Meijing Wu and Di Zhang
Drug development must adhere to a meticulously regulated environment in which regulatory guidance and policies are crucial for providing structured and/or standardized approaches. There is a growing trend of collaboration among industry, academia, and regulators to implement and develop regulatory guidance and policies that promote innovative design and analysis methods.
Here, Meijing Wu and Di Zhang—editors of the Biopharmaceutical Report—interview Ivan Chan from BMS and Bill Wang from Merck about the impact of regulatory guidance/policies in driving statistical innovation.
In your experience, do you observe any discrepancies between the recommendations outlined in current regulatory guidance and the actual practices within the industry? If so, would you provide specific examples?
Ivan: Much progress has been made to harmonize the regulatory policies and implementation around the world for drug development. However, we still sometimes encounter disparities between the regulatory guidance and actual practices. Managing these discrepancies can be complex and challenging, particularly when regulatory bodies from different countries have different requirements to meet their local needs. For example, for designing and analyzing time-to-event endpoints, the US Food and Drug Administration and European Medicines Agency have different recommendations for censoring rules, leading to variations in our approach.
Bill: Indeed, there has been broad collaboration in regulatory innovation in statistics. An illustrative example is the ICH E17 [general principles for planning and design of multiregional clinical trials] guidance, which establishes clear principles for sample size allocation in a multiregional trial. Despite consensus on these principles, the implementation varies by country and region due to local laws and regulations that may not align with ICH guidance.
Another scenario involves regional regulations providing their own guidance, making it difficult to bring regulators together, even when the industry attempts to create unified ICH guidance. The FDA’s IND [investigational new drug] safety reporting guidance serves as an example, where efforts to reduce reporting burden contradict with other agencies that insist on comprehensive data. This highlights the industry’s struggle to align different regulators for mutual benefit, leading to divergent paths in regulatory approaches.
What are the challenges in promoting statistical innovation within the constraints of regulatory guidance?
Bill: There have been many noteworthy innovations proposed by various entities, being academia, industry, or regulatory agencies. These innovations sometimes find their way into regulatory practices and evolve into official guidance.
However, there are some obstacles. For example, many innovations—especially in statistical methods—often come with theoretical assumptions, whether model-based or not. These assumptions, when confronted with real-world scenarios, can be challenging to accommodate comprehensively. Consequently, guidance needs to strike a balance, being general enough to avoid specifying on every possible situation yet specific enough for practical implementation. This creates a potential implementation challenge for the guidance.
Ivan: We’re also dealing with regulatory requirements from different countries and various health agencies. Usually, sponsors aim to adopt the most conservative approach that would be accepted by all agencies because the study will be run in multiple countries and data from the study will be used for global submissions. This conservative approach, however, tends to inhibit statistical innovation. Regulatory departments within companies are often very conservative. To ensure protocols pass every review body, this conservative practice may potentially impede innovation before reaching agencies.
Educating regulatory liaisons in-house to be more willing to take risk is crucial, providing teams with the opportunity to engage with different agencies and experiment with novel approaches to trial design and analytical methods.
On the regulatory side, harmonizing policies across different regions and countries is essential for promoting innovation. A great example is the rollout of the estimand guidance, where every regulatory body aimed to adhere to common principles, showcasing the positive impact of unified approaches across regions.
Do you have any experience/examples of successfully influencing a regulatory guidance?
Ivan: There is a good example on adaptive designs I was fortunate to be involved in. A working group was initiated by a group of quantitative experts from industry in 2005, at a time when not many people besides the interested group would consider the potential for applying adaptive design to phase three trials. The industry working group developed a position paper on adaptive designs and began collaborations with academics and regulatory agencies, primarily the FDA. This effort led to the draft guidance in 2010 developed by the FDA.
Over the years, new methodologies and examples emerged from both academic and industry settings, culminating in the finalized guidance in 2019. This extensive process involved conversations, meetings, and workshops, creating a platform for the exchange of ideas and innovation, which laid the foundation for developing the guidance.
The ICH guidance—such as the E9(R1) [Statistical Principles for Clinical Trials: Addendum: Estimands and Sensitivity Analysis in Clinical Trials] and E17 documents—were recently finalized, as well, with formal industry participants contributing to its development. This model of industry involvement in guidance development is seen as a positive step forward.
Looking ahead, there’s an opportunity for FDA to consider involving industry participants in the guidance development process, similar to the ICH model. This, I believe, would be a significant step forward in the development of guidance.
Bill: In the realm of statistics, we have numerous examples where statistics play a pivotal role in shaping regulatory guidelines. I can share two personal examples—one involving myself and the other related to Ivan.
The first example is Ivan’s work in fundamental research on statistical methodology for vaccine trials. In vaccine trials, where outcomes are typically rare, Ivan and another former Merck colleague proposed an innovative method known as the “exact method” to design and analyze these rare endpoints. Ivan’s method was specifically quoted in a recent draft NMPA [National Medical Products Administration] guidance on vaccine clinical trials.
Another personal example involves my participation in the working group that drafted ICH E17. I was one of the few statisticians among almost a dozen cross-disciplinary professionals representing regulatory agencies and industry associations. In this case, we encountered an issue that, if each region required a specific sample size, it would make the multiregional clinical trial become unmanageably large. The challenge was to come up [with] an approach that could achieve the global study objectives while still being implementable.
In this context, a few statisticians, including me, contemplated the idea of using pooling strategy, which is a statistical approach often applied but not necessarily as a major strategy in global collaborative trial design efforts. The idea received positive feedback and evolved into one of the seven principles finalized in ICH E17.
This example illustrates how a statistical concept, originally just an analytical idea, can transform into a major strategic component in regulatory guidance.
How would you assess the current efforts made by health authorities in fostering collaboration among themselves, industry, and academia for the development of regulatory guidance? Are there any notable initiatives or strategies you find particularly effective or lacking?
Ivan: Bill’s involvement in the ICH effort extends beyond the scope of a particular country. It aims to harmonize policies across different countries and regions, involving industry and developing guidance with participation from both industry and academia. The process invites not only industry representatives but also academic participants, creating an ideal forum for developing regulatory guidance. This approach fosters collaboration across the three different areas of expertise.
In the US, FDA has been collaborating with academic centers, and there may be opportunities to extend this collaboration to statistical experts from the industry.
Engaging industry partners in the development of guidance can be beneficial, as they bring real-world examples, understand the challenges, and can help identify areas for innovation. This collaborative approach aligns well with the needs of the industry and can enhance the quality of regulatory guidance.
There is a good example of FDA involving the industry in initiatives like the Complex Innovative Trial Designs initiative. The FDA initiated several workshops where they invited academic experts and industry representatives. Participants presented ideas and commented on what they were thinking. Personally, I was fortunate to be involved from the pharma side, invited by the FDA to discuss industry examples, potential gaps and challenges, and how to implement complex innovative designs.
These workshops foster open conversations between industry and regulators, addressing challenges and sharing experiences.
Bill: I can provide an example of a notable initiative that has proven to be effective. In the US, there is a public-private partnership initiated by the Center for Drug Evaluation and Research and Center for Biologics Evaluation and Research. The partnership promotes collaboration between industry and regulators.
Three years ago, I co-led the ASA safety working group and initiated a private-public partnership with the FDA through a rigorous application process. Each task force under the working group has designated liaisons from different divisions of the FDA. This strategy of forming partnerships across industry, regulatory bodies, and academia has proven to be a successful and effective approach.
What actions do you believe health authorities can take to enhance collaboration among health authorities, industry, and academia, specifically with the aim of driving statistical innovation in regulatory practices?
Bill: Yes, there’s an example of the tripartite effort in China, which is driven by regulators but closely involves partners from academia and industry experts. The goal is not only to develop innovative methodologies but also to apply them to regulatory practices.
On the global scale, the ICH has involved industry and regulators. There are also academic professors from Europe who are part of the ICH working groups. These have enhanced collaboration among health authorities, industry, and academia, specifically with the aim of driving statistical innovation in regulatory practices.
Ivan: The private-public partnership is an excellent example, especially when working groups come together. I’m thinking that in addition to addressing issues and fostering innovation, we could take a step further in empowering different stakeholders during the process of writing the guidance. In the ICH, they usually have academic and industry experts participating in writing position papers and guidance documents. China is attempting to follow the ICH model, which I believe is a good approach to emulate.
In your opinion, what role can the industry play in strengthening collaboration in the development of regulatory guidance? Are there specific measures or initiatives you think would be beneficial for enhancing industry participation in this collaborative process?
Bill: I can identify a few aspects where industry can contribute. Pharmaceutical companies can leverage our implementation experiences to think about standardization and be part of the topic proposal for developing regulatory guidance. Collaborating with regulators, industry experts can identify areas that require practical innovations.
Another area where industry can improve is sharing best practice and clinical trial examples. Sharing statistical methods isn’t an issue; the difficulty increases when it comes to sharing specific data or detailed implementation, but it’s still possible.
In the complex innovative designs case I participated in, several companies shared examples with the FDA. Around four or five companies shared their experiences on different aspects of complex innovative designs, highlighting how methods were implemented in studies and the challenges faced. It’s possible to share valuable insights without divulging proprietary information.
Ivan: Industry statisticians, with a wealth of trial experiences, understand the practical needs and challenges. They play a pivotal role in developing methodologies tailored to specific studies. By sharing results and experiences across the scientific community, industry experts contribute to a collective knowledge base. This collaboration enhances the voice of industry in shaping regulatory guidance, promoting standardization, and fostering innovation.
Editor’s Note: The opinions provided here are the authors’ and may not represent those of their employers. This article originally appeared in the spring Biopharm Report.
Leave a Reply