Amy Abernethy, a thought leader in the field of evidence generation and the development of real-world evidence, was named principal deputy commissioner at FDA.
Abernethy, who is now the chief medical officer, chief scientific officer and senior vice president of oncology at Flatiron Health, is expected to start at the agency in February 2019 after leaving Flatiron. She will replace Rachel Sherman, who oversees clinical, scientific, regulatory, and operational medical programs and initiatives across the agency. Sherman is retiring after 30 years of service at FDA.
“I am very excited,” Abernethy said to The Cancer Letter. “It’s a bit daunting and also a big step and change in what I do everyday in my career, but overall, it feels like the right thing to do both from the standpoint of how to continue to work on the things that I’ve been working on for the last 20 years, as well as from a career standpoint. And also, I’ve always wanted to be in public service.”
Prior to joining Flatiron, Abernethy was professor of medicine at Duke University School of Medicine and ran the Center for Learning Health Care in the Duke Clinical Research Institute and Duke Cancer Care Research Program in the Duke Cancer Institute.
For more than a decade, Abernethy’s work has pioneered the development of technology platforms to accelerate clinical evidence development, including the development of systems by which aggregated clinical data can support personalized medicine, outcomes research, cancer care quality monitoring, and scientific discovery. Notable among her publications are:
- Journal of Clinical Oncology, Rapid-learning system for cancer care
- Nature Reviews, Patient-reported outcomes in cancer care—hearing the patient voice at greater volume
- Pediatrics, Demonstrating the learning health system through practical use cases
- Lancet, Effect of palliative oxygen vs. room air in relief of breathlessness in patients with refractory dyspnea: a double-blind, randomized controlled trials
- JAMA, Association of broad-based genomic sequencing with survival among patients with advanced non-small cell lung cancer in the community oncology setting
“Dr. Abernethy brings to the FDA deep experience in evidence generation and clinical trial efficiency and innovation, which will help us advance our clinical policy and programs and continue to drive scientific excellence in support of FDA’s public health mission,” FDA Commissioner Scott Gottlieb wrote in an email to agency staff. “She has also put patients at the center of her work throughout her distinguished career.
“And through my interactions with her, and my knowledge of her many accomplishments, I have full confidence she’ll bring this patient-centric approach to her work at the FDA as we continue to enhance our patient engagement and patient affairs programs.”
Abernethy’s appointment comes at a time when FDA is taking the first steps toward incorporating real-world evidence in regulatory decisions. As required by the 21st Century Cures Act, FDA has published, earlier this month, a framework outlining how the agency would apply real-world evidence across its drug and biologic review programs. The report, “Framework for FDA’s Real-World Evidence Program,” is posted here.
“If you haven’t read it, it’s really worth reading. It’s an incredibly well written document,” Abernethy said. “I think that the major takeaways in my mind are consistent with what we were expecting before, which is that there’s a move towards trying to understand how to use real-world evidence.
“And that move is asking, ‘How do we make sure that the data quality and the analytic quality is adequate to address the kinds of questions that are going to be brought before the agency?’ Really focusing on quality is going to be critical.”
Abernethy spoke with Matthew Ong, a reporter with The Cancer Letter.
Matthew Ong: Congratulations, Amy! So, you’re moving to D.C. How do you feel about that?
Amy Abernethy: I am very excited. It’s a bit daunting and also a big change in what I do everyday in my professional life, but overall, it feels like the right thing to do. It will give me the opportunity to continue working on the things that have been my focus for the last 20 years. And also, I’ve always wanted to be in public service.
How did the FDA gig come about?
AA: I received a call from a colleague a couple of weeks ago indicating that, confidentially, there was a potential for a role at the FDA opening up as the principal deputy commissioner. He asked whether I would be willing to hear more and, specifically, could he pass my contact details to Scott Gottlieb, the commissioner.
And it was one of those funny moments where, first of all, I didn’t expect it at all. And secondly, I was not in a physical or mental place to be able to truly contemplate the question. I answered with what just immediately came off my chest, which was “Yes, of course.” And it’s moved forward quickly since then. It was like I immediately knew what was the right thing to do without needing to think about it too much.
What does your job description look like? Or is the agency going to figure that out as you are being onboarded?
AA: I think that we’re all going to watch this evolve together. The FDA’s been undergoing a series of important organizational changes and this includes the fact that the center directors now report directly to the commissioner—that happened a while ago. So, for example, folks like Janet Woodcock [director of the Center for Drug Evaluation and Research]. And as these organizational changes have been coming through, this also is the opportunity for this role, itself, to be refined. I wouldn’t say that it’s going to be completely reshaped from what Rachel was going before, but to be refined.
I anticipate that this role will continue to have a responsibility for many cross-agency priorities, as opposed to those that live, specifically, in the centers. Like the current scope of the role, I will likely focus on activities that have been progressively prioritized over the last number of years, such as combination products and pediatrics. The actual areas of scope and direct responsibility will be announced over time.
Does this mean that you will have direct oversight over how the centers are being run?
AA: No. Importantly, the oversight of the centers is directly with the commissioner—the center directors are responsible for running their centers.
How will conflict of interest management work in your case?
AA: This is a really important question. On one hand, I have been working on the topic of real-world evidence and confident evidence generation for a very long time. This is my area of expertise. On the other hand, I work for a for-profit company in this space.
I am going to be very upfront about it with the ethics officials at the FDA and anyone who asks so that we can navigate conflicts transparently. I am sure that a number of rules will be put into place and I will abide by them. Whatever the rules are will be the rules I follow.
Managing conflicts of interest is also in line with my personal and ethical commitment to patients. At the end of the day, we should ask ourselves what matters when we wake up every morning and how do we make sure that the right thing happens. Transparency in principles and motivations, and how you action them is a critical step.
This obviously means you’re leaving Flatiron. How have you grown in your time there?
AA: I’ve grown personally; I’ve grown professionally. And I also feel that I’ve done at Flatiron what I went there to do. When I reflect on why I went to Flatiron, I went there to address a problem that I had been working since the early 2000s.
While at Duke, I’d been trying to find ways to better harness the data that was already being collected in as a routine byproduct of clinical care in order to meaningfully contribute to and accelerate the process of clinical evidence development. When I say “clinical evidence development,” I mean determining whether a treatment or care delivery approach was effective and safe enough to be implemented into clinical practice.
There are just too many times as an oncologist in clinical practice that we don’t have the answer as to what is the best thing to do for this person sitting in front of me. If you look at the programs that I managed at Duke before Flatiron—that issue was a core underpinning.
And I remember doing an interview, just as I was moving from Duke to Flatiron about why I was making a move. One of the things that I said in that interview was that I felt like I had topped out as to what was going to be possible at that point in time within academia.
I knew that the data was there and we could harness it if we could access it, but I was challenged to build robust enough technical systems to aggregate, clean, store and analyze it in real time. Solving this problem was going to require a lot of software development and technical acumen not easily accessible in academia.
This was 2012, or so. And it was clear to me that if we were to address the problem, the tech industry really had to be motivated to do so, and also had to have the guidance of “What might that look like?”
So, going to Flatiron gave me the opportunity to work directly with the tech industry and provide that guidance. Importantly, “How do we build this infrastructure in service of patients and with the highest-quality expectations possible?” I feel like we largely accomplished that at Flatiron.
I also learned how business and capital works, and ultimately, how to think of the landscape outside of the university. This broadened my scope of understanding in a substantial way.
What is your thinking about where Flatiron is? A lot happened during your time there. We’ve made leaps and bounds in cancer informatics, the Cancer Moonshot and the 21st Century Cures Act happened, and real-world evidence is now, well, real. Where is Flatiron going from here without you?
AA: First of all, I am confident that Flatiron will be just fine. One of the reasons it felt like such a natural time for me to take this job is because, if I reflect on where Flatiron is, it’s really clear that the company is in a good place to continue the forward momentum.
I anticipate that Flatiron will double down on continuing to build highest quality real-world datasets. As a company, Flatiron will likely want to continue participating in the community conversation about advancing the science, and doing this in a transparent and consistent way. I don’t think that any of this will change after I am gone.
I think you’re going to see Flatiron continue to ask the question of, “How do we develop real-world evidence beyond the boundaries of how we develop and apply it right now?” And when I think about the current boundaries, I include international boundaries. So, how do you advance this work recognizing that human health is an international activity?
I think that many of those things that I put in place while I was at Flatiron have now been knocked into place in a consistent enough fashion that I’m not worried that they’ll be able to continue. One example that I very explicitly worked on at Flatiron was advancing this concept of what ultimately got named “regulatory-grade real-world evidence.” When I was at Duke, I used to call it “research-grade.” So now, it’s been re-coined regulatory-grade.
I think that you’ll see Flatiron continue to think about how to do this work well. I can’t be the regulator having that direct conversation with them about it. But I anticipate that that will be something that they continue to focus on.
With all this in mind, what do you think you bring to the table at FDA?
AA: I think I bring four things to the table. First of all, I hopefully bring to the table a new and fresh point of view coupled with a collaborative spirit that allows me to handshake with many points across the FDA, and pull forward all the great ideas that are already there. I will try to couple this with the observations that I’ve had across time.
The second thing I think I bring to the FDA is a substantial fund of expertise that may be useful. Before I was at Flatiron and also at Flatiron, I knew routinely collected clinical data inside-out, in a way that is probably better than most folks. How do you develop it, work with it, clean it up, make it actionable, etc. I have been thinking about and actioning the concept of building a learning health system for a long time, and hopefully that applied knowledge will be helpful.
I also have been working in patient-reported outcomes and clinical trial design for a very long time. So, I think I bring that kind of knowledge to the table. And I don’t anticipate that I’m going to be asked to get rid of my current expertise.
Third, I’m an oncologist, so I also bring a clinical background that will help to put the work that we’re all doing into a clinical perspective, including, what does it mean to take care of people who have been threatened by an illness like cancer?
I think the fourth thing I bring to the FDA is at least a shortlist of personal priority areas that I’d like to pour extra personal energy into to help advance them forward. This is synergistic, because what I worry about day-in and day-out aligns with FDA priorities as well.
These include: accelerating clinical evidence generation; making sure that the work we do is fundamentally patient-centered, whether that means that you’re constantly thinking about patient-generated health outcomes or thinking about patient-informed drug design; and also precision medicine.
How does your background and expertise in oncology—and you mentioned this briefly—as well as evidence generation equip you for the job, and help give you a perspective on how we might be able to look at other fields this way as well?
AA: Oncology is an interesting area that acts like a microcosm, demonstrating where many aspects in human health and medical product development is going.
What are some of the features of oncology? Well, it is a disease area that is inherently patient and family-centric. The patient experience and patient voice is clearly prioritized.
It is a disease area where, in order to make meaningful change in patient outcomes, we have to master the biology. And all of the incredible advances in biology are leading to remarkable product development in oncology.
Oncology is a disease area, also, where the idea of continuous innovation and constantly looking for the next best solution is inherent to how we take care of patients. So, in oncology, we can very naturally embed the process of continuous evidence generation—truly demonstrating what a learning health system looks like.
And, along with innovation in oncology comes the complexity of health care costs and who is going to pay for it. Cancer care is a place where we have to figure this part out, too.
Oncology and being an oncologist brings to the table experiences in one therapeutic area and allows us to ask the question, “How does this translate outside of oncology?” And how can we not only prioritize meaningful work within cancer, but also translate lessons learned outside of cancer?
I can imagine immediate translation into therapeutic areas where precision medicine and biologically informed product development is certainly a huge area of focus right now. Areas like rheumatology, where we are understanding better both the immune system as well as targets, are making a big difference. And even areas such as psychiatry, where personalized approaches are more and more a part of how we think about taking care of patients. And rare diseases. We certainly have a lot of work to do.
And you also touched on this: I know it’s probably early days, but from your years of working with FDA, now that you’re going to be there, what would you like to see happen or make happen?
AA: First of all, I need to learn how to work in a confident and collaborative way within FDA. I want to make sure I prioritize that first.
Then, as I ultimately get there, I think one of the things that is particularly special about this point in time is that we already have a national agenda laid out for us in 21st Century Cures. That, for me, is something that I want to look to since the American people have already said this is important. And FDA is compelled to do that anyway. Cures includes aspects as it relates to modernizing clinical trial design, thinking about how we use all data available, including real-world evidence in a confident and consistent fashion, surrogate endpoints, and many other areas.
Another area that the FDA is focused on that lines up with my personal priorities is developing drugs faster and also making sure that we have multiple drugs in market so that, not only do patients have as many solutions in the portfolio as possible, but we increase competition in a way that allows the free market to work for us.
Finally, as we move into a landscape of biologically informed precision medicine, we need to think about what does confident evidence development look like. This includes judicious decisions leading to initial approval, as well as continued evaluation to understand how to ensure treatments optimally perform across time.
Since we’re on that train of thought, the way the Oncology Center of Excellence is set up right now is disease-specific, whereas the rest of the agency is set up by modality. Is it possible to restructure the agency using the approach that we have in oncology? And does that need to happen?
AA: Such an interesting question. I think the Oncology Center of Excellence, like oncology itself, acts as a proof point. It gives us the chance to figure out what works and what doesn’t for this particular model.
As we figure this out within OCE, and think about essentially trans-center collaboration in order to now bring meaningful products to patients in the most efficient way possible—until we figure this out, we’ll need to ask what’s working, what doesn’t work.
Importantly, oncology should not be the only place where the best of ideas happen and continue to move forward. As soon as we understand where great ideas are and what’s working, we need to translate them. And when something isn’t a great model or shouldn’t be translated, we need to know that too.
I don’t pretend to understand all of the nuances of the application of this model outside of oncology—or within oncology, for that matter. And I want to make sure that as we look at what OCE is doing, we think about how to support and learn. And in fact, I think the legislation has been written in a way that is intentional to try to move that approach forward.
As it stands, does the Office of Hematology and Oncology Products, including the OCE, be consolidated even more? For instance, including having authority over areas such as gene and cellular therapy, where the applications may primarily be oncologic right now?
AA: I don’t know. It’s a good question. I have to watch the space to see how this all evolves.
So, the agency recently completed its report on real-world evidence as required by the 21st Century Cures Act. What do you see are the major takeaways from the framework, and how we will be moving forward with that as a field?
AA: First of all, if you haven’t read it, it’s really worth reading. It’s an incredibly well written document. I think that the major takeaways in my mind are consistent with what we were expecting before, which is that there’s a move towards trying to understand how to use real-world evidence. And that move is asking, “How do we make sure that the data quality and the analytic quality is adequate to address the kinds of questions that are going to be brought before the agency?” Really focusing on quality is going to be critical.
I think that you start to line it up from what we already knew from 21st Century Cures, which is to focus on critical example areas, such as post-marketing commitments and requirements, and ultimately also starting to learn about how might real-world evidence be used with respect to label expansion.
My personal expectation is that we do the science really well—that real-world evidence is not about creating a landscape where we can be sloppy in our scientific work. And I feel like this framework document reminds us that that’s the case. For example, before the start of a retrospective study, we’re going to need to pre-specify the protocol.
Another thing that was in there that I thought was important is ultimately, the focus on pragmatic trials and asking the question, “When can the generation of real-world evidence be retrospective vs. when should it be prospective?” And I think that’s going to be something that, as a whole community, we’re going to need to think through.
As a clinician, as drugs become available, understanding better how to use drugs on a day-to-day basis when taking care of patients means you likely need more information than was available at the time of the original registration clinical trial(s). Sometimes that can be learned by querying data that are being generated, for example, in the electronic health record. But sometimes, you still need to plan a prospective trial. And those prospective trials might look at changes in dosing, drug combinations, supportive care, etc.
The other thing is that it’s hard to write a group of recommendations or guidances for real-world evidence without having enough examples sitting in front of you to be able to say, “Here is what it should look like.”
So, a critical element of this entire book of work is to have a series of meaningful real-world evidence examples that the FDA can critically review, and they can then form how they think about building the guidance and the recommendations.
Actually, I don’t know of anyone better to ask this question: a vocal critic of FDA, whom you may know or have at least heard of, says that the agency loves to measure its success based on the volume of drugs that it approves. Now that the agency has real-world evidence, their criticism is that we’re going to have so many more approvals in multitudes of tinier and tinier indications, with marginal benefits and big price tags, when FDA should be focused on meaningful benefit. How would you respond to that?
AA: I’ve heard this. Actually, I’m going to go back to my same core focus every time, which is that as a doctor and a daughter of a man who died of cancer, and a person who may be susceptible to cancer in the future, what I really care about is that we learn what works and what doesn’t work, and we have a level of confidence such that it allows us then to make judicious decisions in the clinic.
So, if the FDA is advancing our ability to make those judicious decisions in micro-moments, so be it. If they’re doing it in macro moments with macro approval, that’s terrific too. Ultimately, all these things come together.
One of the things I think the FDA does for us is provide a level of authority and sophistication and expectation of doing the right thing in service of public health, that then allows us to know that the FDA has seen this and given a thumbs up, or thumbs up with caveats—now we know how we can translate that into practice.
The translation of this into cost and price tags does matter, so I don’t want to belittle that part of the argument. And it is really important that as a society we address it head on. But I don’t think we should eschew figuring out what works and when to use treatments because we are afraid of cost implications.
How close are we to being able to use real-world evidence and real-world endpoints? I know a number of teams are already doing really interesting studies. But I also understand that, at the end of the day, it’s about having regulatory-grade data, and then being able to show equivalency to existing endpoints. How close are we, really, at this point?
AA: I want to make sure that I disconnect my answer to this question, from the questions about my new job at the FDA. This is Amy Abernethy speaking, knowing what I know about endpoints.
Despite what I just said in the last paragraph, not every single thing that we work on in clinical evidence generation and science has got to do with whether or not the FDA says, “Yes,” or “No.”
Oncology should not be the only place where the best of ideas happen and continue to move forward. As soon as we understand where great ideas are and what’s working, we need to translate them. And when something isn’t a great model or shouldn’t be translated, we need to know that too.
Oncology should not be the only place where the best of ideas happen and continue to move forward. As soon as we understand where great ideas are and what’s working, we need to translate them. And when something isn’t a great model or shouldn’t be translated, we need to know that too.
That’s the first thing we need to acknowledge, that in the development of endpoints, it’s actually a part of moving the field of real-world evidence forward, full stop. It’s not just an FDA thing. I think we are in the process of developing a portfolio of real-world endpoints that’s going to get larger, essentially, every year.
When you think about it, there are five endpoints or outcome types for most cancer research. (I’m just going to talk about this from the standpoint of cancer research, because that’s what I do.) There’s mortality, tumor burden, patient-reported information—patient concerns and experience—toxicity or side effects of drugs, and then health resource utilization. And all of those can be differently addressed by passively collected data that’s available in the electronic health record and in other systems like claims data.
I think what we’re seeing the whole industry do is, in a very progressive and consistent way, start to develop more specificity about what confidence we can have in endpoints, in each of those categories.
Mortality, I think we’re getting better at being able to have source mortality data from multiple places and bundle it to be embedded into real-world datasets. We must constantly measure the quality of mortality data, though, because this assessment tells us the confidence we can put into results such as point estimates of overall survival.
When it comes to tumor burden, a number of groups have been working on this in different ways, whether that is looking in electronic health records for documentation of disease progression, calculation of time to treatment-change, or understanding issues as it relates to direct access to radiology images. I anticipate that, in the next 18 months, we’re going to have a lot of meaningful movement here.
Adverse events, if you start to look to places like Sentinel and pharmacovigilance work, that’s some of our best work around adverse events, but it is less of a focus for many of the groups currently working on real-world evidence in oncology. If anything, I think this is the area that’s been underrepresented in the last two years and one we have to move forward quickly, especially in the landscape of IO drugs.
Patient-reported outcomes, PROs, are critical and we need to get on with it. Real-world evidence is about taking care of the patient sitting in front to of you. And if you don’t have patient experience information, it’s really hard to do that in a confident way. So I see this as something that we’re going to see more and more conversations about.
Then finally, health resource utilization comes often from claims data, which is some of the data that we have that’s the most reliable in RWE systems. We need to keep learning to use it.
Would you say that real-world evidence is probably most developed in oncology compared to most other fields? And do you see the potential for this to be applied across FDA’s portfolio in a reliable and consistent fashion?
AA: I think that there are aspects of the RWE work that have been more advanced in oncology than in other areas. But I think that there are areas outside of oncology that are also similar or more advanced. It’s got to do often with the nuances of the disease and available data.
When you think about real-world evidence within the landscape of precision medicine, oncology is pretty far along, because many of our patients have available biological data—for example, next-generation sequencing and other biomarker data. Oncologists tend to be really crisp in how they document features of disease. So, when I think about RWE in that space, I suspect oncology is probably ahead. I often think about this as the emerging field of small-cohort real-world evidence.
But then, when I consider comparative effectiveness research and truly comparing the performance of Drug A versus Drug B in large-scale databases, other therapeutic areas are very well developed. For example, hypertension, multiple sclerosis and diabetes. Those are areas where there’s a crowded therapeutic landscape and many patients with disease. Large-scale comparative outcomes research can be done in a way that’s often different from oncology. It looks more like large population science.
Will oncology continue to hold a special place in, well, your professional heart, as you expand your horizons on this job?
AA: I’m an oncologist, so I use that as my day-to-day reminder that this is what it looks like to take care of patients in clinic. I use that as also my framing to how I think about problems and illness. But I see a huge responsibility of expanding my landscape beyond oncology.
The other thing to remember is that most oncologists started off as internists first. We actually started off with a trans-disease focus, and most patients with cancer don’t just have cancer. They’ve got multiple illnesses and health concerns. So, I think that one of the things, as an oncologist, I bring to the table is the fact that I don’t just think about the cancer.
The other thing to say is that oncology is an important place where we think a lot about prevention. I think prevention is important across all of health. Prevention is about maintaining health, and ultimately, the best way to reduce the scourge of disease is to help people maintain health. I think that being an oncologist helps to frame things well for me from that perspective.
I will use the fact that I’m an oncologist as a frame of reference, but then, I’ll go from that frame of reference to recognizing that I have a responsibility across the spectrum.
When are you starting the job officially, and when will we start seeing you around town?
AA: I officially start some time after Jan. 31, I think. We need to go through all the ethics clearances and do that work. So, I anticipate that you’ll see me more after the end of January.
Thank you so much.
AA: Thank you very much for your time.
