Q. Would you like to start with some introductions, Joel, and talk about a little bit about your journey so far?

Joel: Sure, thanks very much for having me, Rohit. I’m Joel Klein, the Chief Information Officer at the University of Maryland Medical System. We’re a 14-hospital system here in Maryland and one of the largest nonprofit employers in the state. I think we might actually be the largest.

We’re one of the dominant health systems in the DMV area. Our academic health system in downtown Baltimore provides everything from multi-organ transplants to high-acuity trauma care, bone marrow transplants, and other quaternary care services.

In addition, we have many community hospitals, rural healthcare facilities, and ambulatory sites across the state. It’s an integrated delivery system that operates in close partnership with the University of Maryland’s School of Medicine, School of Pharmacy, School of Nursing, and others.

Although we’re a private company established decades ago and not formally part of the university system, we work very closely with them. We operate very much as an academic health system with strong community components.

Q. That’s awesome. Joel, I understand you studied to be a physician. I’d be very curious to hear how you moved from being a physician, and perhaps a practicing one, to this side of things.

Joel: Sure. I’m an emergency doctor. I went to medical school back home in Texas at the University of Texas Southwestern Medical School and completed my emergency medicine training at Johns Hopkins. After residency, I was hired at one of our community hospitals in the system, where I still work.

I was the “night guy” for many years, working night shifts and taking care of everyone. It was a very busy place—at the time, it was one of the biggest emergency departments on the East Coast, with over 100,000 visits a year, which is massive by ER standards. The hospital was surrounded by nursing homes and assisted living facilities, so I saw a lot of elderly patients with sepsis, hip fractures, and other “bread and butter” community emergency cases.

I was eventually elected to lead my group of 60 doctors, nurse practitioners, and PAs. In that role, I oversaw incentive compensation, which was tied to performance metrics. To manage that effectively, I needed data—something that didn’t really exist at the time, about 20 years ago. Performance was mostly judged subjectively.

To get the data I needed, I began collaborating with the IT team. I worked with them to track metrics like door-to-doctor time, throughput time, admission rates, and compliance with treatment protocols, such as administering the right antibiotics for pneumonia or sepsis. This partnership with IT helped me access the data I needed to objectively evaluate and improve performance.

Around 2012, our organization embarked on a digital transformation journey, implementing a single electronic medical record (EMR) platform across the entire system. This unified platform gave us all the same situational awareness of our patients and allowed us to operate more seamlessly. I played a leadership role in that project early on.

By 2015, I had officially joined the IT leadership team, and in 2019, I became the CIO. This is now my sixth year in the role.

I still maintain my medical license—I just renewed it, actually—and I remain on the medical staff at that community hospital. During the first half of COVID, I split shifts with my old partners, working early mornings, from 5 a.m. to 9 a.m.

However, balancing shifts with family life and the demands of being a CIO eventually became overwhelming. I stopped practicing about three years ago. While I could still go back to it if needed, my focus is now entirely on my CIO responsibilities.

Q: I’m very, very impressed by your early decision in 2012 to go for the single medical record and also start your digital transformation journey. With over five years in the CIO role, could you share with the audience some lessons learned, roadblocks, or difficulties you found along the way? Where do you think you are in the journey? Do you think it’s completed, or is there still a lot more to do?

Joel: There’s a whole bunch to talk about here, obviously, but if I were to pick out one key thread, it’s this: the thing we’re really trying to do in medicine is reduce unnecessary variation in how we care for patients. There is a best way, for instance, to manage things like pneumonia, sepsis, hip fractures, prenatal screening, or how a delivery in obstetrics should be handled. There are best practices in all of these areas.

One of the biggest tools that any healthcare organization has to drive consistency and encourage all members of the care team—doctors, nurses, therapists, and others—to practice consistently is by having everyone on the same platform. A common set of tools and visibility into patients allows us to use data to address that variation.

If you’re all on different platforms—if, for example, the eastern half of your organization uses one system and the northern half uses another—you’re always going to be fighting against that variation. You’ve got to get everyone on the same playing field, not just to have a common situational awareness of a given patient or part of your organization but also because it’s the way to get to consistency.

There are so many tools within electronic medical record systems that can help drive that consistency. If we don’t move towards reducing unwanted, unnecessary variation, we’re never going to get costs under control, and we’re never going to see best practices done everywhere as they should be.

We’ve gotten onto a common platform and are very far along in driving towards consistency, but this is a never-ending journey. It’s still something we’re actively working on as a system.

Q. It is very interesting that you mentioned about reducing variation and driving consistency. How would that come into play, Joel, if there was any situation of, let’s say, acquisitions where you now have to deal with a hospital system that came into the same portfolio or a new set of physicians and they are not used to the same platforms or the tools or the systems that you are driving?

Joel: Yeah. A lot of health systems are dealing with this right now. I mean, this is true for any acquisition, any MNA activity. So I think you could generally speaking, divided up into two categories. One is an acquired entity or an entity that joins, you know, a health system or, you know, a larger company, um, that is smaller and less sophisticated and less mature.

And in that case, usually they welcome the help, you know, part of that’s the whole rationale for why, you know, we joined up in the first place. You know, I, I know. Um, some of the organizations that joined our health system, you know, it’s change and change is always hard, but the opportunity to to benefit from all the hard work we’ve done over the last, you know, 12 plus years in the journey that I just described, you know, they get to benefit from that without having to put in all that sweat equity, uh, that we’ve developed.

So that’s one scenario. The other scenario is harder, and that’s when you have two kind of like sized organizations that join forces to try to get some economies of scale. And the question is, how are you going to navigate, you know, the digital side of that merger? And it’s hard. Um, typically speaking, you’ve got to pick, right?

You’ve got to say, well, let’s go with this one and let’s extend that umbrella out, right? And the challenge is having governance to say, well, how do you do it? How do we do it? And I think the hardest part is, you know, getting that, um, that pattern getting into a groove of compromise. And, you know, it starts with trust.

It starts with dialogue. It starts with the leaders. Uh, not just the I. T. Leaders, but the clinical leaders and the operational leaders, you know, having relationships with each other, um, and realizing that, you know, it doesn’t really matter, uh, necessarily whether we do it, you know, whether the wallpaper is this wallpaper or that wallpaper, because they’re both okay, let’s pick some wallpaper.

Um, and, and that way, you know, we don’t have to spend a bunch of energy maintaining both kinds of wallpaper. Yeah, and you know, that’s some of the hardest work there is. Um, and, um, but it’s, it’s part of what makes the job fun and exciting.

Q: I’d like to shift gears a little bit, Joel, and ask about how you approach innovation. What’s your thought process? In a large system, innovation is usually difficult to do. How do you gather innovative ideas from patients, your nursing staff, physicians, or any other constituents or stakeholders in the organization?

Joel: Yeah, right. I think you have to approach it in two ways. One, you have to, right? Everyone knows you have to water the plants or they won’t grow. So you need a way to listen for innovative ideas, collect them, and then find a way to put some gasoline behind them and get them the resources they need to try to run with them.

Sometimes that involves identifying people out in the organization who are not part of IT and bringing them into IT, connecting them to larger teams that can help them. Clinical engineering is a great example in healthcare organizations, where people are out there repairing things and dealing with biomedical devices. They’re not always part of IT. You need to find talent in the organization that can benefit from the resources IT typically provides, whether that’s project management or infrastructure that isn’t intrinsic to clinical engineering. That’s just one example. So, you’ve got to be listening, and you’ve got to find a way to run with the ideas.

On the other hand, we have the “good idea fairy,” right? When the good idea fairy shows up, each hospital might have two to three projects or pilots they want to run. All of a sudden, you have 50 new projects, and that’s just not possible. So, you’ve got to have some governance.

Some of that governance requires command and control. You need people who have the right eye for success. If you have a big committee vote on ideas, you’ll get into groupthink, and sometimes real opportunities will be missed. You need people who can smell something good when they see it, and you need a process that’s not a dictatorship or some backroom deal. It’s a tricky balance, but you have to balance all of these things for innovation to be successful.

Q: Can you share some examples with us or anything that pops into your mind of something that you and your team have done that is innovative and that you are thinking of taking further along?

Joel: Yep. So, you know, there’s a ton of examples, some of which we’re really excited about in terms of products that we’ve developed that we feel have a serious future on the marketplace. But speaking generally, and running with the theme of clinical engineering, we believe that, just like all of us run around with our phones and use them for so much of our daily tasks, we’ve been spending a lot of time in our organization thinking about the devices in our nurses’ pockets.

How do we get more and more of the technology that patients depend on—everything from clinical monitors, cardiac monitors, fetal monitors, ventilators, nurse call systems, telemedicine, and, of course, the electronic medical record system—into a clinical device? The goal is to make it indestructible, vomit-proof, and in a form factor that our nurses really love using.

A lot of the devices we have to choose from, like an iPhone, are consumer devices, right? They’re not designed for healthcare. Whether it’s from a form factor point of view or user infrastructure point of view, iPhones are built to be used by individual users, not shared devices. That’s not what we’re going for.

So, you’ve got to wrap all of that together. It’s a lot of engineering and it has to unify a ton of vendors. What I just described isn’t the work of just one vendor, right? You need a team capable of putting that engineering together from so many different platforms. That’s an area where we’re really working hard to create something that our nurses will love to use.

Q: So, you mentioned something very interesting that ties into the environment many of us are facing right now, where you have to do more with less, in terms of budgets and resources. You mentioned throwing gasoline on a project, and to do that, you need money, resources, time, and energy to pursue these initiatives. How do you make that happen in an environment where budgets may not be available?

Joel: Yeah, so it’s a bunch of tactics. One, you have to look inside. In any IT shop, there’s probably stuff you’re doing that you don’t really need to be doing. You’ve got to be really thoughtful about that. I’m not saying stop keeping the lights on; I’m not saying that at all, but you’ve really got to be thoughtful about whether the work you’re doing is adding value. There are always people who will try to hide behind work because it’s easy, so you’ve just got to pay a lot of attention to that.

The second thing is, you’ve got to make it clear to your organization the value of what you’re doing. You need to talk about what you’re doing and what’s possible because, you know, we assume the budget is X. In my organization, the narrative from our finance community is that we need to find a way to get IT more resources. Yes, there’s the narrative of doing more with less, but remember, from a capital point of view, we’ve got to build new hospitals and buy new medical equipment. It’s a very capital-intensive industry. So how you frame all of that is key. Do they know what you’re working on? You’ve got to be able not just to tell the story, but also to demonstrate the art of the possible. Some of that involves making sure that everyday users feel the impact of what you’re doing.

And then, the last thing is good governance. You can’t just pull stuff out and say, “We’re not going to do this anymore.” A much better choice is to give the organization options. You know, say, “We want to pull X million dollars in costs out of the organization. Here are five ways to do that. They’re all terrible, but what do you think is the least terrible?” And you want the right people around the table making that decision rather than me making a unilateral one. As a doctor, I could do that, and it might be more credible, but it’s better if I do it in lockstep with my peers.

Q: Since we’re talking technology, Joel, there has been such a rapid change in new technologies like Gen AI and large language models that have really changed the AI landscape. Everyone is experimenting with Gen AI in some form or another, whether they like it or not. So, what are some of your thoughts around AI and Gen AI? How do you see these technologies accelerating innovation or improving patient care and healthcare delivery?

Joel: You’ve got to think about cybersecurity and privacy. You can’t dump a bunch of protected health information into some Gen AI platform and risk that data becoming public or leaked through a malformed or malicious prompt. That’s a real issue we need to address as an industry, and our vendors will have to help us if we want to really lean into these technologies.

That said, there’s a ton that the industry is looking at, and we certainly are, too. Ambient technology is a big piece of this. We’re piloting a number of solutions, and it’s promising, just like many new technologies are. The challenge is whether every doctor is going to appreciate ambient documentation the same way. This involves having a microphone in the room, typically on the doctor’s phone, that records the dialogue between the doctor and the patient. The output is documentation of the visit, which can take 5 to 10 minutes per patient if the doctor’s typing, or maybe only 60 seconds to review if they’re looking over what the AI generated.

It’s promising, but a third to half of providers love it, some aren’t sure, and others aren’t convinced. So, how do we get these providers to adopt it? How do we coach them? It’s like autopilots in airplanes—there will always be those who say, “Back in my day, we didn’t have that.” But it’s a tool for safe and efficient flight, so how do we get people to embrace it? That’s one example.

Another big area is idea generation. Whether it’s for an individual patient or hospital-wide strategies like patient flow, how can we use AI to come up with new solutions? What are we missing? AI excels at thinking about things differently, and that’s incredibly valuable in healthcare. Those are just a few of the thoughts we’re considering.

Q: So, you’re already looking at possibly using GenAI, and it’s crossing functional areas now, so you can actually use it in many different functional areas at the hospital system. You mentioned that you were at a VC conference and that you might want to share some thoughts and ideas from your visit at the conference. So what are some of your thoughts there?

Joel: Yes. Well, I’d say two things. It was an amazing meeting. There were about 250 or so CIOs from all different industries, including some of the biggest companies in the country. I think what’s fascinating is how all of us in IT leadership are fighting the same battles. We’re all dealing with vendors that are engaging in really punitive price increases. They’ll acquire something and then jack the price 500%. It’s not going to work. We will find a way out of that, and it’s a huge problem.

We’re all dealing with AI, productivity vendors, and capital challenges. Whether it’s an ERP vendor or CRM or anything IT-related, we’re all doing the same stuff no matter what the industry is. It’s just fascinating to hear that.

The other idea that I heard was using AI for idea generation. So, you know, here’s a business problem we have—give me five unorthodox solutions that I didn’t think about, or even some really crazy prompts, like, “Imagine the following business problem—how would the Middle Ages have dealt with this?” And it sounds ridiculous, but it’s not, because it gives you a spark, like a little way of connecting things.

What I heard from a lot of my colleagues at big companies is that that’s what they’re doing to try to get an edge on their competitors. I love it, because it sounds wild and fun. It’s a great way to keep us all collectively engaged at the highest levels, solving some of the toughest problems we’ve got today.

Q: I would love to get your thoughts on value-based care, Joel. Are you already in the midst of that? Are you planning for that? And what are some of your suggestions and advice for people who are thinking about that?

Joel: Yeah, so value-based care is not new. This is something that I think the whole industry has been trying to figure out, in some cases very successfully, and in some cases not so successfully. It’s this whole idea of having skin in the game, for some block of covered lives.

I think it comes back to how much we can focus on primary prevention. As long as we have contracts—this is Joel’s opinion— as long as we’re talking in terms of single-year, time periods of responsibility, primary care is always going to take a back seat. So, in other words, if I’m really accountable for getting patients to lose 20 pounds because of the health upside of that from heart disease or hip replacement, there’s just so many ways. But if I’m only responsible for, you know, between now and the end of the fiscal year, I’m not going to worry too much about that. It’s too short a time period.

So, how do we think about that? To me, that’s something that the whole value-based care space needs to think more about. Then, how can we partner better with payers? Who are the intermediaries here? A lot of people say the payers just suck money out of the system and they’re just bureaucracy, skimming 10-15%. But I think it’s more complicated and nuanced than that. They do have the ability to help providers be more strategic about how they apply resources. I think there’s more than just, “We’ll give you some analytics or a list of things to do,” but I think there’s opportunity there.

Q: Would you like to share some final thoughts on any new initiatives around digital transformation that you’re thinking of in the next few months or years? What do you see coming your way or our way as an industry?

Joel: Part of why I love my job is the whiplash. I’m going from this conversation to a meeting about how we can support philanthropy more aggressively, and then my meeting after that is with our CTO to talk about all the projects we have going on there. And then my afternoon is focused on our informatics team and how it’s, I mean, that’s like typical.

So, it’s great because there’s so much to do in our industry. I don’t think I’m alone. And, you know, to sort of point to any one thing would be saying, well, I love this child best. There’s just so much. And to me, that’s what’s so much fun about this job.

We hope you enjoyed this podcast. Subscribe to our podcast series at www.thebigunlock.com and write to us at info@thebigunlock.com  

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.  

Welcome to the Big Unlock Podcast, Piyush. It is great to have you as our guest today. I’m Rohit Mahajan, Managing Partner and CEO at BigRio and Damo Consulting. I would request you to please introduce yourself.

Piyush: I’m Piyush Mathur. I’m a physician, an Anesthesiologist, and work in ICU too. Originally did my MBBS in India from Armed Forces Medical College, Pune, and then subsequently through my journey came to Cleveland Clinic where I have been for more than 21 years now. Did further training in quality and education and for the last many years have been very focused on how can we impact the quality of patient care using artificial intelligence, which is a data driven science.

Q. Great to know that you started your journey as a physician. So, what motivated you? What choices you had at that point in time when you started your journey and how did you come to US?

Piyush:  The motivation was very simple. My dad, actually wanted to be in the armed forces and he always wanted me to be a doctor and always encouraged me to focus on that. I happen to be good in biology too, throughout my high school. So, it was a natural fit. I ended up at armed forces medical college for that reason, and then decided to further pursue learning and training in the US.

I didn’t know about Cleveland Clinic at all, being in India over two decades ago, and it was actually my last application for a residency program. When I matched here, I was really excited. Since then, I’ve pursued my career at Cleveland Clinic, training in anesthesiology and later in critical care. Working here with people from all over the world and learning from them has been an excellent journey.

Q. Piyush, I know that you’re pretty busy as a physician, but you still find time to do other business ventures. Could you tell us about those?

Piyush: Yeah, I always like to learn and grow. Being a clinician is important, but as we thought about applying AI in healthcare, we wanted to innovate and make it something sustainable—beyond a hobby or pure research. That’s when, along with Frank Pepe, chair of plastic surgery at Cleveland Clinic, we founded BrainX, an AI and healthcare company.

We were both very excited about this scientific opportunity for applying AI in healthcare. But there were a lot of challenges and there were a lot of misunderstandings somewhere around six, seven years ago and still continue to be. But that’s what we wanted to demystify, that how can we apply this exciting, rapidly growing science into healthcare to make our patient’s journey, a better journey, how do we empower our clinicians with this wonderful tools that the science represents so that they can be enabled to give better care to our patients.

How can we make the healthcare system more efficient using this data driven science so that the delivery becomes better? So those were all the questions that we were trying to trying to address. And we did believe together, collectively, that AI represented a great opportunity to solve that rather than just trying the tools that had been tried in the past.

And subsequently, we had other founding partners like Dr. Szymanski, Dr. Maeshwari, and Dr. Khanna join us. And we started working on exploring this, and over a time period, we had Raghav and Shreya join us as our technical leads and our research leads now. They are both PhDs in AI. And we have just subsequently grown leaps and bounds in this exploration.

Q. That’s great to know, Piyush. How many years has it been, and which are your focus areas? Tell us about work by giving perhaps one or two examples.

Piyush: We never anticipated that we’ll grow that rapidly. So, it’s just been six, seven years now since we started. Initially started with just our core team, the BrainX team. Eventually, there were so many other people who were asking the same questions what we were asking, and everybody was so excited. So, we ended up forming a community around us called the – BrainX community – so we can engage people from across the world. That’s a 6,000 plus member international community. Also, very active LinkedIn group with the same name called BrainX community. We have a website, a whole host of resources that are available over there. Free to join. All the content is free. As we grew, we also realized that, you know, we need to expand our research team.

We created one of the verticals that we have now within BrainX, our own lab called BrainX AI Research. Now we have people from across the world participating in our research work with all different backgrounds, all the way from undergraduates to PhDs. Everybody working together, trying to research different aspects of AI application in healthcare. All the way from basic data science, we are very focused on all those.

Q. We were very fortunate to have you as our keynote and spotlight speaker when we did our GenAI workshop near Cleveland Clinic. You spoke very deeply about the applications of AI in healthcare and GenAI in healthcare. Could you share with the audience some of the insights perhaps from either that talk or any others that you very frequently do otherwise as well?

Piyush: So, generative AI—or for those who may not know the term, AI like ChatGPT—is an application people are now familiar with. Tools like this have become widely available for public use, and there’s a lot of excitement around them. Our team was already focused on the predecessors of these tools, particularly in natural language processing during the pre-GPT era. So, as these fascinating tools became available, it was a natural transition for us.

It’s like being in a toy shop where developers are putting out all these different algorithms—we’re fascinated and excited to trial them. But it’s also a pivotal moment. Just like Steve Jobs made personal computers accessible to everyone, AI is now talking to people. People can interact with it directly, see how powerful it is, and integrate it into their daily lives.

This is generative AI’s “hello world” moment. But with all the excitement, we need to be careful, especially in healthcare. These tools need validation, testing, and possibly adaptation. One model might not fit all healthcare needs, so we and others worldwide are experimenting to understand the right approaches for applying these models.

What’s exciting is that these state-of-the-art tools allow people to interact in natural language, but we must ensure they are vetted and applied appropriately to help with patient care. That’s the exciting part.

Q. That’s great to know. And I am also super excited to learn that you have launched a new book, Piyush. I’m looking forward to reading it very soon. Would you like to give some insights to the audience about what is the contents of the book and what motivated you to write the book?

Piyush: Thank you, Rohit. We’re trying to put together our experience and expertise in that book. It’s called ‘Strategies for Artificial Intelligence and Healthcare.’ Y You can find it on Amazon. Frank Pepe, Jeff Mangus, and I co-authored this book. The idea was not to provide overly technical knowledge, as that already exists in other publications, but to focus on strategy.

Strategy is something I often find lacking across organizations or individuals—how to approach applying AI in healthcare. That’s why the sections of the book are laid out to address questions like why AI in healthcare, who, what, and so forth.

It’s an easy read and offers a framework for applying or learning about AI in healthcare, whether you’re a startup, hospital, or multinational company. We hope this book provides that key strategy element for mindful AI adoption in healthcare.

That’s great to know. So with that, I would like to finish our conversation for today. And I say this because we might have a follow-up podcast with you, as the landscape is changing so fast.

Q. What do you see in the crystal ball? What do you see in the next few months regarding healthcare adopting more AI and generative AI applications? If that’s what you’d like to talk about, or anything else you’re observing—whether it’s at the Cleveland Clinic or in your practice as a physician.

Piyush: The thing that excites me the most is the opportunity that AI brings. Look at this opportunity—right? You and I are talking about AI here, a physician and someone from the technological world. It’s brought all of us onto a common platform.

And what is it bringing us together for? It’s bringing us together for a common purpose, which is beautiful: to improve the delivery of patient care. That is what I’m most excited about, and that is what keeps me energized all day long about this opportunity.

I think we, on this platform, collectively working together, will clearly build great tools in the future, whether it’s with BigRio or through collaborations with many other companies we’re looking forward to working with.

The other part of this is the evolution of AI itself. If you see how fast-paced it is: we started with natural language processing a few years ago. It was slow for a while, and then suddenly, with the advent of generative AI, things accelerated.

Now, we are moving into multimodal AI. The excitement of a constantly evolving technology is exceptional. I haven’t seen such rapid growth in science or so many people engaged and involved in decades.

That’s what excites me and gives me hope. It keeps me optimistic that we are building a great future for our patients, our clinicians, and everyone involved.

We hope you enjoyed this podcast. Subscribe to our podcast series at www.thebigunlock.com and write to us at info@thebigunlock.com  

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.