This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Today on This Week Health.

We don't want to compete on data, competing on data is the exact wrong thing to do. So the more that we can standardize data where patients own it and it can be shared for their patient experience, if we're gonna compete, we'll do it on quality care. And on a culture of patients and families 

Thanks for joining us on this keynote episode, a this week health conference show. My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of this week Health, A set of channels dedicated to keeping health IT staff current and engaged. For five years, we've been making podcasts that amplify great thinking to propel healthcare forward. Special thanks to our keynote show. CDW, Rubrik, Sectra and Trellix for choosing to invest in our mission to develop the next generation of health leaders. Now onto our show.

  (Main)

Good afternoon everybody, depending on where you're calling in from. It's afternoon here on the East Coast, and we have a great panel discussion lined up for you today, and I'm looking forward to it.

As you guys know, if you've been on any of our webinars, we will give everybody about three minutes to join. And what we'll do during that time, just give you a little opportunity to get to know our panelists a little bit more. I know some of you log in just to hear some more about our panelists.

I'm gonna go back to last year's question. We did this opener last year and I enjoy the question a little bit more. And as I'm introducing these guys, I'm gonna ask them what was the first computer that they owned? And it could go all the way back to when you were a kid. It can be, whatever it happens to be.

And we'll start with Dale Sanders, chief Strategy Officer for Intelligent Medical Objects. And someone who has a decorated data background. Dale what was your first computer?

Well, I think it depends on how you define computer, right, buddy? I would say it was somewhere between a slide rule.

That was pretty cool computer in its day. And I had one that was in this hermetically sealed case, so that it was super precise. You didn't suffer from expansion and contraction. I mean, it was quite the expensive investment. But the first sort of electronic computer was a basic old four function TI calculator.

Yeah. Do you remember when the big thing was, should we let people use calculators on tests? Oh, totally. Absolutely. And, and there were some professors that would say, yeah, you're gonna use a calculator in the world, go ahead and do it. And there was others that were like, no, you need to learn how to do these calculations.

Yeah, right. I think we're going through that again with with ChatGPT. Like, are we gonna let them use it or are we not gonna let them use it? It.

Yeah, for sure. And it's inevitable we're gonna end up using it, so we might as well just get ready for it. Yep. Yeah,

absolutely. Pete Marks cio, WakeMed Pete, what was your first

computer?

1982. I got a Commodore Vic 20.

When I was in high school, Vic. Vic, 20, man. You were before me. I was a Commodore 64. Yeah. Did you do anything with the Vic 20? What'd you do with it? You could program

in basic. I was a, I was an awful high school student. I mean, awful. And and for some reason I just really wanted one and saved up a little bit of money and my parents reluctantly kind of helped me get one out.

They thought it was a passing fad. But it made me interested in school and I think they were hopeful.

Yeah. Passing fad just like that guy that predicted we'll only need four computers in the world. Yeah. We'll do everything we need.

Yeah. My first, in fact, I programmed hangman on it.

That was the first thing I programmed on. Oh, there was only four. There's only three

words though. So, that's fantastic. For those of you who are just joining, we're we're gonna give it another minute. I'm gonna introduce our last panelists, Rick. Shepardson is the Chief Strategy Officer for Clearsense.

And Rick outta curiosity, what was your first computer?

You took mine earlier there. Bill Commodore 64. I wasn't programming on it, but I was playing games and I think there was a shooter game, the original kinda performance Atari there and that's what I remember.

Oh I love the Commodore 64.

That was an amazing computer. It was amazing computer to learn on. Had great games and you did learn some basic as you were playing around with it, but it would take this is not exaggeration, kids, it took like seven minutes to load a game. Like you would load, start out, whatever, and then you'd run the program and then you'd go like, I don't know, watch some TV for a little bit and come back and it would finally have started.

So I Donette tape, right? Was it cassette tape? I didn't have the cassette tape. I had the floppy, but it's still, it was pretty slow. I came with a cassette tape. Oh wow. We can, we can continue this conversation for a long time, but we're gonna get to the topic of modern data strategies.

Just for perspective people. We're talking 1980 computers and here we are in 2020, we're gonna be talking modern data strategies. So in 20 years, 40 years from now, I wonder what we'll be talking about when it comes to data strategies, but I don't know, maybe we'll get there. This one is leader, it's part of our leadership series.

We did a bunch of priority webinars. This one's on Modern Data Strategies. We really want to thank you for joining us and we wanna encourage you to continue to ask questions. A lot of you have already asked questions in the form. We have those, they're incorporated into our discussion today.

So we will be asking a bunch of your questions that you've already given us. We want to thank our panelists. I'm gonna introduce 'em real quick again.

Pete Marks is the CIO for WakeMed Dale Sanders, chief Strategy Officer, intelligent Medical Objects, and Rick Shepherdson, chief Strategy Officer for Clearsense. And Dale, you and Pete go back a ways on this data journey, don't you?

We do. What is it, 10 years or so, Pete, when? Something like that. When we worked together, helped.

Catalyst was working with you at Wake Med and Yeah,

well before that I would stalk him at HIMSS conferences.

Cause I mean, he's such a smart guy and I was like, this, what's

he doing? Well, it was an easy chemistry with you from very beginning friend. It was, and we always had fun with the Army Air Force theme,

and yeah. Yeah, we, I appreciate that. Well, here's where I'd like to start. Dale, I'd like to start with you.

I mean, you've been doing this for a while. Talk about the evolution that we're looking at and when people are talking about modern data strategies, data fabrics and whatnot, what, what are they talking about? But let's start with just a little bit of history of where we've come from.

Obviously we've digitized everything and we I don't know, started collecting tons and tons of data. I would argue without a clear plan for how that data was going to be used back in the day. And that sort of led us to a lot of fits and starts around the data side. But give us a little of your history in this and how you see this thing evolving or moving.

Well, it's I'm a little frustrated actually, because I feel like we're resting on our heels just a little bit. Believing that the electronic health record is sort of the end of the digital data journey for patients and we just have to round out that data. I always go back to my experience in signal processing in the military and satellite telemetry and all that kind of thing.

And, we're still collecting a very tiny slice of data on patients. Right. And I've been saying this for 20 many years, I feel like a broken record. But, a patient comes in to see a clinician three times a year. And if you think about digital sampling, if you think about a human being as an analog creature, an analog signal if you really wanna understand that analog signal, you have to sample that signal at a very high rate, nyquist and those kind of sampling rates.

So conceptually, we're just not sampling enough data about patients to be really precise. We're sampling enough to drop a bill on three encounters a year. So, I'm kind of wondering at what point. And when will we really form a true digital strategy for patient care that starts to round out way beyond what we collect three times a year in an encounter, but also in their lives when they're not in those encounters, the other 362 days of the year.

So I'll pause there for just a second Bill. I could go on forever on that.

Yeah. Well, I think you and I in a podcast episode, we talked about the whole patient profile and you talked about sampling rates, but also the whole patient profile of just, e even in the healthcare journey, we don't collect enough information for healthcare, let alone the whole patient.

We talk about clearly there's a bunch of sensor data. We talk about what people eat. Obviously what people eat is huge when it comes to health and wellness. They're. Ability to move and exercise and whatnot. And one of the questions that came in from one of the people who's listening to this is about all these, we have more sensors and maybe they're not as sophisticated as we want 'em to be yet, but we have a ton of sensors.

But where is that data gonna go? And are we even thinking about how we're going to use that data? And is that even the right starting point? Is the right starting point to really define what information we want about the patient and how we could use it. And Dale, I'll start with you and then I wanna open this up a little bit.

Well,

this is where I do draw on my military background and, working in the intelligence community I believe we need to first create the passion for data that the military has. Military leadership is, Pete can attest our passionate about data, right? You're in trouble if you don't provide a battlefield, battlefield commander the data they need, right?

And so you have this notion of an intel analyst and that intel analyst is responsible for a sector or a topic or a country or whatever, right? But they have this area of responsibility and their job is to synthesize subjective and objective sensor data into a picture of what's going on in situational awareness, and then present that to the decision makers.

I really believe we have a need for that in the future of healthcare. And I've been calling it a digitization for a long time. Some people think it's an informatics person. It's, I think that might be one, but. A physician doesn't have time in these encounter based fee for service models we have now to absorb much information, right?

They're treating and moving on, they're treating and moving on. But think about having an Intel analyst that preceded that encounter, that prepped that physician with a situational awareness, informed by data, subjective and objective about that patient, what it might do for them. So that's kind of my dream in the future, that we evolve that informatics role into that Intel analyst and they have a panel of patients are responsible for, and they're collecting data both manually and with sensors to inform the best care possible.

Interesting. So Pete, I want to come to you because, a couple of the questions had to do with how do you develop that culture where the leadership craves data, understands data what they want to create that data culture. They want to create that environment where data is gonna be used to really move the needle with regard to the care of the people in our community.

Talk a little bit about what it takes to do that.

When we started and again, Dale was kind of here when we started. We, and one of the reasons that we enjoyed working with him and still do is that instead of being technology focused, he was very outcome based, focused, very healthcare based, focused, and our journey started with trying to do the more simple things.

The first thing that we did was in urology and said, What's the right amount of opioids to give pre, during, and post-surgery for a patient? And can we show that and then measure that against do they come to the ed? What's their pain after the surgery? And so to your question, when we take that data and we align it and then demonstrate in that case that this is the right amount of opiates that make sure that our patients don't, readmit, don't come back to the ed, have best outcomes, don't have a lot of pain, we take that data and we show the inputs to the outputs.

When you do that and you do it successfully from partnering with your providers, they eat it up. Right? I think we had some concerns that they would think, oh, gee whiz, you're kind of. Looking at me and you wanna know how my clinical practice is gonna do and we dispelled that right away.

We said, no, we just wanna put this together and you own it. You can own the data. Every provider I know wants to do the right thing for their patients. And so all we did was say, here are the inputs, this is the effect on the outputs. And then that really helped our model take off. So when we talk about data, I think we need to talk about the outcomes.

Yeah it's interesting. I've seen you present now a couple times and you talk very specific use cases, very specific data inputs and outputs and results. Very measurable. And and I think the thing I like about what you've done is you didn't try to boil the ocean. You essentially chose a couple of paths.

Showed a in some cases, a return on investment, some an improvement of quality or reduction in opioids. You showed those things and it's just cotter's model of essentially, look, we are, we're delivering. And all of a sudden it creates that culture of, Hey, we've done it. Can we do it over

here?

Yeah. And it's a flywheel. So once you do it in that area of surgery, you can do it in any area of surgery. Once you do it for anything you're creating the basis on the framework and the flywheel, and then you create it and it just continues to take off. And our C M I O, Neal Chawla, who, Dr.

Neal Chawla, just a great guy. He also comes in and says, I like to not boil the ocean, but he said, we need to be able to command the ride on a skateboard before we can get on a Ferrari. He uses it all the time. It's taken off at Wake Med and everybody comes in and says, we start with a skateboard. We start with a skateboard.

We do something simple and

then we build. Interesting. Rick you're taking the place of Terry. And Terry couldn't make it today.

I'm doing my best standing for you though.

One of the questions actually referred to Terry as the queen of data governance, and we're gonna go ahead and bestow on her the queen title. But that's where I want to take this a little bit in that how are we doing on data governance? I mean, is that concept clearly understood?

Are we making progress in that area in order to do these things? Or are we still struggling a little bit in that area?

I'll tell you, I think that there is a lot of organizations that are a awakening to the needs for data governance. I see that many many groups are looking at ways to empower their stewardship communities, analytics communities, because we have a garbage in, garbage out problem, and we have more and more data produced all the time.

So the only way that we see opportunities to really drive some of that change is to get the individuals, Pete mentioned, the source and the target, and you bring it to the providers that are gonna have the impact. Well, we see needing to involve those types of individuals in the data and increasing their literacy.

And we see as people really wanna be able to use chat G P T, for instance, but maybe they don't trust it yet. A real opportunities to turn that story around and give individuals a seat at the table, be able to engage in. That data journey and trusting that data and bringing a data governance discipline is becoming, I think, more and more top of mind.

So there's a lot of opportunity out there, I think and people are starting to realize that opportunity.

Yeah, you're you bring up chat g bt and we've decided to give $5 to Alex's lemonade stand and childhood cancer for every time it's said on the show. And man, we are off to the races.

It's unbelievable how much money we are raising on that one drive alone. But it does change things. It is changing things in the way people think. And I think one of the things, there's a misnomer that's sort of going around of, Hey, do we really need data governance anymore? I mean, just shove all this data at chat G P T, it sort of figures it out and then you, away you go.

And Dale I want to come to you cuz we, you and I have talked a bunch about. What data do we collect? Obviously there's data we need to collect for regulatory, there's data we need to collect for billing. There's data we need to collect for care. But we've gone back and forth and argued, we collect almost too much data at times or data that's not relevant or not necessary.

And you started this by saying, we, we actually need more data. So I'm curious because, which direction are you looking at? I mean, should we be going through what we're collecting all the time and saying, Hey, do we really need this data? I'm just curious. Well,

I always start off a data strategy with two themes, which is, first is optimize the data that you do have.

And part of that optimization of the data you do have is asking yourself, do we need the data? Right? And are we, is it recreational data collection or is it meaningful data collection? So first theme is always optimize the data that you do have. The second theme is start to acquire the data that you don't have and that rolls into a, what I call a strategic data acquisition strategy.

Pete and Rick, you guys might remember years ago we had IT strategies, right? And it was mostly around computers, networks, right? Infrastructure kind of things that we, none of us really paid attention to anymore. So that was an IT strategy, right? Connectivity. Maybe. Then a few years ago, there was a digital strategy, right?

And that really amounted to people having a website and web presence. But what I've witnessed the last two years in particular is this new uptake of data strategies. Like a couple of years ago, you never saw the title of data strategist. You never really heard people talking about a data strategy.

Certainly in healthcare you never did. So there is this new thing that's now people are starting to realize across all industries in healthcare that data needs a strategy. It's like the most important persistent asset you're ever gonna have. Buildings are gonna come and go. People are gonna come go, but data's pretty much here forever.

And so, yeah. Anyway, I'll pause there friends, but I, but let me make, let me, I wanna make a plug for Pete for just a second here. What Pete and Neil have done at Wake Command, along with, Chris and others that were there, is they did the soft things right around the data strategy. You heard him talk about outcomes.

You heard him talk about that subtle thing. We're not gonna be your watchdog. We're not gonna strip away your autonomy. We're gonna give you the data as adults and let you decide what to do with it. What they did. There is the right thing with the soft skills around data and culture. Enabled by the technology.

The technology is not that important. The technology is becoming more and more of a commodity. The soft skills are not, and that's what they've sustained that at WakeMed and it's super impressive cuz it's not hard, it's easy to sustain the technology, it's hard to sustain the culture.

You, yeah. It's, and Pete, you and I had a call yesterday for something different and you were presenting, Hey, we're going into a new area.

I think it was virtual care we were talking about. We're going into a new area and we're gonna develop a whole set of things around this. And you sort of alluded to this, that, once you have sort of the machinations, it's, it is just a matter of hey, point, all those things at a new target.

And and it sort of comes together and you showed me a picture of everybody was in a room. You're sort of kicking it off. Give us an idea of what that kickoff meeting, what that entails, what's the conversation look like?

So I'll relate it particularly to data. So we, one of the things that was helping us is we were just so bad at data.

I mean, there was nowhere

to go but up.

Well, no, that's the best place to start is when it's really bad. Cause you get Yeah. Measurable improvement very quickly. Yeah.

We thought our analytics program was writing reports and interestingly, When we were in that methodology of writing reports, we didn't realize, a, we could never keep up with it, but B, we could never define how it was really helping our patients and families at Wake Med.

We couldn't measure it. And so we did a couple of things and these are really practical kind of things. We found a good partner, right, who was very outcomes driven. We created a good business case but we brought in a large contingent of the WakeMed organization into the process.

The other thing that sounds small bill, but I think it really helped us a lot. Is we got our board of directors involved and we created a measurement. We didn't create it, we used a measurement to say, how mature are we in terms of our data and analytics program? And we went to the board of directors and we just said, we want this on the WakeMed goals for the year.

And we were zero the first time that we measured. And they got really interested because when you are struggling with something, the board of directors actually really wants to know what's going on. They're not, most organizations I've been in, they're not an organization that is not really paying attention.

They're smart people, they have great experience in history. And so we would go on the board of directors and they would ask the most fascinating questions, and then they were on board with our analytics journey. And we would come back every year and talk to them about it. And they'd say, tell me more about this.

I mean, these are attorneys and bankers and people in healthcare. They're smart. And that helped us continue pushing our program forward.

I was gonna ask you who was on your board? Because a lot of times it's people from other industries and so when they're kind of frustrated.

They sit in that room in healthcare and they go, Hey, can we get this? Can we get this? Can we get this? And we sort of look at 'em and go, yeah we'll be able to get that to you next month or what? And they just look at you like, I mean, these feel like core things you should be able to tell me.

I mean, yeah.

Yeah. We had a, an owner of a number of auto dealers in the area and he would describe the data he uses, but that really made him, somebody who was with us in the circle of this data is an untapped resource. And then my final thing on this is I. We have been asking our care community to put data into a system forever, we owe it to them now to take that data and do something with it that helps their patients get better.

I I don't think we have much more time if we don't start doing that, they're gonna be like, really? Why am I doing this? And so, some great opportunities.

It's interesting. There, there's a couple questions I wanna throw out, and this is more just me wanting to know the answer. And that is when I think about these we're doing a lot of things at the point of collecting data.

So, nuance has their announcement with Microsoft and Epic and, we're doing all these things at the point, is this gonna make the collection of data easier and is that gonna make data governance easier or harder moving forward? And I'm not sure, this is a pragmatic question, I'm just sort of curious if it's gonna get easier or harder.

I, who, I, you who wants to take a stab at that one?

Y I'll take the first stab for folks and then I'll let people maybe play off of that. I think it's only getting harder, I think the more data sources that you have, right? The more that you're gonna have to rely on different actors across your organization or outside of your organization.

You've got more that comes into play and you mentioned change management considerations. Pete mentioned aggregating or getting everyone involved. You've got more people to involve and you've got more data to manage and to, really, as Dale kind of said, bring those 360 s together, you're gonna have more data to manage.

So to get that 360 view of your patient or your provider or the like, so it's only gonna become maybe more important. I hope it's not harder. I do hope some of our technologies and tools maybe help make it easier. Maybe they aggregate data for us and help us validate by exception. Maybe they allow us to collaborate in newer, more effective ways, but we gotta get on this journey.

And if we don't, then we're gonna have a really hard time as a human society, I think as as the ai comes on in, in greater force.

  📍

 we'll get back to our show in just a minute. Our rural healthcare systems face unique challenges in America. Join us for our upcoming free webinar, Rural Healthcare Challenges and Opportunities, on August 3rd at 1 o'clock Eastern Time. We'll unpack these challenges and look for opportunities for smaller health systems to take the lead in the delivery of care to this underserved population.

Join us as we look to unlock the potential of technology to make a difference in the lives of thousands. In rural communities. Remember August 3rd, 1:00 PM Eastern Time and 10:00 AM Pacific time. You can register over on this week health.com. Now back to our show. 📍

  📍

Dale, I want to come back to you. The modern data strategy, one of the biggest things about it is where the data resides and how it gets transformed.

Alright? So in a modern data, in these data fabrics, they're essentially saying, leave the data where it is and we'll transform it when we're ready to use it. And I'm sort of curious cuz it's, that's not a new concept. We're calling it a modern data strategy, but it's not a new concept. How is that different than what we've been doing before?

Well, there's been so many attempts at that over the years, right? There are guys to create a lingua frankwood that would cover all sources of data and you can query it and leave it in residence. I'm telling you, it doesn't work and it won't work, and we can try all day long at that if you want to. In fact, those of you who might remember my history, it's late binding, right?

So there's early binding, late binding schema on Reed schema on, right? And all that kind of thing. And, every once in a while you can pull off a late binding schema on Reed. Lead the data in its native system inquiry every once in a while. Simple use case. You can do that. But the truth is what you get out of those kind of queries is barely useful.

And you never get away from the need to harmonize the data. And so at some layer you have to harmonize the data. Now the good news about what's happened with N L P and the convergence of N L P and machine learning like that didn't happen until about five years ago. Like I've been in N L P since early nineties.

And what we have now is AI machine learning techniques, bringing that , into natural language processing for the first time over the last five or six years. And so we are going to be able to do things with nlp. That we've never been able to do before For sure. And that's happening even though there's too much hype around large language models right now and chat g b t way too much hype.

It is gonna have a big impact on us, but it's still not going to get rid of the need to have some degree of harmonization in a layer that allows you to interpret what's been in those source system. Then again, that's kind of where data governance comes in, right? Data standards common languages, common terminology.

That's the reason I basically got out of analytics, when I exited Health Catalyst and got into I m O at the front end of care and terminology because I got tired of trying to clean up data on the back end and I wanted to make data better on the front end with terminology standards. And that's the, that's exactly why I got into imo.

So we need a harmonization layer in order to get value out of the data that is what you're saying at this point.

Yeah, and you can go back and read things like Shannon theory. Matter of fact, I encourage everybody Study up on Shannon Theorem and study up on some of what Norm Chomsky wrote back in the fifties about information entropy.

And that's what we're dealing with now with all this new data. We're actually introducing more entropy into our information models. And so you gotta have an understanding of how to reduce that entropy. Right. And kind of using some of, Shannon's techniques and concepts. But you know, there's no, if I travel to France tomorrow, I couldn't get along without a translation and harmonization layer between me and the rest of the people in French.

Right. That's the same thing that happens with data. You cannot get away from harmonization.

Let me ask this question and then I will get back to the more pragmatic stuff that I think a lot of people came in here for. Is the healthcare system the right locus of the personal health record?

Is it the health system? Because we have not really shown an ability. To, WakeMed, Raleigh, that whole area. I might see three different health systems, three different providers, and we haven't shown a real good propensity for bringing that all in, harmonizing it and ma and providing value.

A at the point of care. We absolutely provide value, but outside the point of care we still don't have patients using their health record much. I would say it's low percentages that are actually using their health record for value and adding to that health record to create that value.

And my premise is the health system isn't the best locus for the personal health record. That there needs to be sort of an outside entity that's collecting all this sensor data and all this other stuff now. Again, there's no economy for that yet, but I'm just curious if you think the health system should be I'm gonna ask all three of you that question.

I'd love to hear your thoughts. Pete, we'll start with you.

That's deep. So the short answer to the question is what we're trying to do is make and keep people healthy and I've been working in healthcare my whole life, but I don't feel like I have command over my own health information.

It's not bad. And you would think that I'm somewhat of an expert in putting that health information together. So the big picture that you're really talking about, which is a great question, is how do we take that data from all those sources where it's owned by the person who is receiving the care?

Right. It's very transportable. And then your secondary question is, is the locus at the hospital? And I guess I would say right now, today, right? The collection agency is. The hospital. And if you've ever tried to get claims data, I mean, just, wow. I mean, that's a hard problem, which we're trying

to get.

That's one of the que that's one of the questions that's in here, so, yeah. Yeah.

So, you have to start. I have seen a lot of third parties try and condense all that information, and I just haven't seen a lot of success yet, but I still have a ton of hope. So I think we just, we know we have to continue to evolve.

What I think the health systems need to do is say, this isn't our data. Right? This data belongs to the patient, and then we need to be very accepting of. Even organizations that will probably fail along the way as they continue to move towards this is the patient's record. They need to have access to it.

And so do the providers that they go and see like, the blocking and releasing all that information. There were arguments that, that's gonna be a struggle. We're gonna be on the struggle bus, but sometimes you just have to start somewhere. Right. And then see where it goes.

And from an evolutionary standpoint, I think that's not a bad place to start to get people to say, even though I can't understand what it says, I least I have it. And for me, sorry, last thing. I don't have a lot of complex healthcare problems, but when I was trying to figure out one of them, I was reading through my own records.

I, well, I took my name out and everything, and then I just threw what the doctor wrote into chat, G P T, and I was like, what are they saying? Interesting. And I was like, I got it. Okay. Makes sense. So we have a lot of evolution to do. the point is people need to own their own health information.

Rick I'll come to you.

Do you think that the health system right now is obviously the center of gravity for the health data, but do you think there's an opportunity for an outside source?

A hundred percent. I've been tracking the personal health record for some time even believe that it might have to be born in a developing country.

So we're not stuck with our some cost infrastructure that we have, because you're right, we don't have an economic model and we have a giant incentive issue. We don't really have individual personal incentives the way you might think. And ultimately the healthcare reimbursement models and value based models aren't maturing enough, fast enough.

So I think that in some ways the. The pay provider organizations, right, who are taking on risk and also providing care are the best proxy that we have for the incentive models that we might need for individuals, but they're also not quite able to do so. The economic model doesn't support them yet. I think we might end up in opportunity for decentralized autonomous organizations, if you're familiar with the Dows of the blockchain world, who might be able to come in and sweep in and help in some of these transaction based in center, the aware intelligent routing of information.

So lots of opportunities out there. I think we're really just getting started, but we gotta solve that incentive issue first and we gotta help people be able to take control of their own destiny

as humans.

Yeah, it's interesting you bring up PAs with the Kaiser Geisinger thing. And in that announcement they said essentially we're gonna do five more acquisitions and spawn a 30 billion organization.

And it's really interesting and it's based on organizations that have made progress on population health. That's what they're looking for. So it'd be interesting if that becomes a locus for that. Dale I'll give you the last word on this, then we're gonna get back to some pragmatic things on this.

Well, there's a really cool movement afoot thanks to the folks at Levitt Partners and the Karen Alliance around digital patient identity. So I encourage everybody to Google that and in and evangelize the concepts. They ran a really cool proof of concept from the Karen Alliance facilitated by Levitt Partners and Ryan Howles that put the center of the universe for their data.

Right. So we've got 21st Century Cares Act that says you've gotta give patients their data. The only thing we've been missing is that digital identity so that, Intermountain and University of Utah and my primary care physician in Walgreens all know that it's me requesting that data. So there is a really cool technical movement of foot right now, friends, and I think there's gonna be an economic model that might sustain it, that puts patients at the center of their records and their identity, and I'm super excited about it.

The, my favorite in that space, and I have no financial interest in them, by the way, is a company called Al Clear. Id Al Clear id, they've had a presence in Europe and the banking sector for a long time. Bo Holland is the founder. He's really thought this out. Very smart guy. Partnered with a good buddy of mine, Tim Zoff from my c i CIO days at Northwestern.

So I think all of us should get on that bandwagon and make it happen. And I think there's an economic model that's going be merged around that digital patient identity with them being the broker of their data.

Yeah, I get couple more resources for people. Any of the episodes I've done with Anish Chopra.

We go into this topic at length. It's one of his passion projects is to talk about this topic. And then I, I'm reminded of two other conversations, one with Charles Boise and the other with John Holka talking about different countries Charles talked about India, where we're seeing this personal health record emerge and really start to take hold.

And John Laka talked about one of the African countries. That, that utilizes it. In fact, in some of these, in Mexico, I've heard this too, people carry their entire medical record with them. I mean, that's just, that's considered normal. To like, show up with your entire medical record.

That's what we did in the Department of Defense.

Yeah. Right.

Yeah. We're not gonna talk about the D o D project or the VA

No. A long time ago.

It's different now. Implementation. What data is valuable as Pete, as you are moving forward and you're sort of looking at these use cases and you're looking at new use cases and that kinda stuff.

Do you find that there's data that's missing that you have to go out and get? And do you find that have you been calling information out as well? Have you been like, looking at the things going, Hey, you know what? Man, these physicians keep collecting this. We don't need this. And it's contributing to the stress and the cognitive load.

Are you doing both of those? Are you going after more data and culling out some of the data you've been collecting?

We're the culling out Bill isn't what I would focus on. I would focus on kind of like Dale said, can we take all the data that we have access to at WakeMed and put it in a way that we can get to it rapidly and look for outcomes quickly?

So we've done it across our clinical space. Every system that comes in here, we have evaluate not only from a security perspective for, from a data perspective, and it's every Thursday we did it. So I'm coming from that fresh off that meeting and talking to a vendor, well intended person. And we were like, can we have that data?

And they said, sure, we'll give it to you as a common eliminated file. And we were like,

okay, we'll give you PDFs. No,

sorry, everybody's doing that. They, it's still happens. It's. And we turn in a very friendly way and say, look, our table stakes is that data flows right into our model, right? So we appreciate you and the way that you approach this and 99 times out a hundred, they go, we can do that.

It's a little bit more money. But this really gets to your point. So we are very strong on the clinical data side. We are strengthening on the use of data, let's say from supply chain or from finance or rev cycle. But if you're gonna do something like I talked about earlier where you say the right amount of opiates for a patient, here's the readmission you also should say.

And with the, with those lower readmissions, right? What's the financial impact of that? And if you don't have all of that data within that ecology, Right? Then every time that you want to get it, you are reinventing the wheel. So you're gonna go back and you're gonna say, well, how can I get this data?

How can I bring it in? And you're only getting that slice. So it's much better to have almost your periodic table of elements out there and you pick the ones that you want to create the chemistry. So the way that we do it is as you come in as a system comes in, we ask those questions, but we have a lot of systems that we're still trying to put that data together in on the back end.

But perfection for me looks like all that data is easy to get. Then the governance is important cuz you don't wanna embarrass anybody by having bad data meh together and give you the wrong thing.

Right? Is there value in the claims data?

Incredible value. You need it for pop health. You can't do pop health without claims data

21st Century Cures was supposed to help this, but I don't think the carrots and sticks are in place yet.

Are we seeing movement getting easier or better?

Technically it's still pretty hard, but there, there are vendors out there, there are processes right now. You just have to be willing to say that data model and the carrying of close gaps is worth it for our organization to do it and our organization.

It's worth it.

There's questions in here and people are asking about, we have been talking about self-serve analytics, self-serve insights for, I don't know, Dale is laughing we've been talking about for a while. I saw one of the use cases for chat G p T is they're gonna put it on top of slicer dicer.

And give it a a whatever. A human inter, not a human. Yeah. A

natural language. Natural. Natural language. Natural language. Query's been around for a long time. Yep.

Yeah. So, is self-service analytics, are we making progress in that area or is that still a little ways off?

Oh, I think we overkill it now.

That's my personal opinion. I think we have a tendency to overload people with self-service analytics right now. It is become so easy. So I it is, it's not about self-service analytics, it's what I it is more give me the data that I need when I need it in the form that I need it without me asking.

So you wanna find more and more ways to put it right into the workflow.

Yeah. Anticipate needs and deliver that data at that point of need. Right. Instead of, so I do think there's room for AI and machine learning in that space actually to predict when data is needed in the workflow of a clinical encounter, for example.

And what kind of data. Yeah. Hey, before I forget guys, I want to give a plug to Mickey Tripathi at O N C. By the way, onc and that team they've really taken onc to a whole nother level of pragmatism and pushiness in a very good way. Right. Mickey has been a great addition to the o c and the o c team has been right on board.

So Ryan and Steve Pozak and those others, they're doing a really good job enabling and pushing data governance and interoperability and patient access to data. Yeah. I just wanted to give them a plug.

Yeah, absolutely. Mickey's been on the show a couple times. I, and he's a phenomenal, what's the word?

Word, evangelist for for interoperability. But not only that, but behind the scenes, he used, you used the word pragmatism and pragmatic he's already done it in Massachusetts. Yeah. So it's just a matter of bringing that forward. What are we gonna see?

I said I was gonna ask you about this. I didn't never got back to it. What are we gonna see from the federal government? I mean, do we really see their roadmap? Right now? It's essentially, it's interoperability. It's us C D I, I guess we're on iteration three of US CDI I are. We just gonna see all of these things continue.

I think we're gonna continue to see it. I think we're gonna continue to see more standards. I, at least I hope we do. I don't think that we can share data as an ecosystem unless we establish more of those standards. And I certainly hope that we're gonna be able to see maturity in U S C D I in fire more, open exchange of data between all of these different actors in this healthcare ecosystem we've been talking about.

I sure hope we get there.

Yeah. We don't want to compete on data,

right? That is,

yeah. Competing on data is the exact wrong thing to do. So the more that we can standardize data where patients own it and it can be shared to for their patient experience, if we're gonna compete, we'll do it on quality care.

And on a culture of patients and families. Said,

well said

Dale I wanna come to you with this cuz when they were rolling out U S C D I version three and those kind of things, one of the things they'd hoped would happen is the industry would sort of take areas and coalesce around it and drive standards.

I'm not sure we've seen that it's still all really coming out of these groups from within the O N C. Do you think, we'll, do you think that's the way it's gonna continue? Do you think we'll just see U s cd cdi I version 4, 5, 6 7? Or do you think we'll see like a group of oncologists come together and define certain segments of the data?

Well,

you're kind of touching on one of the concerns I have with U S C D I, I mean, first of all, it is awesome that we're now starting to think more broadly about the data we need, right? So that's U S C D I in that regard is super helpful. The problem is underneath those data concepts is the terminology, and ICD is kind of a clinical disaster.

ICD 11 is not any better. It has all sorts of pros and cons. Snomed is a clinical disaster to use. Loin is a disaster, right? So we have these,

I was worried we weren't getting any quotes from you, but thank you.

It really is. I mean, I, and I, I grew up under the tutelage of Stan Huff at Intermountain, one of the leading thinkers in the terminology space, right?

So I've been watching, dealing with terminology since 1997 in healthcare, and even before that, and now with imo, I'm fully embedded in it, and I've really come to appreciate how messy these terminology standards are and how difficult they are for a clinician, especially to use. Not to mention, and Rick can attest to this, they're hard to use on the backend in analytics.

Yep. So that's a part that I'm still not quite sure. It's like, I think we have to have a revolution around the terminology in order for U S C D E I to become successful. I guess that's the summary.

Yeah. I think you're right. And, I've actually done some work with IMO in my history and I know what you're what you've been working on over there in terms of, combining, normalizing some of those terminologies.

And I think you're right on that. That it's necessary. I think There's terminologies in a number of different contexts. We have conversations with payers and providers on an hour by hour basis it seems, and they don't speak exactly the same language if you say patient and you should be saying member all of a sudden, like we gotta stop the conversation, and it you need that translation engine.

And that's actor or context based and, it's also organizational based. So one of the things we believe, and I'll toss this back over at you, bill or Pete, but. We really think a citizen data management, right? So at Clearsense, really thinking about, you talk about self-service analytics, citizen data science is a thing, right?

Why not Citizen Data management allow individual organizations, people and organizations to manage their terminologies that they believe in, map them across terminologies to terminology sets that they believe in, or then allow them to put those to work for them in managing their data, in producing analytics, and being able to do this in a very intuitive way, right?

User-friendly way. Cause let's be honest we don't have a lot of education pillars that we used to have. And we're not gonna have a lot of data scientists, data engineers to do our work. We're gonna prompt engineers, so we better make this pretty easy

for everyone. Prompt engineers does not count.

You did not say the word chat, G p T, but I just did. So there's another $5. Let me go through some of these questions. Rapid fire. Pete, I'm gonna focus in on you cuz a lot of 'em have to do with where you live. With the health analytics team, overwhelmed with requests and leadership pressure to grant end users access to data query tools.

How do you balance data availability with data validity and quality?

I could talk about data validity and quality. So we can't go too deep cuz I know it on time, but we have specifically on the clinical side, we have governance groups that say these are our first priorities for data. Okay. We build right with them, then we turn it back over to them and say, you validate.

And that creates, right, the quality. So, and if they don't validate right, then it really doesn't go into production. So that's a really fast answer and probably not nearly deep enough. But I, these things we try not to have too much technology talk. We're like this, let's be practical about our approach, right?

It's like doing anything else. So the flywheel I talked about is just P D S A, it's all P D S A,

I'm gonna come back to you. How do you strategize around getting the clinicians and frontline folks to be active participants as it relates to data quality, et cetera? It seems we need that correct mindset from those folks in order to be able to drive outcomes.

So I guess the question is how do you get them engaged? And I think we touched on this a little bit. You get the quick wins and you've really put the data governance back in their hands, in the data quality back in their hands. Anything else you would add to that?

At a high level?

Yeah. Right, so providers wanna do the right thing, the and they're incredibly smart people, but there's way too much data for them to accurately be able to say these things that I do create these outcomes. Right. But what we're doing right now, we've been incredibly successful on the clinical side and we're trying to move to the operational side right now.

And it comes back to, I think, an educational session and working with all those folks. So we have gotten permission to bring our entire operations team led by the c e O to a one day offsite where we go, what are we gonna do with our operations data? How are we gonna put it together? What are the things that we want to achieve with a skateboard in mind?

And then we create almost like a one year to 18 month plan. These are the things that we're gonna knock out. And then the most important thing to me, bill, is let's not get too focused on descriptive data. I mean, it's important and sometimes you have to see it, but what we're really trying to say is, What does this data, how does it create better outcomes for us, either operationally, financially, supply chain, you pick it.

And so now we are taking the success we model that we had in the clinical side, and we're gonna adopt that model into the operational side and create an 18 month plan and with the CEO at the table. And that's important. That's important.

I'm gonna do an exit question in a second. I wanna talk about patients and if you have a couple seconds, just address this because there's a question here.

There's a bunch of questions about pace. How can organizations create high quality analytics that serve as many use cases as possible in the fastest manner possible? So clearly there's a lot of use cases, there's a lot of demand. Everyone sees the need everywhere. But the thing I like about, I want you to talk a little bit about patience.

And we talked about boiling the ocean early on. Why have you moved in this methodical pace and why do you feel like it's the right way to approach it.

So why do we move a methodical pace?

Because, and I believe in giving data to everybody, but I think that comes with some education and you have to really change kind of the way you think about it. So if you want to have some fun going to your healthcare system and try and figure out what's the definition of length of stay, right?

That's gonna help you understand that the data definitions are wildly variant. And so if people jump in and just plug things in, they're gonna go, oh, this is the effect on length of stay. And somebody from finance is gonna go, wait a minute. That's not a finance length of stay, or that is not a nursing length of stay.

So I believe in pushing the data down to people, but you have to educate them at the same time and you have to accept that as you go through this transition, folks are gonna be wrong. And it's not a chance to go, Hey, you're wrong. It's a chance to go, Hey let's learn more about what this data looks like.

And then methodically, the other option was to just open everything up. And chaos and suits didn't be clear. That didn't seem clear. And you couldn't keep everybody on the bus. I think we'd lose people. Yeah. They'd fall off the bus.

Alright you each get 45 seconds with this question. How is generative AI going to impact data in the next three years?

Dale will start with you

in the next three years. Give me what I might be able to tell you the next three minutes. Well, I think with all things, it's gonna be very good things and very evil things, right? People are gonna produce a lot of false information, more false information than ever that looks real.

And that's gonna be used to manipulate people. On the other hand I think it, it will accelerate, ideas and drafts and things like that. But I, I gotta tell you, my experience with it right now is it's full of holes and there's no sense of data quality. Like where do, where all talk about large language models what does that really mean?

Like what, what's in your large language model? Is it CNN or is it Fox News?

Right.

Right. And there's no transparency. We have no idea.

No, no transparency. So I honestly don't know, guys where this is gonna go in three years. I have no clue. No clue.

No clue. Well it, it looks like Pete, you actually typed it into ChatGPT what did it say That how's generative AI gonna impact. Data in the next year or so.

It's gonna be, to me, it's gonna be in the hype curve. We just need to keep eye on it. We need to try and do the right things. We need to find use cases. We brought one to Epic and early and they're kind of jumping on it and we're thankful on the personal side.

Our, my C M I O and I wrote our wives Valentine's haiku through it, and that is, that, that was major points. So I would just suggest that it was her,

I'm making a note to

this.

Yeah. I put

her name in it and a couple of things, and she was like, you're so creative. She still doesn't know.

So,

But I

think that people should get in and play with it and then they can understand the scope and then have important conversations. Like Dale said, Fox News, is it CNN Those are important conversations, but. Understand

Rick last word, and then we're out.

Three quick

things, right? It's going to give everyone a lot of templates to start from.

We're gonna accelerate ourselves if we're smart. It is also going to. Make people realize they gotta go figure out how they trust their data and how they build quality in the source of that data. And they're gonna have to go figure out that. And then they're gonna yearn for transparency.

They're gonna yearn for that white box they're gonna yearn for. How does that intelligence generate, how do I interact with it in the future better? And it's gonna spawn last one. It's gonna spawn general ai. We better be careful what we ask for because generative models are, reach their apex. And, next stage is is general ai.

We gotta get it ready for that.

I, my, my three seconds on this is we're gonna go from large models. We're gonna go to narrow models very quickly through Microsoft. We're gonna get an agreement, we're gonna point it at specific data models. Once we pointed at those data models that are essentially medically based, it's gonna, we're gonna take it to medical school.

I don't know where it's gonna lead. But it's going to be interesting as it starts to get very narrow focus on on our data sets. Yeah. What it can and cannot do. And we don't know,

when DARPA started the internet, they had no idea it was gonna get like this. It's

okay. Right. That's a great point.

I was on the internet in those DARPA days. I never envisioned what we have now. You were

there before. The worldwide

web essentially. I was there way before that. It was one of the few sites we were on it. Yeah. And who've ever thought, I mean, there's just, yeah.

Yeah. Well, I wanna thank everybody who joined us.

I wanna thank our panelists great discussion. Really appreciate it. Hey, thanks team. Great discussion. Thank you

guys, honor. Thank you. Thank you.

Bye-Bye.

I love the chance to have these conversations. I think If I were a CIO today, I would have every team member listen to a show like this one. I believe it's conference level value every week. If you wanna support this week health, tell someone about our channels that would really benefit us. We have a mission of getting our content into as many hands as possible, and if you're listening to it, hopefully you find value and if you could tell somebody else about it, it helps us to achieve our mission. We have two channels. We have the conference channel, which you're listening. And this week, health Newsroom. Check them out today. You can find them wherever you listen to podcasts. Apple, Google, overcast. You get the picture. We are everywhere. We wanna thank our keynote partners, CDW, Rubrik, Sectra and Trellix, who invest in 📍 our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.