Transforming Healthcare with Artificial Intelligence with Elad Walach, CEO at Aidoc: Audio automatically transcribed by Sonix
Transforming Healthcare with Artificial Intelligence with Elad Walach, CEO at Aidoc: this mp3 audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.
Saul Marquez:
Hey everybody! Welcome back to the Outcomes Rocket, Saul Marquez here. Today. I have the privilege of hosting the amazing Elad Walach. He is the co-founder and CEO of Aidoc, a healthcare AI startup focused on using deep learning to relieve the bottleneck in medical image diagnosis. He is an expert in AI with visionary business insights in the healthcare space. Since establishing Aidoc which is A I D O C in early 2016 Elad has led the company through three rounds of investments, raising over 60 million dollars, driven commercial availability of eight product lines, seven of which are FDA-cleared. Under Elad’s leadership, AIDOC has managed to grow and install based to five hundred global hospitals while growing the company to over one hundred and sixty employees. Elad began his career in the elite Israeli Defence Force technology program Talpiot and he led A.I. research in the Israeli Air Force where he initiated and led several teams focused on machine learning and computer vision projects. Just a phenomenal leader and also just the thought leader in this space of A.I. in health care. And I’m excited to have him here on the podcast to share more. So, Elad, welcome.
Elad Walach:
Thank you so much, Saul. And maybe just to share the big news first, now that we can share, it’s not 60 million, but we’ve raised another sixty-six million on top of that. So we’ve been happy to announce it’s going to be fairly fresh when the audience hears this, we just also see around six million dollars funding led by General Catalyst and really excited about this as well.
Saul Marquez:
Congratulations on that, man.
Elad Walach:
It’s been a hell of a journey, I would say.
Saul Marquez:
And so the promise of A.I. in imaging and diagnosis is huge, reducing variability, understanding those bottlenecks, so I’m excited to dive into what you guys are doing specifically there. But tell me, what is it that got you on this mission? What is it that lights your fire on health care?
Elad Walach:
When I finished the service after about a decade, so I was in this unique program that you serve for about a decade there I met my two co-founders, Michael and Guy. And when we all finished our service, we actually just knew we wanted to do health care for two main reasons. A, we like solving tough problems. So we wanted like a very high barrier. We didn’t want to fight three, three people in a garage starting like an app. We like this deep, deep spaces. But more importantly, we really want to feel the impact on people’s lives, the direct impact. I think every business can bring value, but it’s more of personal taste. We wanted to see those examples. And I can tell now that getting that text message from a physician that is telling me a story about how I save the patient lives or sometimes even more humorous than that, something they say you just save my ass or and when we get those, that’s really fulfilling. So I think the reason why we want to do health care is to feel that direct impact on people.
Saul Marquez:
Yeah, well said. Well said. And such an interesting beginning there, where you’re really working with just critical intelligence, needing to make time, critical decisions. I mean, for me, there’s such a parallel between military work and also health care work. Right. There are lives on the line. And if you make the wrong decisions, it’s costing people lives. It’s costing the system money. And it’s critical. So talk to us about how that translates and specifically about Aidoc. What are you guys doing to add value to the ecosystem that’s different?
Elad Walach:
Yeah. So I’ll start with my biggest lesson learned from my service. When I led AI research it was on the one hand a lot of work with the academy and know cutting edge algorithms and all of that. But what we learned is that means nothing if you can’t operationalize that. And this is where we learned, I learned with my experience on the difference between something in the lab that’s something that really works in the real world. And I think that’s one of the biggest gaps you have in AI for health care. You have things that look very cool in the lab, but when trying to operationalize them and the workflow make them work at scale, that’s where a lot of companies see challenges. And now what it is that we do with Aidoc, essentially, where this is always on a safety net that helps catch those critical patients with critical findings and help make sure they receive appropriate treatment. So the way it works, every imaging exam in the hospital is also sent to Aidoc. And we search for certain critical findings. And we have a pretty comprehensive set of diseases right now like stroke and brain bleed and pulmonary embolism or spine fractures. And if we detect a critical finding like that, we then flag the patient to the radiologist’s to the physician diagnosing damages so they could diagnose the patient right away and then treat the patient right away, and the whole idea is to increase efficiency and improve quality of care, reduce variability as well, especially across big health systems. And maybe to share just sure one used case just to get a sense of what it is that we do. So let’s imagine a busy emergency department at night. Maybe you have 50 patients at the same time. You don’t really know which patient has what. Maybe you have a patient fell down a few stairs, that patient to get everybody get their heads CT the exam because you have to look at the head CT to understand what they have and they all sit in a reading queue. So, radiologists sit in these dark rooms. They have been they have 50 patients to go through. And unfortunately, patient number 50 may actually be one of the critical findings. Most of them are normal, but number 50 maybe is the one that you want to treat right away. And unfortunately, in a busy environment, that patient can wait hours. With AI, AI runs in the background immediately on every exam. If there is, let’s say, a brain bleed, as an example, you should the patient to the top, the radiologist reviews the exam right away, diagnoses it, and then treat the patient. And that makes a big difference in their outcomes.
Saul Marquez:
That’s fascinating. So in this case, taking a look at a pile of images, digital images that they have to see, it’s risk stratification. So based on what you’re seeing here, the highest risk ones rise to the top.
Elad Walach:
Exactly.
Saul Marquez:
Fantastic. Fantastic getting ahead of those things that really I mean, there is no system to help. I think that’s why it’s aiding the doctor. A lot of people have always said, hey, you know, A.I. is going to take over. The reality is it’s not. It’s a great tool for us to use to augment what we’re doing here. And so this is a great example. And so, as you think about, you said seven-plus indications now. So what are those that have been FDA approved? I think you mentioned a couple of them.
Elad Walach:
Yeah. So we initially really focused on critical Time-Sensitive findings because the value there is immediate. You feel that you can take him by the hand and treat them better. So it’s a brain bleed. It’s a pulmonary embolism. It’s frère in the abdomen, rib fractures, cervical spine fractures, stroke. So those are the type of indications that we have. I think I covered most of them. And again, the idea in all of them is to do this risk stratification. The interesting thing is that our customers did a lot of research about what are the downstream implications of doing the right prioritization because the question could be, well, if you’re prioritizing for increasing quality, how are you improving outcomes? Is that enough? And one of the things that we really encourage our customers to do, unbiased research on how it impacts care, and some of our customers, for example, Yale, New Haven, and Cedar Sinai found they can actually reduce the length of stay. So by the that you catch them at the right time to make the critical decision. So you overall improve their outcomes. And if you improve their outcomes, potentially you could have better stay lower cost of care, et cetera, et cetera.
Saul Marquez:
Totally. Yeah, that makes a lot of sense. And those are the things that when you have a tool like this, you have to think about. And it’s not traditional thinking. You plug this system into your radiologist and hospitals workflow, downstream effects becomes very interesting. It could mean higher profits for the hospital. It could mean getting patients out sooner. That’s a really interesting point, Elad.
Elad Walach:
Exactly. When people think about there is a kind of high-level question of AI why would people kind of asking who is going to pay for this right? Is it going to be a payer play? A patient play? A provider play? And we feel that at least in the early phases, it should be focused on provider value and provider ROI. Why? Because payers, that’s fine, but it’s going to take a while. Patients, I just we’re sort of the portal for physicians doesn’t make a lot of sense. We feel that providers are the one that should adopt the technology right away. And the key to making it happen is to show downstream impact in ROI. And if you can show that that’s what we’re seeing such accelerated adoption, because if you show them it financially makes sense, it clinically makes sense, then it’s a technology that can be adapted well.
Saul Marquez:
Yeah. I mean, and I think we should maybe focus on that. Is there anything you want to highlight within that downstream in addition to what you’ve mentioned?
Elad Walach:
Well, I think what we learned is that even though the radiologist is the user, you have to look at the broader picture of the patient because unfortunately, the health care system today, we all know it’s kind of all say it’s broken. And one of the things that make it broken is that you have silos. the radiologists theoretically care only about radiology in the hospital. So when you build those products, you really have to think about the whole value chain. And I believe that you need to look at the broader picture. When I say downstream, I mean, don’t just look at the impact on radiology. Look at the impact on their time in the ED, for example. That’s something that again, as being able to show that he can reduce time in the 80s, look at the length of stay, look at those broader looking at the broader care continuum for the patient. And there, again, if you do a significant impact, then you can find those metrics and that it makes it a much no brainer decision. Radiologists can love it because it makes their job easier. It gives greater confidence, quality, peace of mind, attracting talent. And when you go to administration, you can show them well, just makes that much sense to the hospital, both clinically and financially.
Saul Marquez:
It makes a lot of sense. And I appreciate you diving into that. And yeah, when you look at this as a provider tool that helps optimize, that helps make more efficient the operations around this. And it makes a lot of sense to start there. And so if you think about this and there are provider leaders listening to this right now, how does the model work know what are you paying for? Can you walk us through that to help them understand?
Elad Walach:
Yeah. So when you go to kind of provider, one of the things we learned is that, is that you have to understand that, ok, maybe to give some history. A lot of the companies around the AI space were built around a single disease state. Typically, you know, you have the whole company around stroke or whole company or lung nodules, and that makes a lot of sense and brings value. Don’t be wrong. I’m not arguing the value because every disease is unique and important has a big financial impact. That being said, when we look at the buying behaviors of hospitals, it’s so difficult for them to buy 50 point solutions to integrate them. And how do you handle that, especially when they’re looking to three years in the future? Right. when you see the proliferation of AI exponentially growing, can I really integrate one hundred different solutions? Right. that doesn’t make a lot of sense. It’s very difficult, even like contracting is going to be challenging. So what we chose to do in terms of building a really like enterprise approach to this is to build a comprehensive solution suite. So we’re not doing one thing. Actually, the two and a half first two and a half years of the company, we didn’t build a single solution. We only built the platform. We only be the platform that can integrate, that can develop AI solutions quickly and just the core competency of developing and integrating the workflow. And over the two years after that, then we really scale up our solution. In two years we have 10 solutions that we can sell. So the whole point in our whole thesis as a company is to build this enterprise-wide coverage. Only if you do that, then hospitals can really consume AI at scale.
Saul Marquez:
That makes a lot of sense. And it’s really fascinating strategy there where you just kind of said, OK, these point solutions are great. But if hospitals are looking to invest, to take care of a lot of the disease states that they take care of, why don’t we just optimize a platform that can serve their needs? And after that, we start creating the solutions. And quickly, now you have these seven to 10, 10, I think you said disease states you could manage. It becomes something more promising for the enterprise of the hospital to address some of these issues that they’re dealing with. As you think about sort of building the company and learning the ins and outs, what would you say is the one setback that you guys have experienced that was a huge learning that’s made you guys even better today?
Elad Walach:
I think one of the things that we. It’s an angle on the downstream impact. So when we think about building solutions for health care, we divided it into three different layers. Layer number one was just built the algorithm. Layer number two is to make a product out of it integrated the workflow, make it a product. But layer number three is build a solution. And when I said solution, it means you have to understand the pain points of the research around, what KPI are you changing, and a way to measure and externalize that. So algorithm, product and solution. We always knew algorithm isn’t enough, but building the product we had the approach of if you build it, they will understand by themselves and they would kind of take a product off the shelf and learn how to make it clinically work and extract the value they want out of it. That was completely false, they did it like when they started. The point is they didn’t understand how integrated workflow is, what value they should be seeing like they could understand anything. It took us a while, but we learned we need to shift our approach o do much more consultative sales or solution-based sales where we target instead of pay, download the software, start using it, seeing the results in your workflow. How about we understand? What are your pain points? Let’s map those out. Let’s analyze. Maybe you don’t know what’s your gap turnaround time, right? So let’s map those out then. Let’s measure that value across a certain period of time. Build a solution out of the product. You think that was a really key insight. It took us about a year to figure it out.
Saul Marquez:
Man, I can imagine you guys were just one of those oh crap moments where you’re just like this is just like not working the. And then finally, boom solutions, piecing it all together, helping people connect the dots because when you have a technology like A.I. where really it’s still on the adoption curve early on, we need that. And so it’s great that you guys saw it. It’s great that you started building out some of those solutions and now the promise lies ahead for all of us. So you just got some additional funding. Congratulations on that. What would you say you’re most excited about here as you turn the corner on what’s next?
Elad Walach:
One of the things I’m really excited about is the continued product growth. So a lot of companies in our state, which is to scale up selling the hospital, is growing rapidly. You kind of I wouldn’t say freeze the product, but you reach a certain product maturity and then you’re mostly focusing on go to market. For us, it couldn’t be more different because we only barely scratched the surface of what I can do. We just start To gain trust. We got this initial entry into the world of health care that is actually being used, but there is so much more that we can do. It can start predicting diseases and do not just for certification but predicting diseases. It can drive workflow in a meaningful, meaningful way. That’s an area we started into. So I think that’s what excites me a lot, is the product development is going much deeper in the human-machine interfaces about how we can really augment and transform workflow. I think that a few years from now, every patient on every scan should have AI running on it. I am absolutely certain of that from the value we’re seeing and the aim is to reach there. So it’s going to be an exciting journey ahead.
Saul Marquez:
I love that. And what would you say the current penetration is? So if the vision is all of them, where are we at now? Five percent? Three percent?
Elad Walach:
I would say we’re about even higher. Now, I would say we’re almost 10 percent of this.
Saul Marquez:
OK, cool, cool.
Elad Walach:
So I don’t know if you saw we have the biggest private practice in the United States radiology partners. There is about 10 percent market share. There are other practices of similar sizes that are also part of the big health system. So it’s got to take maybe time to deploy so I wouldn’t say where the 20 percent yet, but I think we’re about maybe anywhere in the five to 10 percent range. And if you think that it was zero two years ago, then it’s just crazy.
Saul Marquez:
It’s amazing, I was having that conversation with Jeroen Tas over at Philips. He’s the CEO there. He was at Citi before and he was telling me, yeah, banks use AI to know if a high net worth customer is going to leave based on certain behaviors. So they risk-stratify to understand. If finance is doing it. We need to be doing this too with health stuff. Elad, you and your team are doing some extraordinary work. It certainly is exciting to really be absorbed by your vision. I love the way you’re painting the vision here and I believe in it. So just keep up the amazing work. I appreciate you jumping on and sharing with us. What closing thought would you leave for the listeners today?
Elad Walach:
I think that probably the listeners to this are all involved in health care to some extent. But I would also say that as a patient. It’s funny, I was speaking with my cleaning lady the other day and she was asking, what am I doing? I tell her about it. And she said, oh, it’s kind of it’s a technology that’s going to go out there. She kind of felt she shouldn’t care about this yet. It’s in the future. But I think that’s not the case. Actually. It’s weird for us because the adoption curve is so steep in this, but I think we should care about it. I’ll be honest, if I go to the hospitals I want to go to a hospital where they have a doc. Obviously, I’m biased towards us, but, you know, let’s not say Aidoc. Let’s say AI for imaging. It’s such a big difference in terms of care. So I think that we all as patients should start caring about how it’s used and how data drives decisions and hospitals. That’s the wave not of the future of now, like, I think that’s the evolution and transformation we’re going to see in hospitals. And I’m really excited about this space. I think it’s literally taking shape as we speak.
Saul Marquez:
Wow. Well, there you go, everyone. If you’re not thinking about this as a today thing, then you may want to reconsider because it is happening today at 10 percent adoption rate and exponential adoption curve. We’re looking to have this be part of our lives, whether it be a scan that you need tomorrow or a year from now, we’re going to be using it. So keep up with Aidoc. You could find them at Aidoc.com. What’s the best way for them to get a hold of you and your team Elad if there’s interest in learning more.
Elad Walach:
They can always reach out to me simply at Elad@aidoc or via LinkedIn. But if they want to reach the team, you know, our website has a ton of contact depending on what you want, like careers or sales or whatever. So you have all of those centralized there, but always feel free to reach out to me. I promise to be there and respond.
Saul Marquez:
But they have it, folks. A personal invitation from Elad to connect. Hey, I can’t appreciate you enough, Elad. Thank you. Keep up the amazing work. Congrats on the round. See, just phenomenal work by you and your team, my friend.
Elad Walach:
Thank you. Thank you so much. Thanks for inviting me.
Sonix has many features that you’d love including share transcripts, enterprise-grade admin tools, transcribe multiple languages, world-class support, and easily transcribe your Zoom meetings. Try Sonix for free today.
