The Devil is in the Details: Public Hearing on FDA Guidance on Patient-Focused Drug Development
The FDA demonstrated its strong commitment to patient-focused drug development (PFDD) with the public workshop it convened on December 18, 2017. FDA officials, industry representatives, researchers, patient advocacy groups, and other stakeholders participated in this gathering at the agency’s White Oak campus in Maryland. The attendees all offered insight on the overarching question at hand: Who do you get input from, and why? How do you collect the information?
This FDA workshop is the first part of a broader effort to develop better strategies to collect patient experience data and use such information in drug development. This meeting specifically focused on getting public feedback on the first draft of one of the four guidances the FDA has prepared on this topic. These various efforts are in accordance with Title III, Section 3002 of the 21st Century Cures Act. To read the initial draft of guidance one and to get more information on the other PFDD FDA guidances, click here.
We were impressed by the diverse representation and level of engagement at this meeting. Both suggest a solid foundation for future progress on the next three guidances, so we look forward to what the FDA produces in the next few months!
The meeting was broken up into four sessions. Two of them began with presentations from an FDA official, followed by a panel discussion. See below for a detailed summary of each session:
- Session I: Defining Research Objectives & Methodological Considerations for Designing Studies to Collect Patient Experience Data
- Session II: Methodological Considerations for Data Collection, Analysis, and Operationalization
- Session III: Translating Best Practice Into Real Practice – Developing Guiding Examples
- Session IV: Identifying Key Themes and Next Steps
Session I: Defining Research Objectives & Methodological Considerations for Designing Studies to Collect Patient Experience Data
Presentation: Ebony Dashiell-Aje, COA Staff, OND, CDER and Kunthel By, Division of Biometrics V, OB, OTS, CDER, FDA
This session looked at some of the fundamental questions researchers should ask and points they ought to consider when designing a patient study. Many are included in the diagram above, while other recommended considerations include number of subgroups and reporter characteristics. Ms. Dashiell-Aje noted that the FDA guidance does not provide instructions on how to calculate sample size, as that depends on various factors that the FDA cannot account for in one general document. Also, while the FDA encourages study designs that reflect the heterogeneity of the target population, there is no mandate or guidance on how to ensure appropriate representation.
A panel discussion with Kunthel By, Ebony Dashiell-Aje, Steve Cohen (RTI International), Richard Gershon (Northwestern University), Meena Khare (National Center for Health Statistics, CDC), Elisabeth Piault-Louis (Genentech), and Suzanne Vernon (Bateman Horne Center) considered the following key questions:
- Are there are any other factors to consider when defining research objectives and designing studies that should be included in the guidance?
- What are other factors and/or approaches to consider to ensure collection of representative input from the target population (patients with disease of interest)?
- In which situations is it more important to sample patients using a probability-based method? In which situations is it less important? What will be gained and what may be lost?
Panelists weighed in on additional factors to consider like non-response bias, weak health literacy (as an inclusion factor, not an exclusion factor), co-morbidities, and concept saturation – in other words, the question of whether doing more interviews will actually provide more information. One largely unanswered question is when companies should actually submit their data to the FDA. Often, companies are conservative in their reporting, as they aren’t sure how the FDA will use the information or whether it will end up on the label. An FDA spokesperson responded that any data that could be used to support a decision should be submitted even if it doesn’t go on the label.
Alternative research design and methods were also discussed:
- Mr. Cohen recommended considering use of existing databases such as large national surveys. This would avoid re-doing work that has already been done when targeting a wide population.
- Mr. Gershon pointed out social media and texting may be better ways to reach younger populations. He noted children are actually accurate self-reporters all the way down to age eight, suggesting child input can and should be collected along with caregiver input. He also spoke about passive mechanisms of data collection, citing research that analyzing movement data is a better indicator for depression than clinical depression screening.
- Ms. Vernon highlighted the potential of patient advocacy networks to identify trial participants. She further suggested focusing outreach efforts on specific clinics that have built trust in individual communities.
The panel did not have an especially focused discussion on the critical issue of probabilistic sampling. Mr. Gershon did make the intriguing observation that web-based sampling, which is non-probabilistic, may become probabilistic as technology develops and internet use grows.
Session II: Methodological Considerations for Data Collection, Analysis, and Operationalization
Presentation: Selena Daniels, COA Staff, OND, CDER
Ms. Daniels provided an overview of identifying appropriate sites, access, sampling, data collection, recording, resolving field issues, and data management. The key takeaways here were that the research objective should inform the methodological approach, and that data collection and reporting should be standardized as much as possible.
A panel discussion with Selena Daniels, Steve Cohen (RTI International), Sheri Fehnel (RTI Health Solutions), Gary Globe (Amgen), Isabelle Lousada (Amyloidosis Research Consortium), and Kai Ruggeri (Columbia University) addressed the following questions:
- Future guidances will further discuss qualitative, quantitative, and mixed methods. Is more detail (or less) needed in this first guidance about which source (e.g., interviews, focus groups, consensus panel, etc.) to use to collect data? Is anything missing?
- Is more detail (or less) needed in this first guidance on operationalizing and standardizing data collection and management? Is anything missing?
- Are there are any other factors to consider regarding method selection that should be included in the guidance?
- Are there any other factors to consider regarding operationalization of the data collection process?
The panelists agreed it would be helpful for the FDA to be clear about the purpose of qualitative data in the guidance. A commonly heard critique is that qualitative data capture often repeats what quantitative data capture has already found. One panelist also warned of the mosaic effect: Any individual piece of data might well be insufficient to identify a patient, but putting all of the pieces together could violate patient confidentiality. Ms. Lousada asked the FDA to clear up confusion about which tools would be useful during different stages of drug development by including more context in the guidance.
Ms. Fehnel suggested working with patient advocacy organizations, which could help collect data more accurately with nuanced understandings of the condition. Ms. Lousada, herself a patient advocate, agreed with this contention. She added patient preferences and experiences change throughout the journey of living with the condition. Traditional methods of data collection only capture one moment in time. Ms. Fehnel said digital data is appealing but one-on-one narrative information is impossible to replace. She further suggested building a registry to support clinical trial recruitment so as to avoid reaching out to the same patients, key opinion leaders, and advocacy organizations over and over again.
Mr. Cohen called for an operationalized feedback loop so that the research goes back to the patients. As Ms. Lousada put it, the model needs to be “bedside to bench and back again.”
Finally, Ms. Lousada made the powerful comment that the choice between survival and experimental treatment is an extremely challenging and emotional one. She said it’s crucial to consider who should elicit information in these patient populations and how to appropriately interpret feelings and experiences in these situations.
Session III: Translating Best Practice into Real Practice – Developing Guiding Examples
A panel discussion with Richard Gershon (Northwestern University), Telba Irony (FDA), Susan McCune (FDA), April Naegeli (Eli Lilly), Sally Okun (PatientsLikeMe), and Elizabeth Stuart (Johns Hopkins University) was moderated by Sara Eggers, DSAT, OPSA, OSP, CDER, FDA and Megan Moncur, Office of Biostatistics and Epidemiology (OBE), Center for Biologics Evaluation and Research (CBER), FDA. They addressed the following questions:
- What are your thoughts on the examples that are currently included in the discussion document?
- What concepts from the discussion document would be helpful to illustrate through examples or case studies?
- When collecting patient experience data, what are some common challenges seen in study design or implementation that might be useful to address through additional examples? How can that challenge result in data that are less suitable for regulatory purposes? What are practical ways to avoid the challenge?
- Are there novel approaches or exemplars for collecting patient experience data that could be useful? How could someone replicate the effort?
Examples from the discussion document referenced in this panel discussion include:
The panel had varied feedback on the examples included in the guidance draft. For instance, Ms. Naegeli argued the extensive preparation and resources needed to collect patient data means even more detailed case studies would be useful to gauge best practices. Mr. Gershon, on the other hand, had a more fundamental objection to the examples. He felt they were somewhat restrictive and could stunt creativity, as they only covered traditional methods of research and could be misinterpreted as instructions instead of guidelines. Ms. Stuart echoed this point by advocating for examples that highlight how various research objectives require different sets of considerations. She also asked the FDA demonstrate in the examples the presence of “deliberately heterogeneous sampling.”
Ms. McCune acknowledged the need to address alternative trial reporters like children, and Ms. Okun wisely observed that novel does not mean not rigorous. She spoke about the work that PatientsLikeMe is doing with the FDA to validate information collected from the online patient support platform. Ms. Okun spoke to the argument there isn’t enough evidence to use non-traditional research methods in FDA decisions by arguing it is precisely that need to spur further discussion that requires the inclusion of such methods in the guidance.
Session IV: Identifying Key Themes and Next Steps
A panel discussion with Conny Berlin (Novartis), Sonya Eremenco (Critical Path Institute), Kimberly McCleary (Faster Cures), Theresa Mullin (FDA), Elektra Papadopoulos (FDA), Celia Witten (FDA) and moderated by Sara Eggers, DSAT, OPSA, OSP, CDER, FDA and Megan Moncur, Office of Biostatistics and Epidemiology (OBE), Center for Biologics Evaluation and Research (CBER), FDA addressed the following questions:
- What are the three most important messages you have taken away from the workshop discussion that should guide FDA as we complete this draft guidance?
- How can FDA strike the right balance to meet stakeholders’ needs? How well do the discussion documents for today’s meeting strike this balance? For example:
- What is the right level of detail?
- What is the right technical level?
- Would the overall structure and format of the discussion documents for today’s meeting be reasonable for the guidance? Do you have additional structure and format recommendations for FDA to consider when developing the draft guidance?
- Considering that this is a first in a series of guidances that will be developed over time, how might FDA best facilitate stakeholders’ understanding of the big picture and how all the pieces fit together?
Each individual on the panel spoke about their main takeaways:
- Ms. Mullin appreciated the FDA’s vantage point in understanding the evolution of the role of patient groups but warned of likely confusion from stakeholders who don’t share the same perspective. She said the FDA must provide more clarity on how patient information will be used.
- Ms. Papadopoulos echoed Ms. Mullin, saying the FDA can be more transparent about how it will use patient information. Clearly identifying objectives will help avoid redundancy. She said patient advocacy groups have considerable potential as conveners and supporters. She was similarly intrigued by non-traditional methodologies such as leveraging existing data consortia, virtual research, passive data collection, etc.
- Ms. Berlin’s big takeaways were the power of collaboration, the importance of identifying goals before data collection, and the usefulness of benefit-risk analysis in the identification of key objectives. She added that a roadmap of drug development would be valuable so as to track when certain kinds of information need to be collected.
- Ms. McCleary said she was struck by the increased opportunity for patient perspectives to be heard, as typified by this entire PFDD meeting. She contrasted this with the HIV crisis, when there might have been at most a single patient advocate at the table. She emphasized the usefulness of shared vernacular and called for the final guidance to be detailed without being overly prescriptive.
- Ms. Eremenco spoke to the importance of collaboration to create new measures and data collection instruments, with the end goal of reducing burden for patients. She asked the FDA to expand the guidance to include new technology, as well as context for how this guidance will relate to more general FDA guidelines and specifically those dealing with patient-reported outcomes. The diaTribe Foundation is also invested in following these PROs.
- Ms. Witten noted the potential to reduce duplication of work. She suggested data could perhaps be shared informally, especially when it is not going to be submitted formally to the FDA
The panelists had sage advice for the FDA regarding future steps, pointing out that the guidance will have to be released with a strategic communication plan, as understanding of the guidance – especially when it’s just the first of four – requires a lot of background information.
It was such a thrill to be able to witness the FDA welcoming constructive feedback with open arms. We are hopeful that patient-focused drug development will continue to evolve in a manner that benefits both patients and researchers, and we look forward to the next series of FDA guidance meetings!
By Jeemin Kwon