Storming the Data Moats of Healthcare- A Conversation with Niko Shievaski


Electronic Medical Records (EMR) are like operating systems. They do everything, but nothing well. Like windows, hospitals need applications on top of EMRs to get real value. However, the EMR data is hard to get to. Enter Redox, a way to get data into applications. The company now has an “app store” of sorts with over 700 partners including hospitals and applications. Post Covid, Redox saw volumes down by 50%, but also an asks for novel applications for care delivery. In some sense Redox sits in a unique place reflecting the health of hospitals but also where we are headed.

This is Alan Pitt from Healthcare Pittstop, and I’m here with a good friend, Niko Skievaski. We’ve known each other for a number of years, and we were having a conversation about data, and I thought it would be really interesting to talk about how we might storm the moats of healthcare, how Redox is doing that, and what the future looks like in a post-COVID world, where things are evolving. So Niko, welcome.

Niko Skievaski: Thanks for having me.

AP: Can you tell me a little bit about you and why you built Redox? And for listeners who may be unfamiliar, can you give them a brief overview of what Redox is?

NS: Absolutely. I started working in healthcare when I worked at EPIC. While I was there, I saw a lot of health systems adopting new technology, really, to work around and with their electronic health records to do ancillary workflows or bring patients more into the mix with different sorts of user heuristics and applications. I also saw systems work with providers that had unique workflows and specialty areas, and maybe they needed special software applications to function. One of the things we observed was a huge gap between modern cloud-based technology—it’s built in largely AWS—and other cloud providers, and the status quo world of electronic health records. The technologists who work on these softwares really use, fundamentally, different types of technology that, when they speak to each other, they’re really speaking past each other in many cases. A lot of the cloud-based applications didn’t know what HL7 was, or the protocols like MLLP that it’s used to transmit it. And same on the other side: all of the new web-based technologies were not talking to a lot of the electronic health record. IT folks knew how to work with either.

NS: So, we saw a need to build this middle layer in between them, and that’s where the idea for Redox came from. We said, “What if we can actually pull data out of all these health systems, suck them up into the cloud, standardize them, make them all work the same, and then make them available over a modern web-based API? So developers can plug into that API once and then work with any health system, any EHR that they needed to?” And our big vision here is: How can we actually drop the barriers to the technology adoption in healthcare, so the right technology is actually being used, creating this technology meritocracy? That was the idea back in 2014 when we started the company, and it took us about a year to actually figure out who our customer was in this scenario.

AP: That’s always a question.

NS: Right, we sit between health systems and the software vendors that they choose to work with, so we didn’t know if we should charge health systems or the software vendors, because there’s value propositions that we create on both sides. So it’s kind of a classic two-sided network problem, where we have the supply side and the demand side. In our unique situation, we realized that charging the vendor side was actually much better for us to actually grow the network quickly. And growing the network quickly was important, too, because we wanted to generate network effects. The more vendors we have, the more health systems we get, and the more health systems we get, the more vendors we get.

NS: So we tried charging both, and then we eventually determined that charging vendors is a lot easier because the problem is more acute with them. If they can’t integrate, the health system just moves on to a different vendor, so that’s a pretty difficult problem for them to get through. We ended up charging the vendors, and then the vendors use us as their integration partner and bring us into every health system they work with. Then, basically, we make relationships with these health systems such that any cloud-based vendor that they’re working with, they send over to Redox as well, creating a channel partnership for the model with the health systems. So that’s a long, confusing way of saying, “We charge the vendors, the vendors charge the health systems, and we help the exchange in between.”

AP: Well, I’m going to dumb it down so for people who aren’t in healthcare. Fundamentally, I look at the electronic medical record, which many people may struggle with—they can’t get their records out. Doctors struggle with it, nurses struggle with it, patients struggle with it. But I always think of the electronic medical record as the operating system, but it needs applications, because it does everything—but nothing well. So fundamentally, you make those applications work much better with the electronic medical record for large healthcare systems. Is that a fair statement?

NS: Yeah—and that’s a great analogy, by the way! Yes, that’s exactly what we’re doing. The applications really run the whole gamut of digital health. So everything from telehealth to population health management and care coordination tools, even back-office sorts of tools that help out with workflows or billing or other elements like that. The electronic health record is the source of truth for clinical data, which gets turned into financial data for the health systems. So we’re basically taking all of the different pieces that users will interact with and make them work with that source of truth—that operating system, as you called it.

AP: Yes. Now, I would tell you that data is like the lifeblood of everything in healthcare today, but in the post-COVID world, things are rapidly changing. I don’t show up for work as regularly as I used to, in part because I know they don’t need other bodies around the hospital to come and get exposed to the virus. But the volumes are markedly down in my hospital—at least, as of last week or so, where we weren’t doing any elective procedures. I would imagine that given all the applications that you’re helping, you’ve seen some changes representative of what I’m seeing in my clinical practice. What does that look like from a data experience?

NS: It’s actually been really interesting. Right now, Redox is connected to about 700 health systems across the country. These systems run the gamut from very large academic medical centers, all the way to small-scale nursing facilities and specialty clinics. So we see a wide breadth of data across the industry. We’re by no means getting a majority of the data, but we’re seeing a lot of diverse sets of data. And first off, one of the big things we’re saying is that our volumes are down about 50%. For instance, we had a lot of health systems we work with. We get ADT data, which is…

AP: ADT is like your name, your birth date, what insurance you have, admission, discharge transfers, the trunk of the tree of all data in healthcare.

NS: Yes. It’s the data that we get every time a patient moves within the healthcare organization—when a patient is admitted, or moved around, or discharged. Let’s narrow it down and look at just the discharges from the hospital. At the beginning of this year, we were averaging about 12 million discharge messages per week, and that accelerated to about 14 million at the end of February. So it went up a little, and then it took a huge drop. Beginning March 2nd is really when we saw that drop start, and it bottoms out at around 6 million toward the end of April.

NS: These data corroborate a lot of the narratives that we hear about in healthcare. What I see that as, is that at the end of February, there are slightly more discharges than normal as they’re trying to get patients out of the health systems to prepare for the surge that we all expected from COVID-19. And then, because patients are largely avoiding healthcare organizations right now unless they have the disease, the volumes are way down. You told me last week that your hospitals are like a ghost town right now, as both providers and patients are staying away unless they are directly impacted by COVID-19.

NS: It was really interesting seeing that in our data. One thing what we’re looking at right now is that surgical scheduling is beginning to rebound a little bit, because elective procedures are starting to come back. My neighbor lived for about eight weeks with a rotator cuff injury, and he just underwent surgery this week, because they started opening that back up again. So we’re actually starting to see that with surgical scheduling data too. Starting to see an uptick, not nearly to the places that we saw before the pandemic, but it’s a really interesting kind of corroboration of the narratives that we’re seeing.

AP: It’s interesting. You’re like the canary in the coal mine. You might be seeing how hospitals are recovering even faster in some ways than the COVID data, which may lag a bit. You get sick and then you get admitted. I think that’s an interesting insight to be able to see.

NS: Yes. Well, I think one of the really interesting questions that we’re going to be playing with for the next six months or so is, a lot of the health system revenue comes from seeing patients, obviously. So physicians are getting reimbursed from payers, and a lot of the elective surgeries involve things where margins are higher than treatment in, say, the Emergency Department. The financial implications that this drop in volume will have for healthcare organizations for the year to come will be really interesting. Because if they’re struggling more from a financial perspective, are they going to decrease the size of their staff? Are they going to reallocate budgets and stop acquiring new technologies? Are they going to stop growing? How is this going to affect the consolidation trends we’ve been seeing in the industry for a while? I’m really watching that to figure out how our health systems are actually going to be responding based on the new financial situations they find themselves in.

AP: Yes. For some of you who may listen in regularly, you know that I sometimes cover policy issues too. The numbers I’ve seen are something like a billion-dollar loss for hospital systems of 15 or so hospitals. Of course, if you have more hospitals than that—I work at Common Spirit—it’s going to be more than a billion dollars. If you’re a rural hospital, it’s going to be even worse. You might not lose as much, but your buffer is not nearly as padded. And I think it’s really going to cause hospitals to rethink what they are, because they won’t be able to do some of the things they’ve traditionally done. You mentioned to me in another conversation that not only did you see the number of transactions, the amount of data going into and out of hospitals drop, but you also were being asked to do an increasing number of things in terms of applications.

NS: Yes.

AP: Can you give me an example of that, and explain why you think that’s happening?

NS: Yes. At Redox, we are implementing all sorts of digital health solutions. We have about 200 projects that are, right now, in the works at various health systems across the country, implementing different types of software. Initially, we saw a lot of those projects get shelved, so anything non-COVID related was really put on the back burner, and these active implementations were abandoned. So that was kind of this first wave that we saw—as health systems were starting to respond, they looked at the things they had in flight and said, “We’re going to re-prioritize these.” So for a lot of the projects that we had been working on, they said, “We’re going to come back to this in six months,” or “We’re going to pause this indefinitely.”

NS: But we also saw many new technologies being accelerated to meet the new demands of these organizations. The obvious one, the one that everyone is talking about, is telehealth. How do we actually take care of patients in a remote setting? It prevents us from having to don PPE and increases capacity dramatically. For certain types of care, especially in an outpatient or behavioral health setting, telehealth really should be a primary way to see patients. So we saw a huge uptick in telehealth integrations. You can see a bunch of reports from a lot of telehealth companies; these reports are showing massive increases in their business being adopted and number of visits that they’re seeing.

NS: We’re also seeing a lot of virtual triage-type solutions. Gauss, for example, as an app called Apollo that allows patients to stay in their cars, complete a questionnaire from their phone, and then receive a QR code so that the system can basically triage these patients without having to even get them out of their car. They can invite the patient in really quickly if they need to, or if they shouldn’t be brought into the organization at all, the patient receives instructions and is sent home. Another company that we work with, EmOpti, is using a command-center triage style where the emergency room physician sits in the command center and basically drops in via video to each triage room. So of course, they’re not running around potentially exposing themselves and others to the disease, but they are also increasing their ability to see more and more patients and triage people faster. So we’ve seen an uptick in that.

NS: There have also been a lot of remote patient-monitoring applications. When patients are discharged, how do hospitals figure out how those patients are recovering after they leave the organization? For patients recovering at home, how do we make sure that they’re actually getting better and they shouldn’t be admitted, because they need more inpatient-style care? On the public health side, we’ve actually seen a lot of uptick in public health style software solutions—things like contact tracking and contact tracing. That all starts when the patient is diagnosed. When you know that the patient has a positive COVID-19 diagnosis, how do you actually identify all the people they’ve been in contact with? There are many different methods that we’ve seen start to come about from that, whether it’s just recalling from memory, all the way to things like tapping into someone’s cell phone to figure out whose Bluetooth signals they’ve been interacting with. So things like that have been quite interesting to see this increase in, these types of digital health solutions being adopted to respond to the pandemic.

AP: As you’re talking, it’s super interesting for me because I don’t know of any other place that has a more holistic view of the rapidly evolving trends in healthcare in the post-COVID world. I can’t think of another place you would go where you could holistically look at something like how a healthcare business is responding to this. I think that’s super interesting.

NS: Yes. Pre-COVID, we didn’t play much of a role in figuring out what technology is being adopted by whom, and for what reasons. But as the pandemic has affected the economy, we found ourselves needing to look deeper into the data and figure out, well, who is accelerating, and who is decelerating because of this? Where are we placing our bets as far as the companies that we’re supporting to ensure that our business can survive this as well? So, yeah, we’ve definitely been doing a lot more research in this front and I think it’s something that, like you said, is a really interesting vantage point.

AP: No, I don’t know where else I would go to get that view point. Unfortunately, hospitals are trying to own patients—they perceive them as something they want to own. I personally don’t really believe that; I think hospitals have to figure out how to stay in their lane, how to deal with acutely ill patients, and then how to do a better job partnering with all other forms of care delivery. CBS, Reidy, Walgreens, all those kinds of things are probably more the future of healthcare delivery for the consumer, whereas hospitals are going to be in an acute space. What I really want and what I think what most people want is a more holistic view of their healthcare that’s de-tethered from their hospital system. Where I can go on my journey, and how can I understand wellness? You’re hosting all these applications. Are you starting to look at how these applications talk to each other outside of the hospital, to help put together this kind of personal exchange, if you will, or personal health record? Are you starting to think about that?

NS: For sure—that’s the “golden record,” as we call it, the long-term holistic view. That’s the holy grail. If that were possible to collate, it could do a lot both for the patient who is seeking care at the time, as well as for areas like research, when people are trying to figure out what works for different types of diagnoses and things like that. One of the challenges is that, in the past, there has really been no one entity that has the proper incentives, or really the heuristics to make it possible. So if we think about providers and the organizations they work for, like hospitals and health systems, they have a lot of the clinical information that they’ve gathered for the patient when they have seen them, but that is only for that encounter or that episode of care. Patients move around to different care settings, and they move to different regions. They switch between different specialists who are not part of those networks. And so it’s a slice of that data.

NS: Well, what if payers could be the ones to capture all of that data and bring them all together? Right now, payers just get claims—and of course, claims are missing a lot of important information, things like notes, other details in the clinical chart that are vital to understanding the whole patient and what is going on with the whole patient. So you might think, “Well, maybe this should be a public good, the government could help provide that.” But we know that our government is staying far away from doing that at this point. So the only entity really left in all of this is the patient. In the past, patients lacked incentive to gather that information, because a lot of patients are not engaged in their care until they’re actually the ones sitting in the emergency department or they get diagnosed with a disease. So they have this kind of rational ignorance to their historical medical record until it’s almost too late. You have to proactively gather all of these pieces of information. Patients are not really doing that.

NS: However, something that has changed recently—with the pandemic—is that suddenly everyone has more reason to be engaged in their health. We know that those with comorbidities fare far worse with COVID-19. So it’s important to have a strong history of all of that, if you do have chronic conditions, things like that. But additionally, one of the barriers to this was, it’s just hard to collect all that information. If you see patients who have a lot of chronic conditions or comorbidities they have binders that they carry this stuff around in.

AP: I wrote a blog post. The shoebox should hold shoes, not medical records. Because shoeboxes full of papers are what I’m presented with in the clinic. Patients come in with a box full of records.

NS: I guess you’d have to have kind of a wide shoebox to fit the papers in them, but yeah.

AP: Let me ask a related question around that. One of the things that we talk about is social determinants, homelessness, food insecurity. If you actually look at the numbers, about 80% of any individual’s wellness is determined not by their drugs and doctors and hospital, but these other factors. Your organization is basically a data exchange. Your cloud-based way to exchange data, have you thought about incorporating non-traditional forms of data to help even hospital systems understand who they’re serving in a different way?

NS: Yes. We have certainly worked with applications that do a lot of that. There are a couple that come to mind—for example, when a patient is discharged from a hospital, they often need a lot of social services or advanced planning to make sure that they don’t simply come back in the next day or two. They need to end up in a place where they can actually recover from their care and have the right nutrition and shelter and all of that. We work with a few groups, Nott-house is one that comes to mind, where they actually create a database of these services and provide them to discharge planners. That way, when a patient is discharged, they can tell the patient about relevant services that they should be working with, and help the patient access these services. So we’re seeing things like that emerge a lot more. We’re also seeing groups that tackle the access side. How do you actually make sure that patients can get into the organizations? Whether it’s simple things like scheduling and making that easier to work with all the way to groups like Roundtrip, who are helping patients actually get rides to their appointments, even if the patient needs wheelchair access. Or even just doing subsidized ridesharing, because health systems often have agreements with taxi companies.

AP: Yes. I think as healthcare systems evolve, they need to have a better understanding of the true cost of “ownership” of a patient or a customer. How do I manage the total cost of ownership in the best possible care for the lowest possible cost, traditionally do what we’re paid for, drug surgery stuff like that, where the patient really needs these other services delivered to them. I have to think that if somebody’s listening here, that they worry a little bit, because you have all this data going back and forth between these applications and hospitals. Prior to COVID, of course, HIPPA was a really big deal and I’m certain that you’ve found ways at Redox to deal with HIPPA and to keep information very protected, because you’re moving all this data around. But I see a real evolution with HIPPA now, where the walls have kind of come down, and what we’re willing to accept is far different than what we were willing to accept, say, six months ago. I’m hearing people talk about wearing a wristband that indicates your COVID status so that you can go out in public or go back to work, or you want to date people that only have the blue band and not the red band. What do you think is going to happen with privacy with all this COVID stuff? I think presents a massive change in how we’re thinking about things.

NS: Yes, absolutely. Even pre-COVID, we were seeing the industry and the public bubble up around privacy. I think a lot of this was brought on by big tech like Facebook and Google getting in the limelight around things that aren’t technically illegal but that make us feel uncomfortable about them having our data. Even GDPR, what Europe did with privacy, it brings up a conversation around: Well, what about healthcare data? The law of the land right now is that health systems gather data about you as a patient so they can better take care of you. What that means is that their “business associates,” to use the HIPPA term for it, are also allowed to have your data. So they’ll sign up a bunch of business associates to help care for a patient, and all of these business associates need to have some sort of reason to have that data, and that’s largely the paradigm that Redox operates in. All of the vendors that we send data to are our business associates or the healthcare provider, whether that’s in an inpatient setting, or an outpatient setting or skilled nursing facilities or the like. They’re business associates, and of course, Redox becomes a business associate and that chain of data privacy as well.

NS: I think the way in which this new paradigm that is really interesting, we talked about this a little bit, is around patients themselves. If, as a patient, you post your medical information on Facebook (even protected health information, like if you’ve tested positive for COVID-19), which I’ve seen a lot of my friends do, that’s not a HIPPA violation. Patients are allowed to share their data with whomever they want. I think that’s where a lot of this is going in that, right now, if I have your data in Redox, I don’t know if I could give it to you. You’re not a business associate of mine, but you are the patient who has rights to that data. I would direct you to go to your health system and request a release of information. But I think that the future holds, and this is kind of what the President at GDPR said, is that you as the consumer have a right to ask for your information, no matter who has your data and for what reason. You have a right to understand what they have and to ask them to delete it. I think that’s probably the right side of history, whether that’s going to take us six months to get there, or two years, or a new administration to actually write legislation like that, but I feel like we’re moving in that direction. Recent legislation enacted by ONCNCMS mandates that health systems and peers give data directly to patients via an API, for whatever applications that they want.

AP: Yes that’s a very recent change.

NS: Yes. That was enacted, it was the Monday of HEMS, but HEMS, of course, didn’t happen. So that was in March. And we took a tiny step back in that because those deadlines got pushed back about six months right now. But that could be a slippery slope for those deadlines might start getting pushed back more and more depending on how whole systems recover. But the idea is that patients they don’t derive value from data. Patients derive value from the information that can be interpreted from that data, and that’s where the application layer comes in. That’s why APIs are still important here—because an API turns the data into something an application can use. And the application is what a patient is going to be interacting with, and that application is going to be the one saying to the patient, “Hey, you should check out this thing that we’re seeing in your data. Talk to your doctor about that.” Or engage patients in all different sorts of ways, or engage patients via different interactions that they can have with each other, with health systems, with any care provider they choose.

NS: I think we’re on the cusp of this, both giving patients and consumers more rights around their data, but also enabling them to act more like consumers in their own healthcare. Because they have their data, they can shop more effectively, they can understand what how their behavior is actually going to change their health outcomes, because combining clinical data with the Internet of Things—wearables and all that sort of stuff—will hopefully be able to derive statistically significant insight around, “How does my sleep, my nutrition and diet, my exercise behavior affect other things that might be going on in my life?”

AP: Yes, I think you raise a really interesting thing. Patients, I don’t think, really care about their data. They just care about their health and their wellness, and they need an interpretation layer where they can better understand—for example, “Hey, if I do this, then that happens,” or “I do this, then that happens.” I call it a negative good: Most people don’t really want to know about their health. They just want to be well, and pain-free, and not die. Stuff like that.

NS: Yeah. It’s what we call in economics derived demand, and it’s actually a big contributor to a lot of the market failures in healthcare. Basically, consumers don’t demand healthcare, they demand health, and then they have to buy healthcare. And there’s a production function between healthcare and how that contributes to your health, right, and there’s diminishing marginal returns in that. You can only buy so much healthcare. I could spend a billion dollars on healthcare, but I’ll still die eventually.

NS: So there’s diminishing returns. That production between healthcare and health right now, the sort of default way of doing that is that the doctor is the interface between the between healthcare and health. They’re the ones telling you, “If you have this procedure, it’s going to cure this diagnosis,” or, “If you change your behavior in this way, it’s going to help out these other characteristics of your health.” So they are that interface. Of course, in the start-up where we always say consulting doesn’t scale because you can’t scale a human, but you can scale software. And that’s where software can hopefully start to replace that bad interaction layer, and both enable providers to be more effective and reach more patients, but also enable patients to have a much more engaged interaction with their health, because it’s on their phone, it’s on their person, in their pocket. They’re interacting with it daily rather than every—well, for me, every two years when I go to the doctor’s office. I know I should go to more, but I’m young and invincible, so I don’t.

AP: Exactly.

AP: Thank you for starting the conversation on how we scale these modes of health care. I think the next few months will be very interesting as we reinvent the process of healthcare delivery. And I would encourage you with your large eco-system to begin to look at how Redox can extract other value, and help us understand it as a community what’s going on in healthcare as people are looking at these applications. Because I really think you’re in a unique place. Thank you very much.

NS: Thanks for having me. This was a fun conversation.

Reader Interactions

Leave a Reply

Your email address will not be published. Required fields are marked *