2020 was a banner year for Nym, which included signing new clients, such as Geisinger, as well as securing a $16.5 million round of funding from GV (formerly Google Ventures). We took a moment to speak with Sagi Horev, Nym’s Chief Medical Intelligence Officer.
A medical doctor, software engineer and researcher, Sagi served in unit 8200 of the Israeli Intelligence Corps in various technological positions and has extensive work experience in technology, including high-performance computing, networking and cyber-security. His research focuses primarily on computational and linguistic approaches as it relates to psychiatry. Sagi explores the question of language from both a clinical and theoretical prism and is deeply interested in the intersection of language, psychiatry and technology.
Sagi: Nym is an extraordinary group of experts in artificial intelligence, linguistics and medicine who came together to build something that was assumed impossible. I’m not just being hyperbolic when I say “impossible.” If you read the current textbooks in health informatics, you’ll see that even just autonomous ICD-10 coding is considered an impossible task with today’s technology. We’ve achieved much more than just ICD-10 coding with the NYM engine for a long time now, and that is without a doubt because of our remarkable team. It might sound corny but it has been my experience that there is no recipe for magic, it either happens or it doesn’t happen - and there’s magic in the nym team.
Tell us some more about your role at Nym
Sagi: I work a lot with R&D in solving interesting problems of medical chart understanding, trying to come up together with new technological solutions while maintaining our 98% accuracy rate. I also work a lot with the product side, providing the medical aspect of where we should be heading. Lastly, I work closely with clients and prospective clients, providing assistance in our support and shadowing processes.
What’s your favorite part about the position? And the most challenging part?
Sagi: As a medical doctor with a background in Computer Sciences and in linguistics, I couldn’t ask for a better match. I get to deal with the subjects I find most fascinating on a daily basis, building a product that has some of the most advanced medical chart understanding AI on the planet. The most challenging part is traveling in and exploring this uncharted territory. From day one, we only had ourselves for guidance and every time we reached a dead end in our R&D, we had to take a step back to re-evaluate everything.
What were you doing before Nym?
Sagi: I was doing research at Israel’s Sheba Hospital in Tel Hashomer, studying language structure abnormalities in different psychopathologies with advanced ML methods, as well as working on the clinical team for affordable-care psych service for youth who had been expelled from the public education system.
What sets Nym apart from the competition?
Sagi: At the moment, no one else is even close to our level of medical chart understanding. It’s a very odd feeling to see comparable products from absolute software giants fail at things that are basic for the nym engine like negation, subjectivity, confidence or medical process analysis.
Nym made some fairly interesting moves in 2020, can you give us some more information?
Sagi: Absolutely! Clinically speaking, we’ve drastically increased the resolution of data we extrapolate from medical charts. This data granularity allows us to understand exactly what happened to the patient in his encounter with the medical system. I think that’s super exciting. One of the limiting factors of evidence based medicine is how hard it is to automatically understand medical charts – and when we have a robust pipeline that solves that problem, I’m excited to see what we’re going to do next!
What is the culture at Nym like?
Sagi: I know it’s something people say in tech, but we really do have a team of superstars. Each person in Nym is a master of his or her field and a passionate explorer and solution builder. We have an atmosphere of support that allows people to grow into the positions they want for themselves. I’ve worked for numerous startups, mature companies and hospitals, and I’ve never seen anything like it.
What would you say is the coolest thing to have happened at Nym?
Sagi: Those first few runs when we got real world data and everything we’d been building for so long sprung into action and worked exactly as planned. That was such a huge feeling of accomplishment. We’ve been hearing from day one how what we’re trying to do was attempted and failed, and how it’s impossible. We were able to do the impossible quite a few times in Nym.
If you weren’t at Nym, what would you be doing?
Sagi: I’m very passionate about medicine, algorithmics and language, So hopefully something that combines the three.
Where does Nym go once it has cornered the medical coding market in the US?
Sagi: This is such a great question, thank you! We’re building an AI engine that can understand medical charts to a resolution and cross-referencing capability that exceeds human limits. I would love to see this system integrated into medical research platforms, for example – think about a physician that can ask a question about all the patients that have been in his ward, and get a full, data driven answer on the spot. The Nym engine can do this today, and we’re only getting better and more accurate as time goes by. This can change the way we do evidence-based medicine and that’s huge.
What does Nym have planned for 2021?
Sagi: We do data-driven development and that means that we address more common clinical scenarios first, because there’s more of them. We’re now getting to what’s called in data science “the long tail” of clinical data, and our projects are becoming more intricate and specialized. I see how these capabilities will match our CDI product nicely and allow some very detailed analysis of lost charges for the hospital system.
What advice or helpful information would you like to leave our readers with?
Sagi: Don’t be scared to try something other people tell you is impossible just because they say so. Who knows? Maybe you’ll be lucky enough to be part of the team that proves it was possible all along.