This Week Health

Don't forget to subscribe!

June 10, 2024: Sarah Richardson, President of Exec Development at This Week Health, joins Bill for the news. They discuss Epic's groundbreaking move to release an open-source AI validation tool on GitHub, questioning what this signifies for AI equity and usability in health systems. The conversation delves into the challenges CIOs face in proving ROI for AI investments, prompting us to consider how governance and data quality play pivotal roles in successful AI integration. Sarah's insights on Microsoft's Copilot and its potential impacts on organizational productivity lead to a broader discussion about the balance between efficiency gains and cultural shifts within healthcare environments. As they explore the implications of new AI advancements from OpenAI and Google, we are invited to reflect on the practical applications and future directions of AI in our daily lives and professional settings. What does it mean for a tech giant like Epic to open its doors to the developer community? How can healthcare systems ensure that AI models are relevant and fair for diverse patient populations? And as AI tools become more integrated into our workflows, what trade-offs must we be mindful of?

Key Points:

  • Epic AI Tools
  • AI Model Validation
  • Healthcare ROI
  • Financial Planning
  • Innovation and Risk

News articles:

Transcript

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 Newsday.

you have to really be thoughtful about where you put it in and what your desired outcome is, because if it's just for cost savings or just for efficiency, What else are you trading for that everything that's a win has a trade off somewhere in the equation?

My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of This Week Health. where we are dedicated to transforming healthcare, one connection at a time. Newstay discusses the breaking news in healthcare with industry experts and we want to give a big thanks to our Newstay partners,

Clearsense, Crowdstrike, Fortified Health, Healthlink Advisors, Order, Rackspace, Sure Test, and Tausight

Now, let's jump right in.

(Main) β€Š πŸ“ all right. It's Newsday and today I am joined by Sarah Richardson, newly, appointed, is that the right word? Hired? It feels weird to say hired because we've been working together for so long, but

officially recruited.

How about that? Recruited, yes. the longest recruitment period in life.

It was. Like years. It was a very long recruitment period. And we're going to do our first Newsday. I don't think you've ever done a Newsday if I thought about this. Have you?

I haven't. No, I've only ever done like the keynotes, the interviews in action, maybe a couple podcasts way back in the day when you We would joke, you'd be like, Hey, I have a podcast.

I need somebody on it. What are you doing tomorrow? And I remember the time I told you I would never buy an Apple watch. And now I've been obsessed for probably six years.

That's the thing. It's like on these shows, once you put down what you've said, I've had people come up to me and said, you said this.

I'm like, I was wrong. I was like, but if you're expecting infallibility from this show, you're not going to get it. But hopefully. We're looking at this stuff pretty closely. We're talking to a lot of insiders and we try to make the best guesses we can. this could be our warmup.

You and I and Drex are going to start doing This Week Health live. We're going to do our pilot on June 7th at one o'clock Eastern time. It's going to be live on LinkedIn, on YouTube, and on our website. And we're going to give people a chance to send questions ahead of time. We're going to have different segments that we do on an ongoing basis.

and we are going to talk about the news a little bit on that show as well. So this is a little bit of a warm up for that June 7th. And let's start with this EPIC story because this is out of character for EPIC and it's actually fun. So EPIC releases open source AI validation tool for health systems.

And they've made it available on GitHub for global use. The software suite allows healthcare organizations to test and monitor AI models. within their electronic health record systems. This tool, designed to ensure AI equity, can be used to validate models developed by EPIC or other organizations.

Health systems can evaluate performance and ensure fairness, equity across different patient demographics. Organizations like Duke and University of Wisconsin and others plan to utilize the tool for local validation of AI models. This is different. I don't recall Epic ever releasing, this code is in GitHub.

It's actually been released as open source code to the world. I don't remember this ever happening before. This is a pretty big step.

It is a big step, and here's why it matters with them as well. You think about, they tend to get heat for not playing all in, or for being too proprietary, etc. And so when you think about GitHub really is a space where, as you develop things and people want to be able to share more, they're putting it into those types of platforms.

But it gives Epic a chance to be able to say, hey, we are willing to play. We are willing to be some of the first ones out there to make a difference in these spaces. But I also love that they're talking about, AI equity and usability. You tend to hear about the different forms of commonly used AI, but you always hear about some of the applications.

So the equity and usability also starts to create a space where it's comfortable for people to step into it and see what else is happening. So you don't have to be the pioneer. But you can certainly use a highly trusted source, if that's your EMR as an example, in a highly trusted environment, which is GitHub for most developers, and start to think about, if I'm going to bring this forward into my health system, I'm going to go ahead and look at, an imaging warehouse, as an example, and detection models and radiology.

I can do that with Epic at the forefront, and it's not going to be as hard to find people to be able to utilize those environments either. So you're not creating unicorn users and scientists. It's all in the same realm today, and that's a big deal for all of us.

I think the biggest use case here is we have these models being developed, and there's not a lot of transparency into the models and how they're trained and how they're being developed.

And so if you take, let's say you take a national population and you train a model and you say, look, it's been trained on a national set. That still may or may not be relevant to you if you're serving inner city Philadelphia or San Francisco you just have different nationalities, different ethnic makeups, and so one of the things this tool can do is they can look at it and they can say, hey, look, this model is only trained on this ethnicity and this population.

And your population is 10 times that, whatever that number is. And it's model isn't relevant your location. That's a big deal. healthcare, that's a big deal. in some cases, we can't trust nationally trained models. really do need that granular level of data in order to validate these models.

When you have to have a rate that's the match rate, for example, is your accuracy rate. It's got to be in the high 90s. Otherwise, an 85 percent accuracy rate on an AI model is not going to be something you want to utilize in clinical care. So again, to your point, it opens up the ability to have higher accuracy rates, but also drill down to be local or to take a national view.

When you can slice the data in ways that you trust, then you really do create extreme learning amongst different health systems. And that's always what EPIC has endeavored to do, is They get better by learning from all their implementations and it aggregates into a space where they then can utilize it too.

So make no mistake, them doing this allows them to also be at the forefront of going to them for these types of endeavors. So they still stay The most influential in the spaces that they continue to lean into.

validation model. So it's clearly not a core code or anything like that.

I think I love about it though is my gosh, putting it in GitHub means all these I have a community of geek doctors and their term, not my term, but they're coders. They're literally coders who happen to be doctors who are playing around with FHIR and doing all this kind of stuff.

I imagine they're going to be giddy over this. They're going to be like, this is great to have them releasing code into this. And I hope this is the start of more of that kind of behavior. So that'll be interesting. Sarah. So we're going to do one more serious AI story, and then we're going to go all the announcements because I want to take a look at the announcements.

They're just a lot of fun to talk about. But this one, you're a closer former CIO than I am. And this story is from CIO. com. Where's the ROI for AI, CIO's struggle to find it. And they surveyed 700 IT leaders. This was Gartner surveyed 700 IT leaders about ROI.

And they said, this is hard. This is hard to figure out. What do you take from this story?

What I liked about the story though, in general, is when you think about understanding the true ROI, crucial. You have to justify the significance of that investment, and if you have better planning, better resource allocation, and the ability to achieve the strategic goals of the organization has in front of them think about that.

You can maintain executive support, you can secure further funding, and if you can get things right, the ability to demonstrate ROI influences the broader adoption across your either health system or other aspects of the organization. Helps you drive innovation, helps create competitive advantage, but you've got to have a place where you can start.

And so I always go back to that space of like, when you have a digital implementation to accelerate your business models, when you have the right governance and you go from the classical aspects of what you've been using it for to more of a futuristic goal, you can propel yourself way ahead of what some of those early strategic goals are, but you've got to be able to have.

The right structure in place to be able to do it. So I would say, you start with governance, data quality, data modeling. If you stop investing in data governance, you've already lost. So I feel like those teams that have to always claw for every dollar, this gives them exactly the foundation that they need to be able to justify some of the expense.

And you pick things that you can actually tackle. People can see. There's no shortage of different formats of AI. Heck, whether you're doing ML, NLP, you're going after the basics. clinical decision support systems, even your chatbots and virtual assistants, those all aggregate into what you're doing from this perspective.

And that's why we think about how to justify the money. Have the right structure in place. And then you can actually accelerate some of those. goals that the organization has if people are all focused together on the right things.

Yeah, love the fact that you brought up those other models because the one I know that healthcare is not struggling with is the ambient clinical listening.

Nobody's struggling with that. it's doctors can see this many patients. Doctors are spending this much time. You even have the measures within the EHR itself. You can see You know, how much time they're how much pajama time or those kinds of things. So you have satisfaction survey, although I liked in the article that they talked about using a net promoter score, which I thought was I'd like to see that used more in healthcare.

But then I saw a study from, I think it was WellSpan. did a study and the doctors that were given the option of going home or seeing another patient and they all saw another patient because they care about these patients. They don't like keeping them waiting for, two months to get into, have an appointment.

And so they're actually more productive and able to move forward. So that one's easy, but I'm going to have you put your CIO hat back on. Here's the interesting one. So, Microsoft slides that little piece of paper across the table and they say, Hey look, Copilot is being used by a lot of your peers.

And it's really awesome. And we talked to some of the people within your organization and they've played around with some of these features and it can reply to your emails and it can take your meetings and it can record those and it can, give you summaries and it can capture to do's and all this thing.

Think of how productive your organization is going to be, Sarah. Think about it. Isn't that amazing? So here, just check that box right there. And then, we'll multiply that times your number of users and we'll start charging you for Copilot. And look, you can roll it across your entire enterprise.

How do you respond to that?

My favorite answer to everything, as is always, it depends. It

depends.

And and here's why. And you get a space for, can you pilot, co pilot? Where is it going to be most beneficial? I would ask them for the tangible. Where is it making the biggest impact? Where is the cost of the investment worth what it's producing for the organization?

And talking to peers, I'm like, if all my peers are using, and all these amazing people have said it's so great. Let me talk to them. Show me. Let me hear about where it's making a difference. Because some organizations culturally, this goes back to how you manage the human aspect of what you do every day, some people don't want that to replace the executive assistants.

And this was a real conversation I had with a peer of mine who said, Hey, are EAs Are the most knowledgeable people about what's happening in the organization, period, that network knows everything. And if you eliminate them with CoPilot as an example, all of a sudden you don't have access to all of the tidbits of information that make this place hum and run every day.

It doesn't mean you don't implement it where it makes sense to implement it, especially for those that want it, but you have to really be thoughtful about where you put it in and what your desired outcome is, because if it's just for cost savings or just for efficiency, What else are you trading for that everything that's a win has a trade off somewhere in the equation?

Oh my gosh, come on. You gotta sign this document. Look it's only gonna cost You know, times 20, thousand users, the 20, it's only 20 per month per user times 20, 000 users. How much can that possibly be? I didn't do the math, but 240 bucks a year times 20, 000. Hey, that's a lot of, that's like real money.

It could be real money. I remember that I came from CFOs, And so you're always going to back into the number you're looking to get. Here's the thing. If it's really worth it, then you're going to watch somebody else. I have never been completely cutting edge most of the time, nor have I been a complete lagger.

I like to be right in the middle somewhere. So let me see it. Let me talk to people like me who have these same experiences and then really get a sense for where do you try something that has the least amount of risk. Incrementalism at its finest. You create something that somebody has that somebody else wants.

Then it's going to adopt itself. And for the right reasons and the right environments versus just telling people, I just signed an agreement. We just got 20, 000 users on co pilot. Catching up to the ROI is ridiculous, but somebody who really wants it, especially in an influential role, that could be a complete game changer.

So always goes back to how does the culture of your organization run? Where do you put the tools where they're going to make the most sense? Sometimes we don't have a choice. When we do have a choice, let's be really thoughtful about what we're putting where and why.

β€Š πŸ“ πŸ“ πŸ“ πŸ“

All right, it's June and I am wearing my yellow hat. And this June we're championing, uh, A cause close to our hearts, Lemonade Days, supporting Alex's Lemonade Stand Foundation in our fight against childhood cancer. In 2000, a young girl named Alex started a Lemonade Stand to fund cancer research. Her legacy has inspired millions in donations, funding research, and supporting families.

This June, we're inviting you to join us. It's simple. Just visit ThisWeekHealth. com and click on the cancer ribbon to make your donation. Together, we can continue Alex's mission to make a significant impact. Every donation moves us closer to a world where no child has to face cancer. So, take a moment, click on that ribbon, and make your contribution.

Thanks for your support, and let's make this June a month to remember. πŸ“

β€ŠYeah, you know what, when I see CIOs make that announcement like, hey, we're on copilot, we're all in on copilot and that kind of stuff, I scratch my head and I'm like, you know what, are you trying to make a career and a name for yourself?

Are you actually doing what's in the best interest of your organization? Because this article points out two things. And we've made these EHR. It's really interesting. The first one it points out is, What problem are you solving? And you made that point as well. It's okay, great. You're going to do this.

What are we doing? Like, why are we doing it? What are we solving? Are we literally getting more I don't know, productivity out of the general user, because it can suggest an email response in some cases that works, some cases it doesn't. I don't know. And I'm sure there's more sophisticated models than that.

That's one of the things it says. Know your problem set. And I think a lot of CIOs are approaching AI with that. What is my problem I'm trying to solve? The second one though, is the problem we made in the EHR is we did not commit enough money and time. to education of these tools. And so, signing that contract for 20, 000 users might seem like a good idea, but my question would be, what's your framework for adoption?

What's your framework for literacy? We still don't have literacy around some of these AI tools. What can it actually do? What are you doing around training? And if they say, Oh we're going to put that together. I'm like, you should not have spent that much money until you had that plan together and identified the problem sets.

And I think that's where we're in complete agreement that you just don't move forward that quickly. And then measuring the ROI, if you have the problem. I think becomes easier.

And the other thing too, is because all of us are stretched so thin financially, and it's back with my CIO hat on, is every ROI has the, has to be there within a year, or it has to be replacing or saving something somewhere else.

even if margins continue to improve in some of these health systems, you don't necessarily see the CIO. Getting that additional percent added to their bottom line. You're still held to a pretty strict percent of revenue to a degree. And so if you're going to bring something else in, then it's got to be able to have the impact that is also expected by both your board, your CFO, et cetera.

So there's a lot of conversations that need to occur before these things become mainstream.

Absolutely. we'll get into the fun one here. So Google. launch their AI assistance. This is an MIT technology review article. It's one of the better ones. We'll talk about OpenAI and Google, and then we'll also touch on Apple and Amazon, which aren't really covered in this story, but we'll just touch on them.

And just To set the right expectation up here. They're talking about general use of this. This is not an article around healthcare. This is more like, Hey, some really cool stuff just came out. So OpenAI came out with chat GPT-4o, O standing for Omni. And it's really good at listening to you.

It's really good at anything you can point your smartphone camera at, it can see and understand. They have some really cool. Demos, they're talking about 320 millisecond response delay, which is human esque. They have demos where you can cut it off. And it's able to adjust and come back to you.

They have demos where it's actually looking at a math problem on a piece of paper and it's educating a a teenager, I think on how to do trigonomic equations and that kind of stuff. So it can always Omni, it can see, it can hear, it can respond. to things. It's pretty cool. I think one of the biggest things that they did is they have a free version of it out there and they really want to democratize this.

Get this out in the hands of many. Obviously they have paid versions, 20 bucks and for teams I think it's a little more than that 30 bucks. Per user. been using GPT 4. 0 for a little while now. it's better, faster. I don't know the accuracy rate on it's a little higher, but we'll see where that goes.

And then Gemini, Google came out with Gemini in their what was that conference called? The IO conference. And it's the most closely aligned. Although they talked about another model Gemini Live launches in the coming months. It's roughly 20. It has the similar kinds of things in that it can see things.

It can hear things. It can respond. Chat GPT-4o is available and Gemini Live will be available here in a couple of months. And the last thing you can't really say which is better at this point because neither one is really fully out there.

And they talk about safety around them. So, did you see any of those demos? They were, pretty interesting.

They're interesting, and here's what I love about all of them. As they are all iterating and they're all coming into their own, play with both of them.

Honestly, how much do we all spend at Starbucks, as an example, every month? Pick up both. Start playing with them in your own realm. And then when they become part of your mainstay in your day to day life, it's not such an awkward thing. And I think that's a really important conversation when it starts to fold over into other spaces such as healthcare, as an example in the world that we live in.

So, I'm like, as they are and coming up in the ranks, try them both out, and pick your own use cases, and get a sense for how these things work in your life, because You think about all the connected devices you have in your home, or all the ways that you are already utilizing these features in your day to day life.

It just continues to become part of the fabric of how we live every single day. So it's not as awkward when that next leap comes forward. We're actually already anticipating it coming.

the thing I say to healthcare leaders at this point is, I used to have a couple of canned decks. that were always current that I could pull out.

If somebody said, Hey, can you speak to that board? Cause each one of our hospitals had a board. There were 16 hospitals and invariably somebody would say, Hey, can you come in and talk about technology? Sure. What do you want me to talk about? Oh, I don't know. Whatever you want to talk about. I'm like, okay, this is one of those.

I would have a canned presentation on AI, just absolutely would have it. it would cover literacy, like how do we get people to use it? What's the adoption rate would cover policy and governance, but it would really go in depth into use cases.

Like where are we seeing it used today? who's using it to respond to patient queries? Who's using it? What are they finding out? There's a lot of that information out there. UCSD is talking about it publicly. UNC North Carolina is talking about it, Stanford's talking about it. So you can just out in the public, grab these articles and have current things that you can talk about.

This is definitely one of the decks I would have ready to go. there were, there was other decks that I had at the time, but this absolutely would be one. Let me ask you this. So Apple's going to have their worldwide developer conference next month. And this is off the beaten path a little bit.

This is asking you to be. Technologists think about it. And then Amazon has an investment in Anthropic Microsoft has an investment in open AI. I think that one is a pretty good match. I think you'll see, Amazon and Anthropic look an awful lot like Microsoft and OpenAI.

I think they're going after the same markets and they will evolve in similar paths. What's Apple going to do? you talked about the watch. This is where we speculate and people go, Do you realize you said this all those years ago? It's what they're gonna, Siri's kind of antiquated now, isn't it?

To a degree, I will say that yes, do I feel like Apple's slightly behind the curve? I almost wonder if they're not doing it on purpose, just to see where some of the hype dies down, because you know that they're capable, and you know very well that your ecosystem is all tied Everything I own is them.

And so I, why wouldn't I then throw my energy into them in that space. Then do you worry about being too closed of a loop on that? Not so much when you know that they're going to stay current with the others. I am probably most excited about what Anthropic brings to the table right now. It seems to have a better hype cycle.

It also tends to have more awareness from user base right now, but I wouldn't ever discount Apple. Truly, number one, there's an affinity that doesn't go away and they have the ability and they have the capital to do leapfrog truly over some of these areas. So I would say I wouldn't discount them in the race by any means.

Yeah, it's interesting. The, Apple wasn't the first for the smartphone. They weren't the first in the PC. They weren't the first. They're the best. Yeah, they just, and I think you hit the nail on the head we talk about this with EHRs, the cost of change is so dramatic that it's not like somebody's going to pick up and go you know what, they're so far behind.

I'm going to get all new PCs. I'm going with an Android phone. I'm going, the cost of changing that, a new watch, a new, it's just, it's so great. So they'll have a significant amount of runway to get it right. And I think they will slowly move into this. I think what they will do more so than these others is really focus, as they always do, focus on us, on the user.

And it's a good model. They will focus on us and say, how can we make the consumer better, productive, more healthy, more And they will AI in all these little ways That when they first roll it out, we'll be like, that's weird. And then as we use it, we'll go, Oh my gosh, that's going to change my life.

The fact that it's already happening and you didn't know that it was already happening. It's almost like the frog in the boiling water, but do you remember the story of the gentleman from Australia, Cliff Young, 1983? He was a farmer. He basically was an Australian potato farmer and he won this inaugural race from Sydney to Melbourne.

And he was wearing like overalls and gumboots and. He even took his dentures out at one point when he, because they clacked when he ran. So all these ultra athletes joined this race. He wins. And he's oh I just was acting like I was chasing sheep from a storm like I did anyway. So he wins. No one expects him to.

Guy didn't know there was a prize at the end. And then he ended up splitting the prize money with all the other contestants, because he was just grateful everyone finished. am I comparing Apple to a potato farmer? You know what? The fact that he was always in it, he knew it better than anybody else, and he did what he knew how to do, and that's what created the win.

To me, that's them. You don't always expect them to come in first, but somehow they almost always do, and I will keep using them. We are a house divided, by the way, and we're half Android, half Apple. I would not want to switch everything out, to your point. All these things that are happening around us, they're built into how we live our lives every day, and they just become part of how we do things.

Yeah, I agree. It's not

a sea change that occurs, like you don't go buy the device that suddenly does this in your house. It's happening all the time. And we don't have an Alexa in our house, and yet, you know what? We'll have a conversation the next day. You bet that banner ads on someone's machine. So weird.

Sarah, this was the first of many. Hopefully, love doing the Newsday with you. I'm off to Philadelphia for one of our City Tour dinners. You have, gosh, what do we have? Eight of them on the calendar for the fall City Tour dinners. We're going to be going to a lot of different cities and pulling some people together and whatnot.

it's going to be fantastic. We'll give people more details about that as they get closer. But it's exciting time and I love the stuff that UI and Direx able to do in this industry. It's just a lot of fun.

It is a lot of fun.

All right, that's all for now.

Thanks for listening to Newstay. There's a lot happening in our industry and while Newstay covers interesting stuff, another way to stay informed is by subscribing to our daily insights email, which delivers Expertly curated health IT news straight to your inbox. Sign up at thisweekealth. com slash news.

Big thanks to our Newsday sponsors and partners,

Clearsense, Crowdstrike, Fortified Health, Healthlink Advisors, Order, Rackspace, Sure Test, and Tausight

you can learn more about these great partners at thisweekealth. com slash partners. Thanks for listening. That's all for now

Contributors

Thank You to Our Show Sponsors

Our Shows

Today In Health IT with Bill Russell

Related Content

Transform Healthcare - One Connection at a Time

Β© Copyright 2024 Health Lyrics All rights reserved