Episode 106: Technology Driving Clinical Care

Speaker 1: You are listening to Your Practice Made Perfect. Support, protection, and advice for practicing medical professionals, brought to you by SVMIC.

J. Baugh: Hello, everyone and welcome to this episode of Your Practice Made Perfect. My name is J. Baugh. I'm a claims attorney here at SVMIC, and I will be your podcast host for today. The topic that we're going to cover today is clinical IT, and to help us discuss this topic is Dr. Kevin Johnson. Dr. Johnson, welcome.

Dr. Johnson: Well, thank you, Jay. Glad to be here.

J. Baugh: Well, thank you for taking the time to talk to us about this very important topic. Before we get started on talking about clinical IT, maybe you could tell us a little bit about yourself, your background, and that sort of thing.

Dr. Johnson: I'm happy to. I'm a specialist in biomedical informatics. I'm currently the chair of that department, of Biomedical Informatics at Vanderbilt University Medical Center. I was born and raised as a wannabe veterinarian who turned out to be more interested in taking care of little kids at the end of the day than I did taking care of animals. So I'm a pediatrician. I love teaching and I love doing research, especially in the areas of medication safety and health information exchange, and just trying to see whether we can really create a healthcare environment that is as good as the one we always see on television, because that's unfortunately not the one we live in today.

J. Baugh: Well, that's true. And this area of medicine, this area of healthcare, I know is really taking off. I'm hearing more and more about informatics as it relates to healthcare and I'm wondering how you got into this area of biomedical informatics.

Dr. Johnson: Well, it's a funny story. I was in college. I was taking the second semester of calculus and our teacher said to us, this was back in 1988, our teacher said, "For those of you who are interested, we have a new computer science major and I would encourage students who have done well to continue in this space." And I was among the students who had done wel,l but who had no intention of ever picking up a calculus book ever again.

J. Baugh: Okay.

Dr. Johnson: And so we all looked at each other, winked and said, "Yeah, that won't be happening."

J. Baugh: Right.

Dr. Johnson: And then the next year, when I was taking analytical chemistry, we had to go into the lab and use the computer to help us to do this equation balancing work. While I was sitting there, I fell in love with a game that was being written by a guy in the lab. And when he showed me what he was doing, which essentially was this beautiful graphic screen that was being created by words in a program, I just got completely turned on and said, "You mean to tell me you can actually create out of nothing using programming languages?" And he said, "Yeah." And so I started working on my own game, ended up learning enough in my avocation to pass out of the first year of curriculum for a computer science minor, and I continued all the way and got a minor in computer science at my college, Dickinson College.

And as a senior project, knowing I was going to go into medical school, started studying the ways those two fields can be combined and resolved that was going to be something I was going to do with my career. So I started becoming interested in the ways computers can be used to document visits, and the ways computers can be used to do clinical decision support. I got to Hopkins Medical School and by the end of my first year had met a guy named Dick Johannes, and started working in his lab developing tools to document colonoscopies using the computer. And really the rest was just deciding I needed to make this part of my profession getting additional training and off I went.

J. Baugh: Well, that is an interesting story as to how you ended up in this field of healthcare. And as I said before, I know it's a growing field. It's one that is becoming more and more important in terms of delivering healthcare to patients, so I appreciate your work that's being done regarding biomedical informatics. You mentioned earlier that you're currently at Vanderbilt University Medical Center, and Vanderbilt was one of the most innovative sites, yet you made a decision to abandon your advantage and purchase a commercial system and I'm wondering why you made that decision.

Dr. Johnson: This decision to go from a very well-known and popular internally homegrown system, to a well-known commercially but perhaps vilified locally commercial system, was probably one of the toughest things I've ever been a part of. But, for me, it really came down to a couple of key points. Vanderbilt had done amazing work and we were one of the first sites to go live with what's called pharmacogenomics. The idea that patients who might need a particular medication because of a disease, but who have particular variants in their genome, shouldn't receive average care medicine. They shouldn't receive the medicine we typically give. They need an exception based on the dosing or even the type of medicine. And we had done a number of projects of that level of visibility, ranging from work we did in ventilator associated pneumonia and increasing the ability to save patients' lives by simply creating a dashboard, a work that I did to improve generic prescription writing all in our homegrown system.

So everybody said, "This is going to be the end of advancing as we know it." And I said, "Well, let's actually look at the reality here. Vanderbilt is an island. Nobody else uses this system. We have a major training program, biomedical informatics. We send people out into the world and although they might know about our electronic health record, they know nothing about the one at the environment where they're going to land. Furthermore, we have this large group of really talented faculty who are responsible for these systems, and if any one of the leaves then I spend," at the time, "easily six months to a year trying to find somebody else to do that work because these are just really hard to find very expensive people."

So I told people, "Number one, we need to be more relevant, meaning as a training program, we need to be teaching people what they're going to learn when they leave Vanderbilt. Number two, we need to be more resilient. We need to be able to recover when I have outstanding faculty or trainees or staff who leave Vanderbilt to go do bigger and better things." The biggest one I felt was we needed to have the opportunity to do multicenter studies. And, in fact, there had been work done right before meaningful use, a meta analysis, to understand what we thought would actually happen in this country with widespread adoption of electronic health records. But the meta analysis that was done, looking at all the papers that had come out over the last two decades, said that, "In fairness, every positive result has come from one of five institutions." And we were one of those five.

And so they essentially said, "We don't have any idea how this is going to go around the country, because all of these studies are coming from just a few places, and the only way that we could learn more is to be able to do larger multi-centered studies," which meant we had to be a part of the group that used the same systems at Vanderbilt that were used everywhere else. And then the last thing I often say to people is, we were also getting just buried under all of the regulatory based changes that we needed to make in our EHR. And as a point, when I was talking to our over 400 members of our IT group, I said, "Raise your hands if you've done anything interesting in the last five years." And about 150 people who do a lot of development, raised their hands. And I said, "Now keep your hand up if the thing that you did had nothing to do with regulatory change." Every hand went down.

And I said, "Just imagine what's going to happen when another company is doing all the work around regulatory updates, and I can take this same work force of 100 people and focus on the cutting edge things that Vanderbilt's known for." And so, in fact, we are now in the process of doing that. We've gone live with Epic, and we called it E-Star, because our belief is that Epic is now our platform. We actually have data and data lakes as our platform and we have Epic on top of that, but we also have still homegrown parts of that system that are the things we're doing that are cutting edge. And we still have over 100 other applications here that aren't Epic, that are all on top of that other platform. But because we now don't have to deal with a lot of the changes that are built because of regulatory changes or reporting requirements, I believe we are now at the precipice of really doing interesting, innovative work that is relevant to the rest of the world.

J. Baugh: So as we're looking at the topic of clinical IT, what do you see as being some of the most pressing issues in that area?

Dr. Johnson: It's an interesting time, because now that we have more than 80% of the country using IT, people, me as a father, me as a son dealing with older parents, you're starting to expect a lot more out of our healthcare system. And this is one of the things that really propels me forward every day. So, the number one thing that I know I experienced that everybody in the country talks about right now, is this whole idea of interoperability. A fancy word that simply means if a patient is seen at hospital A and then goes to clinic B, and then gets sick and goes to emergency room C, that record should follow them. And starting probably about 30 years ago, this was a recognized challenge. Now that we have the level of adoption we have, it's almost now a travesty. And as it turns out, we are making very, very fast progress in the area of interoperability or record sharing.

I actually did a film about this called No Matter Where, that you can find on Amazon Prime, and we talk about the history of interoperability and electronic information exchange. And it's fair to say that it's had many fits and starts, that the model that we'd like to see for interoperability is now turned on its side, and what I mean by that is, as opposed to us creating these community networks, which largely have failed, we are relying a lot more on our vendors, trusting each other to communicate information without the necessary involvement of the C-suite of every one of these hospitals. So I think that that's been a really big change, and that's probably the number one area, certainly CMS Don Rucker and CMR Verma have a lot of interest in that. Certainly, they've had a lot. CMR Verma and Don Rucker and others with the federal government have had an enormous interest in this area, and have been pushing this idea of promoting interoperability, and other groups are as well.

The other big area, and one of the ones that I think shocks people, Atul Gawande in his most recent story, Why Doctors Hate Computers, talked about the fact that the fastest growing group of users of the electronic health record are actually patients. And it's very true. The portals that we all now have access to as patients are a very new way that we can access the data in our electronic health record. And so with that comes the idea of digital health. Digital health is something that Eric Topol and a lot of other people have written about. I've done a little bit of work in that space through Robert Wood Johnson funding years ago. And that's become the other really pressing issue, is how do we take and democratize access to the data so that we can have appropriate disruptions in the healthcare system - hopefully aspirationally - without negatively affecting the existing healthcare system in any way that is adverse. That is actually the real challenge, and we can talk about that more if you'd like.

But fundamentally, the idea that we are going to use smartphones and technologies that might attach to smartphones, and portals and all of the consumer health informatics environment to move health and healthcare beyond the walls of a typical health system, is one of the most current areas in the field. I would say the third one and the one that probably everybody's waiting for me to say is AI. Everybody's talking about AI and prediction.

J. Baugh: That's right.

Dr. Johnson: I think we are squarely in the hype, hope Gartner Hype-Hope Cycle right now, where if you follow the New York Times, five years ago, this was all about promise. Big data, what if we could do machine learning on this big data, and reinvent the treatment of cancer or discover new medications? And all that's happening. But we are also aware now of some of the concerns of that, including the idea that some of the people who have access to those data might scare people. Patients might be afraid for their privacy when groups like Facebook or Amazon or Google or Microsoft, and I could go on, are the ones who are really leading the charge towards democratizing access to their information, using AI and prediction to possibly help patients to make better choices or help providers make better choices about the way to care for groups of patients.

I would say those are the three biggest areas. There are one or two others that I think I find phenomenally interesting. Within all three of those - interoperability, digital health, AI and prediction - there's also this big concern about health equity and the fact that as I sit here, looking out of the building where my office is located, I'm looking at North Nashville. And I'm looking right by HCA, which is one of the largest for-profit companies for healthcare in the world, and beyond that, an environment of people who have enormous problems with data access and realizing that this equity piece only is going to potentially get worse if we don't come up with ways to educate, to integrate this knowledge into the routine of people who are coming from all different walks of life in this country. And if we don't come up with technologies that help to capture data from all of these different walks of life so that AI and prediction is equally valid in populations that aren't just White middle-class males, but also all of the other groups whose data need to contribute to the types of AI and predictions that we're building.

J. Baugh: Well, that's a good point that you bring up Dr. Johnson. When we talk about things like informatics and clinical IT and these types of issues, if it's information that's not available to the general population and then it's only available to certain segments of the population, or if the data that we're compiling is only from certain groups of people, then it really skews the results of what we're trying to do and might not provide information that's as useful as it could be if it were applied to everyone equally. So that's a great point that you bring up.

Dr. Johnson: Thank you. That's exactly it and it's something that I think most people don't recognize. When we look at developing, let's just say a machine learning algorithm to diagnose a particular condition, say celiac disease, what we have to recognize is that simply saying your patient is at increased risk for celiac disease is the beginning of a very long journey toward diagnosis. So if we aren't careful about the various presentations of celiac disease based on diet, based on age, based on gender, then we may do an enormous number of duplicate tests or unnecessary tests for certain patients we may completely ignore through false negatives, populations of patients who are at increased risk but who had insufficient data in our modeling environment for us to identify the parameters that would be important to trigger it.

So a lot of us don't realize the importance of taking what we're now learning in AI and replicating it in our own local environment then designing, testing, and deploying in what we call a learning health system ways to intervene based on that piece of information, to see whether in fact that intervention results in earlier diagnosis or possibly missed diagnoses all related to this one prediction. So there is an entire long tail of work that must be done for every one of these major advances that we are seeing in AI and prediction. It's not just dropping in the brand new rule and expecting it to work the same way in one's environment. So you have to have an electronic health record. You need to have experts in informatics and data analytics to work it with people. You need to have a relationship with patients and with providers, nurses, physicians, administrators, so that you can build the right infrastructure. It's much more complex than I think most of us are hearing in the news.

J. Baugh: Right. So you mentioned just a moment ago about electronic health records. What do you see as the future of EHR? I mean, that's a topic that so many people talk about. Some physicians have a difficult time with their EHR. I hear a lot of complaints from physicians who just can't get the EHR system to work the way they would like for it to and maybe they want a different system, but there are problems in going to different systems as well. And just what do you see as the future of electronic health records?

Dr. Johnson: That's a great question. Medicine, as we know it, is a field that is thousands of years old and really everything that we think about with the present day electronic health record is decades old, three or four decades old. So it's very new. It's a very new innovation. And what I've often told my students is, imagine what a car would be like today if Mario Andretti had actually developed it. It would be really safe. It would be roll proof. There would be no cup holders.

J. Baugh: That's right.

Dr. Johnson: And so there's a point where we have to say, "The EHR of today was developed by geeks like me who saw major problems in patient safety and in interoperability, and focused mostly on the engineering to improve safety interoperability, the integration of clinical decision support." Even though we might be physicians, we're a different lot. We're a group of physicians who type almost as quickly as we talk, and point and click is a part of what we've been doing since we were between two and five years old for some of the newer people who are coming out. And so you wouldn't expect the first EHRs to get to this level of adoption to necessarily be as good as they should be, and they've clearly demonstrated that they're not as good as they can be.

So what I think is going to happen first is feedback and response to feedback, and having had a chance to watch what the commercial vendors are researching and thinking about, especially frankly in areas like voice recognition or speech recognition technology as one of the newer pieces, and ways in which we can use smaller devices like mobile devices to actually run the EHR, I'm very confident that we're going to see changes in the EHR that make it more useful. Everybody knows it has to happen. There are people who believe that the commercial vendors have no interest in that particular area. And I am quite certain that it's not true. Where we end up, it's anybody's guess. Following the model that Gibson once put out years ago, which essentially is the future is already here, it's just not evenly distributed, there are groups of people who are using the electronic health record in very different ways.

Vanderbilt has had two very exciting projects. One we call Viva, which is an enhanced voice assistant, which I think revolutionizes the way we think about information retrieval. The idea being that if I can make a well-formed vocal query to the EHR, let it process the information, and then give me back a context dependent summary. So, for example, if I'm sitting in front of a computer screen, show me a trend of data, if I asked for something about the data, if I'm in my car, read back to me the trend in the data, that kind of a technology I think could revolutionize the searching capability of the EHR and stop a lot of what Paul Tang has called "foraging" that currently goes on with existing clinicians, trying to find information in front of a patient in the EHR.

The biggest challenge is also the challenge of data entry, and I think there's a lot of work to be done first in the regulatory space to take the pressure off of physicians, nurses, and advanced practice nurses in particular, to document everything that was done in a note to send to a payer to get reimbursed. My personal belief is that we need to work closely with the health committee to pass legislation that allows the definition of what documentation looks like to include things like the video of a visit. Once we have the ability to broaden our definition of what a documented visit looks like, we can start thinking much more creatively about what are we really trying to do with the actual ASCII based or text-based summary of a visit. It may be that we go back to the summaries of the 1950s and '60s, which is 18 month old, painful ear, no other concerns, ear is red and inflamed, diagnosis of otitis media, treatment, amoxicillin per usual.

And that we then look into the EHR, which is the system of record, to see that the physician did do a review of systems, that the physician or nurse practitioner did actually prescribe a particular dose of amoxicillin with some frequency and some duration and some comments, et cetera, et cetera, et cetera, so that the notes can actually take on a completely different role in the world in which an EHR surrounds that note than what we think about right now. I'm really optimistic that if we can just make a couple of big changes in the regulatory environment in which we are building electronic health records today, we will enable people like researchers around the country and the world to redefine the systems that our doctors are using and get cup holders in that car.

J. Baugh: That does sound like a very interesting way of looking at medical records, a very different way of looking at EHR as we do today and perhaps one that will be more beneficial for everyone that's concerned. So as we draw this podcast to a close, I hear that you are starting a podcast and I'm wondering if you could tell us a little bit more about that.

Dr. Johnson: Yeah. It's actually going to be fun. And as many people may know, I love to be creative and as you probably could glean from things like movies that I've created, I didn't share with you that I recently did a screening of our Epic Go-Live to Epic the company, which had about 15,000 people that actually had a chance to see it. That's an enormous impact. And I'm not a blogger, but I've been thinking for a while about ways I can get my crazy ideas out in the world as well as ways that we can translate what we do every day in a field that's got an esoteric name like biomedical informatics out to people like my parents, who try desperately to tell people at church what I do and fail every single time they do it. And so I had this idea that I was going to create a podcast that essentially was a table with three to four people around it trying to explain some of the thorny topics, interoperability, data science, as well as even what is biomedical informatics in a way that my mother would understand it.

And at first I was going to call it Kitchen Table Science so that it could be broader than informatics. But then that title just hasn't quite stuck with me yet, so I'm currently going through the process of having a number of people give me ideas for titles. There are a few that are out there, and the first episode is going to be all about how we named it. I've got some interesting people coming back with interesting names that I would love to share with you, but it will just be so much more fun to hear them talk about their name themselves. So I'll leave that there.

J. Baugh: Well, Dr. Johnson, we want to thank you for spending time talking to us today about biomedical informatics, and I want to let our listeners know that the information about your podcast and other information that you have mentioned today will be included in our show notes. And once again, thank you for being here.

Dr. Johnson: This was a lot of fun, Jay. Thanks a lot for having me and I look forward to doing it again.

Speaker 1: Thank you for listening to this episode of Your Practice Made Perfect with your host J. Baugh. Listen to more episodes, subscribe to the podcast, and find show notes at SVMIC.com/podcast. 

The contents of this podcast are intended for informational purposes only and do not constitute legal advice. Policyholders are urged to consult with their personal attorney for legal advice as specific legal requirements may vary from state to state and change over time.