Data for the Common Good with Dr. Juan C. “JC” Rojas, Associate Chief Medical Information Officer at Rush University Medical Center

[00:00:10] trained as an internal medicine physician, pulmonary medicine, critical care, but also took a informatics path and a public health path, one of the leading voices in augmented intelligence. He'll tell us a little bit more about exactly what that means today.

[00:00:27] And one of the most unique things about his career to date is he has been in the dirt, in the roots of Chicago neighborhoods from a data perspective, combining his clinical skills and community level data he's curated.

[00:00:41] And I'm so happy to have your voice from a different angle of healthcare on our podcast here today, the Associate Chief Medical Information Officer at Rush Medical Center in Chicago, Dr. JC Rojas. JC, thanks for joining me today. Yeah, thank you, Brian, for the invitation and look forward to having a colorful conversation.

[00:01:04] Colorful indeed. I think you are one of the most rare combinations of physicians, at least in the United States of America, let alone some other significant leaders across global healthcare efforts. And to say that in all honesty, JC, because we had met at a biotech conference over a year ago, I believe at this point, in Boston. And I was curious of how you had connected into that space.

[00:01:31] And now looking more at your background, it's pretty obvious. And I do want to start way, way beyond that initial meeting that we had and into more of your origin story. So I think a lot of physicians today that are of the modern era are just starting to take the path or a point of view that you have for many years so far.

[00:01:55] And what I mean by that is you've really built your career on the intersection medicine, informatics, technology, and now elevated at this leadership level. So I'm curious, JC, how did you get here? Did you just come to a point when you were in high school or middle school that you wanted to go into medical sciences and then data and technologies and emerging innovations kind of struck you in your later years? Or did this start really early on? Where did this all begin? Yeah, yeah.

[00:02:24] You know, I think every person's journey into medicine, I think, is a little different. Mine is, I think, both sometimes not unique and unique. And I can kind of share some tidbits along the way. But I think it sort of starts with having kind of familiar role models. And what I mean by that is my father immigrated to this country from Paraguay to do his medical training in 1973. And, you know, decided, hey, I'm going to learn how to be a better doctor in the United States and then ultimately decide to stay and raise a family in suburban Detroit.

[00:02:54] And so that's where I grew up. And so I had a physician in the household sort of being a role model for how you care about your patients, how you care about your community. He kind of gave me those sort of principles early on in life about like a physician's responsibility to sort of care for both his loved ones. But also, I think just broadly speaking, the people he worked with, you know, in his office and the patients he served during his long career. And so I think him as a role model certainly shaped medicine as an idea in my head. You know, certainly I saw it every single day growing up.

[00:03:24] And when I went to college, I think that's when I really sort of clicked that not only was having a career that I could sort of think about myself having like my father's, which, you know, brought stability and joy to our family. But I really got that bug around sort of the scientific inquiry, scientific knowledge, the idea of people. That's when I first realized health care really is a people sport, you know, in my early times doing shadowing and really understanding that like a lot of it is science and a lot of it is art.

[00:03:51] And some of the times it's mostly just like talking to someone and maybe having their best or worst day in their life. And you're just trying to be a sounding board to help people get through that. And so that's sort of when I decided, hey, like I like the combination of skills. You need to be a doctor and some of the knowledge you might need to acquire to become a doctor. But I will say, you know, that at that time I was sort of really heart set on being sort of a doctor like my father who practiced clinical medicine for 35 years in the suburban Detroit area.

[00:04:17] And it wasn't really until my postgraduate training after medical school where I really sort of got this idea or this bug that one, I was finally exposed to what every older physician was complaining about that time, the electronic health record. And I sort of lived it, saw it. And I realized that like that was something that was not going to go away from medicine. It was, you know, at that point, I've already been entrenched, you know, when I started in the 2010, 2011 time was there.

[00:04:43] And some places have had it for longer and some were kind of growing into a new HR. And I realized that this new HR had two things that we could do better. One is, well, now we have digital data. What can we do with that? And that was sort of intriguing early on in my kind of like light bulb moment, my first couple of years as a intern and resident. And then what I sort of realized at that same time in health care, there was this big journey towards and there has historically been for the last 20, 25 years of trying to do better for patient safety and quality.

[00:05:12] And, you know, medical schools were sort of having new quality curriculums. And my medical school was no different. My residency had a QI curriculum. And during my second year, when I was going through that curriculum during my internal medicine residency at University of Chicago, we were doing a project on vaccination rates in our primary care clients. I remember thinking, okay, well, how do we get people to accept the flu shot more? Is it a process issue? Is it, you know, we're too busy?

[00:05:38] But the reality was the first question that everyone always asks during any quality improvement project in health care is what does the baseline data say we're performing at? And I realized in that process that even though I saw this checkbox in our electronic health record every time I was in clinic, getting the data outside the computer into a spreadsheet was a heavily manual, regulated, and time-consuming task. Even when we had a mandate to do this as like a QI project that a residency thought was very important.

[00:06:06] And so that was sort of the light bulb moment for me where I realized I want to learn more about what actually happens when I click on a button in our electronic health record. And how do I eventually capture that to make a decision for patients, not just one at a time, but potentially on a population level or for a research question that involves, you know, 100 patients, right? That's sort of when I sort of realized that I didn't really know at the time where that would take me.

[00:06:28] But that's really that first Eureka moment that every project, both research and quality improvement in health care that requires clinical data has the same blocker. It's the actual data. And so that's sort of for me when I transitioned into fellowship, when I decided to become a ICU doctor or a pulmonary doctor, that's where I really knew I wanted to find training and expertise and mentorship in that space.

[00:06:52] And that ended up becoming now the kind of pivotal years where I really grew and learned some skills with the mentorship of people like Matt Chirpek and who is now at University of Wisconsin, who is kind of an ICU data scientist who really taught me first about coding and data structures and how to get data out of Epic. And actually learning a little bit earlier about stats and biostats and all the things you have to sort of think about with health care data.

[00:07:13] And that was really when I realized what I really like about my doctor hat is when I take the doctor stuff I've learned throughout all my on the job and off the job training and then putting that together with the skill set of being kind of a data person and then kind of being a bridge for people who don't have maybe they have the data skill or maybe they have a great clinical skill. But how can I help bring those two people together to solve health care problems?

[00:07:37] And that's really ultimately, you know, what I've been trying to do and now in different trajectories and hats I've worn throughout my career is really being that bridge kind of what, you know, what a classic bike. I kind of view myself as a informaticist who lives in the center of the bike. And one side is the clinical operational people and the other side often is a data science person or a data engineer, data analyst.

[00:07:58] And how do I help them talk in a better, cohesive language so we get to a better place for our organization here at Rush, better research project, and ultimately, hopefully get other trainees who are interested in this type of work, excited about doing it and not feel like it's just going to be too hard. I'm just going to just go ahead and wear my doctor hat 100% of the time because all this other stuff isn't that important. Exactly. Exactly. To your last point, and I've got to tell you, now I have the view of you.

[00:08:24] You are an informaticist with a white coat and you're really Dr. Rojas 2.0 taking your father's legacy and going far beyond the clinical setting, which I feel like you have to no matter what because of what physicians owned in terms of a hospital system or independent or private practice.

[00:08:48] You have to, you have to be aware of how there is a ripple of data for every point of service interaction that exists. So fascinating to hear that. Great shout out to Matt too. So hopefully he's going to get this vibe here all the way up in Wisconsin. So I love that origin story, JC. And one thing that you hinted on and we didn't go deep into is a public health lens. We share that in different regards.

[00:09:18] And I think a lot of your work has been tied to population health as a programmatic structure, but through the lens of a public health scientist. And there's where the data comes from.

[00:09:31] So being a team sport, public health, in the ICU, team sport, has those experiences shaped the way that your philosophy and care has maybe changed over the years in comparison to those that you work alongside or that have mentored you before? How has that shaped out for you? I think it's a great question. What I would say is I think the ICU is a great example.

[00:09:58] Like what I love about being an intensivist is that you have a variety of different team members who wear different hats and expertise, caring for one person who's critically ill. And that also, by the way, usually means their family and loved ones if they're lucky enough to have someone at the bedside. And so we often are really focused on that sort of hyper laser focus on that one individual or two individuals on our team who are really, really sick that day.

[00:10:20] And what I realized, the more I did critical care and the more I did the clinical side of my hat is that we have a lot of this data that often is either living electronic health record or is happening in sort of the way that we deliver care with phone calls, text messages. And I just always had this idea that, boy, if we had a way to sort of just get that data outside of our brains, either through a phone call or outside of our EHR, because it's always stored for an hour on the storage. Like, I wonder if we could deliver better care in a more future state.

[00:10:49] And this is even before the AI rage that we're seeing today. Now, I think we're in a place now where I'm really excited that there are things like in the ICU where we have tons of data that's been totally kind of not wouldn't say pushed aside, but it has been only used for that one moment in time, not been used to learn about or to make plans about other patients in the future. And on the same token, for even simpler things like, which we might talk about later in the show, we as health systems have these really powerful corpus of data and electronic health records around diseases, right?

[00:11:18] Like I'm in the Chicagoland area and there's many hospitals in this market and all of us take care of Chicagoans or Cook County or people from Indiana, people from Wisconsin coming to get care here in the Chicagoland area. And wouldn't it be nice if there was a way that we could work together and not have these data silos to then sort of think through how can a health system play a role, even if it's, you know, a small role in the public health of the community and actually make that part of its mission?

[00:11:44] And I think that's where I think I may see things a little differently than the person sort of taking care of patients every single day is I always sort of think not only can I make this person, the person I'm caring for right now, potentially healthier and better today, but how can I use the systems in which we sort of abstract data from that patient to then inform how we care for someone like them a year from now and two years from now with all the research it might take to actually get to an insight, right?

[00:12:07] And I think that's really where I spend a lot of my time thinking now about how do we get data to then help inform a lot of this new enabled healthcare world from the different places I spend time in, right? So that's the ICU, that's the outpatient clinic for my pulmonary head, that's just basic EHR data, all depending on what we have available in our inner system. And how do we make that even better? How do we enrich the data to be actually more accurate, right? Something as simple as is someone that we care for in our emergency room at Rush, do they unfortunately have no housing?

[00:12:37] That is not an easy field to find in any sort of modern day EHR. How do we find a way to use AI to sort of say, if someone wrote down homeless or undomiciled, we actually capture that. So that way next time so-and-so comes in, we actually may have something we can provide as far as a social worker or case manager or something like that. More proactive, less reactive. Yes, and very almost prescriptive in the community health sense when you're able to capture that data that is self-reported.

[00:13:02] And there's tons of data troves throughout the United States that are buried within credit bureaus, within other aggregators of data. But having it self-reported and having it real time, there's nothing like that, to your point. And JC, you've done some really cool things. I think there's one big thing that's led to something you hinted at that you gave me a preview of recently that I want to ask you about here. Rush Health Equity Data Analytics Studio. It's a mouthful.

[00:13:30] But I do recall this being quite an amazing project that I think culminated a lot of your skills and a lot of the data that you're talking about in that workflow you're talking about, too. Was that something that has led you to what I'd like you to get into now is health maps, which is live to my understanding. And I was thankful to get a beta, if I could say, preview of what that looks like from a user perspective.

[00:13:57] And this is going to be, I think, a very helpful tool because it's going to be so accessible to a lot of researchers and clinicians in the field. But can you walk us through if that health equity data studio was connected to now health maps being live? And tell us about health maps live. Yeah, no, thank you for the plug. It's something our team is really proud of and we're really excited to talk to people like you about. So, yeah, essentially. So the studio, let's start from the beginning.

[00:14:24] About a year and a half ago, we were fortunate enough through a large philanthropic grant from the Cyril Family and the Chicago Community Trust to really do two things here at Rush. One was to help improve our research informatics ecosystem, move our data from an on-prem SQL server to more of a cloud-based ecosystem for our researchers, and really enrich that data and get it kind of easier for people to get data quicker. Going back to my initial story, my origin story, right? How can we help researchers be more successful here at Rush?

[00:14:53] And one way is to make the data quality higher and to sort of deliver data in a faster way. And then the second pillar of that same grant, which we're going to talk more about, was really thinking about how do we take the core mission of our organization? So one of the big pillars for Rush University System for Health is a core focus as equity as part of its core tenet of, you know, what we think is really important for our health system.

[00:15:14] And I credit our leaders, you know, Omar Latif and others and David Ansel, who really have made that a core pillar of our organizational mission and also both internal and external where we're walking the walk and talking our talk, you know. And so what we thought would be really helpful is increasingly we have a lot of teams across the campus who are trying to say, you know, we have all these different quality and patient safety metrics we're trying to target.

[00:15:38] But oftentimes the vast majority of those metrics are only one aggregate, like say a percent of people vaccinated, like the one I was just talking about before. So wouldn't it be important that like we thought when we started the studio for certain high value use cases, shouldn't we also stratify some of their performance across what we call sensitive attributes, right? So that's something like race, ethnicity as a proxy often for really place based risk. So where you live often does for better or for worse in the United States play a role in your health care outcomes.

[00:16:07] And so we thought it was very important to have a team totally focused to that, that we could sort of be like a hub and spoke system where partners in our cancer center, partners in our cardiac center. If they're trying to work on an organizational data project or a dashboard, they can come to us to help with some of those sort of deeper dives around how are we doing across all the patients we serve? Are we having really true needles being moved for everyone or are some having a needle being moved in a higher direction versus someone else?

[00:16:33] And then if we find something, the idea is to really just turn this data into action, right? So it's not just about finding a disparity. It's actually about saying, OK, here's a disparity that we did find. What are some remediating things that we could try? And that's where the innovation part happens, right? That's where maybe there's an AI tool that you could deploy that might help with access. Or maybe it's something to do with your social determinants of health. And maybe we can think about a way to partner with our social workers or case managers, right? So I think that's really what we thought would be really important.

[00:17:01] But a year in, we've done three or four projects with different parts of the internal campus here at Rush. And I think we're still kind of growing in our role. But we're really excited to have a team of people just focused on doing that. And then on top of that, in parallel, what we thought would be really important, if we're going to stand up a studio, we might as well also say, let's not just worry about what's happening inside the walls of Rush. But let's also think about how can we make an impact to actually deliver something that would be helpful to the whole Chicagoland area.

[00:17:27] And that's sort of what you alluded to, this new website that we've developed called ChicagoHealthMap.com. And when we were thinking about how can we deliver something valuable, what we realized early on in this ideation phase of what ended up becoming this Chicago Health Map project is that in Chicago, we are fortunate enough that we have the Cook County and Chicago Department of Public Health who have resources like something called the Chicago Health Atlas. It's a wonderful resource for both researchers, for community-based organizations.

[00:17:57] And what we notice is, yes, if you want to get a rough estimate based on how the community is doing as it pertains to, say, hypertension, we have great people where they actually go out and survey random people in the community and say, hey, did your doctor say you have high blood pressure? Based on those estimates, the epidemiologists that work at these public health departments are able to say, community X, there's a rate of 44% plus or minus 5. And that is powerful. And that actually has been historically in the U.S. how we have and should do, continue to do public health screening.

[00:18:27] But what I realized was that all of us in Chicago, being a very competitive healthcare market, we all had diagnostic codes for patients we cared for at a different time in their life, whether that be in the ER, the inpatient setting, or the outpatient setting. And so the idea was simple. What if I could partner with health systems, break down the daily silos between the health systems in a way that's still privacy-preserving under the auspice of this new research project that we're calling Chicago Health Map,

[00:18:54] to then say, can we work together to kind of unlock this data in a different way, to sort of say, based on who actually goes to get care, not a survey, but who actually showed up in the hospital or showed up in the clinic, in any given year, how many times did a doctor write down the word high blood pressure and bill for it with an IC10 code? And then we could do that across seven health systems in the Chicagoland area. That would be a big swath of patients that we serve here in the Chicagoland area.

[00:19:21] And by the way, including the counties around the Chicagoland area, which are also in this data set. And so we were fortunate enough to kind of find a partner in Capricorn, which is a Chicago-based research network led by Dr. Abel Coe at Northwestern, but really has been supportive of many other sites in the Chicagoland area now for many, many years. And they've been great. They've been a great partner. And they really have been a great research tool for many Chicago-based public health researchers and researchers in general for many years.

[00:19:49] And when I pitched this idea to their team, they really thought, as long as we could do it in a way that's privacy preserving and, you know, really make sure the data is done correctly, like this was a great idea that we can actually not only study how we think about, how can you use electronic health record data as sort of a rough estimate of a prevalence of disease in, say, a zip code, a community area. So in our case, we were lucky enough that we were able to actually use a Capricorn network to actually geocode conditions all the way down to the census tract level.

[00:20:16] And so you're able to see in a neighborhood in Chicago that may have five to seven tracts, where in those tracts, if you were going to put a blood pressure clinic, like a blood pressure drive, and you're going to have a van go show up to people who get diagnosed with blood pressure, which one of those streets might actually be best to go? Before, you were literally just throwing a card in the wall. Yeah, you may have said, okay, these communities probably have a higher blood pressure rate on average, but community areas in Chicago are fairly big. There's 77 of them.

[00:20:43] And so we were, for the first time, able to sort of drill down on a smaller unit, which is the census tract, which hopefully we think, and we're hoping that we're only three days into this launch, so I can't tell you how it's going to work out. But what we really think is a complementary tool that isn't meant to replace what the great work that our public health team members have already done with the Health Atlas, but is a complementary tool to sort of say, how can health systems help do this? And how can teams like a health system like our studio think about not just serving the internal organizational mission,

[00:21:13] but really, in our case at Rush, really thinking about how do we serve our mission of health equity outside patients we care for? Because obviously, many of the people in the Health Map project are not Rush patients. That's where we can really have health systems really learn, like, where can they deploy resources? Where can community-based organizations deploy resources? And we're really hoping that this will just be another blueprint for how other urban and even rural areas can think about partnering together to provide data, for lack of a better word, data for the common good.

[00:21:42] I think it's really one of those things that we're doing this because we think it's going to help both community members, community-based organizations who are struggling to often find funding, and also inform our politicians, right? So we're really planning to talk to our aldermens or our state representatives about this type of data and sort of say, hey, this is the health of your community. And what are some things you can do as a politician to really maybe move the needle? You know, it's going to be a long journey over time. And a lot of these things that cause some of these chronic conditions are not going to go away overnight.

[00:22:10] But for the first time, we will be able to track changes over time, right? So if there's an intervention made in 2023, our data will hopefully go with the funding we have for this project through this philanthropic grant to 2029, right? So we'll be able to see some longitudinal differences and potentially rates of blood pressure, diabetes, obesity, etc., in the Chicagoland area. So yeah, we're really excited about it. It really is a true testament to, you know, if you put together a team who has sort of a core mission of what's something we can add value that doesn't currently exist,

[00:22:39] just believing in yourself that, like, initially it sounded very daunting, like there's no way any hospital would ever agree to this. But then we were able to find Capricorn is a wonderful partner. And, you know, we have a bunch of both the public-facing website, but now we have a bunch of research projects that are going to stem out of this to really better understand this type of data. And so more to come. And hopefully if you talk to me in a year, we'll sort of see what kind of feedback we got from the community and from the Chicagoland people doing this type of public health work.

[00:23:05] And you know what I'm going to talk to you in less time than a year from now and so forth. And I think you're, dare I say, re-centralizing EHR-oriented data and you're democratizing it in the same light, which is fantastic because you're doing it not for a profit-reaching sake, although there are health economic implications of how this data can be used

[00:23:30] to improve certain qualities that would be safe costs or improve ratings and things like that for hospital systems, integrated delivery networks, etc. And I think the mission of where you're at is, JC, I got to tell you, we never got to talk about this. Zip code plus four analysis plus five. That was one of my first loves in terms of SDOH data and looking at neighborhoods

[00:23:57] and looking at individual household identified data. And it's just a beautiful thing that you've been able to put this overlay of EHR data to that very pinpoint geographic location when we're talking about neighborhoods. And I like how you said that this isn't just an application for a very competitive healthcare market like a Chicago, but it is for rural application as well.

[00:24:25] You may not have the total populace that we're referencing in urban areas, but the value is very much there for those that are researchers and those that are trying to see trend over time, just as what you'll absolutely see come 2029 or even before that. So just thank you for the full story, because you're on day three of life here with HealthMaps, but we're going to definitely have the link to share.

[00:24:52] And I want an update absolutely in a year plus from now on this work. And I think it's so interesting because you've already seen what it can do in a short period of time because you've been building this. What do you think it will ultimately do to a clinical workflow, to partnerships in the community? Do you see Rush as a gravitational effect, having more community partners that can apply food programs,

[00:25:20] medically-tentered meals or mobile delivery to help feed people? Or do you see policy adopting this as a standard review quarterly, annually? Is that in the vision plan when you started doing this? It's a little bit of both. I mean, I think what we're really hoping for is, you know, when we designed this, we really had three archetypes of people who I think might click on the website in the first week. Obviously, one, as a researcher myself, is definitely people who are public health researchers who work in and around the city of Chicago

[00:25:48] at a variety of wonderful institutions that do great work. And I would not be surprised if they're the first people to go running to go look at the data. But then really, another two that we wanted to make sure, we also made the user experience and the data, from a data literacy point of view, equally possible, were both community members, so engaged community members who either are just curious about the health of their community, or community members who also happen to have a community-based organization, whether it be a faith-based or not, in their local area.

[00:26:16] And so that way, if they're putting together a packet for a grant, putting together a packet for some philanthropic donations, they can actually go to this HealthMath project pretty quickly without being super, super tech-savvy, get some stats for their proposal. So that was one kind of use case. And then the last one that you mentioned on the policy side is, I don't want to speak for our elected officials, but certainly if I were an alderman here in the city of Chicago, a state representative, and the mayor office, I think this is just another resource to sort of say, how can we come up with a plan

[00:26:44] that sort of makes sure that we meet the needs of the 10, 30-year vision for health in Chicago? And how do you use this data to help inform that plan, right? And so I think it's all of those things combined, and then hopefully there'll be like a fourth person I don't even know about that will really find this helpful. But those are kind of the three that we designed it for in kind of version one. In version two, we're really thinking about how do we almost be like an aggregator for other data points, right? So the central core thesis of this is that this unique IC10 data set of where people live

[00:27:14] and what conditions they had over a longitudinal period of time is really the key unlock. But we do think that there's a lot of other publicly available data that has never been shown with this data, right? So one example that we started with in version one of this web app is, for example, looking at both food insecurity and kind of food deserts. And so we have an overlay of grocery stores on there and also pharmacies. And that's another big hot topic here in Chicago where there have been certain areas certainly hit much harder in the post-COVID area where a lot of the pharmacies are actually closing in that local community.

[00:27:44] And so you're able to actually overlay pharmacies as it contains to say the rate of obesity in a community. Now, obviously that's just an association, not causal, but I think it gives people the idea to have this self-service hypothesis generating tool that then they can do a deeper dive to probably a different type of data set, right? And so that's what we hope this really is, is sort of a good place to start for public health chronic diseases in the city of Chicago where we can overlay things that people find valuable. Like one thing we're thinking about adding

[00:28:13] that's not there now is eventually air quality data, right? Or the heat index. Oh, fantastic. Right, so these are all things that we're trying to think, how can we add different data sets, overlay them in comparison, you know, a heat map of A versus B and see sort of what shakes out. And so we're excited to be on this, the beginning of a journey and sort of seeing how we can make it as helpful for the Chicagoans and the people we serve. I love how you said that in terms of new data that could be integrated into this to continue to show the different effects

[00:28:42] from a public health perspective on certain populations in certain neighborhoods. And maybe the Weather Channel can pick up this conversation and they would be a willing donor of that type of data because I would think that they have troves of that to be able to apply. So, and other parties, and I think to the broader conversation, everyone has an ear and a voice in this ecosystem that is human health. And we all have to make some sort of contribution,

[00:29:11] whether it's large or small. And I see AI transition here in our conversation. A lot of agentic agents being built here that can reason and act and assist clinicians in and outside of a clinical setting. Do you see that being a key layer that plays into not only your work here as an informaticist, but also as your clinician side of your world

[00:29:40] in Rush University Medical Center? Is that something that is being adopted in stages or in certain places like research and maybe like decision-making or upfront analysis of certain disease causal pathways for physicians to look at? Where are you seeing that play in your day-to-day life? Yeah, no, it's a great plan. I would say right now, I think our organization, like many others, I think is trying to figure out their plan. And in the sense that I think AI is moving faster

[00:30:10] than any healthcare system could ever have moved. And so I think right now, I would say in talking to my peers who do similar work across the country and us thinking about a strategy, in fact, there was a call earlier today on strategy before this one. I think we're going to start with probably the vast majority of places. So what I was lucky enough, not that long ago, to partner with Scottsdale Institute on a national survey of just this question. And so it's a paper that we're writing now. So, you know, results to be determined in the near future, but I can share that with you as a post links.

[00:30:38] But essentially around this survey went out last fall, was answered mostly in the December, January time period of 2025, then going into January 2026. And what we found was that of the around 40 organizations that responded, around 65 to 70% were either piloting or thinking about doing agentic. And I'm sure that number is even higher now, you know, that much longer. And of the use cases that they were using, not entirely surprisingly, it was mostly back office stuff.

[00:31:07] So rep cycle, client, rep cycle, potentially some supply chain automation and agentic workflows. And then to some degree, patient engagement. So that probably is code for call center. And so I think those are the three ones that I think the vast majority of agentic use cases in the United States, if you're a patient and you're going to experience AI outside of Ambient, which by the way, has been the fastest growing informatics thing in my career, went from zero to 100 in, you know, less than a year. And so with the exception of having,

[00:31:37] you know, Ambient scribe in the room, I think the next kind of frontier are going to be more kind of things that patients may not see directly. And then in my point of view, as a clinician, what I am hopeful for, I don't think we're there yet, is how do we start bringing some of those agentic workflows, like almost like a digital twin in the ICU, like in my future state, in the ICU, part of the things that we do is a lot of data in, my brain makes a decision about what's best for the patient. And I would prefer to have an AI assistant to help me do that. Maybe I also realize it a little different because I love AI

[00:32:06] and what it does for my daily workflow. But I would imagine even my colleagues who are not on AI blogs in their personal time would possibly say, have a person who sort of is on your shoulder saying, hey, by the way, did you think of this? Maybe you should think of that. That's really where I see where we're going to go. And I think I'm excited to sort of see how this evolves and where people start that journey in the next five to six years. I think we're going to start, probably my guess, if I had to put my nickel down, is going to be a lot of care gaps. It's the first place. Yes. In the military care gap. Yes.

[00:32:35] Because there's lots of those. And some of that- Even annual wellness visits, I think is one of the biggest gaps. Yeah. And so, you know, there's a lot of things like, you know, people falling through the cracks, whether that be for age-appropriate cancer screening, even just hypertension control and A1C control. These are all things that I think right now in healthcare, even though it's a very expensive, labor-intensive market, we just don't have enough people to do all the tasks we need to do for the patients we serve. And so, if an AI agent can make 100 phone calls and a human can only make 10, then hopefully we can have

[00:33:04] the human watch the AI agent's output and sort of be the, as people are saying, kind of the agent's sort of supervisor, if you will, and sort of scale that time in a different way, right? That's where I think what I am most excited about is that future. What I'm most concerned about is that we are not upscaling or training our workforce to be in an AI-native healthcare environment yet. Very good point. I think if you look around like my organization and I would say most of my peers, like there are probably just like any adoption curve, some of us who are on the bleeding edge for a variety of reasons

[00:33:33] because in my case, I fell in love with coding as a fellow and I love coding and I love being on Codex or Cloud Code or whatever. But there are times when I explain some of the things I'm able to do in 2026 with AI as assistants and I talk to my colleagues and they're like, JC's crazy. Who would possibly want to do that? Like spend their weekend doing that. And I think it's one of those things where yes, there are going to be shades of expertise that need to be happening. But I think what needs to happen is that basic bottom level has to come up, right? Like we are now three and a half, four years out

[00:34:03] from the kind of GPT moment, right? And I would say many people are still in that. Like I just chat with GPT for like, you know, I'm planning. Advanced Google search. Yeah, advanced Google, right? But really we're in a place now where if they just knew what you could do in your workflow for a ton of people who work in healthcare, I think it would be a huge unlock. I think that's where wearing my IT hat, not so much my scientist hat and my equity hat, that's where we as healthcare leaders have to be more responsible for not just a tool in and of itself isn't going to solve any problem.

[00:34:33] But what I think healthcare has done in different degrees is we haven't provided tools. So people are just using shadow AI. And so, and that part of it is because the cost, right? The reality is a lot of the AI platforms out there, if you think about how many individuals work in your health system and how many people you want to pay for per month, or even if you're doing a token-based plan, you have to sort of account for that, right? And healthcare is always trying to find ways to optimize a really thin margin. And so I think that's part of it. But I do think there are places who have been innovating in that space that can be kind of a role model, right?

[00:35:03] How do you bring AI in a safe, compliant, privacy-preserving way, but also at the same time do it in a way that doesn't break the bottom line and hopefully use that journey to then get people used to playing with these tools in your job, right? Not at home, on the weekend, actually doing stuff that actually is helping your patients directly if you're a clinical person, back office person to make you a little bit more efficient and quicker at your job. And that's what I'm really hoping we can figure out in this next three or four years. If it happens to be that I get like an ICU digital twin, that's awesome and that's what I want to get to.

[00:35:33] But I think there's so many lower-hanging fruits before then. Yeah, that is a big milestone to achieve in the arc of adoption that you were just mentioning, JC. And that would be utopian. It truly would be. But getting comfort and folks to be more native in AI is really important. And I think there are some barriers. There is a barrier of literacy. There is a barrier of ROI, just at a consumer level.

[00:36:02] But I think if you're spending $100 to $300 as a subscription, small business or individual per month, it may seem like a risk. But it's not going anywhere. I mean, it's not leaving us, as we should say. It's here to stay. And I think, to your point, just like when you fell in love with coding, you knew in terms of an EHR language, that was there to stay. We're in that next it's here to stay type of moment at least in healthcare.

[00:36:32] And that brings me to a bold bet question here, JC. So obviously, we've interviewed on our show here executives and government, CMMI and FDA and health plans, Opsum, and large and small in between. And you're one of the very few that live in this intersection of medicine and as an information scientist. And I got to hit you with this. So if we catch up again, 20 years from now,

[00:37:01] what are you convinced that we're discussing today is going to be a staple in the future of healthcare work? Yeah, I guess I'm not sure how many practicing clinicians you've had on the call. So maybe I'll skew my answer there. I am hopeful that we will have a situation in which what people have initially started calling a learning health system, which I still think is a great word, by the way, that we will have a self-evolving kind of unique AI-enabled healthcare system for your own organization

[00:37:31] that can help, especially the care delivery part, right? So I'm hoping in a future state I will have to do at least 50 to 60% less clicking, almost entirely voice. And things will happen with voice. I prescribe a prescription, prescribe a lab. It just happens. There's no clicking, there's no signing. Except for maybe at the end I just have to authorize, review the list of things that happen with my voice. People are already thinking about that now, but I think in 20 years, if we didn't use AI

[00:37:59] to really fully, almost reinvent the patient-doctor relationship in a way that kind of like goes back in time where you literally had a pen and paper or a scentoscope and that's it. And even, you know, where you were the true clinician, I'm hoping that we can sort of take that, what ended up happening for that transition period when we turned on EHRs and what has become and still is a true dissatisfier for many clinicians practicing in the United States. How do we remove that pain point

[00:38:29] by AI helping the clinician in a more kind of a seamless way where they're there? It's still privacy preserving. We think about all the right ways that it has to be protected, but then I can go see a patient go room to room and almost never touch the keyboard. Wow. And if I'm in the ICU, I'll have even more support because there'll be a world in which there'll be computer vision models seeing if a patient actually has delirium and is acting different than normal. There'll be more smarter pumps that are actually telling the nurse, hey, like, maybe you should go up on this dose of medication. I'm sort of seeing some signs in the vital signs

[00:38:58] that maybe you start going up on that. Right now, it's all cognitive human burden, but I'm hoping in a future state it really will be a AI plus human clinician working together to hopefully change the way that we deliver care and also, hopefully, the safety and quality of the care. Right? If all three of those happen, then I will retire a happy man. I don't think any of us want you to retire anytime soon, but the vision is definitely on the horizon if we have brains like yours

[00:39:27] in a clinical setting and also in a informatic setting. And with all that said, Dr. J.C. Rojas, Associate Chief Medical Information Officer of Rush University Medical Center in Chicago, thank you so much for joining our show here today. It was a mind-blowing conversation from my end. So thank you so much for dropping some knowledge. Yeah, thank you, Brian, for having me. I look forward to hopefully checking in in a year or 10 years and we'll see how things go. Maybe I'll be way off or way right. I don't know.

[00:39:57] We'll find out. Thank you, J.C. J.C. J.C. J.C. J.C. J.C. J.C. J.C.