This LIVE interview took place with Eirini Schlosser in Las Vegas at HLTH. Her company Dyania Health is one of the most exciting AI companies in health-tech today, and they just announced $10 million in Series A funding.
Watch on YouTube.
Eirini Schlosser discusses how Dyania Health uses AI and large language models to revolutionize patient care, specifically for cancer and heart failure patients.
It's really one of the most awesomely beneficial uses for AI that I've seen yet, and it's a product we should all be excited about.
🚀 https://dyaniahealth.com
- So, did you always know that you wanted to be an entrepreneur? - I didn't explicitly know that. I actually wanted to be a doctor and fall in the family business. However, the entrepreneurial bug was biting me. I had a cookie business as a kid. - You know what that is? - Yeah, I would make cookies. But you know what the worst was? - Yeah. - Is that I would just open up the fridge and take out ingredients and just start making them. And I would go sell the cookies to my dad's patients. - Okay. - And the deal was that they could get, their entire package of both Thanksgiving and Christmas cookies for $100. So they never had a bank. No baking, you're done. Like basically that was my selling line. And one year I made $8,000 and then my parents said I had to start paying for my ingredients. - Wow. - Business model just didn't make sense anymore. - How old were you in that house? - I was like 13. - You were 13 years old. So you know what I was doing? I was taking a hammer and I was seeing how hard I could hit my skull before I started to bleed. Basically equal origin stories there. - Oh, wow. - One of us, one of us. (laughing) (upbeat music) So we're here with one of the baddest women entrepreneurs in the game, Irini Schlosser. Today's a very big day. What just happened today? - So we just announced our series A, which was led by Halifax, TechScore Ventures and pretty significant participation from Cleveland Clinic Ventures as well, which we were really excited about. - What has this been for patients? Why has this, in my opinion, one of the best use cases for AI and one of the best use cases for a large language model? - Sure, absolutely. So let's take a cancer patient population right now. You know, I think that is something that hits home to many people. I've done a lot of work in cardiology very specifically. Heart failure is one of the top mortality that causes in the U.S. Causers, if it's not a word. The causes of mortality in the United States of America. And so basically, you know, we're looking at patients whereby their clinical characteristics are changing on a daily basis. And if their medical record has, let's say, 5,000 characters in it. And that's just the one note. That's not inclusive of their labs, all of the rest of their data that is being generated every time that they go in for a visit. Effectively, their medical record gets stored in a database and never gets read again. Because physicians are busy running from patient to patient, nurses are globally understaffed. There's just not enough humans to ever get through and sit down and spend 30 minutes reading a medical record. So if you've got 20,000 breast cancer patients, by the time a clinical team would get, even just one person getting through all of those medical records one time, you're looking at almost a year of reading for a team of people that would just read through one time, and that's it. Our system is reading through daily for the entire patient population and the changes that those patients have to their medical records. And so we're informing very important clinical questions that a physician would have about that patient and flagging them in the system in real time so that the physician knows what opportunities exist for their care pathways, as well as for them to get better care during the window of opportunity where they would benefit from that care. So we work with clinical trials where there might not be a drug that's approved. And so patients are often looking for how they can get access to a clinical trial drug. But effectively, those drugs will have a protocol whereby the patient must have finished one therapy, not yet started another, have a planned surgery in two weeks. And during that exact window, when let's say the patient's adverse events from their previous therapy have been resolved, that window of time is when the patient would respond best to the new drug. If they miss that window, they've missed the opportunity to get the best therapy at the right moment in time. That's why timing is so crucial. And so our system is reading and automating the medical record review in what is otherwise an hour and a half for a human. We do in half a second for each question that we're asking. So actually it's 0.3 seconds. - Do you see this as a milestone for you that does it make you feel proud or is it really something that just doesn't even register because you're already thinking about the next thing? - Oh, you know me so well. You kind of just like said the answer. Yeah, it's the latter, yeah. To be honest, it just feels like, do you know how when you go to the gym and there's that line of treadmills? - Yeah. - I was going on the slow one. And then now I'm on the treadmill that's like curved. And then I just hop over from treadmill to travel. - Basically that okay go music video. You've seen that one, right? - Yes. - We'll link to that one. I think those are subconscious. - Yeah. - Basically where our clinical team would have, starting with an EMR note, provided a question, answer, and a justification for that answer. That's the defined ground truth. So it's a case that gets fed to the model. We view a set of 10,000 cases as equivalent to an AI residency in a certain disease area. And so we build cumulative residencies onto our AI as it starts to advance and become triple and quadruple board certified. So we would think about this in a very different way than you would look at a generative model that's starting to generate text or summarize text. In this case, we're looking to understand and draw conclusions in the same way that a physician would. - So at an event like this, it's obviously AI in this AI that basically every other booth is AI something or other. And yet we know that it should be something of a bubble before likening it to the .com bubble. What do you think is going to separate the companies that are going to survive this and actually become valuable larger companies from the ones that won't make it? - That's a great question. So I think that it really depends on number one, when you started. I think that there is an element of just putting your head down, getting to work and not focusing as much on the buzz per se. I think there's a lot of fluff and froth in the market. And so I think that it can be sometimes distracting from any companies. And so probably the companies that won't be around later are those that do get distracted by that. And I think that in some cases that can result in from focusing too intensely on only capital raising. And now actually building substantial results. And so I think that at scale with hundreds of millions of patients, that's our mission. And make sure that no one is left behind. And I think that once that's done, we'll look at what's next. And I don't think that all companies are driven in that manner. I think they're kind of building to build and taking advantage of an opportunity in some cases. And so I think that will decide who gets there and who doesn't. So. (upbeat music) - So I know it maybe doesn't feel like it, but obviously this is a huge accomplishment. And it's a big, big day. And I think the world is soon going to recognize how different your approach is to not only your business, but your business in general. How does it feel to be one of the very few women founders on a floor like this? Is that something that you think about at all or never? - I think the world asks me to think about it, but I, it just doesn't really pop into just like go. - Well, the thing is, is we're just doing like it's like, you know, waking up and brushing your teeth. So it doesn't really feel different until you get like a smack in the face and then you're like, oh yeah, that was kind of different than the guy next to me. But then when you're a woman, you've also grown up in that environment. So you've already been trained. And I think to be completely honest, maybe it impacted me in the beginning, but then I just learned to roll with the punches. And so I think that's been thing that made other people potentially underestimate me. And it's always to be, it's better to be underestimated and to be overestimated. It started to be really like a superpower. So I would get started and it was like we're playing chess in a hurricane. - It's been an analogy. - So like, I think that's been one of the biggest capabilities for me to be able to see around corners of what's coming next. Because if I understand how other people think, but they don't know who I am, I'm at an information advantage. So. - Yep. - Well, anybody who underestimates you does so at their own tell, so congratulations again. - Thank you. - They will soon know, the world will soon know what a force you are to be reckoned with. And it's been a pleasure to watch your journey and to be a part of it in some small way. - Thank you. - Obviously, we're gonna do amazing things, no matter what. That's very clear. And the world is about to find out. So onwards and upwards for Diana Ireni Slosser is here. Thanks for joining in. Congratulations again on an amazing help. - Thank you so much for having me. Looking forward to next year. - So. (upbeat music) (cheering) (upbeat music) (cheering) (upbeat music) [BLANK_AUDIO]
