Tatyana Kanzaveli 00:05
Yes, I truly believe that the whole MI and behavioral stuff has to be part of complex disease treatment. The problem is that there are so many problems.
Gregg Masters 00:17
PopHealth Week is brought to you by Health Innovation Media. Health Innovation Media brings your brand narrative alive via original or value-added digitally curated content for omnichannel distribution and engagement. Connect with us at www.popupstudio.productions. Welcome, everyone. I’m Gregg Masters, Managing Director of Health Innovation Media and producer, co-host of pop Health Week. Joining me in the virtual studio is my partner, colleague, and lead co-host Fred Goldstein, President of Accountable Health LLC. On today’s show, our guest is Tatyana Kanzaveli who’s making an encore appearance. Tatyana Kanzaveli is the founder and CEO of Open Health Network, a Silicon Valley-based technology company focused on AI and blockchain applications and digital health including the deployment of private label chatbot platforms. Tatyana’s career spans coding and programming to senior executive roles at nameplate companies and consulting firms, including founder and CEO in the startup space. During her 20 year career, she has been recognized as a thought leader and mentor for her ability to guide fortune 500 companies through a range of business challenges. She is a mentor at 500 startups and Richard Branson Entrepreneurs Center and serves on boards of various private companies. She is also a licensee and organizer of the highly notable TEDx Bay Area conference. She is a frequent speaker at US and international conferences on innovation, entrepreneurship, and digital health. So Fred, with that brief introduction over to you help us learn more about Tatyana’s work at Open Health Network and the launch of Constant Care.
Fred Goldstein 02:14
Thanks so much, Greg, and Tatyana. Welcome to PopHealth Week.
Tatyana Kanzaveli 02:17
I’m glad to be here.
Fred Goldstein 02:19
Yeah, it’s a pleasure to have you it’s been quite a while since we’ve gotten together and talked, I think all pre COVID perhaps the last HIMSS conference. So it’s really great to get you on and learn about what you’re doing. always fascinating stuff. So why don’t we start with a little bit, just give the audience a little sense of your background,
Tatyana Kanzaveli 02:34
oh, my God, talking about diverse background. In my past life, I was coding systems. So as a software engineer, and did a lot of data, data warehousing, knowledge management systems, whatever terminology we’ll use in the past for working with data. And then I was at Price Waterhouse. For a while, as a C-level executives working with fortune 50/100 companies on intersection of technology and business processes in SAP practice and then I decided to be an entrepreneur, and I’ve done a variety of different startups until I got diagnosed with cancer. And then I kind of dropped all the nonsense and decided to do things that matter.
Fred Goldstein 03:29
And so that I guess led you into healthcare, is that right?
Tatyana Kanzaveli 03:31
Exactly, yes. straight into healthcare,
Fred Goldstein 03:34
my understanding is you were also quite the chess player.
Tatyana Kanzaveli 03:38
Yeah, a good portion of my life. I was a professional chess player, love the game, and wish at some point, I will have time to go back and play.
Fred Goldstein 03:48
So once you got in with the cancer diagnosis, and you moved into healthcare, what did you see as the opportunities that led you to start these companies,
Tatyana Kanzaveli 03:55
I think, you know when you get cancer diagnosis, and in my case, it’s like Colon cancer, like, that’s a death sentence right there. So it wasn’t an easy set of news to chew on. So I want of course, through all kinds of different thoughts and different states of minds, but once, but I just kind of at one point, I realized that I was kind of still alive, but I was functioning like I was dead, right? So it’s like, so like, I can still enjoy whatever time I have. And at that time, all the background in technology and data, you know, kind of came back to me and I asked the oncologist, beautiful person, I was with and I asked him so like why me right? And he told me Well, it looks like it’s just bad luck. At that time. I kind of start to figure like looking at what deadlock means in there are a lot of diseases, including cancer cases where we can’t really find out why it happened to us. Even if we do everything right in your life at that time, I kind of outline to create a diagram of 360 degrees of data that can impact your health. And then I started looking at the sources of the data, they realize that we have some data, some data we don’t have, of course, even if we have access to some bits and pieces of health data, they are contained in very different systems. So you can even combine them in one place to analyze in from that I kind of start iterating on this core fundamental idea that we have to be able to bring advanced analytics and healthcare.
Fred Goldstein 05:54
Wow. So you saw this as a opportunity to connect things to provide better information to people and formed is that what really led to open Health Network?
Tatyana Kanzaveli 06:06
Right, so right now, I would say I’m probably running like a fifth company. I haven’t changed the name, but what we do is, you know, drastically change. So the very first version was like social network for oncology patients. And that moved until a product where we created puzzle pieces, and enabled developers and healthcare to develop things that got into Patient’s Sphere that was, you know, data, data, health data marketplace, and that evolved all together into Constant Care, that kind of became the it. For us right now, I can promise there’ll be a new version of that. But we you know, once we deploy things and learn things, it’s natural that you kind of deviate and adjust based on what’s needed.
Fred Goldstein 07:15
It sounds like almost a flow of ultimately expanding it out until you’re at this point with Constant Care, which is sort of like a population health type program or something like that. Is that how you would look at it?
Tatyana Kanzaveli 07:28
Um, I would say it’s more like something that, you know, I have this vision of adaptive, personalized, integrated care management system, right. So basically, I don’t like niche stuff, and I don’t like disconnected stuff. So I’m trying to enable healthcare people to get where they want to be. And when you look at complex diseases, and rare diseases, and chronic diseases, you realize that in, in most of the cases, you have multiple specialists, multiple experts involved in your care. And you you’re like, kind a rabbit. I would say, like, you know, some of you get one, one set of treatments from one specialist, and then you’re care provided by somebody else. You know how it is, right? It’s, it’s hard. So we created, I would say, highly sophisticated framework where you can dynamically create care team, and then they can create completely integrated care plan at a very detailed level, right. So it’s not just you need to exercise and eat your carrots, right. So we’ll go to the next level of details where, for example, for kids with severe cardiovascular issues, here’s how you get on on your bike, here’s your target heart rate zone, here’s how we will monitor that and how long you staying with that target heart rate zone once when you’re done with the survey. And it’s integrated with all kinds of behavioral lifestyle, specific dietary, integrated things. So every specialist can create their subset, if you will, of, you know actions that you need to do. And that includes pretty much anything that you want patients to do. And we automatically generate integrated, personalized plan for you. So you can see exactly what you need to do. You don’t need to open one app for something else. You can write something down on a paper. And we also integrate with a variety of different devices and sensors. The end then when we get all that data captures from assessment surveys, devices, it automatically gets triage and we apply AI and rules that would have been provided to us to notify healthcare team, if there are some patterns that need to be addressed. And all that data goes back to the dashboard in the EHR, at very detailed level, you see, so it’s completely round three kind of integrated from care plan, care management, that patient to devices and back,
Fred Goldstein 10:27
right, so it didn’t. So you’re it allows a practitioner or a healthcare system or others to better manage a population by individually intervening appropriately with whatever their conditions are, is that sort of it?
Tatyana Kanzaveli 10:40
Right? Right. And of course, my vision is that you add different care providers, specialists, they have access to data, and then but the the most fantastic thing is that on the back end we’re starting to get complete data set, your lifestyle, your mental health, behavioral, your heart rate, oxygen’s, sound level and humidity in where you are’ you see what I mean , right, and we can integrate that with genetics, claim, HR, and everything else. And that will, I think, will be just so unbelievably awesome, to analyze .
Fred Goldstein 11:23
So you ingest all those different data sources? And when you say having devices, etc. are you hooked up directly those devices are they coming through Validic or something like that, or individually? How are you working that
Tatyana Kanzaveli 11:35
it’s really the dance, we’re completely agnostic to pretty much anything, we do what makes sense. So we didn’t have to use Validic. In most of the cases, there is integration to devices. we created, for some areas of clinical research and studies, we had to create our own own Data Cloud due to the complexity of data that comes from not just devices, but also from sensors and everything else. If needed, we can work with Validic, our goal is to make systems in everything that actually will deliver value. So whatever makes sense, we do that.
Fred Goldstein 12:13
Got it? And you mentioned EHR, etc. So are you linked into those to be able to both receive data and feed it back into those systems?
Tatyana Kanzaveli 12:21
We haven’t seen any requirements on the receiving data. In most of the cases, just the level of detail data we capture, can you imagine like having every heartbeat push to EMR? Does that make sense? So it doesn’t, right? They, no one will ever look at it. But when we generate the summary for a patient that gets attached to the record, so if there are certain things that need to be looked at very detailed level they have, they have a place where they can look at it. But then most of the cases, you know, I’m like constantly asking health care providers, researchers, in all specialists, what we could do to make your life easier to see your life as complicated as it is, right. So in my view, working with data most of my life, it’s figuring out when they need to be alert and notified, it gave them the easiest way to get the exactly place where that triggered that exception, if you will. Okay. And my goal in life is to make easier for physicians to do what they need to do, and easy for patients to adhere to very complex treatment plan, if needed. So they can, you know, stay healthier or be treated.
Gregg Masters 13:47
And if you’re just tuning in the PopHealth Week, our guest is Tatyana Kansaveli, the founder and CEO of Open Health Network, a technology company focused on AI and blockchain applications in digital health, including the deployment of private labeled chatbots. For more information on open health go to WWW.openhealth.cc or follow them on Twitter, @GLFCEO, and @openhealth respectively.
Fred Goldstein 14:20
And so, you know, having sort of done really early iterations of something like this back in the early, late 90s and early 2000s, before we had quite a bit of this tech, and obviously, your expertise is much beyond mine in this area. Where are you getting these data sources from? And then secondarily, you talk about AI and machine learning, and what are you actually using to do that?
Tatyana Kanzaveli 14:44
So data sources, again, in most of the cases were engaged with people who have complex rare diseases, chronic diseases, right. And we’re lucky enough to work with you know, very well known people across all those, you know, it is that they specify what’s needed for, within the care plan to treat those patients in, in many cases, they come to us and say, well, we need to get oxygen level, we need to get certain heart rate data and other things and help us to find the device that in population health, like inexpensive, at the same time, get, you know, quality of the data that we need. There are a lot of different dimensions to picking the right device. You know, most of the people don’t have the latest iPhone, let’s put it that way. So we need to, in many cases, we kind of look at different devices, get in touch with manufacturing companies and try to get the right set of devices, we can’t compromise on the quality of the data. Okay, but at the same time, we need to look at the price point is to use integration in ability to pull data from that AI. Where do I start? So we in again, I don’t like just using AI terminology just for the sake of knowledge, right? In some cases, we don’t need AI where we can specify a set of rules. So that’s how they are, right? We do a lot of right. Yeah, some people.
Fred Goldstein 16:27
I love that I was about to get there. We use AI. And then I see Oh, well, it’s 1000 SQL queries.
Tatyana Kanzaveli 16:34
exactly right?
16:34
Something like that Right.
Tatyana Kanzaveli 16:36
You and I, we know, that’s not AI
Fred Goldstein 16:38
Right.
Tatyana Kanzaveli 16:40
So yeah, so we, for example, developed algorithms on hundred 20 plus needing of detailed claim data for where we can predict disease progression. Before that has been actually identified, we created a lot of sophisticated AI models that involve natural language processing and machine learning. To analyze conversations, we actually developed a way where algorithms can determine where people ready to change, transition happens from sustained to change language, which is very important to integrate in addictions, but also in complex disease management, because with them to look at adherence a very simplistic way, I’ll tell you five times a day. Thank you, Ned. That’s the adherence. It’s not right. This, in most of the cases, you ideally want to integrate behavioral into complex treatment plan, there’s if person says, I don’t care I you know, I have diabetes, but I’ll eat sugar five times a day, the fundamental issue is not reminding not, but more looking at how you can motivate those people. So in we develop various sophisticated AI kind of ways of helping people to talk through, you know, real reasoning of why they do or they don’t do things, they’re at the point where it’s the right time for intervention, and give them ways to communicate with people who can help them.
Fred Goldstein 18:28
Yeah, I know, our audience can’t see this, but I’m smiling ear to ear listening to that, because I was about to get into, are you looking at personas, readiness to change and obviously you are and that’s what’s so critical if they’re not ready to work on their smoking don’t go hammering them with a bunch of smoking messages, but let’s go work on something else that they are ready to work on. So that’s fantastic to hear that. And so that that also is a piece of your platform that you’ve built in that essentially provides that on the back end. You know, one of the issues that I’d love to hear you discuss is we recently saw last year, this issue, I think, with one of the other systems, I think might have been Optum that had some natural bias built into the system. It was a ML system. How do you sort of think that through and look to work that in your in your case? I know, you said you have 100
Tatyana Kanzaveli 19:10
in most of the cases? Yeah. We even with this standard algorithms that we’ve developed in certain things, I just can’t disclose the nature of all of the stuff we’re working on. But we even the variable to find some biases in certain things we’ve been playing with not playing with deploying, and you need to look at the source of the data, you need to create the ways of determining if there have been misrepresentation or mistreatment of the data and fix it. So we were able to do it on certain algorithms with we build Bill. It’s not trivial thing. right?. For most. For most people, it’s like, you know, they say, well, it’s a black box, we drop stuff in something magic happens and here’s the outcome. So that’s not acceptable because we need to set up frameworks on on the input side to make sure we have representative data, that kind of diverse and will minimize the mishaps of the AI algorithm. At the same time, we’ll also have to look at the outputs and figure out if it’s indicative that something is happening, go back in the retrain models, until we’re pretty sure that its functions the way it’s supposed to.
Fred Goldstein 20:39
Fantastic,
Tatyana Kanzaveli 20:40
no trivial thing, but, you
Fred Goldstein 20:41
No, not trivial at all, you know, fantastic. You’re looking at that. The other thing I know is, we’ve all been living through COVID, you’ve, I believe pivoted into that with some of your services, what did you set up for COVID?
Tatyana Kanzaveli 20:55
So within days, like when it was actually announced, that COVID, which was way later than it should have been. But we immediately using our framework released symptoms tracker, remember those days, no one knew who is sick, who is not. And so with that, we created a heat map. So we did this symptoms tracker, literally, within a couple of days, we were the first ones to put it out there in like 10s of 1000s of people start using, we release it like six languages all free, integrated with physician dashboard with rules that were set up, meaning that people could start documenting their symptoms on a daily or hourly, whatever basis, they could share with physicians, they can look at the, you know, correlations between data sets, and everything else. So that’s been done. Right now what I really want to do, and I found the right partners, to use MI kinds of technologies, motivational interviewing, AI, you know, NLP, everything that we build that I described to you a little bit before to work with anti-vaxxers. So provide point of view, we we blame those people, but we need to talk to them. And they, in most of them are in hiding, right? They wouldn’t tell you, I’m not vaccinated, and then I’m not doing it. And they wouldn’t tell you why so, you know, those are the consequences of that. So I really feel that we have a set of those in my company where we can put together a chatbot where people anonymously can go in and talk through in using our capacity to understand their concerns, talk to that through Talk, talk through those in determine the point where the kind of open to more detail conversation or get vaccinated to give them resources to do that. So this is what I want to do for COVID. And just trying to find the right ways of funding that this I think it’s important for not just for COVID for for any vaccines. And then you know, in the pop health side, engage people who disengaged in health care, most of them are not vaccinated. So there are a lot of I think meat behind that. Yeah,
Fred Goldstein 23:24
that’s that’s certainly sort of the holy grail I’ve been looking for forever. Is this whole population health approach? How How do you put the right message to the right person at the right time to get them to obviously make the decisions to do certain things behaviorally to improve their health? You know, we talked about <Acaw. This app we built years ago, that was the parrot on your shoulder, and I wanted to drill sergeant macaw. But somebody else needs the sweet Macaw with this message like you talked about. So that’s really great. You’re looking now to find funding for that.
Tatyana Kanzaveli 23:52
Yes. So we we found the researcher, very capable, the dean an expert in MI, for this type of areas, we have an engine built. So we need to kind of put the focus effort. It’s very limited funding that’s needed. But I think the impact will be huge for that piece. Yeah,
Fred Goldstein 24:15
absolutely. because like you said, you can use it for this but you can also apply that type of model once it’s proven to any of the other issues that you’re working with. So that’s that’s really great. In the last minute or so, what where do you see this AI and machine learning going in terms of healthcare,
Tatyana Kanzaveli 24:34
it’s a mainstream so, My dream is to work with organizations that will give us access to genetic data, to HR data, to claim data and device patient-reported data. We have capacity of analyzing even if we start with certain condition, you know certain disease area, but who is out there who is willing to go to that level this I think that is what’s needed. It’s not like we candidly analyze genetic data, someone is looking at claims data some ones looking at device data, and so on. So my dream is to find organization who subscribe to the same vision instead of slowly putting those data sources together and build algorithms that can tie it all together and make sense on prevention, treatment, and ultimately the outcomes.
Fred Goldstein 25:32
Well, that’s fantastic. I want to thank you so much, Tayana, for coming on PopHealth Week. It’s a pleasure to catch up with you. And we’ll definitely have to get you back on again to go a little bit deeper into some of these subjects.
Tatyana Kanzaveli 25:41
Thank you. Always nice to be here, guys. Thank you.
Fred Goldstein 25:45
And back to you, Greg.
Gregg Masters 25:46
And thank you, Fred. That is the last word for today’s broadcast. I want to thank Tatyana Kanzaveli, the founder and CEO of Open Health Network for her time and insights today. For more information on Open Health Network follow them on Twitter via @GLFCEO that’s Tatyana’s handle, and @OpeHealthN respectively. And to learn more about Open Health’s Constant Care platform, go to www.openhealth.cc. And finally, if you’re enjoying our work here at PopHealth Week, please subscribe to our channel on the podcast platform of your choice and do follow us on twitter by @ PopHealthWeek. Bye now