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

Welcome to this week, health 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. Today we have an interview in action from the 2023 Spring conferences, vibe in Nashville and hymns in Chicago.

Special thanks to our cDW, Rubrik, Sectra and Trellix for choosing to invest in our mission to develop the next generation of health leaders.

You can check them out on our website this week, health.com, now onto this interview.

A ll right. Here we are vi for another interview in action and we're joined by Michael Berg, who is the Chief Digital Health Officer at University of Rochester Medical Center and Justin Norton, Dr. Justin Norton, GSR Ventures, and gentlemen, welcome. I'm looking forward to the conversation. Thank you for having us.

You have a really cool title, really cool job. Well, tell us about your job.

You know, My, job came into place at the beginning of Covid where we didn't really have a strategy of how to become a digital first health system. And the CEO kind of tapped me on the shoulder and asked could I create the strategy?

And so as the Chief Digital health Officer, it's very high level strategy. Everything from patient portal through doing AI in our clinical service lines. And then I also kind of co-lead the innovation arm of our health system. So Frontier Tech. Working with venture, working with industry, working with retail.

I get to have all the fun.

So since I just talked to Greg, you're Sam and he laid out a five year strategy, which I think is the strategy you're talking about and went into detail. I wanna talk to you guys about establishing an effective investment arm within a health system. This is not for the faint of heart.

We've had some successes. Obviously we can look at U P M C, we can look at asce. And we've had some failures, and I'm not gonna say those names. and I'm wondering what does it take to set it up effectively? I mean, the failures in my mind is when the leadership doesn't have long enough perspective, and they're like, three years in, Hey, we didn't get any exits.

We're not getting a return. We're out. What's the best way to approach it? Justin, we'll start with you.

Sure. So this is a conversation, not to pick at any old wounds, but we can go, we can spend the whole time on. But in terms of how venture can work with health systems, fundamentally, and this is why we're here together, is there's so many ways to work together.

You know, We sit in Silicon Valley at GSR Ventures looking at early stage health technology. Fundamentally, if we're selling to a provider, we need innovative health systems where the leadership is bought in to try new startups. Where do they fit in to that, strategic plan where they can actually add value and it makes sense.

And then where can you help cut down, an 18 to 24 month sales cycle into, three to six months actually get something off.

So this is a partnership. This isn't you setting it up, the University of Rochester, setting it up to try to make. From investing in these companies?

No, initially this was set up to create another pipeline where I could get best of breed technology into my health system and help maybe co-develop help develop these makes sense value propositions.

But it's leading towards, hey, maybe we should take on some more financial risk in these companies. And so we do not have digital health venture. Yet, but it's something that we're actively exploring cuz we have set up a really nice pipeline where we work with a lot of early stage startups that are coming from the west coast, coming from places like Palo Alto and Menlo Park and Seattle, and we're helping these companies be successful.

So, Why not us participating in some of that financial upside of those companies.

Yeah, that makes, that makes a lot of sense. Gimme an idea of when those companies come in, how do they engage? How do you get them to engage with your health system?

So one of the things that makes the University of Rochester unique as an academic health system compared to most academic health systems left in the country is we're fully integrated into its parent university.

Most academic health systems separate. They're affiliated with their parent university. We made a core decision not what we're gonna do. So what makes us unique is I actually get to co-lead a true digital health innovation incubator where I have faculty from the engineering school, computer science department, data science institute, music school, business school, under the same roof as faculty from the medical.

Dental school and nursing schools, digital health startup comes in. We've got the chops to help co-develop. We can do the integration into the electronic health record. We can, our clinical teams, from both nursing to physicians to physical therapists all under one roof, can help think about how do we build you into the clinical workflows, operational workflows of the health system, and then we can partner with our business and data analytic team in a lot of cases.

How do we come. What is your value proposition to a health system? Because what you raise money from this guy over here may not be what I actually want to purchase your product for. And so, Really, the startup companies come through our innovation incubator and we wrap our arms around them and we want to see them successful because the companies that we work with are pushing the larger health system strategy forward.

It's interesting you just described something. That I get a lot of different perspectives on. I get a perspective from some founders and they say, we thought we were gonna do this. We raised money to do this, but then we got engaged with the health system and we sort of lost our way.

Like you could lose your way in that, especially the large academic medical centers could sort of swallow you up and then say, we started as this company, now we're this company Now that could be a good thing or that can be a, muddying of the waters kind of thing. How do you coach them in that process?

So that's such a key point. And, take an example a company like Artis site, which I know we talked, spoke about before. Augmented AI platform really can touch any number of issues across the health system. And each health system will have a slightly different. Priority list of what they want.

Right? And so we need to work with our founders and say, what are the key issues that Michael has? But also, Providence will have, Stanford, will have, everyone else will have such that you're solving one problem that can really go across a number of customers. Because that's the challenge. There's so many problems in healthcare.

Michael's got so many problems on his plate that he would love to be fixed. But what is common? Where can you build one thing that really can have product market fit and area and help you take off? Because you can't just go whack-a-mole one problem after the next. And so, what we think about and how we can help with our companies is, How do we introduce you to as many health systems as can, as many smart people to say, what is the real issue we need to solve first?

Yes, once you're successful, once you're in, eventually you can start to add on more. But what is that one narrow problem to work on? So take for example, what our site's focusing on today, or like the priority for so many health systems across the country. Provider burnout, staffing, and solving those issues.

Focus on that first, then, move on to some of these.

📍    📍 📍 We'll get back to our show in just a minute. We have a great webinar coming up for you in April. We just finished our March. On April 6th at 1:00 PM Easter time, the first Thursday of every month, we're gonna have our leadership series. This one is on CSO priorities for 2023. Chief Information Security Officers, we have a great panel.

We have Eric Decker within our mountain, Shauna Hofer with St. Luke's Health System out of Boise, Idaho, and Vic Aurora with Hospital four Special Surgery. And we are gonna delve into what are the priorities for security? What are we seeing? What are the new threat? What is top of mind for this group? If you wanna be a part of these webinars and we would love to have you be a part of them, go ahead and sign up.

You can go to our website this week, health.com, top right hand corner, you'll see our webinar. And when you get to that page, go ahead and fill out your information. Don't forget to put a question in there. one of the things that we do, I think that is pretty distinct is we. like for today's webinar, we had 50 some odd questions that we utilized, in order to make sure that the conversation is the conversation that you want us to have with these executives.

So really appreciate you guys being a part of it and look forward to seeing you on that webinar. Now, back to the show. 📍

So I'm trying to think where I want to take this. I, part of me wants to be selfish here, and I have the two of you. You have a cool job. You have a cool job. What are you seeing, like what's pushing the needle right now? What's something that you're looking at saying, Hey, you know, I think in the next 12 to 24 months we're gonna see some real.

Innovation that's gonna impact healthcare. We're both,

thinking the same thing

so I'll start. then You can go, but my, attention is right now all in the possibility of these generative AI tools. ChatGPT largely with models on how it can impact healthcare. If we zoom out over time, right? Health, it fundamentally, whether it's our EMRs, technologies or robotic surgery have only increased.

costs to The healthcare system, they added tools. Often cases, many providers will say, have made their lives harder. We have a sea change moment where these technologies fundamentally can and will change provider issues. They'll make it easier to work with. They'll take issues off of their plate. And we're seeing this from startups to the biggest companies of all right, we see the launch of, Acquired by Microsoft now integrated with GPT four today.

I know that's amazing. And the technology works, so this is where my attention is going. I'm teaching a course starting in a couple weeks at Stanford on generative AI medicine. You know what? we Initially we thought, oh, we'll keep it a small seminar, maybe 20 students. Another student in my course said, Hey, professor Norden that filled up in three minutes.

There's more than 50 people on the wait list. This is where the attention is going. This really does have a transformational Potential within healthcare.

A hundred percent agree. It's a complete paradigm shift. What we have seen over the last three months with these large language models is gonna be transformative not only to healthcare, to every single industry, and as an innovation team that has been working on machine learning and healthcare for eight years now. Seven years now. Totally have leapfrogged where the technology is right now. And so now it's like super, super exciting and a lot of the companies that we have been working with, they're tech stacks are like obsolete now of where the AI is. And so it's just like where. Can we go with these possibilities?

And for me, the burning issue is like every health system in the country, as Justin hinted at it's our workforce and can I take things off of my workforce's plate that they shouldn't be doing so they get more time back to spend with the patients, but then also get to go home on time and have their own wellness.

And that's. what's Super exciting for me.

How do you position this conversation when you're talking to clinicians, gp, PT four specifically? I mean, obviously Google has barred, I don't know where that's going. I did apply for the API for GP PT four cuz I've played with it enough now that I'm sitting there going, oh my gosh I want to start, feeding stuff into this.

First of all, you get a larger amount of data you can actually feed into it, larger prompts and that kinda stuff. But how do you position this conversation with the clinicians? Clearly we're not saying, Hey, this is gonna be a clinician, it's not gonna be a doctor, but I'm curious, what's the conversation with clinicians so that they understand the potential and.

The value of it.

So, at least the clinicians I've talked to and I've like you just blown away with playing with GPT four, just blown away, but I'm even more convinced playing with it that it's never going to replace a physician. It's never gonna replace a clinician. I see it as almost like a, virtual care assistant that sits next to the clinician, maybe packages.

Like that super fast and provides that in a way that the clinician can digest it to make a decision, but the clinician is still going to make a decision. Where I see it being powerful and where , my clinicians are already using it and excited about it, or things like prior authorizations with the insurance company.

Nobody enjoys doing that. Like everyone hates doing that, like right off the top. G P T four could probably do that better than my physicians and my providers can do, and. That's where they're super excited. Now, on the overall health system, it's actually not the clinicians that I'm needing to convince of where the opportunities are.

It's actually like my privacy officer and you know, kind of concerns about. You know, Security of these models. And so that is a whole nother culture change and discussions that we're having

within the system.

But to go back to the very beginning, one of the places where people really want to know more about this are medical students.

So I teach this past quarter of Seminar on Digital Health and have almost a fourth of the entry medical students in the Stanford class and Right. They are gonna be full out providers and call it eight. Maybe more from today. And this isn't in the basic curriculum, it's not talking about AI and how large language models are gonna affect their care.

But on the other hand of the spectrum, it's what patients are using, right? There's stories now of people using, chat gpt to diagnose themselves, go through their blood work, well read their imaging.

That's sort of, it's happening today. Like We. We may not have teach it in the med schools, and the doctors may not use it, but it's still gonna get used.

Exactly. And so that's why we need to have, I mean, even conversations like this, because as a clinician workforce, we have to be as knowledgeable about this. Otherwise, how are we helping our patient? Right. We've always had, Dr. Google, someone going through, going to Google search, hearing whatever.

They may find out these tools just level that they just got 10 x more effective. And so patients are gonna be coming in with these requests. And if we're not keeping up as a clinical workforce on top of the state of the art of these tools, we're not gonna be meeting our patients where they need to.

You just said eight years, and I'm thinking the change we saw from 3.5 to four, and I'm going, oh my gosh, what does five look like?

It's crazy to talk about five, I mean, four was released what, a couple weeks ago?

Couple. Let's quantify the change from 3.5. So 3.5 was the base shot G PT model released a few months ago in November, G PT four, a couple weeks ago, 3.5 as people tested it on, us, m e step two, the final exam in medical school was basically at 60%, third percentile just passing without images.

Without images, right? GPT four. As of a couple weeks ago, this came accident. Okay, I need to test this. So we ran it through step two exam. Again, it jumped up to 89%, which is roughly the 95th percentile. We just went from almost the bottom of a med school class to almost the top. This was without images.

Microsoft last week released their own paper, basically shipped, doing the same results and showing it could even do it on images as well. Wow. So what actually comes with G P T five? What does that performance look like? It's gonna be amazing, but even if you've frozen, even if you stopped development today with some of these foundation models, we already have enough to completely transform workflows.

Prior off, potentially patient sent messages.

It's gonna put HR out of bis. Oh, there won't be no more electronic health record.

Thank, thank you for, I thought we were at the top of the H cycle. Thank you for showing me that there's a, an even higher, there's still room to go. Yeah. Oh gosh. Well actually we've gone a little log here, but hey, this is a great, we could do this for another half hour, I'm sure.

That's awesome. Thanks for having us,

Michael. Great. Justin, always great to see you. Oh, really?

 Another great interview. I wanna thank everybody who spent time with us at the conference. I love hearing from people on the front lines and it's phenomenal that they've taken the time to share their wisdom and experience with the community. It is greatly appreciated.

We wanna thank our 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.