Can artificial intelligence and big data be put to work to dramatically improve the treatment of heart disease, the top killer in every developed country? Egnite, a healthcare start-up spun out from California-based medical technology company Edwards Lifesciences, aims to do just that. In this episode of The Venture, McKinsey’s Andrew Roth talks with Joel Portice, president and CEO of egnite, and Don Bobo, corporate vice president at Edwards Lifesciences, to find out how egnite’s digital platform, CardioCare, is harnessing the power of AI and data to identify patients who go untreated after being diagnosed with heart valve disease.
Andrew Roth: From Leap by McKinsey, our business-building practice, I’m Andrew Roth, and welcome to The Venture, a series featuring conversations with legendary venture builders about how to design, launch, and scale new businesses. In each episode, we cut through the noise to bring practical advice on how leaders can build successful businesses from scratch.
Cardiovascular disease remains the number-one killer in every developed country, but only one in ten US patients with severe heart ailments received treatment in 2022. Egnite, a healthcare start-up, has ambitious plans to dramatically improve the treatment of heart disease by harnessing the power of AI and big data to identify patients who fell through the cracks after being diagnosed. Egnite was spun out from Edwards Lifesciences, a California-based medical technology company, in early 2021. Egnite’s digital platform, CardioCare, steadily gained acceptance among clinicians, physicians, and hospitals as the industry rapidly adopted new technology during the pandemic. In this episode, I speak with Don Bobo, corporate vice president at Edwards Lifesciences, and Joel Portice, president and CEO of egnite, to discuss the CardioCare platform, the relationship between egnite and Edwards Lifesciences, and the lifesaving capabilities of AI and big data.
Hi, Don and Joel. Welcome to the show. I’m very excited about this episode for two reasons. First, I believe it’s our first show where we’ll go deep on healthcare. Second, your business has an artificial intelligence aspect to it beyond ChatGPT, so we can talk about what you’re doing in AI from a healthcare perspective. I’d like to start off with the origin story of egnite and how it evolved out of Edwards Lifesciences. Don, our audience is specifically interested in those initial tough decisions you had to make on whether to keep egnite within the company or spin it out.
Don Bobo: I’ll start with a little bit of background on Edwards Lifesciences. We’re very focused on the structural heart business and have introduced a number of innovations that allow you to repair or replace heart valves using a catheter. This allows you to avoid cardiac surgery, making it a very compelling therapy for patients. As we developed those products, running clinical trials and commercializing them, we began to ask ourselves about the patient journey. And we discovered that many patients with the heart disease we treat weren’t aware they had the disease, weren’t aware they were a candidate for treatment, and didn’t know about these new innovative therapies.
We commissioned an in-house study looking at three years’ worth of patient journey data and realized that two-thirds of patients diagnosed with severe disease ended up being what we call “stuck in Middle Earth.” They received a diagnosis but weren’t being moved to the next phase of treatment under the care of a heart team. This really took us by surprise, since we assumed the majority of US patients with heart disease simply got it treated. As we began to unpack the different frictions in that process, we identified a number of very surprising learnings. First, it turns out there’s a lot of chaos around patients exhibiting symptoms and how those get diagnosed and attributed to their heart valves.
Andrew Roth: And is it correct that cardiovascular disease is still the number-one killer, at least in the United States?
Don Bobo: It’s the number-one killer in the United States and in all other developed countries. Even though there have been remarkable advances in this field, it’s still the number-one killer, with cancer being number two.
Andrew Roth: Please continue on about the chaos you discovered in the patient journey.
Don Bobo: One of the most surprising learnings was that patients, who are typically over 70, ended up experiencing a tremendous variation in their symptoms being properly attributed and then getting a confirming echocardiogram and ultimately being sent to a heart team. And as we began to take this data apart, we discovered a couple of places where we thought we could run some pilots to make a difference.
We also had a couple of forward-leaning clinicians who believed this was a problem, which was interesting, since a lot of physicians didn’t think so. Many physicians just thought, “Look, if patients are in our system, the right outcome will happen every time.” And it turns out that wasn’t the case.
So we began to seed a small group that was focused on digital solutions that would take variability out of symptom diagnosis, variability out of the echocardiograms, and variability out of the referral process. And all this was not part of our core business, so you can imagine the internal conversations around investing in this effort, where it was going to lead, and how it would work.
Andrew Roth: I can imagine the war stories on answering those questions. And I want to return to some of those war stories, because our listeners are fighting similar battles and answering similar questions. But getting back to the core problem statement—and this is not your core business—you’re talking with clinicians, looking at the patient journey and seeing variability in echocardiograms and how symptoms are tracked, and getting to the initial hypothesis on how to solve that.
Don Bobo: Exactly. And we took an interesting approach, because when we saw the first data set, most of our physician customers and referrers didn’t believe us. So we said, “What if we build a digital system, take your data, and do a retrospective analysis? We’ll show that you have the same variability issues with most of your patients as we have found elsewhere.”
That turned out to be the right approach, because it stimulated them—in many cases, for the wrong reasons—to say, “Yes, come in and look at our data. We’re sure that doesn’t happen here. We’re a top-shelf academic center and treat hundreds of these patients a year.” And in every hospital where we examined the data, we found the same problem.
That became the next stepping stone for us to ask, “What if we made a product out of this?” We realized we’d probably have to charge for it because we couldn’t give this stuff away due to regulations. That set us on a several-year journey interacting with physicians in hospitals, showing them their own data, and then providing solutions that allowed them to flag these patients instead of losing them along the journey.
Andrew Roth: I’ve heard you say that we treat only one in ten patients who have cardiovascular disease, and the rest go undiagnosed. Please keep me honest here, but isn’t the mortality rate of patients who show symptoms something like 50 percent within two years?
Don Bobo: That’s correct. We’re talking specifically about patients with heart valve disease, aortic stenosis, which has a 50 percent mortality rate at two years. And in 2022, we treated only one out of ten US patients who met American College of Cardiology/American Heart Association (ACC/AHA) guidelines for severe disease.
That realization was part of our motivation to think that while these data sets and digital tools were not our core business, there had to be something we could do better as a company in this industry. Because if you put yourself in the patient’s shoes, quite frankly it’s embarrassing that a majority of cardiac patients in the United States end up without the right recommendation and right action to get the disease treated.
Andrew Roth: So part of the problem here is diagnosing at the right time, and part of it is the actual process of treating the patient. We’ve set the stage on quite a scary, alarming problem. Joel, I’d love to hear about some of those initial customers and when you felt like you were onto something and reaching product–market fit.
Joel Portice: My background is in software and data analytics, and I spent my career building companies that use data and software to solve complex problems in healthcare. When I started talking to Don and the rest of the Edwards Lifesciences leadership team about this idea they had invested in and built out for six years before I became involved, we discussed whether it should be a stand-alone business. We agreed this was a digital health company that we should carve out to be able to think about cardiovascular patients in a broader way, to identify those patients and elevate them to improve the standard of their care.
It was, I thought, a wonderful fit and had the foundation of a great solution. The thesis Don and his team had was: If it has done well within the walls of Edwards Lifesciences, how much better could it do outside as an independent, autonomous organization? The idea was that we would be able to run faster, we would be nimbler, and we wouldn’t be competing with internal corporate entrepreneurs with other priorities. We were going to control our own destiny. Big credit goes to the Edwards Lifesciences leadership team, who did not interfere with the business, kept us at arm’s length, and treated us as a separate, independent entity so that we could go solve these problems for all cardiovascular patients.
We spun the business out in February 2021 when it already had customers, revenue, and a product. But we were forced to do all the corporate and organizational development during the pandemic, and a lot of meetings were done virtually. It was interesting, because the pandemic accelerated adoption of technology in healthcare, resulting in a paradigm shift where hospitals and device companies have put technology investments at the forefront of their strategic initiatives over the next three to five years. And so egnite has positioned itself to be at the intersection of healthcare and technology, where we’re able to use this proprietary technology to transform care delivery in the communities we serve.
Andrew Roth: It’s clear what the problem is—misdiagnosing people with cardiovascular disease—and then treating them in a way that gets them to a positive outcome. I would love to hear how CardioCare solves the problem.
Joel Portice: Think of us as an artificial intelligence, big data solution. We solve three central problems with our CardioCare digital platform, which helps hospitals manage their cardiovascular patient population by identifying and elevating the most vulnerable patients for appropriate follow-up therapy. Last year, the average growth in structural heart disease procedures in the US market was about 6 percent. The average growth of customers using CardioCare in structural heart disease procedures was 25 percent. So we had a significant above-market increase. Why? Because we use technology, we use AI, and we use data.
And the data is critically important because we have about 150 hospitals and 3,500 different site locations contributing data and medical records into this de-identified database, which we analyze with AI to identify vulnerable patients and benchmark across the country. We also have the largest database of echocardiograms in the country, more than two million, which is constantly being fed. And we use natural language processing (NLP) to read the notes on those images and combine that with the AI results from the procedural data, lab data, and all the other data we collect, so we have a real 360-degree view of the patient.
We’re constantly analyzing all that information, which our customers can view with a very elegant dashboard that’s repopulated every night, where they’ll see patients categorized according to priority. This is a fully automated and systematic process. When our users go to this dashboard, they can efficiently identify high-risk patients who haven’t been seen by anybody on the heart team, either ever or within 180 days.
Andrew Roth: Let’s stick on this point because I think it’s fascinating. You have access to a large database of patients. You’re basically, at a national or even global level, starting to set a base rate on the symptoms and profiles of potentially at-risk patients. And you’re baselining all the variable notes and comments from a wide array of different doctors and clinicians and summarizing that into who may be at risk and who hasn’t yet seen a specialist. Is that the right way to think about it?
Joel Portice: They’re either in the system or have remained undetected and unevaluated. What we do is identify these patients, whether or not they’ve been seen by the heart team, and point that out. We also help the industry organizations, device companies, and pharmaceutical companies accelerate the population of clinical trials. We do this through the identification of the inclusion and exclusion criteria, and we can comb through hundreds of thousands of patients to narrow it down to the 100, 50, or 10 of them who would be good candidates for a clinical trial.
What absolutely amazes me every time we do one of these is the number of patients who qualify as a candidate for a trial based on the screening of the data and criteria required. We find anywhere from one-third to half of the patients who are candidates for a trial have never been seen by anybody in the heart program. As Don pointed out, they’re walking around with this death sentence that hasn’t even been identified.
That’s the beauty of AI. That’s the beauty of technology and the application of data, and how it can elevate the standard of care and reduce variability of care. That’s what got me into this when I was talking to Don and the team, because I thought, “Oh my gosh, this should be in every hospital in America.”
Andrew Roth: I like that visual aspect of the dashboard. Because for someone who’s not an expert in this space, it helps me situate this. That dashboard is basically saving time, and saving time is saving lives.
Joel Portice: It’s very elegant, it’s very direct, and it’s very intuitive.
Andrew Roth: Don, it sounds like you were operating the business for a number of years. I met with a client just last night to talk about funding a business, and we discussed things like which business unit will fund it, if there’s a need to raise money over the next five years, and whether to pull in outside investors and spin it off. Can you take us inside the boardroom on some of those tough conversations? What were some of the considerations you were perhaps debating in the boardroom about spinning it out?
Don Bobo: This was an interesting one, because when we got into this and decided to fund it, build the team, and run pilots, we were singularly focused on the patient impact. We thought it would be helpful if we could remove a bunch of these wrinkles that patients experience trying to get from symptoms to diagnosis to treatment.
What we realized as we stepped into the world of digital health were the important considerations it raised, like the product iteration cycle. It’s very different for a digital product than it is for a replacement heart valve, which can take ten to 12 years to develop, do clinical trials, and finally launch. With a digital product, if you aren’t updating the software every month, you’re behind. Our quality and development systems weren’t built for that. So we carved out a space inside Edwards Lifesciences and said, “You’ll be separate; go do what makes sense.”
The business model for how you charge and the entire commercial relationship are also very different. And the digital business models have evolved dynamically in the past six years.
Then you need to figure out your approach. Our business was typically rewarded for having a great product and great evidence, which the market adopted over time because it was the best answer for the patient. But digital doesn’t quite work that way; it often ends up being a land grab. Whoever can get out there and get everybody on their system ends up with an advantage, because the cost of switching is high.
All of that was new to us; we were students, building while we were flying the plane. But it forced us to shift from a pure patient-benefit focus to something that’s broadly scalable. We started having candid debates around who would be the best owner and engaged a couple of outside groups to help us think about it. While we concluded this was a great idea and a great product, we also realized Edwards Lifesciences was probably the worst possible owner. We’d slow it down. We’d plant a tree, and then we’d pick the tree up every day to see how much the roots grew, which really impairs growth.
So we made the hard decision to move egnite outside of Edwards Lifesciences. We were willing to fund most of it but knew we had to let the company go and be a digital innovator. They needed to run their own business model, team, and processes in a way that’s tailored for a digital health company, not a large implant company.
Although we said we’d take a couple of board seats and fund it, we intentionally separated egnite from all of the Edwards Lifesciences governance, systems, and structure. That was a journey. Several of us, including myself, the CEO, and the CFO, were convinced, but then we had to make the case to the board. And they were excited about what this could do for patients and, indirectly, for the penetration of these therapies in a very needy population.
Andrew Roth: Being able to make that decision to spin it out, to help it grow, is not an easy one. Joel, please share a little bit about where you are in that journey. Among our audience, and especially our clients, there’s no lack of incumbents launching new businesses or even spinning out new ones. But there is a lack of product–market fit and scale. We’d love to hear where you are on the journey when it comes to scale and how you think of the business based on your expertise in software.
Joel Portice: I would advise everyone never to launch a business during a pandemic, especially a healthcare business. But in all seriousness, that has clearly been one of the biggest headwinds. Hospitals have been under tremendous financial pressure over the past few years; we’re certainly seeing them coming out of this, and we’re sensitive to their operating expense pressures.
We’ve tried a couple of different sales models, and we’ve made a couple of adjustments, but we know the value proposition of the platform is extremely compelling. And our biggest group of evangelists in the marketplace are the clinicians and their nurse navigators—the folks who work daily with patients. The clinicians, in particular, recognize how this platform enhances their practice.
Then it gets into finding the right economic arrangement. You can’t give the product away, because it’s very valuable, so it’s finding that right price point and working with your hospital customers. That’s gone very well for us. We’ve been out now for just a little over two years, and I would say there’s been much stronger momentum in the past seven to eight months because we’ve had more time to demonstrate how this platform works.
The proof of the effectiveness of our solution has now been demonstrated many times, so we have lots of references from providers willing to call their peers to say this has changed their practice. I’ve been on some of those calls, and it’s really inspiring because we’ve got a smart group of developers and a smart product team, and they’re working every day to see this penetration happen.
We’ve also developed a two-sided network as we move into relationships on the device side as well. That changes the narrative as we work with hospitals that want to be affiliated with certain industry partners. It’s also a big responsibility to make sure we’re delivering extreme value and exceeding all expectations for both segments of our customer base.
Andrew Roth: So on one side, you have an ecosystem of device providers, and you’re starting to serve those device providers in some shape or form, while you’re serving the hospitals as well.
Joel Portice: And in some cases, the same patient benefits. That’s really what this has always been about for the hospitals, for our industry partners, and for egnite. The patient is our North Star and always will be.
Despite this increased adoption, we recognize the reality of how long it takes to sell to hospitals and work through the different processes and functional areas in order to complete these arrangements. But we don’t ever hear no, so we just have to be patient as we go through the process. We’re very deliberate and methodical about how we qualify these institutions to make sure they’re committed to deploying a digital solution. And in every case, they always are once they understand the impact it has on their practice.
Andrew Roth: Keying off your mention of the patient always being the North Star, whenever our clients launch a new business, we always stress the importance of understanding the “North Star metric.” We’re not talking about some vanity metric like the number of registered customers, but a core product value that is expressed in the function of time. I’m curious what that is for you. Going back to the dashboard visual, is it “We set up CardioCare within X number of months and we get a snapshot of Y percent of patients who we could put on a better path to a positive outcome”?
Joel Portice: Yes, we see that all the time with all of our customers. When we go into our customer sites, we have four distinctive qualities of our business. First, we have the patient as our North Star, and around that we have this workflow integration, which is part of the automated and systematic approach to identify the patients, elevate the patients, and show the patients in this very highly efficient, easy-to-use tool.
We’ve also got this massive data set that’s unique, because we found that you can’t detect heart problems with the same level of precision if you’re just looking at electronic medical records (EMR) data or echocardiogram images or notes. You have to have all of it. And from what I know, we’re the only company that does, so we feel this responsibility to make sure we get it right.
In addition to workflow integration and data, we’ve got subject-matter expertise through a number of heart experts on our team and our medical advisory board. We’re also committed to this speed to market of innovation. For example, we looked at a smattering of 25 different sites and found about 3,000 patients who had not seen their heart specialist within 180 days, which is that “Middle Earth” population Don mentioned earlier.
Then as we move forward and the platform transitions, we think we’ll be looking at different points of care and new sites of care to treat these patients. That’s going to be important as we think about how we go after all of these patients. We don’t want to find just the very sick; we also want to find the ones who have moderate to severe illness, because the data shows that those individuals have the same mortality rates but are half as likely to be treated. In most instances, it’s only the severe cases that are getting treatment.
Andrew Roth: So using NLP and artificial intelligence and having access to a wide array of data, you’re able to bubble up patients faster to clinicians so that they can treat them more accurately and quicker.
Joel Portice: When you have technology that’s learning from looking at millions and millions of patients, it becomes incredibly proficient. Our NLP, for example, has a 99.9 percent accuracy rate. We believe very strongly in what we’re identifying, and that’s the beauty of this: it elevates patients and helps physicians understand how this helps improve their practice and how efficient and effective it can be in finding those patients and prioritizing them for immediate care.
Andrew Roth: To summarize and wrap up, I’d love to hear from both of you about what’s on the horizon. What are you excited about? Maybe you’ve got tailwinds here with how artificial intelligence has made the mainstream in the past few months. What are you excited about for egnite?
Don Bobo: From my perspective, there’s actually a lot to be excited about. Every month, Joel and I get together, and we are just overwhelmed with the learnings of hospitals adopting CardioCare, how it’s improving the patient experience and elevating the quality of treatment. But what’s super-exciting to me is how great it would be, if through our efforts and those of others, we could demote cardiovascular disease from the number-one killer to number two. I think that could be done if we were able to bring this kind of sophistication to how we manage, treat, and communicate with patients.
Andrew Roth: That’s obviously a worthy cause and purpose. Joel, what keeps you excited when you wake up every morning?
Joel Portice: What keeps me excited is the fact that we’re having a real and tangible impact on patients. I see and hear the enthusiasm in our customers and our prospects. We’ve never shown our platform and our dashboard to anybody who hasn’t said, “This is what we need.” Nobody has ever said, “No, thanks.” What gets me excited is this increased adoption of our platform and the physicians’ acceptance and embrace of using artificial intelligence. I get asked probably every couple of weeks whether we will move into other diseases. My answer is no, not now. We’re going to continue to solve this because I want to be on the team that waves that flag and says we’ve gone from number one to number two.
Andrew Roth: I think that’s a great place to end, this mission to demote cardiovascular disease as the number-one killer. As an outsider in the healthcare industry, it’s exciting that digitizing parts of the patient journey is going to have such high impact. Layering in the advantages you have with access to data and artificial intelligence is also encouraging because whenever I think about healthcare in the United States, it’s overwhelming to try to figure out how to fix it. Like Don mentioned, it feels like chaos, but it sounds like you’re bringing order to some of the chaos out there. Thank you both very much. Appreciate your taking the time.