This Week Health

Don't forget to subscribe!

November 17, 2023: Andrew Rosenberg, CIO for Michigan Medicine, delves into the evolving landscape of healthcare technology and its profound impact on modern medicine. As Rosenberg unpacks Michigan Medicine's journey, he poses thought-provoking questions: How is cloud computing reshaping healthcare infrastructure, and what does this mean for the future of health data security? The conversation then shifts to the intriguing potential of gene editing – could this be the key to eradicating diseases entirely? Rosenberg also contemplates the role of AI in healthcare, exploring both its transformative potential and the ethical considerations it raises. This episode not only reflects on the current state of healthcare technology but also challenges us to consider its future trajectory and impact on patient care

Key Points:

  • Cloud Computing Journey
  • Gene Editing Potential
  • Prioritizing AI in Healthcare
  • Cybersecurity in Health
  • Healthcare Infrastructure Resilience

Subscribe: This Week Health

Twitter: This Week Health

LinkedIn: Week Health

Donate: Alex’s Lemonade Stand: Foundation for Childhood Cancer


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.

(Intro)   we potentially are here. Over emphasizing the change in the next couple years and under emphasizing where it's going to be in five and ten years

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

(Main)   All right. Here we are for another keynote episode. And today I am joined by Andrew Rosenberg, CIO for Michigan Medicine. Andrew, welcome back to the show. Great to be here. Thanks, Bill.

it's been a little while, though. We caught up this summer and talked for a little bit. And I do want to talk to you about your cloud journey at Michigan Medicine. All those things, but we always start the show with the same question that everybody gets, which is tell us about Michigan Medicine, scope, what you guys do, and how it's changed since the last time we spoke.

it's changed a little bit. So Michigan Medicine is the University of Michigan's health system and medical school, and What I generally tell people that distinguishes it a little bit is not the medical school, not the health system, nor the university. It's the combination of all three.

It's administrative, financial, political integration, and geographic integration at the level and scale of the 19 schools and colleges of the University of Michigan, and the physical proximity of all of these schools with the health system that makes it fairly unique, as well as. Michigan Medicine's role in the state of Michigan.

As a constitutional entity, University of Michigan existed before the state of Michigan existed, so we're written into the Constitution. While we're not a government branch, we have a role that supports from the education, from the health, from all of those aspects the citizens of the state of Michigan in a unique way compared to pretty much every other state in the country.

That's what makes Michigan Medicine an interesting place.

You guys have grown a little bit since the last time we spoke. There's been a couple of integrations and more of a reach, than the last time we spoke. Right,

and I think other health systems around the country, you're still seeing this next wave of integration and consolidations of healthcare across the country, and we're part of that as well.

About six years ago, we acquired A very good health system in the western side of the state called MetroHealth. It's now U of M Health West. And about seven months ago, we acquired another great health system Sparrow Health System in the Lansing area, which is now U of M Health Sparrow. And that. Is also part of a very deep affiliation work that we've done with Trinity Health, the St.

Joe's Hospital, Chelsea Hospital, where these deep partnerships, and it's this network of care that we're building across the state within that common mission that I've mentioned before, that also is something that we've been doing over the last several months and several years.

I interviewed BJ Moore recently and as these things start to scale, it's amazing to me how much more complex systems get, the data, the integration and whatnot.

And it was interesting to talk to him because I think he's been in the chair at Providence for five years. And I said so, how are things going? He goes, I feel like we finally got to the starting line after five years. we kept bringing new things in.

We kept bringing new systems in. And we had to keep doing all this work of integrating the systems, integrating clinical workflows and integrating the communication and all those things. What does that look like? What does that feel like from your chair these days?

I heard Mike Pfeffer say something that I think a number of other CIOs talk about, and we certainly are seeing it too.

In my opinion, uS Healthcare digitized over the last decade with meaningful use and with high tech and with all the E H R implementations and then all the associated health it that goes along with it. Now we are trying to digitize. Our work.

And one could flip digital and digitize, but the idea is taking digital tools now, and we're now trying to change the way we work from very analog methods to these digital methods, really in service to then transform, to do that digital transformation we've talked about for the last decade, but in a way we really haven't done.

Healthcare right now, it's just. up rounding in the ICU last week. And I can tell you that health care delivery in the ED and the ICUs, the work in the ORs is almost indistinguishable from what I did during my residency 20 25 years ago. It looks the same. Now, you'll see people rounding in cramped hallways with Computers on wheels.

You'll see a lot of data that's now digital. Fundamentally, how healthcare is delivered, including, interestingly, after the COVID peak of telehealth, most health systems are still seeing a fairly steep decline in the amount of virtual care they're delivering, with a return to much more, I'll call it pre pandemic, inpatient in person clinic, care, surprisingly, at least at Michigan Medicine, but I think it's elsewhere, even in areas that we thought were optimized for virtual care like mental health, we're still seeing an ongoing decrease in those visits.

So I say all this because I think we're going from analog to digital, we're trying to digitize, and we're getting Into areas where we could do digital transformation, true hospital care at home, more use of robotics, really changing the way healthcare is delivered. But that latter part is still not that common.

Instead, what's happening now that we've digitized, we're seeing this Continued wave of health system integrations and smaller health systems and stand alone hospitals and certainly stand alone practices for the most part, not only becoming part of health systems, but health systems now acquiring smaller health systems and that consolidation of the industry, I suspect, will have at least two more waves over the next Thank you.

10 or 15 years. 10 or 15,

wow. Has the way medicines being practiced changed? I mean, we went to UGM and we heard about, doing searches based on symptoms and then collaborating with people in other areas as you find these rare diseases and those kinds of things.

Has the practice of medicine changed or just the normal activities? In that ICU, are those the things that really haven't

changed much? Yeah, I think it's more the activities. let's take that example that every one of your listeners would know. Prior to the EHRs, there was no such thing as an in basket.

Now, it's the source of numerous conversations related to, most interestingly, burden and problems. Not great... advantages, but there's an aspect to the in basket work that is pretty universally loved and appreciated, especially from the patient side, which is the patient portals. So the patient portals also didn't really exist prior to the EHRs.

And I think most people, certainly patients and their families, really like the portal. And in that regard, that is a change in how healthcare is being delivered. But it has some real advantages, and then it has this massive disadvantage. That's all those messages that, frankly, if I wanted to talk to you as my doctor, I had to call, leave a message, maybe send an email.

But now I just portal and that message comes right at you. And that's both. A wonderful thing for patients, and frankly, I think even for providers, and that it also has this new burden attached to it. But that's true with most, in a way, disruptive technologies. Until we change the way we work, these technologies have both some great features and then they also have some negative ones, but that's one example.

The other one, though, that I would add would be, from a true change. This is the area that I'm kind of interested in. I just don't see happening that much. Where are fundamentally new data being used in patient care that transform the care? I would argue that the ability to do virtual care was an example of it.

The new data or the new method. was the ability at the peak of pandemic to fundamentally see a lot of patients without significant harm and in some cases real advantage. But that's what's interesting is, follow the money. Why hasn't that stayed high? I have my own opinions, but I think there are people with much better insights than I on why haven't those levels of virtual care Which use new technology and in many areas change the way care was done to a different way of doing it.

For example my wife was a speech pathologist. She still, half of her clinic now is virtual. Whereas before the pandemic, none of it was. Routine post op visits, routine follow up with established patients after having done the things that need to be done in clinic, like scoping patients, and you have the data, then you can go forward, but why hasn't that stayed at high levels?

I think people want to be seen in person and Thank you. Health systems may not want or getting reimbursed to do virtual visits in the same way, and they're looking to still maximize the revenue. I'm not exactly sure, but the data, I think, are fairly that the virtual care volume has dropped down to roughly about 20 percent from peaks in the 60s and 70s.

It's still a lot more than previous to pandemic, but nowhere near the peak. But the real interesting thing would be this. Where are the new data that I was mentioning in genomic, proteomic experience, social determinants, the things that don't typically get captured, where are those now being included to fundamentally change?

Healthcare, immunotherapy for cancers, some degree of biologics that are fundamentally changing inflammatory diseases or oncologic diseases, things like that. Where's Pharmacogenomics to use those types of data to fundamentally change the type and class of drugs and agents that people are on.

I don't see that shift happening quite yet, and I know, for example, even at Michigan Medicine, where we've turned on a tool, in this case, an EPIC, to capture structured Discrete genetic data in the genomics module that then can be used in new ways to fundamentally change the way patients are cared, that really hasn't taken off yet.

I think it will, but it hasn't yet. And those are examples where new types of data, I think, could change healthcare.

I love talking to you because we can really go in a lot of different directions, but I want you to choose the direction. What's top of mind? What are you working on, thinking about right now at Michigan Medicine?

Let me pick three things that I think are reasonably new in the IT and informatics space. Let me start with our cloud work that you'd mentioned. Let me then go to cyber security and assurance, but as it relates to these broader and bigger systems that we talked about. And then let me end in something that in a way is not health care, but I think it's still important around continuity and resiliency.

All right. So notice by the way, I didn't mention AI once .

Yeah, that's sort of a badge of honor these days. It's like, I might take you there over time, but I

I'll go there, but I'm not leading with it.

purposely, and I appreciate it. Cloud computing, so, okay.

The last time we sat down and had a conversation, you were talking about developing a model for moving to the cloud, which I thought was really interesting, and love to elevate that conversation a little bit and just get that out there So give us an idea of how Michigan Medicine is approaching your cloud journey.

Yeah, so it starts with, I think, one that most people would recognize right away, with some very clear articulation of business needs. And that's an interesting one in healthcare, because The more typical ones in other industries around business agility burst capability, expansion capabilities, new code, or a new company creation, or an industry where development, particularly software development and software tools and services, are fundamental to the business itself, don't land as solidly in traditional health care or big academic medical centers.

We still have a lot of wet labs. We still do a lot of synchronous education. We still, as I mentioned before, take care of a variety of patients in real time, in person. So some of the drivers that you find in other industries where cloud is absolutely critical to the beginning, I find are not as conducive in healthcare.

One example, you're not going to trade current budget, current financial investments to fundamentally change many of the things I just mentioned to do them new and in the cloud as you would potentially if You are in a different industry. So starting with what the business cases are is a tricky one.

New development of predictive algorithms and AI to fundamentally change the way we take care of patients are just not ones that we're going to be investing. millions of dollars in the cloud yet. It's unproven. It needs to be more deliberate. So where do we start? Really basic logistics. What I've started with was actually more around the security and the resilience imperative, secondary data centers, offload

Legacy, on prem, bare metal, rack and stack kind of behaviors to potentially less expensive taking advantage of cloud workloads. That's where I started. In fact, in some ways that's been the most conducive to say, we want to get out of our secondary data center business. Not our primary, but our secondary.

DR. And some of the things I've mentioned, but we're taking it a step further with Tim Callahan, our CTO, to move some of our primary production systems also to the cloud. One of which, by the way, is to get it out of our secondary data center, we still have to be able to run full production EPIC in the cloud.

And a few places have now done it. I think it was fairly rare about four years ago. Now you're starting to find more and more places. Number one, some of them are doing it having Epic host it. That's fine, but we're talking about public cloud or hybrid cloud using the large hyperscalers. So the first business use case would be getting out of the data center business and how do you take steps to do that?

We're doing it through our secondary because our secondary really hasn't been able to keep up with the massive expansion of our digital footprint. Yeah, it's

hard to get that investment for the secondary data centers.

Right, so right now if we had a problem with our primary, we would not be able to run literally hundreds.

Of applications that are just not duplicated in our secondary because we haven't made those investments. We will back up all the data, and take a true ransomware event, take a primary FHIR or some other disaster in our primary data center, and resiliency to operate well. And therefore, I think the, Cloud workload offers, from a DR perspective alone, an immediate use case and immediate ROI.

And one thing that we're doing is looking at secondary resilient redundant ISPs for our network, a COLO for non cloud Native or cloud able systems and applications. Take PACS, for example, that's a usual pretty big one, or lab information systems. Those tend not to be as transferable to cloud native as well.

And then, how do we start to make some fairly big statements by things like run EPIC fully in the cloud. What does that mean? There are some workloads in EPIC that are more amenable to that than others. And I would say the hyperscalers now, because of some of the other large health systems, have chipsets and increasingly capabilities to run the amount of GREFs that we actually need for the concurrent users at the scale we're talking about.

But one can envision a fairly deliberate roadmap of Steps To move large platforms like EPIC our document management and then even things like our PACS and imaging systems and many others to cloud. Because part of what we're also doing, recognizing increasingly more and more of our strategic platforms are moving to SAS no matter what.

So part of our cloud strategy immediately starts with recognizing more and more of our large platforms are becoming cloud native and SaaS, no matter what. So, everything I've said so far, I would say is still relatively routine discussions among large health systems or others that have done this.

The part that I'm getting a better understanding about is And I'll try to articulate this concisely. We took about a year to study our applications. the servers they're running on, the data that are being used, the type of data, the type of compute needed for those applications, and the amount of data, and the type of storage needed for those applications, and the data, and those servers, so that we would get a very granular view of the application to server to data to compute to storage needs at a fairly granular level.

And you can imagine the matrix of what I just said, so that one could take an application, the amount of data, when it's being used on what type of server, what type of CPU, what type of VRAM, and the type of storage. Is it fast? Is it slow? Is it expensive? Is it not? And map that out in our current state. So that as we now go with our RFPs to the large hyperscalers and other cloud providers, we have extremely detailed, actionable data for them to evaluate and to give us best.

Terms, conditions, services, capabilities, and pricing. So that we can then not only find the best partners, the best services that we will need to migrate and to do that work, but we can then also develop a very careful plan that a doesn't bankrupt us and ideally demonstrates the value and the return on investment that we need.

Prior to that level of granular work that we did, and that took two very good vendors to help us do that kind of evaluation, I would argue anyone doing it before that is very likely to just have a very macro view of what they need, and frankly don't really know what they're going to get until they do the migration and start to see both the services.

How the data, how the applications are being supported, and then most importantly, what are the costs? I'm concerned and reluctant to dive in, but I also don't want us to just dip our toes when I know we need to be moving at scale with a certain degree of speed because of some of the business cases I mentioned before, the need for resiliency, the need to have workloads so that as we do this larger work with our integrated broader network, we can bring more workloads into this new paradigm, not try to figure out how to get data centers to talk to each other or consolidate one data center into another, that old style, but instead how do you shift workloads to modern methods, i.

e. cloud, within a very resilient, and capable framework that meets the Latency needs for some of our workloads, and that meets the investment capabilities that we all have, and not just bankrupt us. All of those pieces are forming up into this matrix, and when we're done, when we release these RFPs over the next month, my plan is to publish that matrix.

Just to make it available to people to see what we did that took us a year to do. We just want to make that available for other people so that they can start to see the outline, if you will, that table one that I've talked about. Like medical journals all have a table one that they can see the framework and then they can just plug their numbers into it.

That's a little easier said than done but the level of this granular detail, we don't have just one storage type. But we don't need the 22 or 25 storage types we currently have right now. We probably need five, and those need to be married to particular types of compute. Some of our workloads don't need super fast, super high available compute and storage.

They can actually be done in the cloud with different storage and CPU sets at a much more affordable price. And then you have things like, what do we do about a lot of the cold, the storage that we know we just want to throw into some form of cold storage, as opposed to keeping on our servers that we recapitalize every five years.

Those, I think, are topics that every major modern health system is either engaged with or is going to need to engage with in the next five years to stay modern and current in how IT is run.

It's interesting. curious how you position this conversation with the executive team as you're sitting there and you're talking through the budget and you're like, look, hey, we're going to be making these investments and there's a resiliency factor.

There's a. Obviously ransomware is an issue and those kinds of things, but at the end of the day, it never really pencils out, like the whole cloud thing, you rarely look at it and say this is going to cost us less, but if it coincides with a cycle, right? You're upgrading your data center, you're upgrading a lot of infrastructure and whatnot.

That's when I've heard people making those moves. But I also hear you making the case that modern practices, modern technologies, and this is one of the things I've heard as CIOs have moved to the cloud. What they're talking about is access to these resources access to architecture that is more, DevOps or DevOpsSec than what they have today.

And to be honest with what's really interesting to me is You know, in the data center, we might have 25 different types of storage, but when you move to the cloud, they only give you the option for like five. They don't give you the option for 25. And so, when you go for images and that kind of stuff, they give you like six or ten images to choose from.

Whereas in most of our data centers, we have, in some cases, we have hundreds of images.

right. And on one hand, that constraint I happen to think is good for all of us the industry matures in the digital part of health care. And again, I would argue we're still in many ways at our nascency because of that earlier comment I made about a decade of time Was just digitizing to really be digital. I think I flipped those when I first mentioned it, but I think that we did that cost billions and billions of dollars. And 1 thing that's not, I think, fully appreciated by many of our.

non it, was it's very difficult now, certainly as a CIO, where our run grow transform, when we want to do something that's transformative, we now have a massive run that just chews up OPEX and CAPEX every single year just to replace all those endpoints, all that server, all the maintenance and license growth.

All the use of legacy that it's very difficult for us to get rid of it. Everyone talks about it, but App Rationalization is fundamentally an emotional, social business project that either your leadership gets and helps to do it well, or you don't do it well. Right now, we've really not done it well, at all.

We just keep growing applications. We're a very large, complex, academic medical center, and there are all sorts of wonderful arguments for why we need the old stuff. And I, on one hand, get it, but it sucks up capital, and most more importantly op ecs now, to do that new stuff. So, so I say this in answer to your question that While the hyperscalers may constrain us, I happen to think it's more helpful at this point because of that standardization.

And by the way, within that standardization is all sorts of ability to innovate, all sorts of ability to work with the hyperscalers around very focused laser type work that truly do need new methods. And I think the hyperscalers are eager to get into those kinds of discussions. But in the meantime, I think as a service is extremely affordable and a great example of very positive ROI.

are your end users starting to push you to the cloud in that they're saying, look, you're an academic medical center, you're a lot of research going on at University of Michigan. And would assume that there are tools that they have access to in the cloud. There's collaboration they have access to in the cloud that they wouldn't necessarily have on prem.

Do you feel like the end user community is pushing you in that direction?

Yes, and they're going there no matter what. And again, this is where you see a lot of SaaS. And part of our cloud work is to architecturally from a governance point of view, from a policy point of view, from a security point of view, from a privacy point of view, to at least...

Organize and keep an eye on those efforts. Not to overly control some of them, and in fact that's where some of that edge innovation occurs, but everyone, by the way I would include most of our innovative folks, understand if they have a big grant or if they have a large amount of money and they go out and they just buy a SaaS or even, they in themselves go all the way down the chain into infrastructure as a service and essentially recreate their entire stack.

They're fundamentally using the organization's data with all of its rules, and they can't just operate independently, or they're not allowed to. And therefore, many of them want the freedom to innovate, but they essentially want to be told where the guardrails. And so part of my view as the CIO is to, from our cloud perspective, there are things we're doing in enterprise IT, but there are also things we're doing to help give...

At least frameworks and guardrails for the people who are going off and doing it themselves. So, yes, they're both asking for us to do things. They're also demonstrating what we need to be doing by going out and doing it. Sometimes poorly, sometimes not at scale. But, hey, what Bill did was really cool. We now want to do it for the entire department.

Great. We agree. This is, for example, where AI is playing itself out. I like finding categories for AI, and one that works for me is all the vendor based AI, and then all the innovation that we want to do on our own. The innovation we want to do own, that's a great use case for us with these hyperscalers that we're looking to partner with to say we want to build out our own environments where we can use the latest LLM and latest deterministic models and other types of AI so that we can experiment and if something works, we can then figure out how now do we scale and make it enterprise grade and implement it.

That's a very different discussion than what we're doing with Epic, or Microsoft, or any other vendor around the AI that they're using around another, workload. So that's an example where the innovative side of AI, the environments, the on ramps, all of which would fit into work with either our hyperscalers to use their native tools, Because I think a lot of our folks don't quite know what could be available to them from the large hyperscalers that they'd say, my God, I can't believe this already exists.

I don't need to find a optical character recognition algorithm or a voice recognition. I can just use native tools. Now I can really do some cool things in the protected HIPAA aligned. Frankly, regulated environment that we would build with the hyperscalers to let people try things out.

Right now, they're trying to build out sandboxes or get access to things and then they know wait a minute, I can't use GPT 4 for patient care. Right? And so that's where AI and some of the nascent work that's going on fits into this larger cloud vision as one example.

I wanna be real clear you brought it back here, but I do wanna talk about, I said I was fine because I love talking about these topics with you. again, I wanna talk about the pool on this. it's almost daily now that I open up the various journals and things that are out there.

And I read a study from. People we know in the industry and some physicians who are utilizing these tools and whatnot. And the reaction is all over the board. It is from, hey, extreme caution, be very careful. And it's not hard to identify the areas where these AI models fall short to extreme optimism.

Oh my gosh, look, it can pass this at a 70 percent scale. Therefore, it operates the same as a medical resident who just graduated. I mean, we've read those kinds of things. And you and I sat in a room where we were talking about the whole notes thing and AI, generative AI specifically responding to notes more empathetically.

It's the study everybody likes to talk about it more empathetically and in more detail than a physician would mostly because. Generative AI, a word doesn't cost a minute, it costs a millisecond. So it can generate more words and it can be more gracious with its words, whereas a physician is taking up their time.

we're seeing a lot of different responses to this, but almost every day I see another use case. I see another new tool, a new SaaS tool, a new, it's interesting. I think Nuance picked up like five competitors. in two weeks, because people are now figuring out, Oh my gosh, to do a transcription, not that hard, and to turn it into a medical note with, the open source models that are out there, not that hard.

And now all of a sudden you have all these companies just going, Hey, we can do this. I know there's a lot more to it. Nuance didn't just pop up overnight. They've been doing this for decades. but we are seeing that kind of enthusiasm and we're seeing that kind of innovation at this pace. I'm wondering, what you're seeing and what you're feeling in terms of the adoption and how you're managing the expectations of the use of this kind of technology.

I find myself frequently resorting back to my inner Luddite. It's a funny thing because I got into this job, we talked about this a long time ago, and I created the CMIO role for Michigan Medicine back in the day, and then this system role, having been a tenured faculty who was doing innovation in the ICU around data and health.

Predictive analytics and things like that. And I kept coming up against this big monolithic IT organization that wanted standards and wanted reproducibility. And I kept fighting it until I got to know the people and then realized, no, there's a, there's obviously a meeting of the minds and a blending.

And yet as the CIO, I find myself naturally falling back to some of those less. out of the gate moments. One, I think most of us right now would agree, and we talked about this together with Ciro and Donna and some of the other CIOs about the joy that I see being derived by the AI discussions more than some of the other big stuff we've talked about before.

Blockchain, virtual care, EHR adoption. There's something about remote process automation Robotics, personalized care, whatever these large themes that have really gotten people interested and attracted within health systems and other industries. There's something going on right now about the AI work that I think is different because it is Accessible in so many other ways.

It brings this kind of joy when you play with GPT and you get things back right away. It's very different than when you try to build out a public Contract in a blockchain model and show, immutable records that there's just a very big difference between the AI and that joy piece.

But also, I think the joy is also part because people see just. Numerous examples of where, when this thing is working and working well and working in a secure and an audited and a non hallucinatory manner, it is very likely to change paradigms. And I think most people feel that way. But I feel that we're very much in that Bill Gates moment where we potentially are here.

Over emphasizing the change in the next couple years and under emphasizing where it's going to be in five and ten years, although I think many people actually do get that. So the Luddite in me is purposely Not moving at speed and haste to get an instance of GPT 4 going, put it in a HIPAA line, letting people start to work on it, and innovating and building out great new apps, because on one hand, most of our most pressing problems, the need to do summaries, the need to reduce burden for a variety of providers, not just doctors, but nurses, I think we're more likely to get success through partnering with vendors.

The use of ambient voice with a couple of very good vendors now they are much more likely to advance how we can use ambient voice in anything we could do on our own. Which, by the way, is interesting because I met a really cool chief innovation officer from a really good academic medical center at one of my meetings about six months ago, who both super impressed me with their innovative thinking and style, but also said, I have no doubt.

That we can build our own ambient voice with all open source and infinitely less expensive tools than any of those large big vendors that we've mentioned, and we can really undercut that incredible cost that we see. And I said, wow, great idea, but I don't quite seeing you and your dozen or so developers.

Just disrupting some of these multi billion dollar vendors who are also working in the space, who are there Now, I do think there's some potentially cool startups out there that will disrupt it, but I don't think we will. it's not our business in IT, even where we do some that space.

I definitely think that there are startup vendors that might do it, but not us, inherently. So in problems where vendors can use AI to solve our problems, I like to do that. It's then, where are there areas that we don't see vendors either currently working or potentially not, or where do we want to stimulate our faculty and our students who want to do new things in places we haven't probably even really even thought about yet?

That's where those Supported enclaves, supported sandboxes to allow them to try things out. That is important for us to build out. That's how I'm separating innovation and disruption. But my job is to try to create the on ramps, to create the environments for people to do the work. I don't see it. As my and my staff's job necessarily to do the tip of the spear innovation, and some of them may get very upset to hear that, I think we can do the tip of the spear innovation within that enterprise IT that we are supposed to support for other people, if that makes sense.

Give me an idea, biggest problems that you would like to solve in healthcare with technology, given as you're looking out across the landscape of Michigan medicine, and we're only really talking about that domain. wHat's the problem you would like to focus the technology resources and the technology innovation on that you feel like would have a.

a meaningful impact on care and in the communities that

you serve? Let me pick three again. One super basic, one kind of in between, and then one just totally out there. The super basic is, many of us talk about this, is where do we find more and more opportunities for a variety of automations that really do reduce work?

Summary. And AI, I think, is a pretty good example, whether they're administrative summaries or back office, rev cycle mid rev cycle, document management, or definitely on the front end and even in the clinical space, that concept around summaries, I think, is something that would significantly reduce time and effort that people do in Relatively basic work trying to collect and summarize key data and find key differences that then one could act on.

The example I like to give in my area in particular would be the ICU. When a new patient is transferred in, especially a really sick patient, multi organ dysfunction, we spend a lot of time the Interns, the students, the residents and fellows, the nurse practitioners, the nurses, the physicians trying to collate data and then collate it in meaningful ways that we would act on like, significant organ dysfunction that's occurred within the last 24 or 48 hours compared to long standing and chronic disease organ dysfunction because we can act on that and potentially reverse changes before they really sink in.

That's just one of dozens of examples. Think of the incredibly busy primary care provider who's seeing a patient for a specific complaint, but really this patient has a number of chronic diseases and chronic issues and the PCPs are just burdened by trying to sort and sift through data summaries.

That's one example of many automations prior auth to reduce denials people who are working in Medicare Advantage and other value based systems where you really do need to try to change the way you're doing care, find gaps in care, the population health type problems automations

is. Whether they're AI, whether they're robotic process automation, or things like that, or even robotics themselves that I mentioned earlier in the care of patients. Lifting much more efficient. movement of people and materials in the heterogeneous healthcare environment. You can see some of that starting to be focused in hospital care at home programs, where, for example, we're now working with Medically Home to do the logistics, that incredible ballet.

Of things that have to occur for something that we would ordinarily only do in the hospital to now not only feel comfortable doing it at home, but showing superior outcomes. So those are automations. The middle area where I would love to see us still working in some of the new types of work is what I mentioned around continuity and resiliency.

You could pick ransomware. You could pick. The disruptions caused by cyber events, DDoS attacks, and other things, breaches, the amount of money and the amount of thousands of hours spent. to manage the after effect of those items. But I would also argue that unified clinical communications the transport privacy data in a variety of ways to communicate with providers, families and providers, to have Resiliency in both our infrastructure, but in also a variety of our care delivery, the ability to manage events as they occur, power, all that digital that we talked about.

The number one concern we have is not cyber attack. It's power. It's good, clean power. Because if our power goes out, and if it stays out, it doesn't matter. It will look like a cyber event, but it's power. It's outages in our data center. It's outages in our hyperscalers. It's now increasingly power and outages at home and non traditional work areas as we have a much more distributed and hybrid work staff.

Increasingly places, especially those that are in incredibly competitive markets, have staff working in different states and in some places different countries. On one hand, maybe that's kind of a cool resiliency, but it also brings all sorts of issues with how do you provision tools? How do you continue to have that access?

How do you identify and authenticate and make sure that this distributed workforce are the real people and can work? So we have a need for business continuity and resiliency, communications, a variety of communication tools, a variety of tooling and automation I mentioned before for provisioning and de provisioning tools, and not just computers, but increasingly clinical tools that connect to networks.

That broader and more heterogeneous ecosystem requires business continuity, resiliency, and cybersecurity. And then the third way out there is A discussion I had with Seth Hane at Epic that I love to talk about, and a few others, and a few of your, frequent guests. Like, I'm not sure, should I be naming names?

Whatever. It doesn't matter. I already named some names, but

whatever you wanna say. You're the guest. .

Yeah. Anyway how do we completely eliminate disease through gene editing, not just genomics. not just actionable genetic alleles to help predict and or modify current treatments or to make even more precise treatments that are given, but how does one, through very carefully, very ethically, very equitably done gene editing, eliminate diseases?

That, that's an interesting idea that is not just future thinking. There are absolute examples of the entire chain of events that are needed to replace defective genes, eliminate proteins that are causing disease, or promote proteins that restore health, and eliminate diseases entirely. And then if you really want to get kind of crazy, then you get into germ cell and eliminating inherited diseases over time.

That has, Incredible opportunities. It has all sorts of risks attached to it, but fundamentally, there is a deep technology need underneath every single thing that I just mentioned. And how do we prepare large health systems, innovative health systems, to be able to do that type of work? That is a, an incredibly complex and incredibly expensive ballet of.

Health systems, medical schools, large vendors to work together. To me, I don't think it's hyperbole to say it feels like when the ARPANET became the NSFNET and started to inter network and we got the internet going. All of those pieces start to come together towards something like, how about we eliminate diabetes, not just treat it better and better.

That, to me, is not only a really cool health system, public health, large industry need, but fundamentally, those who work in IT, information and informatics. are at the core of what's needed to do that kind of work.

I can always count on you to push my thinking even further. It's interesting that you used the internet there.

I've been, people have been asking me about the hype cycle around AI and those kinds of things, and I equate it to the internet. I'm like, I don't think it has peaked, and I think we will be realizing just how transformative this time frame is. I mean, you talk about gene editing, you talk about genomics, you talk about AI and other things.

I mean, there's just so many innovations happening all at the same time that have been enabled by this massive amounts of compute power, massive amounts of data, and just all the work that we've done on, on. Digitizing the medical record and you're right, the experience is still not digital for a lot of people, for the clinicians and for the patients, but the digitization and the capabilities that are available now just seem to be we're right on the cusp of this almost renaissance in terms of how Healthcare is delivered, and hopefully 25 years from now, you won't walk through that ICU and see the same thing you saw 25 years


There is no question we won't, because I think there are a few themes going on, and I don't know if you can see what's behind me, but I got a little interested in maps. About nine months ago, and one of the reasons I got interested is, and there's a gap up there, but what's going to get filled in by that gap are three images.

The first is, are the original four nodes of the original ARPANET 69, the UCLA, UCS, University of California, Santa Barbara, the SRI out of Stanford and Utah. That original one. Node, and if I left anyone out, I'm sorry, I don't remember my NAT but, and then the next one is going to be when the NSF NAT was fairly developed in 1984, partly because Ann Arbor played a Pretty big role in that part.

And then the next one would be one of those new Kapurski maps of the internet and what it looks like now and all of its colors. What's interesting is you dive into the history of how really innovative people just wanted to start solving basic technology, basic computing, basic infrastructure in order to exchange messages.

Started to see the opportunities of a much, much more expansive thing, but it took public private partnerships, it took standards committees, all the things we've been talking about with the EHR development. In my mind, history is going to reflect on the high tech era and the rough decade of U. S.

healthcare digitizing as one of those very early moments. In and of itself, moving to the EHR was nothing. I mean... It was expensive and took a lot of work, but it provided very little of the true transformation that we would say the internet has done. Those early nodes in the ARPANET, to where the internet became, even in the early 1990s, let alone the 2015s, let alone now, no one really saw what the internet was going to be.

They understood the value of communicating and passing information, but in the same way, digitization of U. S. healthcare and what it's going to offer, and we've mentioned a few things, when you start looking at these maps, this growth, and one of the other maps, by the way, that one high up in the middle, those are the early transatlantic undersea submarine cables for telegraph.

And then this massive map behind it, which is from a couple months ago are now all of the undersea cables that of course are transmitting this massive amount of data that's now part of the internet plus other stuff, not just the internet. So I think when you start looking at this history and by the way I just completed a 300 slide history of computing.

I think I'm going to put together a course, but I was doing it in part because my interest in maps got me thinking about the internet, which got me thinking about some of the early computing, and then, of course, all the massive computing and the transistor and where microprocessors went and where we are now and where we're going to be.

tRying to understand that breadth, I believe, helps us try to get an idea of what you've been asking about during This hour. Where do we think things are going? I think a lot of people have some really cool ideas. What I'm interested in, where we are right now, from an IT perspective, is what are those infrastructure pieces, those services, those next steps, an example, are moving to the cloud so that we're preparing ourselves and doing work for that new type of work we've been mentioning.

That's what I'm interested in in spending our time on. And that's what we're doing.

Dr. Rosenberg, always a pleasure to catch up with you. Thank you very

much. Thank you.  📍

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

Thank You to Our Show Sponsors

Our Shows

Newsday - This Week Health
Keynote - This Week Health2 Minute Drill Drex DeFord This Week Health
Solution Showcase This Week HealthToday in Health IT - This Week Health

Related Content

1 2 3 258
Transform Healthcare - One Connection at a Time

© Copyright 2024 Health Lyrics All rights reserved