Welcome And Guest Introduction

David Klonoff 0:03

Welcome to Diabetes Technology Report Starts. I'm Dr. David Klonoff. I'm an endocrinologist in San Mateo, California. We have a very interesting guest today who's doing something quite unusual. I'll introduce my co-host, Dr. David Kerr, who will begin the interview.

Lukas’s Type 1 Origin Story

David Kerr 0:30

Thanks, David. And hello everyone once again. David Kerr speaking to you from Santa Barbara, California. Our very interesting and special guest today is Lucas Schuster. He's speaking to us from Austria about his new company called Syntactic. Welcome, Lucas. It's great to hear from you. Thank you very, very much, and thank you for the kind invitation. Excellent. So we like to begin these by saying to you, why diabetes? How on earth have you ended up where you are as head of this new startup, this venture? How did you become interested in diabetes?

Lukas Schuster 1:08

I'd say I'd say diabetes found me more than the other way around. So I was diagnosed with type 1 myself when I was around I think 17 years old. And during the the whole onboarding process, where in Austria you were in the hospital for two weeks, actually. And I heard it's not the same in every country, so that's why I'm bringing it up. I was handed a lot of, in my view, very ancient tools for managing diabetes. And being in IT already, I tinkering a little bit with products and making my own life easier, starting with a food database, actually, out of all things for type 1s. And then I actually got to know Frederick de Bong through a friend who was running or briefly after they founded my sugar and the diabetes app company in Vienna and got to know the team. And I was I was blown away and felt this is a good place to be. I think it helped a lot for me managing managing the burden and dampening the roughness of being diagnosed. So that's how I got into the into the well reason.

Why Sensitive Data Stays Unused

David Kerr 2:13

It's actually surprisingly common where we find that people in the diabetes technology space have personal experience of diabetes. So this brings it up to the modern day, syntactic. What is it? What is the problem you're trying to solve?

Diabetes Cockpit And Data Donation

Lukas Schuster 2:29

Yeah, so you see, I spent since 2013 I've spent, I started at my sugar and then I left after Rush Acquisition 23 mid. So I spent 10 years working on regulated devices, and a lot of that time I spent as head of data science there. So I was doing a lot of data wrangling, a lot of data analysis. We did many, many publications with partner universities around the globe. And one thing that struck me always was that there were so many great companies that had, or organizations more broadly, that had such a wealth of data accessible to them. And where we were in our knowledge of how to treat it, what how to best provide advice to people who live with the with diabetes, there was so much uncovered, like not yet uncovered. And at the same time, the tools and analyzing the data was so bothersome. And this problem became worse. I saw over the last couple of years, because technology moved forward so quickly that tools that are available are so far ahead of what often is used in research nowadays. And so to bring it in one sentence, I'd say we were very bothered by the fact that often the data that matters most is the most difficult to make sense of today. Because it's sensitive, it sits on servers where it cannot be shared, and it's often barred from the best means of analysing it. And that that is a problem for everyone living with diabetes, because it means insights on new technologies and new products are going to come to you much later than they could. That's kind of what we don't like.

David Klonoff 4:04

Well, because that's at syntactic, what are you doing to collect this type of data so that it's usable?

Lukas Schuster 4:11

Yeah, that's a good point. So it was twofold. The journey of syntactic started with a mobile with a mobile application called Diabetes Cockpit, which I built when I still was at Drosh, I started already with a project. And it was about making sense of all kinds of contextual data around your diabetes therapy. So it was one of the first applications that I knew of broadly collecting sleep, as well as workout types, as well as glucose, insulin, and carbs, of course, but also location and calendar entries and really trying to get an actual picture of what daily life with diabetes looks like. Because a lot of solutions focused on insulin, carbs, and glucose. But I mean, we all know that this is not what life consists of ultimately for people. And so we had this, it was a free application, and we gave people the chance to donate data. And uh this was kind of the foundation that we then started iterating on and building syntactic on. So this data set was given to a lot of research institutions around the globe and and is currently actively used to publish and push knowledge forward.

David Klonoff 5:12

I understand that your company uses artificial intelligence, which requires a very large data set. Are you able to collect enough data from patients donating to you, or do you have any other sources? It's a good question.

Lukas Schuster 5:25

So we we do collect quite a bit of data. It's it's hundreds of years every month that uh is being donated to us and by patients. And we use this primarily to analyze it and to create, to help other organizations and people to create new algorithms, to create new insights and new publications. So we do use anonymized data of those, of this donated data set to help enhance the platform we're building that people then can use to generate new insights. But it's really, I would say, there is so much left to discover at the moment that it's this is kind currently more than enough. We don't need to collect any more data than we already do, I would say, at least for the for the next two or three years for sure, to have so many things to do that need research, I think.

David Kerr 6:12

So, Lucas, just give us an example of the kind of question that you're able to answer. I mean, this brings us into what kind of data do you have? Is it all from devices? Do you have any social data? Do you have any economic data? And just give us a an example of the sort of question that you get asked.

Privacy Choices And Real-World Use

Lukas Schuster 6:32

Yeah, so this really depends on the type of partner we work with. I can give you a few examples. The first the first one that we worked with was around sleep, actually, particularly in the realm of closed loop algorithms, for example, of people who use them. How is sleep, sleep phases, sleep duration actually impacting your time and range, your clustering performance, and things like that? I think it's important to explain though that the platform we're building is not like particularly or only limited to the data set we collect. So we we also support other companies making sense of their own sensitive data. And we also give them kind of our data set as well, so that they if they miss something or if they have kind of blind spots, they can also fill them. And is this only type one or are it's type one and type two that we that we offer for our partners, but then also we work with many companies that work with EHR data, that work with Omics data, that work with other clinical trial data sets. And so we kind of help help them make sense of the data they already have and combine it with our Lucas.

David Klonoff 7:36

When people donate data to you, what do you do to keep it private? Yeah.

Lukas Schuster 7:40

Being stemming from so at my sugar, one of my roles was kind of being responsible for for analytics and also for some of the data collection. And Cockpit, the diabetes cockpit app that we also have for At Syntactic, has a bit of a different approach to privacy in a sense that we actually don't really collect any identifiable data at all. You don't have to sign in, you don't have to give us your name, you don't have to give us your email address, you don't even need a user account that you can forget your password to or something. And so we only ever collect clinical data, which of course there is a certain well, if you have enough of it, it is identifiable. Because if someone else has the same trace knows that David had the same, the same steps or same bolus at the same time, then there's a fingerprinting thing going on. So we collect the data plus the content that we are allowed to use it for research purposes, but we never actually collect names directly. So I don't want any data that I don't need, and to be frank, you don't need names or or other things like that to do proper research with the data, usually. At least for many use cases, you don't, I'd say.

David Kerr 8:47

Okay, Lucas, what about this is a mischievous question here. What about are you able to compare the performance of devices or the accuracy of devices between companies? Or is that just a no no-go area?

Lukas Schuster 9:02

So we do know for a lot of data that we get which device produced it, meaning which device was used by the patient. So far we've we've refrained from getting into basically comparing solutions against each other, primarily because I mean there is definitely some use to understanding performance of those devices, but we do tend to work with those companies. And so we think there's a lot of other things to be discovered that are equally worthy and not a bit less contentious, I'd say.

David Kerr 9:32

Okay, less contentiously. What about practicing clinicians who are not David Klonoffs of the world, the non-experts here? Is there an advantage for them? I'm thinking particularly about the people at the coalface in primary care, or people with diabetes between their clinic visits, is there an advantage for them to kind of be your next best friend?

Lukas Schuster 9:56

Yeah, so I mean, we we've long believed that diabetes happens between doctor visits. I I myself have kind of skipped a few of those anyway. So you take most of the decisions alone, I think. And insofar that we want to help patients directly, I'd point more to the diabetes cockpit app where we try to make sense of correlations, for example, correlating behavior like more steps or less sleep or more heartache if you if you just had had some heartbreak to your glucose essentially or your stress level, and kind of try to surface those things for you. We also offer, for example, experiments where you can try a new thing. Like let's say you think that your diabetes might be easier to handle if you take 5,000 more steps per week or per day, or whatever it might be. But there's a it's it's often not really a good way to test that and to see what actually happened. And so those are all things that we provide on our like patient-facing application that we also maintain.

David Klonoff 10:54

How much of syntactic work is involved in providing data to those who need it, whether it's companies or patients, and how much of your work is you yourselves using the data to create your own algorithms or your own decision support processes?

Prevention Research And What’s Next

Lukas Schuster 11:11

Yeah. That was a very interesting question we want we asked ourselves, Frederick and me. Um, I think it was late night after a conference on the couch in the app. And we, after a bit of discussion, more opted for providing this data to others that we felt are doing great work because we knew we couldn't probably generate a few good algorithms doing it ourselves, but we probably can spark tenfold the impact if we deliver and share this data with others that we think are on the right track and are doing great work. So I'd say primarily we're offering the data platform, which is our core business for institutions, organizations who do research on sensitive data. And then also we work with I think over 15 university partners globally that work on the data and do publications and generate new knowledge. And often those two things overlap. So sometimes those organizations also use data platforms, sometimes they don't. That's a bit up to them.

David Kerr 12:06

Lucas, can you just give us in the audience a glimpse into the future here? I guess you spend your whole life dealing with people who are living with diabetes, but I'm always wondering whether you're getting a sense that prevention, better than you know, prevention is the future in type 1 as well as in type 2 diabetes, as well in other disease states. Is that an area of interest for you now or in the future, or are you just keeping compartmentalized within the space that you're in at the moment?

Lukas Schuster 12:38

So to make sure I understand the question correctly, you're you're wanting, or you're interested in if we also want to move into preventing and research in that direction.

David Kerr 12:49

Yeah, I mean that's especially prevention of type 1 diabetes, the data about the potential value of continuous glucose monitoring, antibodies, other genetic aspects, you know, all of that stuff. And then the type 2 space, of course, looking at lifestyle interventions and so on and so forth.

Final Thanks And Where To Listen

Lukas Schuster 13:06

Yeah, I mean, that'd be the best if if no, if less people would get it or no one in this case. So, yes, I think we're we're we're working in on prevention as well, mainly through partners. So, for example, we we recently partnered with a with a large European research organization or consortium who is researching exactly that in among other things in Taiwan. And so that's that's where we have the easiest lever in our perspective to support research and prevention. Um, I'd say we do both. We offer our data set for for finding for finding better ways to treat the people that already live with it, and we offer our platform and our knowledge and and our technology to people who also research prevention, which obviously would be best.

David Klonoff 13:51

Lucas, you're doing really advanced work with your large data set and uh 50 universities working with you. That's fantastic. I hope Diabetes Technology Society can work with you someday. So I want to thank you for being here to be interviewed today. We're completing Diabetes Technology Report starts. We've had a great interview. This uh podcast is available on Spotify and on other uh websites and on the Diabetes Technology Society website. So until our next Diabetes Technology Report, I'm gonna say goodbye and thank you very much. Thanks. Goodbye, everyone. Very, very much.