5

Challenges and Opportunities in Regulatory Science

Over the course of the workshop, many participants discussed the new sources of information, technology, and techniques that are pushing the boundaries of regulatory science. Many of these new techniques, Stephen Ostroff noted, could be applied not just to help make regulatory decisions, but also to advance the scientific community in other expertise areas, including basic science and translational research. Ultimately, this progress will help make decisions that will serve to advance the health and well-being of the broader population. However, determining when, if, or how new techniques and data should be incorporated into the regulatory science process poses numerous challenges. Alastair Wood touched on a few of these challenges, including organizational acceptance of (or resistance to) a new paradigm, the continuing need to assess the performance of new and traditional data models, and the human capital and infrastructure needed to implement these new information sources. Other key challenges identified by multiple workshop participants involved questions around the data themselves, such as ownership and access or curation and its associated costs. Finally, some participants struggled to fit the tools and the challenges into the overarching principle of improving regulatory science—balancing innovation with rigor to enhance the efficiency, accuracy, and applicability of clinical trials without sacrificing safety to obtain the most efficacious and reliable treatments for disease.

Ultimately, for regulatory science to be improved most effectively, it must evaluate itself in the same manner other scientific disciplines do, said Martin Landray. Despite advances in technology and available tools, Martin Philbert cautioned that precision and scientific understanding can-

not replace good judgment. Many participants thought that regulatory scientists are most efficacious when they remain focused on the public health aspect of their work.

In the closing panel sessions, various workshop participants and speakers outlined the most relevant themes that they identified throughout the workshop and discussed priorities to advance the regulatory science agenda. Statements, recommendations, and opinions expressed in this section are those of individual presenters and participants and are not necessarily endorsed or verified by the Forum or the National Academies of Sciences, Engineering, and Medicine, and they should not be construed as reflecting any group consensus.

REGULATORY SCIENCE LANDSCAPE CONSIDERATIONS

Role of Public Engagement in Regulatory Science

Patients are more empowered and better informed than in the past; both the creation and use of data by patients could ultimately drive clinical or regulatory ecosystem behavior. One workshop participant noted that the regulatory science field should not underestimate the degree to which the most innovative users of data will increasingly be patients; in light of the current progress toward and investment in personalized medicine, there could be a corresponding drive toward patient-level approaches in regulating science applications. For example, a patient’s knowledge that he or she has a gene mutation might incentivize the patient to seek participation in a clinical trial.

Data Ownership, Control, and the Precompetitive Space

Many workshop participants also discussed issues of data ownership, patient privacy, and best practices for communicating clinical trial results and disease or treatment risk possibilities. Another challenge addressed by numerous participants was defining the “precompetitive space,” or the stages of product development in which competitors collaborate and share information. The quality of the data, as well as its curation, ownership, and control, are still debated and there are many problems and opportunities that could be addressed.

Summary of Individual Participant Remarks

Participants were encouraged to think about the major priorities that influence the regulatory science landscape, and how to incentivize adoption of changes. Their remarks are summarized below.

• What are the major priorities for advancing the regulatory science landscape?
  • Defining the precompetitive space so that results from early investigations can be shared and published without compromising intellectual property. (Abernethy, Amur, Philbert, Wagner, Wood)
  • What are the perceived problems in setting the boundaries of the precompetitive space? (Philbert)
  • Deciding who owns and controls the data. (Ostroff, Wood)
  • Deciding how the data are used. (Ostroff)
  • Posing and addressing the right questions will affect the field. (Landray)
• What investments and incentives are required to encourage consideration or adoption of these priorities?
  • Consider how regulatory science can be published in high-impact, peer-reviewed journals to increase visibility and reach. (Wood)
  • Obtaining funding for regulatory science research is an obstacle; neither academia nor FDA currently has the means to invest in the research in a comprehensive way. (Weichold)
  • Investment and incentives for undertaking the laborious process of data curation will be necessary to encourage it. (Participant)
  • Articulate, and reward, public health impacts of regulatory research. (Wilson)
• How can we bridge the gap between regulatory science knowledge and regulation and practice? What approaches could be considered to advance the discipline?
  • Principles for responsible data sharing are critical; it is unethical to not share clinical trial patient data with collaborators because that necessitates repeating the study for every related question that arises. (Wood)
  • Patients voluntarily contribute their own data for these studies and could considerably add to the data ownership debate.¹ (Participant)
  • Because pharmaceutical companies define their own risk thresholds according to their individual cultures, policies, and legal considerations, they would be key in setting data-sharing policies. (Lavezzari, Sauer)

___________________

¹ The Academies recently published a consensus report titled Sharing Clinical Trial Data: Maximizing Benefits, Minimizing Risk. More information can be found at: www.nationalacademies.org/hmd/Reports/2015/Sharing-Clinical-Trial-Data.aspx (accessed June 20, 2016).

CONSIDERATIONS FOR APPLICATIONS OF REGULATORY SCIENCE

Biomarkers

There is currently a “confluence of efforts” in developing and characterizing biomarkers, said John Wagner. Government, pharmaceutical companies, nonprofit groups, and consortia are all working to identify and develop biomarkers. One challenge will be to make sure the efforts complement each other and coalesce, rather than compete.

Clinical Trial Data Integration

The plethora of data now available allows for both changes in the way clinical trials are designed and the mechanisms by which treatments are studied postapproval. Ostroff summarized that it is a remarkable time period with regard to the conduct and assessment of clinical trials, particularly due to the computational tools becoming available. The biggest challenge will be in not compromising the ability to find the best, safest, and most effective answer as the methodologies evolve.

Next-Generation Surveillance

The ability to perform rigorous postmarketing surveillance and obtain reliable answers about medical product effectiveness and safety is not new, said Brian Strom. The novel aspect is the size of the databases and the amount of information available for analyses. Given the increasing number of databases and the potential to aggregate them, the shortage is no longer in hypothesis generation, but in exploring hypotheses with rigor. The tools being developed in surveillance may better be considered as ways to generate hypotheses, not as ways of identifying causality.

Modeling

Modeling, like other aspects of regulatory science, is undergoing immense change in response to the availability of big data. Innovative trial design, master trial protocols, Bayesian methods, and using consortia to understand disease are all dependent on the collection and use of data for the creation of knowledge and novel applications, said Darrell Abernethy. A significant opportunity for regulatory science to evolve is now centered on strong pharmacologic and biologic mechanistic approaches to inform population science.

Summary of Individual Participant Remarks

Participants described many obstacles facing the various applications of regulatory science, from considerations for biomarkers to postapproval surveillance. They also discussed the types of investments that could help realize any improvements in regulatory science, and current gaps between knowledge and practice. Their remarks are summarized below.

• What are the major priorities for furthering innovative applications of regulatory science?
  • Lack of predictability, a defined evidentiary framework, and a timeline for regulatory decisions for biomarker development. (Wagner)
  • The quality of information available and the ability to curate it, as well as rigorously determined acceptable thresholds for variability. (Ostroff)
  • Determining the correct balance of innovation and rigor for surveillance techniques and the data they can generate. (Strom)
• What investments and incentives are required to encourage consideration or adoption of these priorities?
  • Funding, in terms of both time and monetary investment, to allow development of biomarkers. (Several participants)
  • Developing ways to recognize researchers who share specialized software and techniques may encourage more information exchange in the field. (Myers)
  • Identifying current systemic discouragements that lead to a preference for pseudo-efficacy trials in Phase II that may be inappropriate and preclude movement to Phase III, contributing to the high rate of failure to progress to higher trial phases. (Angus, FitzGerald)
  • Removal of barriers to communication of information among sectors (e.g., sharing results with patients in easily understandable terms). (Krall)
• How can we bridge the gap between regulatory science knowledge and regulation and practice? What approaches could be considered to advance the discipline?
  • Innovative trial design is the key to good regulatory science, and encouraging those techniques in addition to focusing on biomarkers may bring personalized medicine to realization. (Krall)
  • Properly defining and especially validating biomarkers will be critical in order to use them for trial design without leading to false conclusions or erroneous results. (Wood)

• A consortium approach to developing biomarkers may address key gaps in both development and defining evidentiary standards. (Wagner)
• Cell-based therapies are likely to be the future of medicine, and planning now for how to move basic discovery into robust regulatory systems will be critical. (Greenberg)
• Apparently biased large datasets could be curated to reveal subsets of data that are suitable for scientific discovery; however, this must be considered as a hypothesis-generating rather than a definitive exercise. (Landray)
• Consider the potential applications for artificial intelligence in the future of medicine, given that medicine will become ever more data intensive. (Burch)
• Increased focus on determining personal effects from the application of population effects. (Krall)

INFRASTRUCTURE AND WORKFORCE CONSIDERATIONS

Building a Regulatory Science Workforce

Attracting proficient scientists and recruiting the right expertise to bolster the field’s regulatory foundation can be a challenge. Owen Fields expressed concern that the current career path for regulators can all too often be undertaken as a result of a random or unintentional decision. Fields observed that regulatory science is not an encouraged pursuit and by some it is even considered a “dead-end” career. Frank Weichold encouraged participants to consider the needs of existing regulators, especially as new technologies arise. It would be advantageous for current regulatory scientists to have access to ongoing training to be able to use and apply these new tools and stay current in the field, he said. Additionally, several participants thought that encouraging the broader workforce now considered to be regulatory scientists to collaborate synergistically with each other is as critical as it is difficult.

Summary of Individual Participant Remarks

Considerations for building, training, and rewarding an enlightened regulatory workforce generated much discourse and debate among workshop participants. Their remarks are summarized below.

• What are the major priorities for furthering the development of an innovative regulatory science workforce?

• Attracting the best and the brightest to the biomedical sciences in general, and to regulatory science in particular. (Abernethy)
• Developing a new system to evaluate regulatory scientists on their own terms. (Altman, Landray, Philbert, Ostroff, Wood, and others)
• Emphasizing that the regulatory portion of the title implies a social obligation and encouraging development of multidisciplinary regulatory scientists. (Rogers and others)
• Clearly defining expectations and responsibilities for scientists individually and within a group, with an emphasis on leveraging complementary skill sets and encouraging collaboration. (Abernethy)

• What investments and incentives are required to encourage consideration or adoption of these priorities?
  • Regulatory science cannot be evaluated fairly using traditional academic metrics, and a reward strategy tailored for the field will be key. (Several participants)
  • Retaining and rewarding data scientists and data curators will become essential, given the juxtaposition of population-based information and personal information now coalescing in the available datasets and medical knowledge base. (Landray)
  • Currently, many regulatory scientists are not recognized or rewarded, and more support for those actively engaged in translational pursuits is necessary. (Potter)

• How can we bridge the gap between regulatory science knowledge and regulation and practice? What approaches could be considered to advance the discipline?
  • Use consumer industries as a model for incentivization, with metrics accounting for both quality of work and time to adoption. (Myers)
  • Articulating the broad public health impacts of the work may help draw scientists into the field. (Wilson)
  • Support rotations within industry and academia. (Cannon)
  • Mechanisms for senior academic scientists to support FDA in pharmaceutical research. (Cannon)
  • A broader consideration for partnerships in general, with less emphasis on the differences among government, industry, and academia, combined with aligning the incentives for collaborating parties. (Shekar)
  • First priority should be exceptional scientific training, then teaching scientists to recognize problems with high potential for regulatory impact. (Altman)
  • FDA needs a clear hiring process. (Abernethy)

This page intentionally left blank.