Empowering Patient Organizations with Real-World Data

February 4, 2022

One of the challenges various healthcare stakeholders face is making decisions based on limited and lagging data about the changing landscape. Komodo Health has collected a broad range of real-world data that allows it to capture a comprehensive view of patients moving through the healthcare system along with next-generation analytics to derive meaningful insights and drive decisions that improve patient outcomes. The company recently announced that it had entered into an agreement with the Chan Zuckerberg Initiative’s Rare As One network to provide software and analytic tools to help patient advocacy organizations in the network accelerate diagnoses, improve care, and advance research. We spoke to Web Sun, co-founder and president of Komodo Health, its platform technology, its potential to improve decision-making in the healthcare arena, and how members of CZI’s Rare As One network will be able to leverage its real-word data and analytic tools.


Daniel Levine: Web. thanks for joining us.

Web Sun: Thanks for having me, Danny.

Daniel Levine: We’re going to talk about Komodo Health, what its data platform enables, and a recent agreement with the Chan Zuckerberg Initiative’s Rare As One to leverage your data platform. Perhaps we can start with Komodo Health and what it does. As you think about the health data landscape today, what are the challenges that exist with the available data various stakeholders have access to and the delivery of this data to provide an accurate sense of the landscape and provide a good basis for decision making?

Web Sun: Well, for starters, Komodo Health’s mission, our North Star if you will, is to reduce disease burden. And so we’ve built our business and we operate it at the intersection between data and technology, and the way we deliver value to our partners is really largely through the way we’ve built our software solutions. Through the partnerships that we’ve formed, including those with patient advocacy organizations like the Chan Zuckerberg Initiative, also known as CZI. Now, when you think about Komodo Health and what it does, let’s start with healthcare. When we founded the company we made the observation that healthcare has always been incredibly siloed with every part of the ecosystem operating with its own narrow focus. Think about payers focusing on payer needs, life sciences companies focusing on life sciences company needs, and every one of these entities and stakeholders using their own data, their own systems. When it comes to the data, what that’s meant is that anyone looking for insights, anyone looking for patient level insights, was buying data piecemeal, often receiving it in the form of long spreadsheets, full of claims codes, CSV drops, and then manually sorting through that information, oftentimes with the help of consultants to try to draw conclusions that would inform and shape strategy, research, and business decisions. Now the underlying data itself has always been fraught with gaps. You know, laginess—long, late lag times, all of which really limits the insight and never really gets to the heart of the patient journey. And this is particularly challenging for rare disease patients and what the Rare As One initiative is focusing on: where the patient journey is oftentimes defined by what we call the diagnostic odyssey—where patients go from primary care physicians to specialists, back to primary care physicians, to other specialists—where there’s no clear diagnosis to be made or test or treatment to be administered. So, there’s always been a massive need in the market for an alternative to that legacy data aggregator model and legacy approach to understanding disease and understanding patient journeys. So, at Komodo we’re continually investing in what we call the “full stack” thesis. We start with a healthcare map where we’re consistently pulling from a variety of data sources and data types containing patient level insights, and then bringing that information together via our platform capabilities to enable internal and external solution engineers, and then really all the way through our enterprise workflow solutions. This full stack approach really reduces bias and gives us a much nearer real-time view of what’s happening across the entirety of healthcare. Now, more importantly, this means we can see not just individual data points or encounters with the healthcare system, but we can actually see fully complete patient journeys for hundreds of millions of patients in a privacy safe, fully compliant de-identified manner, which means going back to what I was just referencing around the diagnostic odyssey—we can meaningfully shorten that diagnostic odyssey to get the right patients, the right interventions, at the right time. And this holds true even in rare disease.

Daniel Levine: Well, let’s take a deeper dive on one of the concepts you just brought up, which is your healthcare map. This is kind of at the center of what Komodo does. Can you explain what the health map is, how it works?

Web Sun: Absolutely. I’ll start just by reinforcing that Komodo Health is a technology company. We were built on the premise that there must be a better way for our healthcare system to derive insights from the ever growing availability of healthcare data. Now, we have a stated mission to reduce the burden of disease, and we set out to build our healthcare map, which is now the country’s most connected and compliant view of longitudinal healthcare in the U.S. as the foundation for all the software that we build and deploy to market. So, if you think about the healthcare map itself, we are tracking the de-identified anonymized longitudinal patient journeys of just about every single person touching the healthcare system. We call it tracking census at this point, and there’s always been this huge need in the market for an alternative to the legacy data aggregator model, which never really gets to that heart of the patient journey. If your methodology and approach to defining disease is centered around pill tracking and thinking about disease as market baskets of therapeutics, you’re often going to be missing visits, missing encounters, and really only seeing snapshots of a patient’s journey at a moment in time versus our desire to truly understand the entirety of each individual’s patient journey. What we observed was the industry has had this need for a data asset that really captures the vast volume of information that’s being created every day, pulling from different entities, payer types, hospital systems, provider systems bringing together medical and pharmacy claims, lab insights, EMRs to really provide that comprehensive insight into the breadth and true patient experience, what we refer to as full patient journeys for hundreds of millions of patients. Again, that’s just the starting point where we bring together the underlying healthcare map, i.e. data assets, through our software and through our platform capabilities.

Daniel Levine: And what is someone able to do with the healthcare map?

Web Sun: That’s great. So the healthcare map is the foundation for our platform map and applications. The two ways to think about it are—I’ll start with the platform. On the platform side, we think about it as companies interested in looking at any kinds of insights, really centered around patient costs and outcomes. If you think about this current explosion that we’re experiencing and observing in digital health, the explosion of healthcare IT innovators, and providing and enabling them to leverage our healthcare map and platform capabilities, what we’re doing is dramatically lowering the barrier for all of these entities to get started in terms of building and commercializing new applications, new capabilities for patients. This is what we call, “Built on Komodo.” This capability allows our partners to leverage our nearly eight years of investment into a truly innovative and transformative technology stack that allows them to capitalize on the underlying data assets in our healthcare map, all the way through to the platform and analytics APIs, and built in analytics modules that we’ve already embedded into our platform capabilities. So, this allows more digitally sophisticated or more cloud native technologists that are capable of leveraging our tech stack and insights to get started on building and commercializing anything that they’re interested around—exploring whether it be patient level insights or analytics, and driving those analytics against our underlying healthcare map. Now in addition, we also have applications that we bring to market. And so, if you think about it, life sciences companies use different combinations of our solution suite to quickly address a wide range of use cases. I’ll start with just a couple of examples: being able to look at different patient cohorts and understand trends in diagnoses, trends in treatment pattern, changes in care interventions. All of these serve as example use cases of how life sciences companies partner with us and the value they derive.

In addition, if you think about it, clinical development teams use our platform and our software to improve patient recruitment for clinical trials by really being able to do everything from feasibility studies, changing inclusion and exclusion criteria to assess what that does to patient enrollment, and whether or not they’ll be likely to be able to simply enroll a study on time. That’s an example of value that our partners extract from working with Komodo. And obviously CZI is the most recent example. Patient advocacy groups partner with Komodo to help find patients, find doctors across the entirety of the U.S. who have experience and more importantly, success managing oftentimes challenging conditions that are hard to identify, hard to diagnose, hard to treat. And as one last example, digital marketing agencies and other kinds of what we refer to as technology intermediaries, they partner with Komodo to optimize everything from patient and provider outreach through non-personal initiatives, and through online efforts through their ability to leverage our healthcare map and accompanying platform and software capabilities. And again, these are just a select subset of examples of how we’re helping our partners across healthcare and life sciences and really don’t even begin to scratch the surface of the breadth of the use cases that we address with our partners.

Daniel Levine: Are you able to overlay other data than healthcare data to gain insights from what you’re analyzing?

Web Sun: Absolutely. I think that’s been one of the most exciting parts of this. We saw this coming explosion of data liquidity, data availability across a wide range of what we refer to as patient level features, and in our platform capability, what we’ve built in is the ability for entities with first party data to essentially join it and backbone it with Komodo’s healthcare map so that, if you think about it, every entity that has some first party data, oftentimes only sees what they see. So if you’re a small provider group you see what happens to your patients when they enter the four walls of your clinic, and then when they go back into the community, you have no idea what’s happening to them. If you’re an aggregator of social determinants of health, you see your member population, and you understand the social determinants of health associated with your member population, but you have no idea what’s happening to them otherwise, clinically, medically, nor do you have any understanding of what’s happening to other patients outside of the member population that you are tracking. The way we’ve built our platform capability really allows entities with first party data to bring their data into our platform and now have it essentially enriched with the entirety of the patient journey and all of these insights around patient costs and outcomes being appended to the initial first party data, those insights that they’re bringing into the platform to begin with. Does that make sense?

Daniel Levine: Absolutely. I think one of the frustrations in the rare disease landscape, when it comes to data, is the lack of ICD codes for many of these conditions. How do you handle that in terms of if I’m a patient organization looking to identify patients with a specific condition or physicians who treat, is there a way around that without an ICD code?

Web Sun: Great question. This is actually where longitudinality and our ability to track the fully complete patient journey really matters. Now, if you think about it, the healthcare map, because we have such a depth of data per patient, and because the data is brought together with all these other data assets, what we’re able to do is go in and build and define patient cohorts, even when a reliable ICD code disease, there’s only a couple thousand patients in the U.S. every year so the epidemiology is there’s very few patients. So the beauty of this is, if you think about that diagnostic odyssey that I was referring to earlier, these patients oftentimes don’t get confirmatory diagnoses, confirmatory labs until, on average, I believe it was five to seven years into that diagnostic odyssey. Why is that? When you look at it, you see, they initially present to a PCP who isn’t familiar with HTTR hereditary amyloidosis. So, what’s happening to these patients is they’re seeing this constellation of symptomology, this constellation of comorbidities that’s evolving over time. So the PCP will refer that patient to a cardiologist, and who’s only, by the way, looking at cardiovascular comorbidities, And then he’ll send that patient back to the PCP. Now, a couple months later that patient’s being referred off to a nephrologist, who’s really focused on endocrinology related disorders and not really thinking about the bigger picture. That patient then comes back to the PCP. And this happens over the entirety of that five to seven year window before someone looks at it and hopefully eventually says I should probably test this patient for HTTR.

Now what we did was we were able to take patients that had received confirmatory diagnoses of HTTR, we reverse engineered what that diagnostic odyssey looked like over that five to seven year window, we identified the constellation of comorbidities that they were originally presenting with, and what we did was twofold. Number one, we identified patients that fit the profile of a patient likely to receive a confirmatory diagnosis years down the line. And then number two, we identified who were seeing those patients, where they were presenting, who was managing them, and through partnership with a number of different stakeholders what we were able to do was drive massive awareness and medical education initiatives around HTTR to meaningfully improve testing much earlier in that diagnostic odyssey. And as you can imagine in situations like this, this can have a material impact on a patient’s outcomes—everything from life expectancy to better quality of life, and it’s been very exciting to see that come to life.

Daniel Levine: As a patient touches the various stakeholders within the healthcare landscape that collect data, is there a unique identifier that tracks them through all those different aspects? Or is there something you have to do in your system to identify one patient through different pieces of their journey?

Web Sun: Yeah, great question. That’s really where the power of technology and the intersection of data and software really comes to life because compliance and privacy is at the forefront of all the work that we do. If you think about what I referenced upfront at the beginning of the call, the entirety of our healthcare map is de-identified and anonymized and tokenized. So, all of those 340 plus million patients that we’re tracking on an ongoing basis across every touchpoint and encounter with the healthcare system, every one of those patients is de-identified, anonymized, and tokenized. As we are tracking them, the beauty of it is that as we’re bringing in more and more different sources, different types of data, different types of features on those respective patients, the platform itself is algorithmically linking the data and also maintaining our cert so that we are managing risk of re-identification, which is, as you would imagine, even more important when we’re talking about conditions like rare diseases.

Daniel Levine: From an end user point of view, how sophisticated does someone have to be to really get value out of what you’ve done here? You mentioned you have apps. Are these pre-programmed reports they can generate? Are they check-off-the -box of the data you want? What’s the approach, or do you need to have mastery of a programming language?

Web Sun: That’s a great question. And I will say we’ve designed our entire technology stack with different audiences in mind. So, to your exact point, we have user personas and end customers that want the easy button, right? They’ve literally said to us, I don’t really need to know what’s in the black box. I just need you to push a recommendation to me, or push an action that you are advising to a specific part of my team. And that’s all I’m really looking for. So that’s one audience. Another set of user persona is really centered around, “I have very predictable questions that I need to answer in my day to day, it’s part of my everyday workflow.” The software that we build for these customers is more in the line of what we call consumer friendly, easy to engage insights, workflow solutions. Now, to your point, what we’ve seen in healthcare is really an explosion of investment in healthcare data scientists, healthcare data engineers. As more and more people realize the transformation that’s needed in order for us to innovate across healthcare, more and more large healthcare entities are investing in those in-house capabilities. For those more technology sophisticated, more cloud native, digitally sophisticated data, sophisticated audiences—we deploy our platform capabilities, where they can essentially access all of the features and insights in our healthcare map and do everything from leveraging our built-in analytics modules as a starting point for them to then iterate on, for them to build bespoke visualizations, bespoke predictive analytics, and apply them against our healthcare map. These are two different types of audiences that we’ve essentially engineered our software for.

Daniel Levine: And what’s the business model. Is this a subscription service? Do you sell the data? Do people buy the product?

Web Sun: Yeah. Great question. We’ve built software atop our healthcare map that allows users to unearth insights from that data to help them do everything from spot gaps in care, address unmet patient needs, improve physician engagement, drive more strategic day to day business decisions. We commercialize that software via standard business to business software as a service, i.e. SAS subscription model. We have partners that license our solutions or license our platform on subscriptions ranging anywhere from 12 to 60 months. And really to that earlier comment that we were just discussing with software and platform capabilities that can be deployed to different user personas. We actually work with clients across nearly a dozen market segments. So, you can think about patient advocacy as what we would define as a market segment, risk bearing entities as a market segment, clinical research organizations as a market segment within pharma, an individual segment would be health, health economics, and outcomes research. Every one of these represents an individual segment and one of about 12 different that we operate in today to reduce disease burden. Now, examples of that include Janssen R&D. We partner with them to accelerate clinical development using real world data. Our partners at Picnic Health work with us to deploy solutions and real world evidence programs for complex illness like MS and hemophilia. Our partners on the digital side at Click Health tlicense solutions called Sentinel and Prism to drive insights for highly targeted digital healthcare marketing programs. And then, obviously, all of the excitement and work that we do across patient advocacy.

Daniel Levine: We talked about the fact that we’re here to discuss the recent announcement about what you’re doing with Chan Zuckerberg and the Rare As One initiative. Walk me through that agreement. What exactly are you providing these organizations?

Web Sun: Wonderful. First off, I just wanted to say we’re incredibly grateful to the Chan Zuckerberg Initiative for their willingness to do something different, do something innovative in an effort to help patients, particularly rare disease patients via their Rare As One network. The agreement and the partnership that we formed with Chan Zuckerberg and their Rare As One network provides all members of the network access to our healthcare map and our software. These Rare As One CZI grantees will be able to more quickly surface insights from our healthcare map on very specific, narrow, rare disease populations to understand everything from nuances and disease patterns and care trends. An example of what that might look like is you think about disease research with going back to that hereditary amyloidosis example or case study that I highlighted with longitudinal data that shows complete patient journeys. Our software will empower researchers and these specific patient advocacy groups to find new symptom patterns that can be used to dramatically shorten time to diagnosis and given the challenges of getting an accurate diagnosis as a rare disease patient, this could give patients back years or even decades of their life with an opportunity to meaningfully improve patient outcomes. In addition, if you think about it from a trial planning and physician outreach point of view, these partners will now be able to quickly surface insights on their specific patient populations to identify the providers and healthcare organizations most likely to see and have to manage those rare disease patients of their interest. And from there, these same patient advocacy organizations will be empowered to plan trial sites for effective recruiting or outreach, and medical education strategies to best engage rare disease patients and their providers much earlier in what’s often a complicated, prolonged diagnostic odyssey. The one last point I want to make is this partnership really allows us to capitalize on the fact that Komodo works across a number of discreet market segments that we are able to bring together. Now, when you operate your business at the intersection of life sciences, clinical research organizations, patient advocacy organizations, you’re really uniquely positioned to connect patients with innovative trials and potentially lifesaving novel therapeutics.

Daniel Levine: One of the costly and time consuming things that are so critical for really advancing research is putting together a natural history study. You’ve got all this historic data. Can this serve to replace that? Could it put out that kind of detailed patient histories and provide aa natural history of sorts?

Web Sun: Yeah, that’s a great question. That’s actually a use case that we are in the process of exploring with a number of our partners that operate in different discreet market segments, because that’s always been the holy grail, right? The way to short circuit challenges around enrolling patients is really to be able to leverage the data itself, to create patient cohorts, to create synthetic control arms, to create the understanding of how that population would behave and respond to different interventions, and that is something that we’re exploring in great detail through a number of our existing partners.

Daniel Levine: There’s also a lot of talk about the need for regulatory grade data. Any sense on how FDA might look at the quality of the data that you’re able to put together, and would you be able to construct a synthetic arm for a clinical trial with this?

Web Sun: I am very excited that we have some things in the works that will be announced in very short order that will speak to exciting developments on this front that Komodo has invested in very aggressively. So, I would be on the lookout for everything from publications, all the way through to co-authored PR with our partners.

Daniel Levine: Is this an economic arrangement with CZI? Is this something that you’re doing as a charitable initiative? What’s the relationship in that regard?

Web Sun: From my vantage point as a co-founder, I’ve really been struck meeting with so many patient advocacy leaders across a wide range of disease areas by how personal this mission is for them, and oftentimes by how little experience and expertise so many of them have in navigating the healthcare ecosystem, as well as how little support and how little funding they’ve oftentimes received. It really speaks to how uniquely positioned Komodo is to support them and their patients across a wide range of use cases ranging from shortening the diagnostic odyssey all the way to connecting them to potentially lifesaving clinical trials. Now at Komodo, we’ve done extensive work in rare disease, both with patient advocacy organizations and across life sciences and it really comes back to our mission to reduce the burden of these. We do not generate profits off our patient advocacy segment and we’re still incredibly excited to support these partners because of how mission-aligned these relationships are for Komodo.

Daniel Levine: So, while there, I imagine new ways that these groups will discover on how to use this data to answer questions that you may not even have contemplated yet. I thought we could just run through a few examples of how you would expect them to use this data and what the results would be. Start by explaining how they would use it to identify patient cohorts.

Web Sun: Absolutely. We have a solution called Prism, which was engineered from the ground up to address this exact use case. Now via Prism end users have this, and it’s really exciting for us because, going back to your earlier comment, you don’t have to be a clinician, you don’t have to be an informaticist, you don’t have to be a data scientist to use Prism. Prism is an offering that provides end users with the ability to very quickly and dynamically create personalized patient cohorts of interest across a wide range of features that we have on these respective patients in our healthcare map. That capability is paired with what we call Reporting Studio. What that allows us to do is change different patient features in order to dynamically create a custom cohort, and then see what that looks like in terms of everything from epidemiology to where those patients present, to how those patients are flowing through the system, to what that patient journey looks like. All of that is visible, knowable, in-app and end users are able to dynamically adjust that cohort definition, that disease definition on the fly and see all of those insights around those specific patients. As an example of this approach, our software allowed another partner to connect patients with a rare liver cancer. This was our partnership with Cholangiocarcinoma Foundation to multidisciplinary specialist providers, and to really identify referral patterns and diagnosis codes as patients traversed the healthcare system and navigated across community oncologists and the specialists practicing in the top tier academic medical centers. This allowed the Cholangiocarcinoma Foundation to deploy a nationwide database of specialists with systems and programs specifically designed to proactively identify and address gaps in care for these specific patients. It’s really powerful. It’s really exciting. And going back to my original point, you don’t have to be a data scientist or an informaticist in order to get to that level of insight.

Daniel Levine: Another way groups are expected to use this is to understand the diagnostic odyssey. How will this help enable that understanding?

Web Sun: Actually, it’s very easy for end users to see these insights because within the Prism software solution, after you’ve created that patient cohort, it shows you the entire diagnostic odyssey for those patients. You can actually see for this cohort that you’ve created what they’ve experienced longitudinally over a given time window that you get to define. So you can say, I want to see what’s happened to this patient cohort in the last 12 months, 24 months, 36 months, et cetera.

Daniel Levine: One of the things I think we often hear about is the rare disease patient odyssey and what it is in terms of times, but often those numbers come from very small studies and really surveys of patients. I’m wondering, could you use your system to get a more accurate figure on what it takes to get a rare disease diagnosis?

Web Sun: Absolutely. And this is where whether you’re looking at the HTTR example, or whether you are just looking at the underlying insights in our healthcare map, the beauty of this is we’ve always believed that patient reported outcomes, self-reported outcomes—these are a very important part of the patient experience, and that experience can be joined with the actual data, and we’ve proven that out with some of our existing partnerships where, going back to the original example, there are a couple companies out there that have member populations and they’re constantly surveying them and being able to backbone patient self-reported outcomes with what’s actually happening to those patients. As I traverse the healthcare system and go through their respective diagnostic odysseys, that’s incredibly powerful and illuminating for the partners doing that kind of work.

Daniel Levine: I think one of the other key uses for this that you would expect groups to do is connect with healthcare providers and researchers. I’m wondering if you could just explain that?

Web Sun: Yeah, no question. I think the exciting thing here is that within our healthcare map we stitch together, on an ongoing basis, all of the entities. All of the entities within our healthcare map are interconnected. Patients are connected to all the different providers. All the different providers are connected to all the different settings of care that they operate in ranging from your PCP that’s working in a strip mall all the way through to your top academic centers. Those provider systems are connected to payers, payers are connected to other entities. And so the beauty of this is it really gives you a dynamic living, breathing, understanding of all the different actors that are working in concert to serve the needs of those respective patients.

Daniel Levine: And ultimately, what do you hope comes out of this relationship with CZI and what do you think this will do to demonstrate the value of the platform to you [and] rare disease patient advocates?

Web Sun: My ultimate hope is we can help put breakthroughs into the hands of rare disease patients faster. The rare disease space, the diagnostic odyssey that so many patients suffer through is really ripe for advancements with the advent of deeper, more nuanced views into the patient journey coupled with software and advanced analytics. When you think about it, we’ve talked about this a bit today. Traditionally rare diseases have been notoriously challenging to identify; by some estimates patients in the U.S. spend roughly six years and 7.3 different providers on average, before their disease is correctly identified and diagnosed. Now rare diseases tend to follow a similar pattern where patients present with that wide range of symptoms that don’t fit neatly into a single box or a single code. And researching these diseases is complicated by the challenges of everything, from what you just referred to very small patient populations and the need to address this challenge of limited and siloed and small data sets alongside variability in clinical definitions, procedures, geographic challenges, barriers to patient recruitment and identification. As a result, these rare disease patients just end up bouncing from provider to provider racking up costs and not really addressing the real problem. We believe that the collaboration between advanced technology partners and patient advocacy groups will be a game changer for reducing the burden of rare disease and unlocking the types of insights needed to drive meaningful change for these patients, their providers, and their caregivers. And that’s my sincere hope as we as we get moving on this partnership.

Daniel Levine: Web Sun, co-founder and president of Komodo Health. Web, thanks so much for your time today.

Web Sun: Thank you so much, Danny. Appreciate the time and the opportunity.

This transcript has been edited for clarity and readability.

Stay Connected

Sign up for updates straight to your inbox.