September 18, 2023: Bob Klein, CEO of Digital Scientists joins Bill for the news. They explore the complexities and advancements of AI in the field of healthcare. How might the landscape of medical services change as AI becomes integrated into daily operations? How will AI make waves in areas as significant as clinical decision support systems? Is AI an aid or a hindrance, a tool for efficiency, or a risk to job security as argued by nurses? Klein grapples with these questions and more, providing a thought-provoking look at where healthcare stands in the wild waters of AI development.
Join us for our webinar "Interoperability Outcomes: A Discussion of What’s Possible" on October 5th at 1 PM ET/10 AM PT, discussing challenges in healthcare interoperability. We'll tackle key issues like fragmented technology systems, data privacy, and cost-effectiveness. Engage with top-tier experts to understand the current landscape of healthcare IT, learn data-driven strategies for patient-centered care, and discover best practices for ensuring system security and stakeholder trust. Register Here.
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.
patients need more information. They they need to know what their scores are, they need to be able to engage with their own healthcare and their health record, and they can't do that in a vacuum.
And I still feel like they're in a bit of a vacuum.
Welcome to Newsday A this week Health Newsroom Show. My name is Bill Russell. I'm a former C I O 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 Newsday show partners and we have a lot of 'em this year, which I am really excited about. Cedar Sinai Accelerator. Clearsense, CrowdStrike,. Digital scientists, Optimum Healthcare IT, Pure Storage, SureTest, Tausight,, Lumeon and VMware. We appreciate them investing in our mission to develop the next generation of health leaders.
Now onto the show.
(Main) 📍 all right. It's Newsday and today we're joined by Bob Klein, CEO for Digital Scientists. Bob, welcome back to the show.
Great to be here, Bill. Good to see you.
So I just conducted essentially a webinar around healthcare's AI journey so far.
We had Brent Lamm Mike Pfeffer and Chris Longhurst. You were one of the attendees of the webinar. I wouldn't mind starting the conversation around what we heard and what you heard. I mean, because this is pretty fresh for me, pretty fresh for you. What jumps off the page at you as you listen to those three?
Leaders talk about their AI journey.
Well, one of them said we're experimenting, right? So that's what jumps out. People are testing it out. Is this thing real? Is it believable? Right? And they're looking for low hanging fruit, ways to leverage it to get their work done, right? And some things don't sound especially complicated considering the complexity of the services, but...
They're kicking the tires. Is this a capability that we can incorporate into actual care delivery? I think for most things, and this is where a lot of the documentation tools are, is just transcription is awesome, and creating notes. With NLP. And then the other piece, it's going to be a longer road to get to actual clinical decision support systems.
Right, there's much more in there as you start to build personalized care models. So all really good feedback. I thought that was great, that they're involved, their teams are involved, and I think the challenge for, I would say for health IT is to keep up, right, to remain relevant, to engage with the clinicians and providers and be there alongside them to guide them and also to validate, okay, how would we actually operationalize this?
So it's not a one off thing. It's got, of course, comply to HIPAA and security regulations and all of that, but it's it's leverage, it feels like people testing it out for leverage in their own daily work and goodness knows we need a lot of that.
It's interesting you had three leaders on and I, I picked these three because I knew they were bullish on it but they see the potential.
I mean, they see a lot of potential in the technology. That's out there, not only generative AI, but the other AI models that they, as they rightfully noted we've been working with AI for quite some time, actually, in healthcare. And we have different models that exist throughout the environment, but it's the generative AI, the chat GPT of the world.
That the large language models that really catapulted this into the public consciousness. And as soon as doctors started playing with it, they saw the potential and, and administrators are seeing the potential and others. And and I think what you found is in those organizations and in fairness, they're academic medical centers.
They're fairly well funded. They have access to grants and other types of resources to do the research. And quite frankly the IDNs and the others rely on them to lead the way in some ways in this space so that, we don't have the resources to do all the validation that's required.
But I love the fact that they're bullish. The only thing I heard that concerned me was architecture. And you're probably a good person to talk to about this. Are we going to end up with 50 LLMs throughout our environment? And are we going to have to worry about networking the LLMs into a, a multimodal model that.
At some point where we have, one l l m to rule them all because there's 50 of 'em throughout the environment. I didn't hear much talk about architecture and maybe we're not at that point yet.
I don't know if we're there, you know, like, what flavor to choose, at this point, because it'll definitely be vendors out there willing to sell it to you.
But I, I feel like, for the larger regional health organizations, what is their job going to be when it comes to building these actual personalized care models? Like, how much, how great is their role in defining what the model is? And it's honestly, we're missing pieces of architecture to do that today.
All right, so if I have remote patient monitoring, where does that data go? And there are just a couple new tools that are out there that let you aggregate that information and let clinicians define a new model that helps them with the monitoring of these patients, let's say, post acute.
And that are at home, to prevent readmission. Let's say, it's a cardiac surgery patient or something.
Bob, don't you think there should be a lot of skill sets that, I mean, I guess we could press the easy button and just rely on the big vendors to come in and essentially give us these models and we train them and there should be models that we train, models that are pre trained, models that have that kind of stuff.
Don't you think there should be at some point the need to build very specific models with Very specific information. Will there be a propensity to just buy the technology or do you think we'll see some of these players start to head down the build path? I
gonna need plumbing, to build their own, but the data and the insights are probably theirs.
And it comes from interacting with patients and getting really close to the patient because I don't know about your model, Bill, but my model might be different from yours, and what I want in a health organization that I trust to take care of me is one that doesn't just, you know treat me like a number, I've got comorbidities and my age and I get, there's SDOH data and other things that are relevant for me and my care.
And also you know, What's stuck in the EHRs could be a year old or worse. You know, A lot of my family history didn't even get migrated into the damn thing. And it's what's current, what's relevant, what's today, what's real time, all the time. That's what's changing, is like a tsunami of data.
Like, there'll be no shortage of data. What they'll miss are the models that are tested and validated. And I think Most of it's around running an ACO, where there's now, we've got aligned incentives to essentially keep people out of the hospital. And if you don't do a good job of evaluating, folks with a risk assessment and then their wellness visits you're going to get penalized.
You will lose money. Right. And I think the value based care is going to drive... the need for building your own models and understanding that you have to have your own models. And honestly, that could be differentiating. At least that's what I've seen in more of the post acute space.
We've got a couple of articles. We'll hit them real quick here. We've got Kaiser Family Foundation, artificial intelligence may influence whether you get pain medication. And I'm not sure if this is an actual AI use case or not, but that's what it's being attributed to. These stories usually start with a narrative of a person going, seeing a doctor.
And in this case, it was somebody in pain and they said, Hey, I'm going to need some pain meds because my hip hurts. And the person said, well, your NARC score is so high. Yeah, I can't give you any narcotics, and the NARC score, I mean, could be AI or it could likely just be, just an algorithm, right?
That's sort of what's happening in this world. It's like... It's just a score, it's just a number. Every algorithm in the world is now called AI, but, I mean, you and I both know it. We've programmed these things where it's like, no, it's just a math equation. That's all that's happening here.
It's just numbers, yeah. But it's interesting. I mean, these things, even though I think they mislabeled this, we could stay in the AI realm. How much authority are we going to give AI? And we saw this with the payers, that they said, oh, well, we ran it through the AI algorithms, and all of a sudden it spit back denial of claim to like, Just this massive amount of people, and it's like, okay, is that the final?
Like, we ran it through an algorithm, we ran it through AI, and it says no, and we're,
like, we're hosed? That's Cigna, right? And they've got a class action suit in California, and the legislation reads that they're supposed to be thoughtful, make a thoughtful review, 1. 2 seconds. For each one, that example, and I don't know that's gonna be litigated.
Yeah, well, is AI a thoughtful review?
I'm not so sure, of my claim, so I have to write up the claim and say that this shouldn't have been... mean, I submit the claim and then they deny it automatically. In a way, I think their providers need these same tools, right? So the payers have all the money and all the data and all the AI tools, right?
So now it's the providers need to build their own because you don't over code or under code. It needs to be just right for reimbursement, but there's, that's some of the challenge, but to this point, the NARCS score, I wonder if. For me, that's a question more about privacy or patient access.
Does the patient deserve that score? And if that score is being used to deny, access to to care, then I got a real problem with it. Right. And the pharmacist, I mean, I understand the need to create an internal score and measure, but I can't see my own score. And I'm like, well, what is it?
Right. I think it's a slippery slope.
Well, by the way, we're not saying that's a bad score because we essentially are in the midst of an opioid epidemic. Yes. This is one of the ways that you sort of track how much in that space of drugs that people are gaining access to and it creates problems.
So clearly that's good. We're talking about transparency. Transparency is interesting. Like it's not only, Hey, I should know my NARC score, but if you're going to deny my claim, I AI model was trained on. I want to know a whole host of things about like, Hey, what's going through The large language model or whatever you're using. How is it being processed? And that's, what's going to be litigated. I think it's like, how are you processing this? If you're saying a human didn't look at it, fine. I need to see your code. I need transparency.
Before they had all these tools, of course, there were business rules that would be applied and they would look at it and say, this might be different, in terms of a similar claim where it's been denied before. It's not as typically part of a standard of care. And so no, it's not typically approved.
But now, to your point, you don't get to see any of that. It's just denied. And they're, California, I don't know what the outcome's going to be, but I would worry about it, and like I said, since I'm in the business of helping people build some of these things I've got providers in the same place, where they they see sometimes that they're delivering care that they're not getting reimbursed for, or they're not even coding for, right?
So, there's a gap between care delivery and what was actually coded. I mean, something was coded, but not necessarily coded properly for reimbursement. And everyone's, of course, watching every penny.
Well, it was alluded to today in the webinar that, we could end up with an AI arms race. You have on the payer side, they write their AI algorithms.
And on the provider side, they write their AI algorithms. And essentially, they go back and forth. And I'm sort of okay with that. As long as, it kicks out at some point, the denial is, hey, this wasn't coded correctly. And the AI on the provider side, smart enough to say, okay, great. and it kicks off an automation that recodes it correctly and then resubmits it, that's fine.
I just don't want two machines talking to each other. All of a sudden, six months later, a patient gets a notice that, Hey, you didn't pay your bill. It's like, I had a
problem. It's like this
Rock'em Sock'em robots, right? Just going back and forth. See what happens. But yeah, I think the transparency is going to be tough because no, no one knows, why or how something is working. And that's some of the challenge of it. It's like there's, you're not provided the reference information and let's be clear, the patients aren't really provided much to begin with at all.
Right? So they're starting from scratch. You're lucky to be able to make an appointment. And, I don't know, and again, back to the ACO model or VBC, like patients need more information. They they need to know what their scores are, they need to be able to engage with their own healthcare and their health record, and they can't do that in a vacuum.
And I still feel like they're in a bit of a vacuum.
So that's the slippery slope, and one of the other stories, and it might sound under related, but I don't think it is, Nurses join striking writers, actors to voice AI concerns. And culturally, I mean, look, it started Thanksgiving last year, and people would stop me.
And it's not like they were, technologists who were stopping me. This was like my parents stopping me. Did you hear about this chat GPT thing? And they were getting on like, cause it was free access and they were using it. And they're like, Oh my gosh it knows this, it could do this and that kind of stuff.
And that was a neat little phase that we were going through. But are we ready for this? Are people ready for this? Cause I read the nurses, they're concerned about the use of AI. And if I read this article correctly, essentially what they're worried about is nurses being displaced as a result of AI, and I'm not hearing anyone say those words.
I'm hearing everyone say, essentially. Nurses are telling us they're overworked, too much time. They're tired of being clerks and just typing things into the system. We're going to bring this technology alongside of them to support them. But are we not communicating correctly? Or well enough, so that you have this case.
By the way, I understand the writer's strike 100%. But the nurses, I'm not sure I understand it as much. Because you're always going to need a person there.
I, I agree. What comes across to me is fear. Right? Fear of the unknown. And AI is a bit of the boogeyman. This worry about smart robots.
The country's in desperate need of nurses. There's like, there's, there's no one saying, we don't need nurses. And honestly, nurses need leverage. They need help. We're burning them out. No one wants to do the job, right? So, how do we make their job better? And this is a lot of these tools, so, they don't have You know, there's missing these whole systems of engagement inside the health organization. Like, they're extremely manual, and lots of running around and standing at the PIXIS machines, and double entry into things, and nothing's integrated. It's a terrible job, they're there to work with patients, and we don't provide them the insights or tools that they need to be more efficient.
The same goes for the other providers, right? So, That's, I think, to your point, they hear about technology and the threat, and they see, yes, it is a bit of a threat to the writers and the creative folks that are in the movie business, where, yes, a lot of this can be automated, but it doesn't replace creativity.
It's, it has to be leveraged for creativity in their world. But in the healthcare world, a lot of it's to give them insights and tools. and make their job better. But no one's making that sale bill. No one's telling them. So they see comrades in arms, but I don't know, to your point they seem completely unrelated.
And, but I, who's going to make that case to the nurses that we need them to engage. So tools are created that help them at work, right? So if you're a CIO and a health organization, you should be asking yourself, let me engage with the nurses and cut through some of this kind of fear and say we could, how can we help make your job better?
Like, what are the things that we're These tools can give you better leverage and let you spend more of your day doing what you want to be doing. Yep.
We'll get back to our show in just a minute. Everyday families worldwide receive devastating news. A child they love has cancer. From that moment, a whirlwind of fear, uncertainty, and challenges begins, marking the start of a journey that no family should have to endure.
The stark reality is that childhood cancer impacts over 400, 000 families globally each year, but there is a way for you and I to stand up and foster change, transform lives, and give children a fighting chance against this devastating disease. This Childhood Cancer Awareness Month, which is September, we invite you to stand up with This Week Health in supporting Alex's Lemonade Stand, a beacon of hope against childhood cancer.
Celebrating our fifth year, we set out with an ambitious goal to raise 50, 000, and you did not disappoint. Our community has exceeded that goal. Here in September, we are up over the 50, 000 mark we believe we can do even more because we know that every donation, big or small, brings a spark of hope to those families and a possibility for a brighter and healthier future for a child with cancer.
We ask you to join the movement. Visit ThisWeekHealth. com today and click on Alex's Lemonade Stand logo, it's in the top right hand column. Become a part of our community. And a community that believes in the future, free from childhood cancer. Now, back to our show. 📍
Well, let's get outside of our swim lane a little bit and talk about the AI with regard to the writer's strike. My kids are all out of the house now, but I grew up in a time where the Disney Channel was big.
And I remember Cody's something. They lived in a hotel. Then you'd watch. But you'd watch these shows and, my kids loved them, but I was sitting there going, Oh my gosh, it's like, I love Lucy and I dream of Jeannie and all those things from when I was a kid. I'm like, it's the same storylines, like they take a new set of things, they put them in a different environment.
And then it's the same storylines for the next three to five years till the kids grow up and leave Disney. And then they start over with a whole new thing. And I'm sitting there going, you know what, AI could be given. Hey, here are the archetypes of characters. Here's the plot lines.
Here's the storylines. And essentially, here's the genre, go. And I'm not saying all, Aaron Sorkin is not going to be out of a job. I mean, he created a new way of a dynamic way of characters interacting on the screen and a method for dialogue and that kind of stuff. That kind of creativity is not going to be...
replaced. But the Disney Channel storylines, I don't know. I, it feels to me like I, you could program that stuff in and spit out a 20, 22 minute episode pretty easily.
Well, you could certainly test it out, right? So it brings a whole new meaning to formulaic, right? I would say the For a lot of TV, it's what we've gone to as long form content, right?
So, really complex stories that go on and on. They're quite difficult to write. And, I think about there's a foundation on Apple TV, for example, which is basically an adaptation from Isaac Asimov's writings, but they're innovative, they're thoughtful, they really feel like movies.
Every night of the week. And I don't know. I think, in a way, that's what technology's going to do, is automate things that might be somewhat transactional and formulaic. If there's a formula to them, then I can recreate it. But it doesn't, it's commoditizing, in a way, and I'm not sure that's good or bad, that's just a fact of life.
If it's not distinct, differentiated, truly creative, unique I don't know. Then a machine can do it.
Yeah. By the way, I don't think the writer's strike is about AI. I think the writer's strike is about money and royalties and the long tail of that kind of stuff. When I was reading about how much they get paid on something That has this really long life cycle, the actors will get royalty checks for the next 20 years and the writers get like nickels and you're like, I don't, it doesn't value the writing of the story as much as it does the acting of the story.
And I understand where that's coming from, but I'm not sure that the strike doesn't feel to me as much about AI as it does really about the money.
I mean, let them be a stakeholder in the enterprise. Right. And so if the movie does well, but everybody wants upside risk and nobody wants downside risk, right?
So, we're not all Harrison Ford either. No,
a lot of the richest actors out there took the downside risk. Keanu Reeves, when they did the Matrix, essentially said, just give me a percentage. And there's a three part series on Arnold Schwarzenegger On Netflix, I think it was.
And it's worth watching cause the guy is fascinating. I don't know, whatever you think of his politics and whatever, but he has had three lives, Mr. Universe, and then he was the actor, the leading actor, and then he was the governor.
I saw those, Bill. what struck me is just how driven he has been his whole life, right? And how dedicated he would be and just. I don't know, he has a a joy of living in some way, like this comes across, where it seems like he'd be a great guy to share a beer with, and just, somebody that probably has, knows
The movie he made the most money on, there's a little snippet in there, movie made, so you think, oh, Terminator, oh, you think, oh no, he's, one of these other movies. Oh, no, it was the movie he did with Danny DeVito and it was like they were brothers or something and They did that same thing.
They took the risk. They essentially said hey, you don't have to pay us Just give us a percentage of the movie And it's the movie he made the most money on and most of us It might have been called brothers or I forget what it was twins, right? And that I mean You think of all the movies he did that's the movie he made the most money on because he took upside and downside risk and It's it's just really interesting.
Well, Bob, as always, we are way out of our swim lane, but it's a lot of fun to to go over there. And the reality is AI is going to impact every aspect of our lives, from healthcare to finance to I mean, I can't think of an area it's not going to touch. In some ways, it's going to annoy us when we have to deal with technology like, like we used to have to deal with press this button, say this.
But in other ways, we're going to be like, Oh my gosh, I just got that answer in 30 seconds, and I used to have to wait on the phone for, 25 minutes to talk to somebody and try to get through that. So in some ways, it's going to be fantastic. In other ways, I think we're going to curse it.
I mean, Bill, who do you think should be driving it inside health organization?
Who's responsible for driving its adoption? It's leverage as a tool for the enterprise for return, right? So, for efficiency return gains or actual improvements in some way care outputs, all these things, who do you see or what do you see? Yeah. Who's driving the
The, who's driving the bus?
That's a good question, who's driving the bus. Even today. We had three people on there. Two were CIOs. One was Chief Medical Officer. Right. So you have Chief Medical and Chief Digital Officer. So there's a reason that person is driving it. A lot of them talked about governance and they say, multi discipline governance.
Here's my direct answer to your question. The CEO and the executive team need to be driving it. I think the CEO needs to be the head of digital transformation for the organization because every organization's being... digitally transformed, and that is the role of the CEO. And when people ask me, it's like, give me an example.
I say, Jamie Dimon at Chase. When you read his annual report, he is the Chief Transformation Officer, Chief Digital Officer. He is the person who's leading. those initiatives. Now, clearly he has technologists, he has programmers, he has all those things, but when you talk about the future and the vision and that kind of stuff he will have those conversations, he will internalize them and he'll say, I know banking.
I know technology and I know banking, and this is where it's going to play out. I think we need more health system CEOs that can internalize all the things we're talking about with regard to generative AI and AI models as well, and be able to turn around and say, Yes we're going to get this team together, we're going to start talking about how we're going to do it in on the revenue cycle side, and we're going to talk about how it gets implemented on the clinical side, we're going to talk about how it gets implemented on the administrative side, and we're going to determine as a leadership team, which areas we fund immediately that are going to have either a short term impact on the metrics that we're looking to drive, whatever those happen to be for that leadership team, or give us the foundation to make sure that we're ready for what's coming.
And all three of the guys today essentially said in five years. transformative. What we see today is not going to be what we see five years from now. I mean,
if all, at what level the CEOs have taken that charge, right? So, and what size of organization and a lot of them are swim, swimming pretty hard just to stay steady, right?
And to take on a new initiative. At certain size and scope, I'm sure they have, and it looks like the organizations you talk to are going down that path, and that is, that does feel like what's needed because it's going to touch all parts of the organization. And I guess, that's another challenge to help IT, to the CIOs of the organization to support all those other areas, right?
So it's not only IT's charge. Right, to make this work because that should be part of that equation, too, and driving what the solutions are. And it's complex, so it's probably a big staffing challenge internally as well. It's one thing to be entering these things in and getting results out, but it's a whole thing entirely differently to manage it programmatically and, or to build some of your own solutions.
Right, so that leverage these tools, and which model? Right? Are you going to use Google, AWS, Azure? What's your flavor? They're not all the same. So it's an exciting place to be. So, but I'm okay, either in that lane or any other lane you want to talk about.
Well, Bob, thanks. Thanks for your time today.
It was great discussing the news with you. I look forward to the next time we get to do it.
All right. Thank you, Bill. You take care.
And that is the news. If I were a CIO today, I think what I would do is I'd have every team member listening to a show just like this one, and trying to have conversations with them after the show about what they've learned.
and what we can apply to our health system. If you wanna support this week Health, one of the ways you can do that is you can recommend our channels to a peer or to one of your staff members. We have two channels this week, health Newsroom, and this week Health Conference. You can check them out anywhere you listen to podcasts, which is a lot of places apple, Google, , overcast, Spotify, you name it, you could find it there. You could also find us on. And of course you could go to our website this week, health.com, and we want to thank our new state partners again, a lot of 'em, and we appreciate their participation in this show.
Cedar Sinai Accelerator Clearsense, CrowdStrike, digital Scientists, optimum, Pure Storage, Suretest, tausight, Lumeon, and VMware who have invested in our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.