TechExchange Podcast: Data Fabrics and the Healthcare Industry
Welcome to our podcast, TechExchange! In this episode, we explore data fabrics and its impact on the healthcare industry. Joining us for this discussion is Chris Maresca, a veteran Silicon Valley Executive and currently Chief Technology Officer for Cyberionix.
00:00:00 – 00:05:08
STEVE STEVENS: Hello and welcome to TechExchange podcast. I’m your host, Steve Stevens. I’m the Chief Solutions Officer for Sysazzle. The Sysazzle TechExchange podcast was created to bring you highly relevant topics from some amazing technology leaders. Today, we’ll be exploring data fabrics and their impact on Health Care. Joining me to discuss the data, data fabrics, and the impact on the health care industry is Chris Maresca. He is a veteran Silicon Valley executive and has joined Cyberionix as Chief Technology Officer. Cyberionix is a leading data fabric provider for healthcare. Hi Chris, welcome!
CHRIS MARESCA: Thanks, thanks for having me.
STEVENS: Can you tell us a little bit about yourself so the audience understands your background?
MARESCA: Yeah, I’m a technologist based in San Francisco, California. I’ve been in the tech industry now for twenty-five plus years. I’ve done a dozen startups had six exits including one IPO. I’ve worked with companies of all types and sizes all the way from small business, to consumer startups, to Fortune 100 companies, helping them figure out their technology strategies and place in the world, especially as technology has been shifting industries around. And now I joined Cyberionix last year and we’re focused on the disruption that technology is doing to healthcare, like it has to many other Industries.
STEVENS: Well, speaking of disruption in a good way, there’s been a lot of focus on data and data fabrics over the last ten years. But most recently, what it can do for the health care industry. So, if we take a look at data fabrics, I mean, are we looking at a function or a technology layer that includes data integration and virtualization or just what is it?
MARESCA: I mean I think the reality is that all of us in whatever industry we’re in or even personally as consumers or just living your life in a modern country, are swimming in a sea of data. The problem, however, is that for most organizations and people it’s really hard to pull all of that data into something that’s useful and actionable. And the idea of a data fabric is to attempt to organize that data and make it similar enough much like, you know, the threads in fabric are woven together so that it becomes useful and delivers actionable insights into what is going on because it’s the needle in the haystack problem. We produce way more data than we can actually consume. There’s an insane statistic from the national Oceanographic and Atmospheric Administration which has been a big data generator as well as consumer, that’s the core of their work. One of their data scientists said even they only use 1% of the data that they generate. So, when you think about those things you think, well, what’s the matter with that picture? We’re producing all this data, and nobody’s actually using it. So, the idea of data fabric is to bring toolsets and technologies and layers of integration and virtualization, in such a way that for an end-user consumer data, who wants actionable insights, they don’t have to worry or understand where all of that is coming from and how to massage it into something useful.
STEVENS: That’s amazing. 1%. 1% and the rest of it is just noise and being collected?
MARESCA: Yeah, it’s not necessarily noise. I mean, I’ll give you another data point around that. So a few years ago, maybe four years ago, I was talking to the Facebook team that does data analytics at Facebook. So you can say, well no as a government agency so they just don’t have the money to process the data. So let’s look at Facebook. So I was talking to the Facebook data team, this is maybe four years ago, and they said yeah we process 500 million data points per day from our mobile apps and on that we run 20,000 queries per day to understand what users are doing to build better applications. And that’s very impressive and that was four years ago. And they said yeah but that’s only 1% of the data we have available because that’s all we can handle. I think that’s a really common problem. And the problem is also that you don’t know what you don’t know. You don’t know what is in the other 99% of the data that you’re missing, that may be even more useful than the 1% you’re sampling.
STEVENS: That leads me to the next question, which is just how well does the market and industry understand the data fabric?
MARESCA: I think that intuitively most people understand that the data can be very useful, but it’s problematic to consume it at scale. And especially in sort of critical areas like Health Care, you kind of want to know what’s hidden in the data.
00:05:08 – 00:10:23
MARESCA: We’ve all seen from COVID how the imperfect data leads to imperfect decisions about health care policy or you know things like lockdowns and stuff like that and it’s easy to argue one way or the other. But the reality is part of the reason why these things seem schizophrenic is because the data is imperfect and humans that are making decisions have imperfect access to the data. So everybody intuitively knows about this often times. The solution is unclear and the other reality is that it’s not easy to implement a data fabric. That’s going to give you a hundred percent coverage that might be unrealistic and impossible. But even a high percentage of coverage such that you have enough confidence that you can make cogent decisions, it takes time and effort to implement that. And I think one of the issues is that on the technologies side we understand this and certainly people who are responsible for strategic decisions inside of businesses understand this. The disconnect comes from the idea of investing a lot of money into data fabric with a view to having results in the future at some future point.
STEVENS: And that future is uncertain because they’re not really sure that they’ll be able to achieve the outcomes that they’re looking for. Where would you say people are at in the adoption curve for the data fabric? Are we still in the early stages or is it going mainstream?
MARESCA: I think the interesting thing is that a lot of industries have adopted this you know over the last ten years, it’s become common in a lot of places. Health Care is quite a conservative industry and is still largely lacking in this area. The whole pandemic has lit a fire, however, underneath of data because one of the things that’s become clear is that the fragmented landscape of data within Healthcare is not helping people make decisions around public health in particular, and that COVID has ripped a bandage off of that problem. And we’ve seen that healthcare authorities have difficulty both making decisions but also communicating to the public what’s going on and underlying all of that is data fragmentation and data isolation and the nonuniformity of data across organizations is underlying that. And it’s become clear that the industry can’t keep going like this. But, at the same time, the heavy lift that is necessary to implement a cogent data fabric, not just at a single organization but potentially across a lot of organizations in a region or even nationally is quite high. And with all the pressures on healthcare, there is substantial push back on that. But it’s coming, it’s inevitable. And recent moves by the government have been pushing Healthcare to standardize the data across systems, which is the first step to getting a more cogent way of pulling data out of the sea of data we’re all swimming in.
STEVENS: So, federal government is moving towards encouraging the industry to standardize and be more willing to share the data. But how will the data fabric benefit them? And what challenges would be associated with that?
MARESCA: Well, one of the first things, the government has done, and the federal government that is, in the United States is, as of, you know, spring of this year, they have mandated that all Health Care Providers must give patients access to their own data in such a way that the patients can then share it with others as they see fit. The issue with that, of course, is that the patient data may be in 20 different systems, all of which have different data structures or different ways of storing the data or whatnot and health care providers all of a sudden have to make this available in a standardized format defined by the government to patients so that patients can have ownership over their own data, which is a good thing. So the upstream effect of that is that the solutions providers that provide medical systems, medical technology systems, that store the data are now forced to provide that data in a standardized format so that it can then be given to patients. So, there’s sort of a chaining effect if you will. If a patient needs to have all their data in a standard format, then every other system that holds patient data needs to provide data in that format or at least there needs to be transformation tools that will extract the data out of various systems and then provide it to patients in a single format. And that is effectively what Cybex and the Cybex platform that’s built by Cyberionix does, it takes data out of systems that hold health care and patient data and then translates it into formats that are standardized and defined by the government. And by doing that, we can also give people within organizations access to that same data so that they can run analytics and other kinds of data analysis tools that will give them actionable data points that they can use in other places.
00:10:24 – 00:15:02
STEVENS: So having the rest of the world have been focused on healthcare for the last year, and I have personally witnessed, you know, a general practitioner who has his own records in his own system working with specialists who have their own systems working with hospital systems and larger systems that have their own records. Then there’s labs and imaging centers and so forth. Are you telling me that the date of fabric solution could potentially tie all of these folks together in a secure and HIPAA-compliant fashion?
MARESCA: Yes, yeah, that’s the sort of desired outcome of building a data fabric. And when you tie all these systems together, you have several potential access points so we can talk about clinicians having a single pane of glass view on all of the patient data across all the systems. We can, you know, with the appropriate access controls so that they can only look after things they’re authorized to look at. We can talk about analysts in the corporate head offices that can use that to understand where to allocate resources in the organization, again, with the appropriate access controls and data anonymization, and so on. We can also talk about population level metrics of the effectiveness of the health care being delivered on a per organization basis, but also on a regional basis potentially. So there’s a lot of potential usages of that, and of course, the current drive by the US government is to give patients access to all of that data in a standardized format. So, that is enabling all of these other functions potentially to happen within that data fabric. Of course, there’s still a heavy lift to glue together all of the various systems and normalize the data in such a way that it is available to everyone who has the appropriate access rights.
STEVENS: It sounds like a huge uplift, a huge endeavor. And I imagine in other industries, well, all industries typically are aligned with something called a value chain, where they add value to whatever inputs that they have and produce an outcome that is more valuable and hence, they are profitable and so forth, but Healthcare seems to be so fragmented. Is this the big difference for healthcare vs. other industries? Is helping to bring this fragmented value chain of healthcare, delivered to the patient and making it patient-centric, is that the big difference?
MARESCA: Well, right now, healthcare…there is a kind of a disconnect between the way healthcare providers are compensated and the outcomes for patients. So right now, most health care providers are compensated on the procedures that they administer, whether that’s lab tests or operations, or visits to an office or whatever. Each time there is an event with a patient, they’re compensated for that event. It doesn’t say anything about the outcomes for the patients. So, these compensation points are regardless of the outcome to the patient. So, there is a sort of a misalignment between what is best for the patient and what is best for the provider in terms of revenue. And there is a move to move compensation to outcomes instead to base a compensation on outcomes instead of procedures. But in order to do that, you have to have enough data to understand the outcomes effectively. Was this a good outcome or a bad outcome for the patient? And that requires data. And so when you’re getting compensated on per procedure, it’s pretty simple. You just bill for the procedure but outcomes is a little more complicated because it requires analysis of the actual medical data to understand what the outcome was for the patient and that shifts the way that Health Care organizations deliver care and it’s extremely important because health-care is insanely expensive in the United States. It’s something like 30% of the GDP is health care. And so there is an incentive for all of us to really pay for outcomes and not pay for just procedures. And so the vehicle for that again is data that will tell you that the outcome was positive or negative and tying that to compensation. So yeah, with other industries, the value chain is based around margins of various kinds often. But in healthcare, the value chain for the provider is different than what should be the value chain for the patient and that disconnect has been a problem for a long time.
00:15:02 – 00:20:03
MARESCA: It’s part of the reason why Health Care is so expensive in the United States. There’s a lot of sort of unnecessary procedures.
STEVENS: So, and a lot of those unnecessary procedures…I imagine some of that is due to, you know, legally covering yourself and I imagine some of it is due to not having access to the patient’s global records and not knowing what lab results have been. And so having to repeat tests because you don’t have access to that data. Would that be a correct statement?
MARESCA: Yeah. Absolutely. As a patient moves from provider to provider, the lack of sort of current record portability means that doctors have to request records from another organization. That’s assuming the patient even knows what the other organization is, right, that they got services at. That’s a really, really big problem. There’s another problem too is that, you know, medicine is for the sort of a not a perfect science. In other words, there’s still a lot of uncertainty in treating people. There’s some things that are relatively straightforward, that the medical profession knows perfectly well how to handle. But often times people, patients come in with very vague kind of symptoms. And so, the reaction of the medical profession is to throw a bunch of tests at them and part of this is because they don’t have really good data of what is the typical thing that we see in our region for patients having these symptoms. There is a lot of research, of course, clinical research, but that’s done, you know, in an abstract faction, you know, sometimes many years ago and, but there’s no real-time data, you know, if there’s an outbreak of when coronavirus, which is a perfect example of this, broke out in communities, then a lot of local providers treated at like the flu. And, of course, now we know that it wasn’t the flu, but because there was no system for, you know, even a hospital to track the data across various patients and have sort of real time analytics around that, it took a long time for the system to pick it up. It relied on local Health authorities and their disease tracking people to actually gather up that data from the local providers, which is a time-consuming and relatively manual process, at least until Corona it was. So, there are a lot of issues around this. I think one of the other things that we really haven’t discussed, that’s one of the other problems is that as a patient, you generate so much data. Some estimates are upwards of a terabyte of data per year per patient that the amount of data is quite frankly enormous and very difficult to analyze and the traditional solution to this is to pull all the data into a single data warehouse or a data Lake and then to run analytics or whatever on that. But if you have this much data, it’s impossible. So Cyberionix, when we were building our platform, we took the decision early on that we were only going to deal with real time data and streaming data. So, rather than trying to aggregate the data to a central place, we leave the data where it exists and we pull it together virtually in real time for people to be able to consume. But this is really important because as the data grows, moving huge amounts of data around is really unsustainable and it really doesn’t deliver this sort of real-time data that clinicians need at the point of care. And that’s one of the things we really want to try to achieve.
STEVENS: So the fragmented nature of the healthcare industry and it’s value chains, the fact that the weak link in a lot of this is the patient themselves who is a person not understanding the medical treatments and so forth and is expected to brief the next person in that line, you know, as to what tests were taking and everything, what their symptoms are, as opposed to that expert looking at what the expert did previously because the patient said yes, here’s the records. Yeah, exactly. How do we get the desired outcomes? You know, how do we get that data fabric in place and get the desired outcomes that were looking for?
MARESCA: Well, I think, you know, it’s the question of, you know, how do you eat an elephant, right? And the answer is one bite at a time. And this is very true here. And I think that for a lot of organizations really, it’s about starting with a small sort of encapsulated project, you know, maybe integrating a couple of different systems in with you know Cybex real-time platform like Cyberionix has and then iterating on that over time, you know, slowly adding more and more stuff in there and more and more pieces of data and eventually you will get a very comprehensive view of all the data in the organization. The one thing that’s difficult for a lot of organizations is isolating a starting point that will deliver something of value…
00:20:03 – 00:23:26
…but it’s not so large that it is requiring a huge lift and that for a lot of organizations is difficult. There’s also the other aspect is the clinicians are already overtaxed and they don’t want to take on yet another thing that they need to deal with. So you need to do this in sort of a transparent way that doesn’t impact their current workflows. So on that side, there’s also lots of other data in enterprises. We’ve seen, you know, billing financial data being pulled in for various analytics. We have been talking to one organization which has food banks, so they’ve been looking at tying in clinical diagnostics and prescriptions with nutrition via their food banks. So when you have a unified system that is providing access to data and moving data from one system to another, you can do these kind of powerful things where you can reach out at the community level, in the example of food banks, to provide a better outcome for patients. So, patients who come in who have chronic diabetes instead of sending them home with yet more drugs you can send them home with a prescription to go get specific items at a food bank where the food bank knows that they need to provide these things to that person. And that’s a really interesting way of looking at a holistic view of health care across all of the things that patients do and not just the 12 minutes a year they see their doctor. At the end of the day, what we are trying to achieve is better outcomes for patients. I think everyone in the medical profession is trying to do that. It’s just it’s challenged by the way the system is structured and the lack of data to even understand how to change things for the better.
STEVENS: Chris, this has been a really fascinating discussion, very informative and just amazing. I want to thank you for your time today and joining us, and appreciate all the Insight that you shared. All of this that we’ve just discussed are things that most people don’t think about, they don’t realize about, you know, all of that data that’s out there or how much of it’s actually useful at this point in time, how the fabric can be applied to fix some of these issues and the impact on our lives. Well, I want to thank you for your time today.
MARESCA: Well, thank you Steve was pleasure, glad to be here. And, you know, I think it’s an interesting challenge to solve these problems to get better outcomes for people. The health care industry is definitely lacking, but we’re also spending an enormous amount of money as a country on this. And a lot of this is about better use of resources that we were not going to drive the system into the ground because of the cost overruns and so on. So, I think it’s important and it’s a big challenge.
STEVENS: Thanks Chris. To our viewers, please subscribe here for additional Sysazzle TechExchange podcasts and please visit us online at www.sysazzle.com. That’s Sysazzle, S-Y-S-A-Z-Z-L-E.com. This is Steve Stevens, Chief Solutions Officer for Sysazzle. So long!