In this podcast, Dan Housman and Craig Lipset discuss innovative and alternative approaches where Real World Data approaches are becoming increasingly integrated into clinical trials. They explore the increasing opportunity to electronically source Real World Data for clinical trials including integrating patient voices into consent and feedback of results. Craig shares a vision of how these new approaches can enable improvements that can lead to both better patient satisfaction and more efficient clinical research as a result.
Dan Housman: Hello, this is Dan Housman with the Novel Cohorts Podcast. I’m here with Craig Lipset, and we have some exciting stuff to talk about. You’ve been involved in lots of real-world data projects and R&D. What got you interested in Real World Data and tell us a bit about your background.
Craig Lipset: I’ll start with the second half. In terms of background, I was most recently the head of clinical innovation at Pfizer. I left there in 2019. I have my own advisory practice, right now working with pharma and biotech companies, growth companies, tech companies, investors, and those that are looking to bring new solutions and approaches for clinical trials and drug development. I’m on the faculty at Rutgers University in health informatics. I teach in their graduate school. I’m on the faculty at the University of Rochester in the Center for Health and Technology. What led me to this interest around Real World Data is actually a different side of my backstory, which is my own personal journey as a patient with pulmonary sarcoidosis. It’s a rare pulmonary condition. During the course of my journey as a patient 15 years ago, I started to find the value and importance of my own personal health data. At every step of the journey, I was making sure that I had access to my electronic health data, imaging data, biopsies and so on. I wasn’t sure what I was going to do with it, but I knew there would be value in that for me and it proved to be just in terms of how I was managing my time as a patient. I brought a lot of that experience into my work in R&D when I was in clinical research at Pfizer, as well as at a small venture-backed biotech company before that. It certainly influenced a lot of my thinking and appreciation for not only the impact and the value of diverse data around a patient, but also around the impact and power of a patient when they have control over their data and the ability to share.
Dan Housman: It sounds like you’ve gotten your fingers in a lot of different places. I know you’ve done a lot with R&D innovation and you’re really interested in clinical trials. How do you see Real World Data intersecting with clinical trials?
Craig Lipset: Well that’s a great question and I think it’s misleading when sometimes people post about this battle of Real World Data versus randomized clinical trials. We always have to keep in mind that most of our clinical trials are for investigational medicines, for which there is no Real World Data. They are not approved and they’re not on the market. So, what is that role for Real World Data in the clinical development and the drug development in the clinical trial process? Well, I think it starts in how we can use that data to help improve the way we’re designing and planning our studies, whether it’s testing protocol feasibility, optimizing studies, improving how we select investigator sites or improving how we’re able to potentially find individuals who may be eligible for our studies. I think the impact for Real World Data then starts to trickle into study conduct. As we think about the ways that we can replace our outdated approaches for data capture, the vast majority of clinical trial data today is still captured by having a physician or a study coordinator. They enter data into an electronic case report form that already existed in its source someplace else. It’s really very little difference today from how we did it 30 years ago, where they were transcribing the data onto paper case report forms. Today, they’re mostly just entering it into electronic case report forms. The opportunity for us to source our data electronically from Real World Data sources can be extremely powerful in terms of bringing better quality, better efficiency and lower cost into our trials. I think there are opportunities for Real World Data to improve our engagement with patients and studies. We think about strategies to include the patient in that data flow and to make it bilateral so that any of the data we’re collecting from a patient in the trial, we’re able to give back to them. There are some additional Real World Data use cases in clinical development today that are a little more on the edge but also very exciting. Our ability to use synthetic control arms and to reduce the need for patients to be randomized to control arms in a study by leveraging existing Real World Data rather than randomizing everybody in a one to one way. Otherwise, half of the patients we are finding in our studies won’t even get access to the new medicine. We can imagine how we can pull that forward even further. We’re already seeing ways that Real World Data can be used for a medicine that is approved, being able to support new labels who potentially can eliminate the need for a clinical trial. We’re able to get access to Real World Data to see if there is a new indication or if there’s enough evidence of efficacy and safety to support a new label.
Dan Housman: That’s an awesome tour de force of Real World Data applications and studies. I’m curious if you’ve seen things that are getting great traction, or any interesting projects you’re involved today with, just to give us the basic concept.
Craig Lipset: I think that the most practical space that we’re seeing Real World Data in the context of our clinical trials today is that very first use case around planning and designing for our studies. More and more pharma organizations are heavily resourcing their feasibility and optimization groups in their development organizations with a mix of clinical people and data scientists. That part of the organization is becoming extremely data hungry, whether leveraging EMR data, claims data, historical study performance data, competitive intelligence data and patient insights. Organizations want to bring together to make sure that the study that they’re launching is the right study and that it’s designed right the first time. Too many clinical trials go live only to then wind up, pausing to have a protocol amendment to correct things that probably could have been preventable. Those delays cost millions of dollars, both in terms of the delay as well as the direct cost associated with all the different resubmissions that are required for all the sites around the world. Getting it right the first time is worth a lot of money. People are investing and leveraging access to diverse Real World Data coupled with the right know-how in their organizations, to be able to use that data to tune and optimize their studies and hopefully select the right sites and locations around the world in which to run those studies.
Dan Housman: Here at Graticule we spend a lot of time trying to free up data like free text notes, genomics, radiology data. Do you see anything interesting that people are doing in the advanced Data space or things people want to do within life sciences sponsor organizations?
Craig Lipset: Yeah, I think that’s a great question Dan. I can think of two areas that jump to mind. Certainly, some want to be able to use that data for research purposes beyond a prospective clinical trial. It’s around understanding patient journey, natural history of disease, the ability to explore and try to identify new types of endpoints and measurements, to help understand why some patients may respond differently, and to help identify new ways that we can measure patient outcomes. I think we’re also trying to find ways to reduce or eliminate redundant data entry and to allow us to source our study data more directly from the place in which it was originally captured. This is what we call electronic source data. It has a name because the regulators have written guidance documents supporting this type of a data flow. For it to work, we have to gain access to it and have to be able to structure it along the way because the source data may live in the clinical world in a highly unstructured way. Our clinical trial data sets are really clean and are tightly structured. For that to work we need the processes and the validated tools that can help us to map that data.
Dan Housman: I think we talked about what’s happening today, where there’s good traction and supporting a lot of trial feasibility and understanding the planning for trials. But what do you think is the edge with the things that are far out there but maybe can become real soon?
Craig Lipset: I would say that the edge for many is the idea that we could really reduce or eliminate the need for human volunteers to participate prospectively in trials. One of the ways that we can reduce the burden on patients, accelerate development programs or to reduce costs is by leveraging existing data and leveraging artificial intelligence coupled with that existing data. Examples are ones which I started to hint at earlier. We can start with control arms and start to explore if we’re able to reduce the number of patients in a control arm so that we’re not randomizing one to one. When we talk to patients, this is a huge win. When patients know that they’re less likely to get the control arm in a trial, and keep in mind if your listeners aren’t familiar, most studies don’t have a placebo, most studies have an active control, meaning patients would still get the standard care that they would probably get from their doctor. The reason why they were taking the burden of participating in a research study is usually to try to get access to that new investigational medicine. If we’re able to reuse Real World Data to reduce or eliminate control arms, if we’re able to use AI coupled with that data to create synthetic patients for digital twins of the patients in our active arms then I think things start to get very exciting. We can start to envision more and more ways where we can be creative using data to continue to reduce the number of patients needed for our trials. A win for patients, a win for sponsors.
Dan Housman: So, Craig I love that you’re so patient centric and the reason why you got involved with Real World Data and clinical development is you yourself were a patient. What do you think the patient experience might be like in the future if we can really connect back to them? How could they experience a different world than they experienced five years from now when linked into the program of clinical development?
Craig Lipset: I think we’re asking patients to share their data more and more, which is not a bad thing. To me, patients are the ideal aggregator of their personal health data. I might have data that lives in an EMR in one hospital and more data in another. I might have claims data in different environments and you could run around behind my back and try to match that all together or you can just come and ask my permission and invite me to aggregate my data from its different sources, also asking my permission to share it. What’s nice is, thanks to initiatives like the Health and Human Services API, I as a patient am going to have more and more access to my data and hopefully in more and more consistent ways. What’s also nice is that we’ve seen pretty consistently that around nine out of ten patients, when asked to share their data for research are willing to. They just want to do so on their own terms as it relates to privacy and the protection of their personal preferences. So, as I think about that next step around patient engagement and their data and studies, first, it’s around coming to me and inviting me to share my data, and second, it’s around making data sharing a bilateral experience. I’m happy to share with you, but you as a research sponsor should share what you learned back with me. If there’s other data insights that you’re able to capture and if there are algorithms or diagnostic data that are generating more data about me, where do I click to get access to that data? Sometimes, that’s not going to happen during the trial itself if it’s a blinded randomized study. We have to manage the integrity of that. But eventually, I should get access to my own data just as I can in healthcare, through various patient portals. So, to me, that’s the direction we’re heading. More data aggregation centered around the patient and more effort to make that a bilateral experience.
Dan Housman: I feel like this patient driving their own research initiative has been an idea that’s been around for 20 years and people keep promoting that. Why hasn’t it surfaced yet? I’ll add into two years ago, people were talking about Bitcoin and blockchain and what happened to all those companies that were going to solve the world by giving people digital currencies in order to participate? Why aren’t we there yet?
Craig Lipset: I think that there is a stepwise journey that we have to take here and it’s hard for us to leap from step one to step three. I would say that step one is around transparency and access. I should be able to access all of my data wherever it resides and I should be able to share that data when and how I wish. That is step one and I think we’re really getting there. I think step two to me is around permission, that if you want access to my data, you should have to ask my permission to do so. If you are accessing my data without my explicit permission, I should have transparency in order to be able to see that. I get that with my credit report right now, I can lock access to my credit report if I don’t want others to see it and I can see audit trails in terms of who has accessed that credit report. Why can’t I do that with my personal health data? My health data, which in many cases unfortunately in the United States today is the number one source of bankruptcy, is the cost of me getting my health care. The data that comes from it is precious to me. The third step to that journey is, if people are going to monetize my data, I should be able to participate, get transparency and access, and I should get my permission. Once I have access to my data, I can share it where I wish. Once we have an agreement together that the permission is the right way to go, adding on monetization where it’s appropriate makes sense in a free market. It doesn’t mean that there needs to always be dollars attached to every transaction, but simply as a part of data permission, I think that’s going to naturally evolve and it’s not a bad thing.
Dan Housman: I am aware of this debate that comes up in Real World Data which is, should you always be asking permission? Is it okay to use what we’ve been doing for the last 20 years, which is a HIPAA waiver to allow access at scale to patients’ data or should every person in the world be asked for every bit of data? What’s your thought about navigating the transition? It’s almost a transition from fee for service to value based care. Is it that complex?
Craig Lipset: I am yet to see a headline in the news where there is a leak or an announcement that somebody is using data without permission. Anytime there is an instance where data has been used without permission, those people become the villains in the story. My biggest fear is that Facebook-Cambridge Analytica headline that’s waiting out there in our future, is the one that says your personal health data is a part of a deep web market that you don’t even know exists, where there are keys to access your data that others have in order to monetize it. You don’t have a copy of the key, and you didn’t even know the key was out there. To me, this isn’t incumbent on buyers in our marketplace like Life Science companies and others to seize control and define the future that they want as buyers. We see other large buyers in retail exert their buying power and authority. Walmart is a great example of this. If they want certain behaviors, they are the big buyer, and they get to decide that and can shift the marketplace. So, I’ll give an example of what the buyers in this market could do right now. In instances where two data sets are otherwise equivalent and both available to answer a research question, set a policy that will always opt the dataset that has the most transparency and permission attached to it. Signal to investors, innovators, entrepreneurs, and data scientists out there that if two options are available, you as the buyer will choose the option that has more permissions. We don’t need a government mandate for that to happen. It’s an opportunity for the environment and buyers who leave this environment to make the change that makes sure that our future is sustainable. The data that we’re relying on for business, for healthcare delivery from improved patient outcomes, isn’t going to try up this assumption that the data flow will last forever. It isn’t going to work against us. Instead, by defining a policy that lets the buyers set the terms doesn’t mean that you’re getting cut off from the data that exists today. We can say in instances where two otherwise equivalent have to be available, and if it’s not available, you’re still okay today. I think investors, entrepreneurs and others need to know that the buyers care. If they do, they’ll seize that opportunity, and the market will meet this
Dan Housman: Awesome. I’ll change gears to COVID-19. It’s been big this year and I’d really be interested in your thoughts. How does COVID-19 affect this space of Real World Data, and clinical trials? Has it moved forwards, backwards? What’s your thought?
Craig Lipset: Well, certainly. There’s been a lot of shift in how we’re running our studies this year. Our prospective clinical trials have been relying much more on remote monitoring and on telemedicine, just as we’ve seen in health care. Clinical trials are certainly evolved in that direction, as we see our studies start to shift more and more to locations outside of the clinic. When patients are able to participate more and more from home or other locations, we start to rely even more on finding other sources for study data. We’re not going to be able to rely as much on having a physician-patient encounter in the office where the data is going to be scribed into that case report form. We call these types of studies decentralized trials as we start to move the center away from just a hospital or clinic. As we shift that center, we start to rely on far more diversity and how our data is sourced. That could mean sensors, wearables, remote monitoring devices, and electronic blood pressure cuffs at home. It also might mean more dependence on other types of Real World Data sources and how we might start to tokenize patients to be able to access and source more of their Real World Data from various places. I think that certainly during COVID-19 we’ve seen the shift in how studies are being run. I think that the shift in general around telemedicine has been certainly catalyzed by changes in policy and reimbursement, breaking down many of the barriers that existed around state medical licenses as an example. I don’t see anything similar on the Real World Data side, but what I have seen is certainly the tremendous spirit of collaboration and sharing that’s been happening as we work together to address this global pandemic. Just as we see in initiatives like operation warp speed and how we’re developing vaccines and therapeutics faster than ever before, I think many people are starting to wonder what’s the next operation warp speed? What’s the next big data collaboration challenge we’re going to go after? COVID-19 was a great catalyst for us all to work together this year. What about Alzheimer’s? What about that next disease area with tremendous unmet medical need? It could be the next space that we light up the same spirit of collaboration and urgency together.
Dan Housman: That’s Awesome. Well, I think that you’ve covered a lot of space. Having thought over our talk today, what things do you think people should take away? What do you want to leave them as key thoughts to think about, and maybe even to act on?
Craig Lipset: The opportunities for using diverse data in the context of clinical trials and drug development programs is fast. There are some that are right at your fingertips that organizations can use today to better design and plan their studies. There are others that the regulators are very receptive to engaging around, we’re seeing that certainly this year. We’re seeing it in the words from leaders like Amy Abernethy and others. In terms of embracing innovative data sources as a part of how we’re developing our studies, to share those plans with the regulators, and to really think through our process for assuring data integrity and otherwise, I would say that for those that are looking at other options for use cases around Real World Data should consider ways to build policies that put a priority on data permissions and start to lay the framework, so that together we can have confidence in a sustainable future for accessing Real World Data to improve the lives of patients.
Dan Housman: Craig thanks so much for taking time with me today. And, I hope a lot of the things that you’ve been talking about come to pass soon.
Craig Lipset: Thanks so much for having me Dan.