May 22, 2023: Geoff Blanding, EVP at Optimum Healthcare IT joins Bill for the news. What are the potential benefits and challenges of implementing a privacy-focused version of ChatGPT in healthcare settings? How can Microsoft's partnerships with Epic Systems and EHR vendors impact the integration of AI in healthcare? How will the closure of rural hospitals and the expected wave of hospital bankruptcies affect the healthcare landscape? How can the data and insights gathered from telehealth experiments be leveraged to drive efficiencies in healthcare reimbursements? What factors determine the nature of the impact of AI on healthcare, including both positive and negative effects? How can clinicians be convinced to accept general AI in healthcare, considering the potential risks and integration challenges? What steps can a CIO of a health system take to address the use of AI tools like ChatGPT and ensure they are used safely and effectively?
"The Future of Care Spaces" is an upcoming webinar that explores the latest healthcare technologies and solutions transforming care spaces in America. Hospitals, clinics, and at-home treatments are all affected by the advancements in healthcare technology, which can improve workflows, treatments, and patient outcomes. What are the latest remote monitoring tools, advanced telehealth solutions, and other innovative technologies transforming care spaces across America? Join us June 8th, 2023 1:00 PM ET and Register Here
Alexβs Lemonade Stand: Foundation for Childhood Cancer Donate
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.
estion of can you get through: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. Clear sense crowd strike. 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.
ββ (Main) π All right. It's Newsday. And today we're joined by Geoff Blanding, executive Vice President at Optimum Healthcare It. Geoff, welcome back to the show.
Thanks, bill. Good to see you. Thanks for having me.
Yeah, I'm looking forward to the conversation.
There's a lot going on from big time things like Kaiser and Geisinger coming together and also announcing as part of that new entity that they're going to. Acquire up to five additional entities and spawn a new 30 billion health entity that, that's no big deal. I mean, that's, a new 30 billion health system.
I remember our health system was about seven and a half billion at one point. That was considered, moderately large. What's a large health system now? I
mean, that's like a, that's like a jumbo health system, right? Like, like maybe 10 billions. The cutoff for large,
I think 30 billions bigger than Providence.
Yeah.
Same neighborhood. At least. It'll also be interesting to see, like, does that get through regulatory
approval? Yeah I think it does. There's almost no overlap. I think so, they sort of said, Hey, we're focusing on population health and. Organizations that look like that and there's a bunch of 'em that they're not gonna get, like sharp healthcare matches the criteria it would never get through because Kaiser and Sharp make up too large of a percentage of the San Diego market.
Yeah. And so, so that would never happen. And there's probably a couple others that might fit the bill, but they'll look at it from a regulatory standpoint and go, yeah that's not gonna work. Whereas a Geisinger, I'm not aware of any overlap. I don't think there's any overlap at all.
I
think so. Is there Kaiser's in like Maryland, Virginia, DC area, but I don't think they get into even Philadelphia, Pennsylvania at all. As far
as I know. Well, and Geisinger barely gets into Philly. Their health plan gets into Philly, but their. Their delivery doesn't. Let's get to the tech side. So four healthcare moves from Microsoft.
This was posted two hours ago, so you had no chance to read it. I love this. I'm like, quiz. Quiz at the end of the day. You ready? So four healthcare moves from Microsoft. Number one, Microsoft is reportedly developing a privacy focused version of chat G P T for healthcare providers. The product which is reportedly aimed at preventing data leaks.
From sensitive industries could be announced by the end of the quarter. That's pretty interesting and actually I talked to a CIO today that is actually in conversations with Microsoft about this cuz they. They want to use it. They really, well, first of all, they acknowledge it's ChatGPT T is being used in their health system.
Yeah. They're educating their users, they're putting some guardrails around it and that kind of stuff. But what they'd really like to do is they'd like to take ChatGPT t to medical school, and not only medical school, but medical school on their healthcare data. And then essentially utilize the tool.
To I, I don't know, do all sorts of things. I'm hearing some interesting use cases, and one of the interesting parts of the conversation with the CIO is He said, bill, if I gave them 10 use cases, it would be a disservice to our institution because just letting them use it, they're coming back to me and it's like, I get like two or three every week of new use cases.
Like, Hey, have you thought of using it this way? And so people are, they're getting a hold of this thing and looking at it and going, this is pretty cool. So Microsoft's saying, Hey, we understand the needs of your industry privacy. We're gonna bake something that's, A little bit more of a a closed system so that it's not there's no risk associated with it.
What do you think of that one?
I think it's interesting. I keep, sorry. I read all these articles on ChatGPT and then I keep pulling up the the hype cycle diagram and being like, okay, here's, okay, here's where we then, so like, what are we like six months away from the trap of disillusion, a year away from there?
And maybe, it's obviously moving very fast, so maybe that's a shower trough then typically you see in the net cycle, or maybe we just get through it quickly. But I think I think it would be really interesting to see like what are the boots on the ground or the low hanging fruit that we really find for it.
And I, I feel like it's, some of the things I haven't heard a ton about are around, like after visit summaries, patient education, stuff like that where you can. Have it kind of within guardrail. So guardrail, so it's still gonna be, hey, it needs to be about this diagnosis, but then if we can have.
Put this avs in language that a five-year-old's gonna understand, or, yeah, do this for a, a millennial who's a visual learner and have kind of like a, just infinite possibilities. But then kind of still being within that safety net of, we know basically what it's gonna say.
It's gonna say it different ways, but it's not, this black box is spitting out potentially anything.
Well, there, there was a, an article written last week, I think it was in jama, and it was a bunch of the health systems doing a study and they threw a whole bunch of messages at inbox, messages at it and said, write the responses.
And so they took the physician written ones and they took the ChatGPT one's, and they essentially assigned a score for it around empathy and quality of the response. And as you would imagine, the empathy score was higher. I know it's hard to believe that the computer's empathy score was higher because you could tell ChatGPT respond with empathy and Yeah.
And knows what words are empathetic kind of thing. So, and a doctor's moving at a very fast pace. And so, a lot of times there'll be shorter messages and shorter messages sometimes are seen as Kurt, or at least not. I don't know, and not caring or whatever. And it doesn't take ChatGPT anything to put a couple caring words in there.
So empathy scores were higher and the, oddly enough that the quality of the responses was higher. Yeah. But at the end of the day, they're going into a draft folder, right. We're not at a point where we're dispensing. Medical advice through chat G P T. I hope no one's hearing that. And so it's going into a folder.
It will be human mediated which one goes out, which one doesn't go out, what needs to be modified. And that kind of thing. But it's but I do
think that's where some of the use cases are probably gonna be, it's like the things we have now that are basically rules engine driven.
So like the, what's going into the avs or what patient education's getting assigned, like, I think there's gonna be a middle ground there where it's still within that rules engine of making sure that this is the content, but then, Having that then go through chat G p T and then maybe that's where it hits that healthcare specific checkbox of, we need to be a hundred percent is accurate and safe.
And so it's not just not just kind like rephrasing a note where it could get it wrong, but saying like, we already basically know what we're trying to say. Say it in this particular manner with empathy or whatever the case may be. Yeah.
Well lemme give you the other three. So Microsoft's partnering with the HR vendor Epic Systems to develop an integrated.
Generative AI into the EHR software. They're doing the inbox thing. They're also doing a front end a natural language front end to slicer dicer is what they're doing. Number three was Microsoft said through its teams EHR connector, that organizations using Epic or Cerner EHRs can launch telehealth appointments directly within their systems.
And then number four on April 12th, Microsoft rolled out its new cloud capabilities for payers that focuses on unlocking unstructured data for better member outreach and care management. Microsoft is moving in the healthcare direction big time. I mean, they've always been a big time player.
When people say to me, oh, is Microsoft. I'm like no. Microsoft's always been a big time player. Like one of my largest contracts I signed every year was the Microsoft contract as a cio. But what do you take from this? What do you take when somebody's asking you about Microsoft's intentions and where they're going and how they're going to serve?
Healthcare moving forward. What's your response to that?
I think They see it's a big market. It's always been a big market. I think we've seen them even before this, just like in the way they were structuring contracts. You probably saw this like. Pushing clients to Azure. Like we've seen a lot of clients over the past couple years start looking at those contracts and realize like, oh, I've got all this.
I'm already paying for all this. And so they're starting to get market or they really already had the market share, but now they're kind of stealing it from others. So I think you'll and continue to see that growth, and obviously they're right at the front with ChatGPT t here.
So that's gonna keep 'em pretty hot for a while, I think, and be attractive both to the health systems, the payers, and then also for the technologies like Epic and Cerner who wanna be associated with that.
Yeah, let's go to next article. This is from April 14th.
tatistics that matter most in:It's interesting cuz I'm hearing that the. Financial crisis is abetting a little bit. The cost of traveling nurses has gone down. They've curved those costs. They're not making, I don't wanna paint the picture that these health systems are making a ton of money, but they're returning back to some semblance of normal.
And breakeven, at least is one of the things I'm hearing. But I have heard this, rural health systems are struggling. They didn't have a big nest egg to really rely on and at best, their nest egg is gone because it's been a tough year and a half, two years. At worst they are, they're closing down services and those kinds of things.
All right. So one and two. Any thoughts on those two?
it's still gonna be a painful:estion of can you get through:But if you can't, if you don't have the nest egg like you said, or anywhere you can cut to get there, then you're kind of in trouble for the next eight months.
Yeah. Next one. So soon more Americans will get their healthcare from government than from private firms, and they're essentially saying Medicare and Medicaid.
sen every year since the year:Or do you think it's probably gonna remain stagnant or even go up?
I can't
imagine anything that would make it go down short of additional large scale reform at this point. Best case scenario, maybe someday it matches the pace of inflation instead of running it. But that's still not a great, that's still not a great place to be.
becoming more widespread. For:So we're seeing more and more of these tools are you seeing that in, in the work that you guys are doing? Are you seeing more remote patient monitoring?
Yeah, definitely. I think a lot of systems are doing more with it and doing some cool stuff. I mean, it ties back to the nursing shortage and if we can have fewer nurses looking at more patients, More of the time.
And that's obviously a cost savings and a labor savings. And I think I mean it, it seems like most of the folks I talk to are doing something and I'd say reasonably mature at this point too. Like it's, we're not in a pilot phase anymore. It's where we've got stood up units and folks in command centers monitoring.
And the devices have gotten better, they're more sophisticated, they're easier to use. Let's see nurse shortage, 275,000 additional nurses will be needed within the next eight years. Drug overdoses are down slightly. We've got an aging population, which we know just sheer demographics and probably heard of a few mental health care statistics and it's not a rosy picture.
One in five American suffers from mental illness. Almost six in 10 people. With mental illness, get no treatment or medication. So those are the top 10 statistics that that this article anyway says are important to keep an eye on. Let's go to the next one. Let's see what,
it's
kinda interesting too, cause if, like, if you read all the hyper on ChatGPT T or AI in general, right?
It's all talking about how AI's gonna, AI's gonna fix healthcare in the us but yeah, it can certainly be a tool in most of those areas that you just went through. But it's not fixing the, some of the fundamental stuff that's going on there.
We're going through the hype cycle longer than I anticipated going through it.
I think we're gonna see different applications of the hype cycle. I think once we really get into healthcare, we're gonna notice. It's not as accurate as we want it to be in certain ways. We've gotta train these more narrow models but we're just not there yet. And so we might still be in the hype cycle.
Another three months from now because people are finding, the edge use cases where it is really helping them to be more efficient as an administrator or caregiver. And so I. But man, it, the hype, I almost want to like make a new hype cycle chart that starts on one page and you put like three pages on top of it.
It's just the chat G p t hype cycle. Let's see, public health emergency ends on May 11th. Here's what will change, and this is. Too many things to read. Let me see. Uninsured individuals may long no longer have access to free covid 19 vaccines. Insurers will no longer be required to provide access to free covid tests.
will remain the same through:Providers could face higher penalties for accidental HIPAA Violations that occurred due to telehealth care providers will be more restricted on treating patients in other states via telehealth, and reimbursement rates will change for telehealth visits, some telehealth visits, depending on the insurance provider.
So that's, that's right around the corner. May 11th, we knew this was coming. Actually, we've been talking about this coming for over, over a year, so it's not like we, we didn't know this should not be come as any huge surprise to people. I think the thing I'm most excited about is the telehealth experiment that we've just done over the last three years or two and a half years.
We now have a ton of data on what works and what doesn't, and I think there's an opportunity here. To rethink the reimbursements for Medicare, Medicaid even commercial payers and saying, Hey, I think we can drive some efficiencies by really implementing it in certain areas because we can no longer say, well, we're afraid of fraud.
We're afraid of this. We're, we did a two and a half year experiment. We have tons of data. And I just feel like now we should have a good foundation by which to make decisions and really take telehealth to the next level in a way that's very pragmatic. Any thoughts on that?
Yeah. It'll be interesting to see what the commercial payers do.
On this. They'll have all that evidence, obviously, like a piece of it's what's the data say? A piece of it's also what's the most profitable
and the problem is in a healthcare system, I want complete, I
want full reimbursement. Like as if they came in the front door. Yeah. And that's just from a commercial payer standpoint, it's not pragmatic.
You're just sitting there going, wait time out. You do a 10 minute call with them, or you have to utilize this facility. In theory, not only in theory, in practice, that 10 minute call should be reimbursed at a lower rate than them coming to your office.
Yeah.
But you still gotta run the hospital. The hospital still cost the same whether the patient's there or sitting on home on the phone.
Yeah, it's true. It's true. So have you done telehealth visits and what are your thoughts on
Yeah, so We were pregnant during Covid, so we did several of our checkups for that. That way, I mean, I've
Are you gonna announce that you delivered your baby via telehealth?
Not that I'm not that brave.
Not that brave. That seems like a good a good call for an in person visit. So we, but we did a bunch of our, probably two or three of our checkups at least, that way of just kind, Hey, here's your test results, or, how are you feeling? How are you feeling today? That kind of stuff.
And I mean, Worked well. Like we, I'm excited to not go sit in traffic for 20 minutes and then wait in a waiting room for 25 more minutes at least, and to have the same conversation. I think we actually, we did one recently too for our now our successfully born daughter for an ear infection,
which was great cuz otherwise you're going into the doctor to basically hear what you already know, like Yeah. She's pointed at her ear, she's got all the symptoms. It's pretty, pretty clear. So being able to jump on with Teledoc or whatever is a pretty good way to
go.
Yeah. In fact, our I've said this on the show before.
Our commercial payer for my company encourages people to go to the payer. Website and see one of the payer physicians because obviously if I don't go to the health system it does reduce the overall cost to them cuz they don't have to pay for anybody and they recognize the savings.
By the way, this is an article that's the reason for the Hype Fast company, how AI can be used to cut $1 trillion. Out of healthcare fast company telling people, Hey, we're gonna be able to cut 1 trillion. So the expectation here is we're gonna cut 1 trillion out of a 4 trillion, that's a quarter of the cost.
So the next time you go to the doctor, it should cost a quarter or less.
Yeah.
They're still gonna be using fax machines, but ChatGPT everything else.
Yes. But the machine learning can read the faxes as they come in, interpret that's, and do all sorts of things. And by way that's actually true.
It some of the stuff they're demonstrating is, The computer vision linked with chat G p t that essentially can read a fax.
Yeah.
That seems like a kind of a mindboggling workaround of, you know, paper being printed so we can scan it back into the, back, into the computer.
Yeah, absolutely.
We'll get back to our show in just a minute. I am excited about our webinars this year. They have been going very well. What I've done is I've gone out and talked to people in the community and said, what works in webinars?
And they came back and said, look, this is what we want. We want a webinar that is not product centric. It's really focused in on the problems of health care. And we want people on there that are actually solving those problems. And so we have done that. And the response has been fantastic this year. We have another webinar coming up.
It is the future of care spaces. Where care is being delivered is changing rapidly. Even the care spaces within the hospital themselves are changing. Technology is being added in different types of technology. A. I obviously computer vision and whatnot is changing that modality as well as what's going on in the home and whatnot.
So we're gonna have that webinar June 8th at one p. m. Easter time. We usually have it on the first Thursday. Happens to be a little too close to my anniversary. So we're going to do June 8th at 1 p. m. Eastern time future of care spaces. We would love to have you be a part of it. If you are interested in being there, go ahead and hit our website.
Top right hand corner. We have a card. You can click on that card and go ahead and fill out the form and get registered today. We would love to have you join us we look forward to seeing you there. Now back to our show.
π β
Let's see. Last article. Let's do this new England Journal of Medicine Catalyst, getting generative ai, right rapid government action combined with careful use case selection, can minimize healthcare risks posed by generative artificial intelligence and put it on a successful path. Let's see, rapidly of AI will impact healthcare, the na nature of its impact.
Including its positive and negative effects will be determined by a combination of market forces clinician acceptance and government action. Pragmatically focusing development on low risk, high impact capabilities is important. Multiple government agencies are in a position to incentivize this focus and mitigate likely the likelihood of harm.
It's interesting, and they go on, clinicians will reject general AI if it harms a patient, produces flawed content or does not fit easily into their workflow. Which is the same for every technology we've implemented for the last decade, right? I mean, it's our decade for forever. It's gotta be accurate.
It cannot do harm period. And and it's gotta fit into their workflow. So, if you were coaching, if you're sitting in front of a CIO right now and they're saying, Hey, how do we get in front of this thing? Like every day somebody else in the health system is subscribing to me and saying, Hey, I'm using ChatGPT to do this or do this.
How do you get in front of this? What are some of the things you would you tell a CIO of a health system right now?
Oh yeah. I think that you mentioned at the beginning with that, I think it was the Microsoft announcement about setting up that privacy restricted version of ChatGPT like that's gotta be the first thing, right?
Is like we know people are using it. Today, so let's make sure they're doing it safely. And it's the same thing with text messages and things like that. Like, we know they're sending text messages about patient care and HIPAA protected stuff, HIPAA restricted stuff. So let's make sure that we're giving them a pathway to do it the right way.
And then at least when, if I'm saying to ChatGPT, hey, write my soap note about a patient with these these diagnoses or whatever, then at least they're doing it in a way that's not gonna backfire and result in a breach or or anything like that. So I think that's gotta be probably ground zero.
And then, I think from there you wanna try to probably set up a task force of some kind, set up a structure and process for folks to come through and say, Hey, here's a use case I wanna vet out and make sure they're doing it again, within view of you as a CIO and understanding what are they trying to do, that they're doing it the right way, that they're doing it with support that, if there's gonna be.
Integration, things like that, that they're asking for. They can be scoped and. Budgeted appropriately. Otherwise, folks are gonna be doing all this, on the server that lives under their desk, that classic example. So I think those are probably two top things I'm thinking
about.
Yeah, I, again, talking to the CIO today and actually CIO yesterday and they were saying, they, they actually caught because they're using a secure browser. Within their environment, they actually caught some different use cases where they were putting p h i and whatnot into it. And he said, having that kind of tool in place to capture because you really shouldn't have a way from your clinical workstation to exfiltrate data.
Yeah. It's like, I mean, that should be the most controlled environment you have anywhere. Yeah. And so, email, you should know if PHI is going across there. You should know if they're putting into the browser, you should be able to see it. And if they're putting the ChatGPT, you should be able to see it.
So they said, Hey, we found it. We're able to educate people and that kind stuff. So the guardrails are so important on this? Yeah. Of just don't let people hurt themselves. It's like, yeah. Right. So I mean, that's,
If
you don't do that, people are gonna be doing it anyways. Like you, you can't just turn a blind eye cuz you're.
You're in trouble if you don't look at it and you're in trouble if you look at it. So you gotta look at it and put those guard rails in place.
please.
Yeah. It's interesting because I think we all recognize the world's changing as we watch right now. I mean, you just sit in front of chat g p t and say, Hey, here's the dialogue from a conversation with a patient.
Yeah. I'm not telling you who the patient is or anything like that. Write the soap note. I'm pretty sure it could write this. I haven't done this yet, but I'm pretty sure it could write the soap note.
Yeah,
I'll be right.
I'll bet you're right.
Well, let's see. I've got a tab open. I have a tab. Well, and that's what I'm finding more and more when I talk to people.
One of the questions I asked 'em is it's like, Hey, do you have ChatGPT open on your browser? And almost everybody says yes. It really has become pretty pervasive. So I think we're looking at a new way of working. This thing has been adopted by a hundred, over a hundred million users in the shortest timeframe bot in history, in the history of the internet.
And it's gonna be interesting to see I was reading an article yesterday, or not yesterday, last week, and it, and I know the author and I'm reading the article and I'm reading it and I'm going, that person did not write that article.
I read something on LinkedIn just today that I was like, I'm gonna get to the end and it's gonna be like, the punchline is, this is written by Chad g p t and it never said it, but I'm like 95% sure.
Well, that becomes one of the things, I mean, you have so many of these new tools and you have like mid journey version five and whatever. It's like, how do you know an image is really an image and we're gonna start doing it with videos. How do you know a video is really a video? How do you know an article was really written by that person?
And I don't know. I don't know that we're going to go anymore.
Yeah.
And someday I'll probably be a differentiator. It'll be like you've got real people writing. Like once ChatGPT and its successors are widely adopted, then you're gonna be able to differentiate yourself on not being ai.
Yeah. and right now we're differentiating ourselves based on prompts cuz I'm teaching my team how to use it and they'll come back with some stuff and I'll be like, Now, no, you need to be more specific with the prompt.
Like write it in a fifth grade level or write it, in this form, or tell it who it should be. You're a professional writer or you're writing your dissertation, so, use that tone or that kind of language. It can, and it does. But yeah, it's really interesting.
What do you use it for?
Like once you have it open, what's
kind of your go-to use case?
Well, I will tell you what will happen. So you and I just had this conversation. I will give it to my team and what I'm teaching them how to do is we've been generating the transcript with AI for a long time. We have a tool that we drop the video in, it takes the audio and it turns it into text.
Cool. What they're doing now is they take the text and they drop it into chat, g p T, and they say, chat G P T. Here is the transcript from Jeff Blanding and Bill Russell having a conversation on health it. And so it reads it all in, and then we go write the show notes. Show notes. Should have four sentences, summary, five bullet points, and two quotes.
Two or three quotes. And it goes there it is. Now again, what I keep telling my team is it's, it generates and we have to review it. Are these the best quotes? Are they the best? We're not trying to, you can't let it, you can't get, become lazy with it. We are still smarter than it is.
It can write better than me at this point. But anyway, so it'll do the show notes. Then we could say, Hey, social media posts. It could potentially write those. There's some things we haven't experimented with, but we could do a lot. It's a pretty powerful tool. Hey, we're over. I've we're just rambling now.
I'm sorry about that, man. I know. I know you have other stuff going on. Jeff. Always great to catch up with you. Thank you very much. Yeah, you too.
Thanks for having me, bill. Good to talk to you.
β π 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 CrowdStrike, digital Scientists, optimum Pure Storage. Suretest Tao, site Lumion and VMware who have π invested in our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.