‘We can use this data for good’: Flatiron Health CEO Carolyn Starrett

Founded in 2012, Flatiron Health helps cancer centers to thrive and deliver better care for patients. CEO Carolyn Starrett joins Maria Whitman to talk about reimagining the infrastructure that supports cancer care and research around the world. She also discusses how the company’s mission has evolved from building software to creating data sets to accelerate and advance new oncology therapies. Listen to hear how we can close the gap between research and cancer care, the role of real-world evidence in improving clinical trials and understanding outcomes, and why data plays an important role in addressing health inequities. 

Transcript

Carolyn Starrett: 

What’s really incredible, I think, when we think about the success and the possibilities we see ahead, is just we have the data and the tools and the technology to really do all of these things within our reach, in cancer, across all of healthcare.  

Maria Whitman: 

Welcome to Transforming Biopharma by ZS. I’m Maria Whitman, and this episode is part of our series on the future of healthcare. Today, I’m talking with Carolyn Starrett, CEO of Flatiron Health. It was founded in 2012, and Flatiron has built the leading US oncology real-world data source. Before Carolyn became CEO in 2021, she led the company’s community oncology business. Welcome, Carolyn. Thanks for joining me today. 

Carolyn Starrett: 

Thank you so much for having me. It’s great to be here. 

Maria Whitman: 

And happy 10-year anniversary to Flatiron. 

Carolyn Starrett: 

Thank you.  

Maria Whitman: 

10 years, Carolyn. For our listeners, can you talk a little bit about the mission of Flatiron, how the company evolved? I especially think about it since the barriers you must have faced with a fragmented health system in trying to get this data together. 

Carolyn Starrett: 

Yeah. Our mission has always been, and hasn’t changed, our mission is to improve and extend lives by learning from the experience of every person with cancer. And that’s really the mission that drove us as we were founded and it’s continued to be true today. What’s been interesting to observe over the last, well, six and a half years since I’ve been a part of Flatiron, is just how much opportunity remains to really reimagine the infrastructure that supports cancer care and research around the world and think about how do we tackle that mission in new ways and what new possibilities have come to bear. We started with a real focus on software and technology for community cancer centers in the United States and then in building the identified aggregated data sets for research to accelerate and advance new therapies and to understand how those therapies work in the real world for real-world patient populations. 

Over the last few years, I think we’ve really started to broaden our ambitions, and so we’re thinking about how can we blend integrated real-world evidence with prospective evidence generation, making clinical trials much more efficient and representative, and then how can we think about really carrying through this mission to the point of care and helping to support improvement in patient outcomes, driving better clinical treatment decisions and clinical decision support? 

I think you asked about the barriers we see, I think the first one, and the one that really was foundational in creating the company, was just that the access to the necessary longitudinal data and outcomes was not there. I think we’ve made a lot of progress there. The two that I think about a lot now are just how do we get to a better place with greater alignment of the incentives, whether it be around what we pay for and how payment models work and the really important evolution towards value-based care that I think is in incredibly important, but also just how do we give patients the role that they need to play in the center of their care and of their health. 

And longer term, a lot of the opportunity, the potential, will rely on doing things that we’re used to doing in one way using different tools and different methods and different approaches, and so the biggest barrier, honestly, is just that natural change aversion that I think we all have to getting up in the morning and changing our routine and looking at a new project and thinking about what are the potential ways that I could do this differently and better? 

Maria Whitman: 

I love that point. I think the change aversion in health is pretty high and, even when we try to change, we end up fragmenting further, but there’s also this reality that we’ve built processes around what we believe is the best, the best for the protocol and the trial, the best for the patient care that any one institution’s going to give, and one of the things I really love about where you’ve prioritized, as a company, is closing the gap between research and care. We are people, we become patients, we’re patients through the whole spectrum, and yet there is a very sharp drop-off, I guess, and the real-world data and evidence can really support that transition in some ways. What would success look like in closing that gap between research and care, and where are we in that journey? 

Carolyn Starrett: 

Yeah. I honestly think we’re still in the very early days. Yeah, what would success look like? Starting from just unpacking the set of words that you described, closing the gap between research and care, well, let’s define research as all of the learning and the innovation that goes into thinking about what new options should exist, and how do we learn which patients should get which therapies and what types of treatment, and the clinical trials that tell us what are the new alternatives that we want to bring into the market. There’s all of that learning and innovation happening on an ongoing basis, but, right now, it takes far too long to figure out how we translate that learning back to patients who are treated any given day, whether it be understanding the right therapy, whether it be participating in a trial, whether it be integrating new standard of care and new guidance that comes out of all of the trials that are running around the world. 

What we really think is that there is a possibility to dramatically tighten that gap, and what would success look like? It would be an ongoing learning system that helps us to continuously reduce the time it takes to bring new therapies to market and to understand who can benefit most from them and then to make them available to the patients that need them, and success would be integrating clinical research and trials into the everyday care experience. Today, whether you are being treated in clinical standard of care or participating in a clinical trial means that you’re going through two different, totally siloed systems of care, and it shouldn’t have to be that way. We should be able to integrate research and learning into the care that all patients receive. 

And I think, at the end of the day, it means ensuring that every person with cancer receives the best possible care no matter where they’re treated and whom their doctors are. And what’s really incredible, I think, when we think about the success and the possibilities we see ahead, is just we have the data and the tools and the technology to really do all of these things within our reach, in cancer, across all of healthcare. And coming back to where we started the conversation, I think the real opportunity is overcoming those historical barriers to progress and shifting our mindset, really embracing the power of the new approaches that are available now to us. 

Maria Whitman: 

I agree that real-world evidence has not ... we’ve not even scratched the surface on what’s possible to do to change patient care, to change an individual’s experience and care throughout their journey, whether it’s cancer, other chronic illnesses, in the system. 

You’ve done a lot at Flatiron too work towards improvement of clinical trials. What is the big next opportunity, if we have to pick one, to get past these barriers? Is there another one or two places where you think we should be much further ahead using real-world evidence than we are right now? 

Carolyn Starrett: 

I think, in some ways, if you unpack that question, historically, with clinical trials, we’ve been able to use real-world data to figure out how to design better trials and figure out where to open trials and figure out how to screen patients and help enroll them. The real opportunity over time, though, is, I think, getting smarter about how we use real-world data to understand treatment patterns and outcomes after drugs are approved. We’ve all seen the trend towards more and more drugs getting approved on smaller and smaller data sets, and what that means is the results of the trials ... certainly, we can understand a baseline set of safety and efficacy, but a lot of the learning needs to happen after the trial is done. 

A lot of the learning needs to happen when we bring those therapies into the real world and we can understand populations that we’re not able to study in a clinical trial setting, where we can understand how to optimize for different dynamics, where we can follow patients, actually, using the real world data in clinical care after the trial is done and start to understand how to optimize treatment much faster and much more rapidly. We’re seeing a world in which there are growing number of post-marketing commitments and research questions that we really want the opportunity to study. And, right now, running a Phase 3, Phase 4, clinical trial is a really important tool that we can use, but it’s slow and it’s expensive, and often it’s still not representative. And so I think we have a real opportunity to use integrated evidence to learn much more quickly about real-world outcomes and real-world opportunities. 

Maria Whitman: 

And that’s really what people want. In our survey, just under 50% of people don’t believe the health system cares about people like them. And, when you go globally, it’s almost the same percentage in most countries we studied.  

And it says that, when you dig in further, “I’m not getting treatment that’s specific and personal to me. We’re not taking the best of what we know and applying it to my health in person,” and that breeds into a lot of questions around, “Do I trust individuals in the system with my data?” And what’s interesting is some of the population says, “I’ll trust you if I know it’s going to improve my care.” Others say, “I’ll trust you if I know it’s going to improve population care.” But trust has been a big issue out there in terms of health and health data. How has Flatiron navigated that particular challenge? 

Carolyn Starrett: 

Yeah, I could not agree more. I think it’s trust, security, protection of patient data and patient privacy has to be at the core of everything we do. All of this opportunity is reliant on our ability to build the trust and stay accountable to patients for how we use their data and how we use it for good, what the benefit of the work that they are doing and the learning system that they’re contributing to is. Flatiron, first and foremost, we adopt the strictest security standards to ensure that patient data is protected. We’ve got robust processes to anonymize and de-identify all the data we learn from. We do expert determinations to ensure that every data set we share externally to Flatiron is individually certified as de-identified, and we prohibit linking and pooling of data sets without further certifications of that data. That’s just the foundation. 

I think, as an industry, we need to get much smarter about how we talk about and build transparency with what data are being collected, how we plan to use that data. Some of the things we’ve been doing over the last few years at Flatiron is really spending time learning from patients about what they care about. We have a patient advisory board here in the U.S., a patient voices panel in the UK, where we spend time regularly talking to patients to understand what they value, what their concerns are, how we can work together to tackle those topics. 

And one of my favorite parts of the work we do at Flatiron is the time we get to spend with patients. We do a patient panel at least twice a year at our everyone team meetings, at our all-hands meetings, a biased sample, of course, but what we find is that, once patients understand how their data are being used and the power of the research and the questions that can come from that, they are incredibly supportive. We always ask the question about how they feel about sharing their data, and I have yet to talk to a patient who isn’t passionate about helping others to learn from their experience. I think it really comes back to that point of building trust through education, transparency, awareness, and then just continuing to be really careful stewards of the data and ensuring that we adhere to the highest possible standards there. 

Maria Whitman: 

That’s great. And I agree with you, patients, many of them, are very eager to help the next patient, as well, not go through everything that they had to go through. And I’m really curious, in all these patient panels, has there been some insight that struck you in particular about what patients saw as possibilities in the data or how they saw the opportunity in front of us to change care? 

Carolyn Starrett: 

I think it’s exactly what you said, it’s that we can use this data for good. I’ve had melanoma twice. I, thankfully, haven’t gone through, really, the intense struggle that so many people with cancer go through, but I think we all have a desire to want to give back, to make sure our stories count, to make sure that, as you said, that someone else can use the experiences that we each have and, hopefully, make that better. And so I think there’s an inherent desire in all of us to want to do good. The real opportunity and challenge is to make sure that we stay accountable to the very appropriately high expectations that patients in the system overall have to make sure that we continue to have the right protections in place and help be as transparent as possible. 

Maria Whitman: 

Carolyn, in the few times I’ve met you, you’ve shown to be one of the most passionate people I know about this, and I didn’t even know about your cancer. Thank you for sharing that with me. Wow. I think one of the biggest opportunities we have, and COVID definitely shown a light on health inequity, and you think about the disparities, particularly in cancer care, can be very stark, just the difference in women succeeding in therapy, Black women versus white women in breast cancer, for example, very, very stark discrepancies. And I’m curious what you think are the most important steps that life sciences can take and help with data, like what you have at Flatiron, and the solutions you have to really try to boost health equity to change what we’re seeing in cancer care? 

Carolyn Starrett: 

Yeah. We’ve had a long-term commitment to health equity at Flatiron and, when we talk about data for good, right, I think that is one of the core and important obligations that we have if we want to be good stewards of the data that we are so privileged to get to learn from. One of the moments we’re most proud of at Flatiron, one of the publications we did was selected to be presented at ASCO’s plenary session a few years ago. It was right after the ACA in the U.S. helped drive expansion of Medicaid in many states, and we were able to look at what is the impact of the extension of Medicaid on time to treatment for cancer and show that it had a demonstrable benefit in terms of how quickly patients were going from diagnosis to treatment that was measurable and meaningful. And that’s the kind of thing that is really hard to learn without looking at real-world outcomes and without data sources like the ones that we’ve cultivated. 

Maria Whitman: 

That’s incredible. That’s incredible. 

Carolyn Starrett: 

A slightly more challenging dynamic, however, is we gotta continue to systemically monitor those real-world outcomes so that we can figure out what those levers are. We also published recently ... we looked at all the data, through COVID, of what happened, and one of the really positive and exciting things is we all saw this tremendous increase in the utilization of telehealth solutions, right? Everyone came together and rallied and figured out, “How do I do as much care in virtual settings as possible?” All great. 

Well, many thanks in healthcare, when you step up, we looked at the data and the reality is that access to telehealth isn’t evenly distributed. And so we find that patients who are in lower income or have less access to resources actually have a harder time accessing telehealth care and so, as a result, that impacts the quality of the experience they’re going to have. And so we have an improvement in one area which leads to setbacks in others, and I think the opportunity is to continue asking those questions and looking at where and how we can address those challenges. 

In terms of what life sciences can do, I think you asked about that, the one that I’m most passionate about right now, and this has been, I think, a growing stream of conversation across the industry, is just that our trials are not nearly representative enough of the patients who ultimately are going to use those treatments, and we really need to make an ongoing commitment to diversity in trials. We’re seeing guidance and regulations, which similarly support this here in the U.S. It’s a really hard thing. There are not easy solutions here. I already talked about it. Trials are too slow and too expensive, and they’re incredibly burdensome to all involved. 

Yes, we can make them more efficient and more friendly to investigators and clinicians and patients and sponsors, but we have to do that so that we can then, at the same time, also improve the representativeness of the patients who are actually enrolling in them and make them more accessible to those patients. And I think that’s one of those things where we all have to come together and wake up and say, “We’re going to do this differently. We’re going to prioritize differently,” if we want to see it really change. 

Maria Whitman: 

You probably saw a lot of the access issue in your role with the community oncology setting too. One of the things I think about is there’s the access, there’s also this amazing fragmentation that’s happening of the care journey now because we’ve disaggregated the “where” of care. We’ve got more alternative sites of care. We have 29% of the population not having a primary care physician. We have many, many more people going to the pharmacist as a first port of call, really trying to navigate care in the way they need to or when they need to, unfortunately, most not until they’re sick. I’m thinking about this and, okay, there’s the people who don’t have access, then there’s this heterogeneity of how do we even find people and bring them in at the right time to the system? I’m not sure if you have any thoughts on that as we try to recruit the right people to the trials that we need. 

Carolyn Starrett: 

I started talking about this earlier, I think some of it has to start from reorienting towards designing trials that are aligned towards the sponsors and the sponsors’ needs, and that makes sense historically. They were the ones paying for it. They’re the ones who are going to stand to benefit if the trial is successful, outside of the patients who will benefit from those therapies. But, actually, do we need to turn that on its head and think more about how do we make it so much simpler and less burdensome for the sites that are going to need to find those patients and enroll them and the patients that need to give up some of their lives to participate in that learning? 

And I think, as you mentioned, we do really close work with almost half of the community and independent cancer clinics in the U.S. and they don’t have the resources to open the types of trials that they would want to open because it’s so burdensome. We still see that 20% of the sites that open a trial never enroll a single patient, which is just mind-boggling when you think about that. I think the things that we can do are a couple fold. One, build better technology to take just the terribly inefficient manual duplicative effort out of this process, right? 

Rather than manually screening patients for all the different inclusion/exclusion criteria, how do we use AI and technology to automate that process? How do we alert clinicians when their patient that they’re seeing the next day may be eligible for a trial, instead of expecting them to memorize the trials that happen to be available and all these different complex criteria? How do we put that right in front of them in an easy way? How do we then eliminate the duplicate data entry? 70% of the fields that we’re trying to collect, I would guess, in most trials are actually already captured as part of the EHR and part of the clinical care. How do we just reuse that and automate that data management flow so that we’re not creating 3x the work in order to make these trials happen? 

Maria Whitman: 

That’s an amazing number, 70% is overlap, and think about the money saved and the hassle for a patient saved in having to relive and rehash all that information experience. Wow. Wow. Go ahead. 

Carolyn Starrett: 

I think about times I’ve sat in clinic and shadowed these processes, and, as far as we’ve come, it’s just shocking how inefficient our systems still are. And that’s on us because the technology hasn’t supported workflows that make sense for those trying to use it. But I watch people, literally, writing down written case forms and then going from one system to the next and then re-transcribing them and duplicating that entry, and that’s not only inefficient, but it’s also error prone. And we risk real errors through each step of the journey when we allow for that to continue to happen. 

Maria Whitman: 

I want to pick up on what you said about making it easier and giving the tools to the physicians. In our study, we were asking more about the data available through connected care, not only what’s in the EHR and helping use that, or if there’s real-world data available more broadly, but also even just remote patient monitoring data. Now there’s more wearable data that can help us understand how a patient’s doing at different points of time, and a lot more EMRs are allowing them to be uploaded. But when we talked to oncologists, almost 80% of them said, “I don’t know quite what to do with it all yet. I don’t know how to use it for personalized care. There’s not standards yet to help me do that.” Are we close to helping to solve that? Because the data’s there to be used. 

Carolyn Starrett: 

Oh, I think it’s a very real challenge. We now have so many new streams of data that we have access to. We’re getting closer and closer to being able to connect those systems in a meaningful way. Very excited to see the new ONC interoperability rules, or APIs, finally roll out, and I think those are going to unlock at least a meaningful starting point for standards that enable a type of data interchange that is just so important. And it’s a step one, it’s not going to be the end-all, be-all, but we’ll continue to get smarter. I do think doctors are just buried under the weight of all that we ask of them, the complexity of the treatment considerations in cancer, all of these new sources of information and data that we might want to consider to truly personalize care, the payer requirements, and all of the different mandates required to then ensure that the patient’s financial constraints are understood and that they ultimately get paid. 

It’s an enormously complex landscape, and I think, when you step back, the only way to deal with that is, really, smart technology that is integrated into workflows in a way that serves the clinician, not, quite honestly, the payer. And we have an electronic health record that we developed at Flatiron, and the reality is those systems, historically, have been optimized for reimbursement, not for clinical treatment decision support. And I think that’s an opportunity that is core to how we are thinking about the evolution of the system that we run, but I think we have to think about that holistically. How do we build technology that really supports decision-making, not just reimbursement? 

Maria Whitman: 

I completely agree. We’ve pivoted. We have more available to make the choices smarter. What you are seeing now is a lot of algorithms and clinical decision support being developed by pharma, by digitech companies. There’s a lot going on out there, but, at ZS one of the things we believe is that, in addition to the technology, this is only going to be solved through partnership, because partnership leads to scale, and we don’t want to further fragment with a lot of technology out there. You’ve had so many successful collaborations at Flatiron with a number of various partners across the ecosystem. I was curious if you have learnings or thoughts about how the biopharma industry, or healthcare in general, can do better to use data and collaborate with each other to advance some of these causes? 

Carolyn Starrett: 

Yeah. I think, honestly, this has been a really bright spot for me. I see a real commitment across the ecosystem to connecting the work that we’re all doing and coming together on some of these topics. We were part of launching the Real-World Evidence Alliance, which is a coming together of players across the field who are working to build awareness and alignment around best practices and policy advocacy and the use of real-world evidence. And that’s, in some cases, competitors, in some ways, coming together and working on practical solutions that we hope will advance and move us forward together. I’m also really excited to see the momentum we’ve seen recently in public and private partnership. I think the increased engagement we’re seeing from Biden’s Cancer Moonshot initiative in the U.S. and the ways in which they’ve made a concerted choice to try to engage all sides of the ecosystem, the private players, the health systems, the payers. 

Similarly, we’ve seen a lot of great collaboration with NICE in the UK on the real-world evidence framework they’ve launched, and I think those types of public-private partnerships matter a lot. And then I think, across both life sciences and the healthcare delivery ecosystem, standards-based approaches are where the real opportunity lies, which talked about data interoperability and the ability to connect systems together. That’s true in terms of care delivery and all of those new data sources and elements that you talked about and how we make the most of them. It’s deeply true in clinical trials as well. And, when we think about the point in which life sciences touch patients and sites and research sites, that is going to be an ongoing area of focus that I think we’ll continue to look at. 

Maria Whitman: 

Absolutely. Those are amazing examples. They all come with some sort of a catalyst too, a catalyst and a partner who’s thinking about the next thing, whether it be a public organization or a collaboration like you’ve created. I do think, as we think about clinical trials, we’re going to have to figure out how do we band together, how do we band together as an industry? Do we need a collaborator or someone to help us get there? But it does seem like we’re getting to a point, especially in oncology, with so many trials, so few patients to really participate, that we have to figure something out.  

So Carolyn, here’s a question I like to ask all my guests. If you could change one thing about healthcare, what would it be? 

Carolyn Starrett: 

Oh, I love that one. Let’s talk about the word healthcare. We spend so much time focused on the care part of that, the tests, the drugs, the labs, the procedures we all do, and I’m really optimistic about what the future holds there, all the things we just talked about, breaking down data silos, better evidence faster, aligning incentives around outcomes. I think we also, equally, need to invest in the health part of it. I get really passionate about improving health holistically before patients are sick, and that comes back to really addressing access and education on the full journey, mental health, nutrition, fitness. How are we truly making our healthcare system work for patients? And that is something I think it’s going to be really important to change. 

Maria Whitman: 

That’s really insightful, and I couldn’t agree more. So what’s next? What’s next for you? What’s next for Flatiron? 

Carolyn Starrett: 

We have an ambitious agenda for the next decade. I think we just talked about it, so much possibility to reimagine the infrastructure of care, transform how we research new medicines and how we optimize the care and treatment available for patients. And we’re just really excited to be on that journey and see some of these possibilities we talked about truly come to fruition. 

Maria Whitman: 

Come to fruition for the patients and for the system, and it’s a really wonderful ambition. Carolyn, thank you so much for joining me today. 

Carolyn Starrett: 

Thank you so much for having me, and thanks for all that you are doing at ZS to advance these goals. 

Maria Whitman: 

Thanks for listening to Transforming Biopharma by ZS. We invite you to subscribe and leave an iTunes review. To learn more about connected health and ZS’s research, visit zs.com/futureofhealth. Thanks for listening, and I’ll see you next time.