Welcome to the big unlock where we discuss data analytics and emerging technologies in healthcare. Here’s some of the most innovative thinkers in health care information technology, talk about the digital transformation of health care and how they are driving change in their organizations.

Paddy: Hello again and welcome back to my podcast. This is Paddy and it is my great privilege and honor to have as my special guest today Dwight Raum, CTO of Johns Hopkins School of Medicine. Dwight, welcome to the show.

Dwight: Thank you very much for having me Paddy. Honored to be here.

Paddy: Thank you so much. OK. For the benefit of our listeners maybe you could tell us a little bit about Hopkins and your role at Hopkins.

Dwight: Sure. So, Hopkins is a pretty well-known name in healthcare. It’s long and storied history in clinical research and care. We’re also a University and one of the first real teaching institutions in medicine in the world and indeed in the United States. So, my role at Hopkins is CTO and I oversee all the infrastructure operations for both the University and the health system. I am also the executive director of our technology innovation center and most recently I’ve been doing a lot of work focused principally on precision medicine and how we use big data and platform technologies to really change the way we’re doing medical research these days.

Paddy: That’s awesome and we will definitely talk about the precision medicine initiatives. So, let’s start with some high-level stuff. How would you describe the technology stack at Hopkins?

Dwight: Yeah. You know friend of mine from a healthcare standpoint. It looks probably fairly traditional. We have a large EHR vendor and we have deployed most of that through on-premise services. We have two major data centers. We’re quite a large enterprise and we’re fully integrated onto one single EHR system. So, as you can imagine we have really good centralized controls over all our endpoints and all the hosting that we do for our EHR system. But from innovation and from a precision medicine standpoint I think we start to look pretty unique fairly rapidly where we’ve been pretty early adopters of cloud. We’ve been had it for now about six or seven years. But things really start to ramp up about three and a half years ago and we started building out this precision medicine platform which is all now cloud native and it really exploits the capability of cloud providers. You know leveraging elasticity and really large compute as well as data storage that’s been pretty pivotal in our transformation of research over the last couple of years.

Paddy: Right. Can you talk to us a little bit about how you’re defining digital and digital transformation and how would you fold in your precision medicine initiatives in the context of the transformation?

Dwight: Yeah. I mean I think digital transformation is one of those buzz words that we hear a lot of in our industry these days. And you know I think everybody has a different definition for what is actually is or what digital means to them. But from my perspective it really means, it’s very simple. Quite frankly it’s how do you employ technology not as the driver of change necessarily but how do you use technology to drive your mission. So, it’s the marriage of new technical capabilities whether it’s cloud or more advanced things like ML or AI. How do you marry those two. The traditional methods that you’ve had for be in your mission area. And then really use that as a way to accelerate it. And it comes across really two different domains. You know I think about it in terms of what does it take to run my enterprise. That’s not super sexy but there is certainly room for digital transformation there where we become more efficient and more thoughtful about how we’re going to use all of our resources. But then there’s this sort of second phase which is how do we reinvent what we’re doing using all these technologies that are now available to us. But recognize that there is this highly symbiotic relationship between the two as you do re-enter, as you do innovation, as you do reinvention. You have to be thoughtful about how you take what you learn, and you turn it back into that that operational enterprise and really achieve scale with it. So, I think digital is really is transforming the way we do things. But it’s not as always as sexy as the buzzwords are out there.

Paddy: Yeah. But I think you made a good distinction as it relates to deploying technologies in the context of automation and efficiencies versus really transforming the way you deliver care.

Dwight: What I’ve seen over the years is that there’s almost as Moth syndrome where everybody’s drawn towards the bright and shining light of a new technology that comes out. And I think they do so at the at the detriment of innovation within that operational space. You know there are great opportunities to exploit there. And if you don’t look towards driving efficiency it’s awfully hard to, I think build sustainable innovations. When you’re reinventing the way, you do your business.

Paddy: The moth syndrome. I think I’m going to use that as a title for the podcast. I love it. Well you know obviously it just begs the question of how you’re going to pay for all of this and how do you really build a business case. You know in both intangible terms as well as its strategic value. Because we all know that healthcare is one of those sectors which carries a lot of technical debt in a long year of underinvestment. I’m not saying Hopkins in particular but in general as a sector there’s a lot of technical debt that the sector carries. And we’ve done some research internally in my firm which indicates that even though the focus is now towards all the emerging technologies of transformation and new technologies, enabling the transformation, fact of the matter is that a lot of the budget dollars available today are going towards maintaining the status quo whether its electronic health records or any other legacy technologies. So how do CIOs and CTOs balance the two? What are your thoughts on that?

Dwight: Yeah. So, I think there is one sort of general principle which I’ll share with you and it’s kind of goes back to my previous comments regarding focusing on the operational enterprise and looking for opportunities there. I think one of the techniques that we’ve been successful in deploying is using those innovations in the operational space to gain efficiencies and using the resources freed up by efficiencies to focus on new innovations and deriving new ways of doing the business. So that’s one strategy. It doesn’t really solve though I think for the larger problem of keeping the lights on versus finding new ways to innovate. So that second point I think is kind of unique to Johns Hopkins in that we are in an essentially capitated state. The state of Maryland has very tight controls on how reimbursement works and because of that we’ve already made this transition to value-based care. So quite frankly it’s a strategic imperative for the entire of our institution to focus on delivering care to the right individuals, at the right place, at the right time. And the only way that you can really quantify or even have a shot at doing that efficiently is to deliver using technology. So, the very future of medicine I think value-based care is coming whether or not we like it that is our future. And the only way we’re going to effectively deliver on that future is if we figure out how to use technology as the it’s the scaling arm as the lever arm for the efficiency that we need to get.

Paddy: Yeah. So, what you say that capitation is actually the shift towards a capitation model is actually an accelerator and a driver for.

Dwight: Absolutely. Yeah. I mean I think it’s a necessity is the mother of invention and without these constraints we won’t have the proper incentives to actually drive true change. You know in a lot of ways the technology pieces while they’re interesting and sometimes really exciting. The organizational changes that you are required to pull off real transformation are often a lot harder. So, I think that alignment of incentives that transition to value-based care drives is really the key to catalytic transformation.

Paddy: Yeah, I like the comment you made that you know you use innovation in order to gain efficiencies and you use the dollars that are freed up to reinvest towards the future state. I love that. Now you mentioned that you’re also the executive director of the Hopkins innovation program. Tell us a little bit about the program and some of its accomplishments and how you actually measure the success and how you actually measure the returns that you talked about?

Dwight: Yeah. So, we actually launched the technology innovation center at Johns Hopkins about four and a half years ago. My co-founder Dr. Paul Margie and I really saw this opportunity to say that we had incredible talent within our IT organization as well as within our department of radiology. We merged our organizations and really bootstrapped it as almost like a startup type culture within Johns Hopkins. So, the entity itself is somewhat standalone and we run it as almost an internal consulting business back to the mothership. But what we’ve done over the last four years or four and a half years is stand up a number of programs that foster that organizational cultural shift that I was describing it as so integral to digital transformation. My co-founder Paul is really brilliant at pulling people together and we’ve launched a number of programs around data analytics precision medicine, around entrepreneurship, as well as data science. So, these programs have really started to bring in a lot of faculty members as well as a lot of staff members and we have brought them into this cultural shift and they now become the apostles if you will of digital transformation change throughout all of Johns Hopkins. So, there’s been this wonderful shift that’s occurred because the technology innovation center is really active as this convening center for change to occur. We measure success in a lot of ways. You know one of the key measures is the number of people that we’ve influenced through these programs but then also because we’re a software development shop and we do these entrepreneurship cohorts every year. We also measure success in terms of number of products that we developed, in number of companies that we’ve helped launch out of Hopkins and into Hopkins. So, over the last four years we’ve had some really great successes with technology innovations. We were really the first institution to deploy a Research Kit app that used the Apple Watch. We’ve made really great innovations around clinical communications just the list a few of them. So, it’s been a really great journey seeing the talent that we have here in our innovation center. But coupling our talent with world class research teams and then seeing what comes out of that mix with technology married to two medical intervention.

Paddy: Fascinating! Well let’s switch to the topic of precision medicine which I know is one of your focus areas now precision medicine needs a lot of data and with emerging sources of data. There’s a lot of ifs and buts about the data. Some of it is usable and some of it is not. And there’s a lot of questions around. So, what are some of the data sources you are using besides obviously the electronic health records. What are the emerging data sources you are using to drive precision medicine and talk to us a little bit about what are some of the challenges you see in harnessing these new emerging data sources to generate insights?

Dwight: Sure. So, as you mentioned we are using EHRs as our backbone of the precision medicine system because essentially is the system which provides patient identity. It allows us the crosswalk through a lot of other really rich data sets. And our approach to precision medicine is really wide. It’s multi disease and the way we’ve done that is we’ve launched these centers of excellence. Each center of excellence is focused around a specific disease and disease area of study. But because there are so many different diseases that we are actually going after simultaneously. We have found that there are some very common high value data sources that we connected to the medical record provide really deep inside. So, these are kind of the usual suspects in terms of data imaging data is obviously incredibly valuable. Marry that with genomic mix data that’s incredibly valuable, physiological monitoring data as well so capturing real time streaming data of in room patient monitors that data when connect it all into context can provide a really complete or complete picture of that patient and their current disease state. One of the other things that we’ve done is as we’ve we stood up into these centers of excellence almost every single center of excellences has had a number of locally created research databases. And these locally curated decent research databases. They’ve taken really elaborate notes and captured the phenotypes of these patients in a depth that our medical record doesn’t just have. So, a big part of our process for standing up a center of excellence is to ingest all that data I mentioned around imaging, genomics and physmon (physiological monitoring) and the medical record but then we also married up with this really rich set of data from the phenotype that’s captured by the clinical researchers themselves. Some of the challenges are as you can imagine in a big data. It is truly big data. We have pretty good cross-reference information on all this data. One of the challenges though is really matching up semantic interoperability. That’s one big challenge. The second big challenge is that those rich phenotypes that describe are often narratives. So how do you drive a feature from that narrative that can be the basis of classification or categorization. That’s a really big challenge and we find that to be a very iterative challenge where we work with each center of excellence to refine whether it’s some sort of natural language processing that we use to extract a feature from notes or maybe it’s image processing to extract a feature from an MRI series. That data once it’s all presented into a curated database that we provide to those research teams. The next challenge is having the right team members available and able to work on that data. So, a lot of times the data is sufficiently complex that you really do need a PH.D. level data scientists to help interrogate and model that data and become part of our relationship with the Applied Physics Laboratory which is one of the institutions of Johns Hopkins University. We have access to some of the world’s best data scientists and we’ve brought them in to really support us in our precision medicine program overall. So, every one of these centers of excellence has one or two data scientists that have been part of that team to actually work side by side with the clinical investigator to interrogate that data but to also take the insights that are derived from data and match that back to the biological model of the disease itself.

Paddy: Yes, that’s what you mentioned semantic interoperability. You know I want to kind of dig into that a little bit. Of course, interoperability has been a big topic for the last several years. You know we put in all these EHR systems, but we didn’t take care of the interoperability. So, we’re kind of playing catch up. Now we’ve made progress with technical interoperability especially through the fire standards. Interoperability is still very early stages. Even something as basic as information with your peer is a huge challenge because of the semantic interoperability problem is where they define and classify data in their claims management systems very different from the way you do it in your electronic health record system. Just a starting point. So, two questions, should we say that the technical interoperability problem has pretty much been resolved and is just a matter of adoption because all the tools exist and it’s a manual will. So that’s a first question. And secondly you know how long you know. When are we going to get to semantic interoperability or some semblance of semantic interoperability that makes us feel like the data is truly fungible between different setting up?

Dwight: So, I do think that technical interoperability is not 100 percent resolved yet, but we will on our way, and I am fairly confident that we will achieve full technical interoperability. I do think there are challenges around vendor implementations and openness. It’ll be interesting to see how the rules shake out in terms of information blocking but that remains a concern of mine. Despite the fact that we have technical interoperability there are still barriers between data exchange. Some of which are artificial, and I think we need to really unleash or knock down as many barriers as we can so that we can at least achieve the interoperability of the data exchange level. From semantic interoperability standpoint boy, it’s that’s a tough question to answer because you know there are so many different interpretations of what any given data point may mean and it’s almost like boiling the ocean. So, I don’t know that there’s any way that I ever see us getting to full semantic interoperability. However, that being said if you think about interoperability around specific diseases there are some really good models out there for what is the most appropriate way to collect the data. And I’ll give an example that we have had recently with precision medicine. One of our centers of excellence is for multiple sclerosis and we identified in partnership with our clinical faculty members who were doing the research. All the elements that were really important in terms of capturing data around that patient on a visit. And we developed a set of intake forms that really standardized rich data collection. Well we are now sharing that with other research institutions. So, when we collect data around MS patients it’s now interoperable and we have the ability to not only share data from a research standpoint but also the ability to be consistent with how we practice care. So, I think there is great hope around specific diseases and the data that we collect and the structure and the meaning behind which we collected that will help us achieve interoperability in a way that maybe we won’t be able achieve in a technical standpoint.

Paddy: Yeah. So, from what you’re saying basically the solution is to approach it by addressing it at the source of the creation of the data as opposed to trying to fix it after the fact.

Dwight: Yeah. I mean I think it’s always going to be a bit of a combination of both because you know Gosh knows we have a ton of data that is sitting there, and we can we can apply all kinds of intelligence to extract meaning from that and normalize it. But that’s in my mind that is a goal that will never be completely achieved we’ll just continuously do it and we’ll get value from it. But to the extent you can move that data capture to be consistent upstream and maybe anchor it around a specific disease. You end with end up with much better results.

Paddy: Yeah that’s very insightful. So, switching to a different topic. You know there’s a huge what we might call a last mile problem in healthcare. So, we do all this innovative work in the background. You know we’re coming up with all the insights and so on and so forth. How does it really make a difference at the point of care and how do you deliver all the benefits of all these technology innovations and insights to patients and caregivers at the point of care. Can you talk a little bit about that and how you’re addressing that at Hopkins?

Dwight: Yeah, I mean so you’re right on and that it is the last mile problem. There is so much research and so much insight that’s already out there that’s been published and presented but it never makes its way to actual delivery and implementation for patients, so the impact is really constrained. So, I have two thoughts on that. One is sort of an alignment of incentives perspective and it goes back to how Maryland is a special state and how I think could serve as a model for the rest of the country as we shift to value-based care. The incentives because we’re looking at treating the patient not on a fee for service basis, we’re looking at their total overall health. There is an incentive to want to practice these insights at a scale that goes across the entire to the population. So, let me give you an example with our precision medicine program we have a center of excellence for prostate cancer and there is a model that’s been developed by a number of faculty here that is able to project out the risk of a man who has prostate cancer who has indolent prostate cancer. So, it’s important to stand that about 40 percent of prostate cancer diagnoses are actually in the line and that’s when the prostate itself is. It does have cancer but it’s unlikely to kill you. And in those cases, about 40 percent of diagnoses are actually indolent. But for many men they actually get treated even though the cancer is indolent. So, our center of excellence actually created a model that helps not only the provider but also the patient to understand the probabilities associated with not treating their cancer or into cancer. Well the upshot of this is that there’s a huge savings to not intervene prematurely and there’s a great benefit to the patient. The patient doesn’t suffer all the consequences of more radical intervention. Well when you look at this across the entire population of men for instance in the state of Maryland you know the impact is its tens of millions of dollars. So, there’s a great incentive to create these precision medicine tools but then also make sure that they’re really delivered at the point of care. Now the second point I want to make about this is a bit more of an engineering one is, and I think is a unique aspect of what we’re doing with our precision medicine and health program that is that discovery and delivery or two very distinctly separate disciplines. Discovery is very much around free thinking and understanding the in depth the disease, but the actual practice of delivery is one of engineering. How do you how do you take the technology and make it accessible and adoptable by the frontline point of care clinicians as well as the patients. So, we all I think have really good models for how to do design thinking, and how to build user centered applications. We need to ensure that when we make it a discovery that we employ the engineering to come up with a product that actually can be scaled to actually achieve those outcomes that we want of a broader population basis.

Paddy: Right. That’s such a great analogy. I think it’s almost like saying this it’s the difference between pure research and applied research. And I love that contrast you drawn between the discovery and the delivery of innovations or insights or whatever the case may be. It’s fascinating. So, one of the things I like to do on my podcast is what is known as a lightning round so I’ll read off three or four terms that are commonly used today in the lexicon of health care and information technology and maybe you can share your top of mind thoughts on those. Ok?

Dwight: Ok.

Paddy: We already talked about cloud. So, let me pick a couple of others. Artificial Intelligence.

Dwight: Incredibly great promise, must be balanced though with experience and the human must remain in the equation.

Paddy: OK. Blockchain.

Dwight: Interesting computer science concepts way overblown in the market.

Paddy: OK. 5G networks.

Dwight: A lot of hype but great promise.

Paddy: Ok. Voice enablement.

Dwight: The future. I think that you know our traditional when we define digital earlier in our conversation today, we probably were thinking in terms of mobile and web. But I think voice is kind of the most obvious next step of user interface that is going to radically change the way we interact with technology. In many ways I see technology starting to become less front and center and dissolving more into the background of our everyday experience and voice enablement is probably that happens.

Paddy: It’s almost like zero UI.

Dwight: Yeah exactly. Yeah.

Paddy: OK so we’re coming up to the end of our time your couple of quick last question here. So, what is your advice for tech firms big and small that want to be part of your journey at Hopkins.

Dwight: So, I think we have to embrace. Well let me back up for a second here. You know I think many of us know that health care in the United States while we are not in boom times in many places across the country. It’s just not sustainable. There is over a trillion dollars of waste every year in health care spending a lot of that is government funded. I think we fundamentally have to ask ourselves as technology leaders are, we going to be part of the solution that changes the way health care is trending in the US. And can we actually improve this country as a consequence of the decisions and the innovations that we lead as technologists. So, I actually think it’s an incredible time to be in the health care industry and in and in technology because I think we have the opportunity to influence the US in a way that very few other industries can. And I think we can do so in a way that will radically improve the lives of future generations. So, I think that’s a very high-minded way of saying we have to be thinking about the challenge in front of us should be much bigger than our individual selves or institutions. And the transition to value-based care and to population health management and technology driven or data driven decision making. These are all going to be the ways that we achieve those great outcomes that I see on the horizon.

Paddy: That is actually so well said and it actually quite inspiring. Thank you for that. So, my last question to you how do you stay on top of all the technology trends. How do you keep yourself up to date and relevant?

Dwight: Yeah. I mean I think there’s always a challenge. One of the great things about being involved with our technology innovation center is that I have a lot of access to really savvy people who are at the leading edge of a lot of technologies. It also doesn’t hurt that I’m at a great research institution and there are brilliant minds much smarter than me around me all the time and just through osmosis and proximity I’m able to gain access to a lot of really amazing ideas. On a personal note I am a bit of a hacker on my own and I do have children and we spend time together playing with new technologies and learning new things. So, I think it’s hard to be in technology and really have a passion for it. If you can’t have some sort of reward from actually from doing it all the time. So that’s the way I stay on top of it. And it’s imperfect but it does keep my keep my fire lit.

Paddy: That’s wonderful. But Dwight thank you so much for that. It Has been a fascinating conversation. And I look forward to catching up with you again in person sometime soon once again. Thank you.

Dwight: Thank you very much for having me. I appreciate it.

We hope you enjoyed this podcast subscribe to our podcast series at www.thebigunlock.com and write to us at info@thebigunlock.com