right product, in the right place, at the right time, for the right individuals.” In this regard there was discussion of pre-positioning tools (such as diagnostics or mobile manufacturing capability).
Rex added that in the end, for any countermeasures to be effective, the public must accept and use them. In this regard, there is a need to educate the public about advances in regulatory science (e.g., approval and use of products that have not been tested in humans, benefit versus risk during an emergency versus routine medical care). There was interest in leveraging social networks and developing educational apps.
Carl Peck of the University of California, San Francisco, emphasized that in all processes, there needs to be a change in mindset or a “reset” regarding benefit-risk criteria. Benefit-risk assessments must take into account the fact that MCMs are intended for use in extreme public health emergencies (not for treatment of, for example, chronic, nonfatal conditions). Hatchett added that the development of rapid diagnostics could help facilitate the reset of benefit-risk assessments, allowing FDA to better define for whom the use of a product would outweigh potential risk (e.g., only to be used for those who test positive).
Participants discussed whether this reset of benefit-risk assessments would mean setting a level of safety that is not necessarily the same as the
|Process Gaps||Regulatory Science Opportunities Cited by Panelists|
|Acquisition||Predictive in vitro systems (e.g., MIMIC, “liver on a chip”) Bayesian study designs|
|In silico systems biology and computational biology models Reset of benefit-risk criteria to be relevant to an immediate threat situation|
|Integration and Prediction||Bayesian, model-based integration framework
Meta-modeling: integrating all data in PK/PD and/or PBPK/PD models
Linking of systems biology and PBPK/PD models
Mimicking adult-pediatric PK/PD dosage paradigm for animal-human prediction Reset of benefit-risk criteria to be relevant to an immediate threat situation
level of safety demanded of a drug destined for the commercial market. It was noted there is a lot of precedence for benefit-risk decisions at FDA, and the agency has a strong record of making good benefit-risk decisions. The challenge for MCMs is that they are being developed for potential use in the future, for events that have not happened.
Ed Cox of CDER noted the differences between benefit-risk assessment for prophylaxis versus treatments. For a compound to be used for prophylaxis, the benefit-risk benefit calculus is complicated by the fact that while a significant number of potentially exposed people will receive the product, only a small portion may actually be at risk for the disease.
What is lacking, Peck said, is a set of processes for rapid, efficient acquisition and integration of all in vitro, animal, and human data (mechanism of action and PK/PD) that would permit prediction of a reasonably likely favorable clinical benefit-risk ratio in humans. Many of the cutting-edge technologies and methodologies discussed throughout the workshop could be leveraged to help close these gaps so that MCMs could be successfully approved under the Animal Rule (Table 5-1).
Alan Shaw of Vaxinnate noted that FDA generally approves products or therapies on a case-by-case basis; however, throughout the workshop there was a lot of interest in platforms. FDA does not approve platforms,
but tools used may be submitted for qualification.3 The qualification process for a drug development tool is product independent. The intent is to evaluate these tools and make the data publicly available so others can use them in their development process without the need for validation every time the tool is used in association with a new product. The focus has thus far been biomarkers, patient-reported outcomes, and clinical quantitative disease progression models, but FDA is working to include animal models as well.
Biomarkers qualified by FDA for a specific context of use can then be used within this specific context use by multiple companies for multiple products. If, for example, a set of biomarkers were qualified for use in a rat model, one could now look at those in clinical studies and then, as data become available, submit another qualification package to FDA. This concept has been called “rolling qualification.” A qualified tool could be considered a modular element of a platform, Rex suggested.
There was much interest in whether there could be a formalized platform qualification process. A participant explained that while the agency only approves products and not processes, a platform approach can help streamline approval of future products. For example, if there is an approved vaccine based on a vector into which appropriate genetic material was inserted, a subsequent vaccine made by inserting different genetic material into the same vector would still need to demonstrate safety and efficacy and validate manufacturing processes, but the process would presumably be faster, as much of the previous work would be relevant. The manufacturing process of the platform is, in essence, qualified (DARPA has used the term certified). One does not have to develop a whole new validation package. For example, the egg platform used in the manufacture of influenza vaccine is essentially a qualified manufacturing platform. As such, a new flu vaccine can be approved and manufactured within a 6-month time frame. Rex added that an adjuvant would be another example of an element or tool for which gathered data can be relevant to subsequent submissions. A participant suggested that the qualification of an animal model as being suitable to submit efficacy data could possibly be considered a platform qualification.
With regard to biomarkers, Richard Hatchett of BARDA noted that in oncology, the co-development of therapeutics and biomarkers is becoming the norm, particularly for trials of targeted therapeutics where the clinical trial population needs to be preselected based on the targeted pathway.
3 See Guidance for Industry: Qualification Process for Drug Development Tools (Draft Guidance) http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM230597.pdf (accessed June 9, 2011).
An underlying theme in almost all regulatory science aspects of the MCM enterprise is communication (e.g., between sponsors and regulators, between funders and regulators, among stakeholders and collaborators, or from the enterprise to the public). Throughout the workshop participants were encouraged to communicate with FDA “early and often.”Phyllis Arthur, Director of Healthcare Regulatory Affairs of the Biotechnology Industry Organization, emphasized the importance of communication between sponsors and FDA to establish a series of agreed-upon goals, with potentially a set of agreed-upon metrics, and accountability for all of the parties to actually achieve those goals. Ed Nuzum of NIAID said that companies should not be afraid to meet with FDA, but when they do, they need to be organized, present their data packages, and have well-formulated questions. Arthur emphasized that companies already deep in the process would benefit more from some agreed-upon accountability on both sides and more clarity and transparency as to what needs to happen along the pathway.
A challenge for small companies that are funded by BARDA is that BARDA encourages its grantees to meet with FDA to discuss moving forward with the next step, often when the company does not feel it has the data or the time necessary to prepare a meeting package. To keep to BARDA-established timelines, manufacturers are often willing to accept a certain level of risk and move forward before there is a complete dataset that could be discussed with FDA. It would be helpful, the participant said, if there could be some agreement or better clarity of the roles of the two organizations (BARDA as funder and FDA as regulator).
One approach to advancing the use of new testing methods or tools in regulatory science decision making is through precompetitive collaborative consortia involving scientists from industry, academia, and government, as well as regulators and patient representatives;a number of meeting participants expressed interest in developing such precompetitive collaborations. One example that was cited by a number of workshop participants as successful is C-Path, which, explained Marietta Anthony of C-Path, advances the development of new testing methods or tools for medical product development through establishing precompetitive collaborative consortia involving over 1,000 scientists from industry, academia, regulatory agencies, government (NIH and CDC), as well as patient representatives. Anthony explained that C-Path is a neutral, third-party entity that is able to forge partnerships and facilitate consensus development on precompetitive science between industry and regulatory authorities. Consortia activities are not product specific. Over 35 member companies in five consortia have signed a legal agreement that addresses issues of confidentiality, intellectual property, materials transfer, and anti-
trust. Tools/methods developed are made publicly available and are used by industry in development of medical products and by FDA in regulatory decision making. As an example, Anthony described the PSTC. The consortium includes several working groups that assess data on candidate safety biomarkers for various organs and tissues. The consortium is not focused on discovery of new biomarkers or sponsoring new research, Anthony emphasized, but on critical evaluation of existing data, conducting additional studies to fill the gaps, enhance the database, and facilitate scientific consensus. Promising biomarkers selected by the consortium are then submitted to FDA4 and other international regulatory authorities for qualification within a specific context of use.5
Hatchett noted that, in addition to C-Path, the Foundation for the National Institutes of Health (FNIH) is also a successful model for collaboration between FDA, NIH, academia, and industry.
Many participants noted that data sharing across FDA centers should be increased. Nuzum suggested that government contracts and funding agreements with MCM developers should include provisions for sharing of grantee data across government agencies. The ability to share preclinical and clinical data across agencies, de-identified and pooled as appropriate, could facilitate needed meta-analyses and should be accompanied by assurances to MCM developers that their proprietary data will not be released or used to benefit a competitor. It was noted that funding initiatives from the various MCM enterprise partners (e.g., the DoD, NIH, other HHS operating divisions) may have regulatory science components, and there should be an effort to incorporate FDA input into the initial requests for applications or proposals, especially those efforts that are targeting product development.
Gail Cassell of Harvard Medical School and the Infectious Disease Research Institute in Seattle recommended that FDA forge closer ties with NIAID-funded Regional Centers of Excellence in Emerging Infections and Biodefense for access to local expertise that might be brought to bear in a public health emergency.
4 See Guidance for Industry, Qualification Process for Drug Development Tools (Draft Guidance) http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM230597.pdf (accessed June 9, 2011).
5 Additional information is available in a special supplement to Nature Biotechnology,The Predictive Safety Testing Consortium 28(5):May 2010.
Although both the National Biodefense Science Board (NBSB) and the IOM have previously issued strong recommendations for increased resources for FDA, in the current fiscal climate the significant increases that are needed are not likely to occur.
Resources are specifically needed to support the following regulatory science-related activities:
- The process of evaluating an increasing number of new technologies, and research to resolve new regulatory science challenges.
- Leadership development (scientific and professional) to strengthen science within the agency.
- Recruitment and retention of scientific talent and technical expertise.
- Establishing collaborations and partnerships.
- Academic Centers of Excellence in regulatory science to promote a better understanding of the development of MCMs and provide training.
- Information sciences.
- Synthetic biology.
Cassell noted that many of the challenges for FDA regulatory science have roots in the agency’s infrastructure issues. Cassell highlighted some of the key issues that were identified in the report FDA Science and Mission at Risk (FDA, 2007) as well as were mentioned throughout the workshop (Box 5-1).
Nuzum summarized the sentiment of the day, saying that FDA is doing great science and agency staff are conscientious and competent, but the tasks before them are daunting and their resources are limited and are not likely to significantly increase. As such, new paradigms are needed to find ways to work within the resources that are available.
Hatchett suggested that the regulatory science needs discussed at the workshop can be sorted into tactical, operational, and strategic concerns, that is, how to deal with data in its acquisition, sharing, or management (Box 5-2).
Tactical Level—Getting the Data
- Developing animal models for specific applications.
- Defining appropriate models or methods to collect data (including for at-risk populations such as children, pregnant women, and others for whom absorption, distribution, metabolism, or excretion may be altered):
- platform approaches,
- qualification of tools (e.g., biomarkers), and
Operational Level—Sharing the Data
- Cooperation, collaboration, partnerships, and sharing of data among stakeholders.
- Internal data sharing and collaboration across FDA centers.
Strategic Level—Managing the Data
- Maintaining competencies as new areas of science unfold (e.g., systems biology, computational biology, biostatistics).
- Benefit-risk calculus, communicating risk data.