Subscribe
Share
Share
Embed
The Scopewell Podcast features candid conversations with leaders transforming how organizations think, operate, and grow. Hosted by James Leuthe, CEO of Scopewell Solutions, the podcast explores leadership, AI, digital transformation, and the operational realities of turning ideas into measurable impact. Each season goes deep into a different industry and the challenges of driving meaningful change inside complex organizations.
speaker-0 (00:05.294)
Hello, and welcome to another episode of the Scope Well Podcast. Today I'm joined by Dr. Juan J.C. Rojas, Associate Chief Medical Information Officer at the Rush University System for Help. So today you work at the intersection of frontline medicine, data science, and AI. Before we get into everything that you're leading, can you walk listeners through your journey into healthcare and what you do at Rush?
speaker-1 (00:31.16)
Sure. so you know, my journey I think is maybe not dissimilar to a lot of other people who got into medicine. you know, at an early age I sort of found a passion for you know, science and biology, and sort of thought through, you know, what are some careers in which I can apply some of that scientific knowledge I really think is interesting, and learn more about, you know, just not just humans, but you know, animals and all the things that are going on in the world when I was a kid.
And then, you know, I was fortunate enough to grow up in a household that had you know, two really loving parents that were really supportive of you know, whatever I wanted to do in my career. and my father is now a retired physician. he was a a heart doctor in the Metro Detroit area for I think like forty something years before he retired. and so I did have sort of a role model, I guess, at at home. But certainly I think we've taken different journeys, even though we're both in the big circle of medicine. I think we've taken very different different roads, if you will.
and so for me it was really that that light bulb moment that you could sort of combine you know, your passion for inner, you know, just working with people and then applying kind of that science or you know, biology background. And then, you know, that's sort of what drove me to to medical school is the combination of the science behind medicine, but also I think the art of medicine, which people sort of say is the people part, right? And I think even now where we are today.
as far as how AI may augment the patient doctor relationship. you know, I think that that human the part that's more human or humanity, I think, is either be even more important as we sort of go into this next twenty, thirty years of medicine.
speaker-0 (02:10.476)
And can you touch on your current role just at a high level?
speaker-1 (02:13.6)
Sure. so I I am currently the Associate Chief Medical Information Officer at Rush University System for Health in Chicago. and so in that role, I spend a lot of my time thinking about kind of AI innovation and and research that help support the mission that you know our health system has, which is around you know great patient care, producing world-class research.
And really investing back in our community with a huge focus on health equity is one of the core pillars of the organization. And so a lot of the work I do both within the office of informatics, working with our chief medical information officer, Dr. Bina Desai, and now also working within our our BMO, Rush BMO Institute for Health Equity, as the director of this new data analytics studio that we set up in the last year. I'm sort of kind of trying to find my way to contribute to those three different parts of the mission, the clinical care.
when I'm, you know, wearing the doctor hat, the research mission for helping researchers get better higher quality data, and also doing my own research. And then also really trying to advance the equity mission that we have at the organization to find ways to close and or mitigate health disparities that exist within our own walls, but even now, more broadly speaking, trying to find ways to close those gaps outside the walls of our traditional health system.
in more in the community in the Chicago land area.
speaker-0 (03:40.334)
Thank you so much for that context. Let's talk a little bit more about the health equity data and analytics studio. What was it built to address and how did it come together?
speaker-1 (03:50.752)
Yeah. so essentially it was built to address really one core problem that we hope to contribute to, contribute to the solution, I should say, which is the the known death gap here in Chicago. so in a this is not dissimilar to other urban areas, but certainly something that has been well documented in the Chicago area. it has been getting generally better over time, but still is persistent. so as of the last year.
there are communities in Chicago that are roughly two to three miles from each other as a crow's fly, if you're just flying directly like a bird. but then if you look at what their expected life expectancy is, those differences between neighborhoods could be as large as twenty, twenty-two, twenty-three years. in in the past, even as high as thirty. and so what we are really focused on as a as a studio is how do we think about
ways to mitigate health disparities that are contributing potentially to that life expectancy gap and kind of doing stuff but within our own walls to make sure that we're doing right by our patients at rush that are we are seeing every day in our clinics, our hospitals and emergency rooms, but also trying to think of how can the studio as an entity of like-minded people who like data, who have a passion for working with healthcare data and also have a passion for really our mission at rush around
you know, health equity as being a core pillar of our mission, then how do we think about applying that, you know, to to other organizations, whether that be clinics we work we work with to send us patients, or community-based organizations that serve our patients, like food pantries, you know, shelters, et cetera. And so I think that's what we really s we're s thinking about when we we set up an organization where we're like, well, do we need to exist? Like anything in an any organization, you always want to ask.
Is there a reason for us to exist? And we sort of felt like we had this right sort of niche where we could apply our our data analytic expertise as a team for specific high value things that we're working on as an organization at Rushd, kind of applying what we're calling an equity lens to problems to make sure that we sort of dive below the surface. And, you know, sometimes something might look like it's really doing really well, but if you dive beyond the surface, patients may have.
speaker-1 (06:10.656)
varying levels of success. One maybe will be driving going up by 10%, but maybe the other one's flat at 2%. So why is that? Can we drive it to the data and then create a plan that might lead to sort of action or improvement for those patients that aren't maybe doing as well as the others. so we're trying to do that both locally and then now trying to also do that in a more collaborative way with other engaged stakeholders across the city from a researcher, public health or government point of view.
speaker-0 (06:38.04)
So one of the outcomes of that work that you're describing was the recently launched Chicago health map. For listeners who aren't familiar with it, what is it and what does it allow people to do?
speaker-1 (06:50.252)
Yeah, yeah, we're really excited and proud about this being kind of our first sort of public-facing tool, if you will. so the Chicago Health Map Project, so at Chicago HealthMap.com, for those of you who want to visit it, is really our first foray into this kind of idea of data to action. in my time as a physician working in data, what has been very apparent to me is that for a variety of legitimate reasons, mostly patient privacy and competition.
Healthcare systems typically sort of have their data live in islands. and we've we figured, okay, well, we know that in Chicago we have five to ten large health systems that care for the vast majority of people that are in the Chicagoland area. Wouldn't it be nice if we could partner with a data partner that could provide us with data to better understand the health of those patients without in any way disrupting privacy? And so we were fortunate enough to work with a research based organization called Capricorn.
and Capricorn is a research network here in Chicago that has been around since around 2014 or 2015, that was sort of stood up by investigators, with the lead investigator being at Northwestern, Dr. Abel Coe, but now is a kind of consortium of investigators who work together collaboratively on sort of shared research projects across the Chicagoland area. And so
the a lot of the health systems that we were interested in maybe partnering with ready were in this network. So we I went over there and made a phone call to to Dr. Co and sort of said, hey, this is my idea. What do you think? and we had some legitimate research questions. We're still trying to answer now what the data we got. But we also said if we're gonna get that data, would it be okay if we also turn this into a public resource, a map? And so what the health map project has evolved into is for those health systems who are contributing data to Capricorn,
We are now making that more available publicly where we can say, if someone has high blood pressure in somewhere in the middle of Chicago on Main Street Chicago, can we take their address and sort of say, in 2019, you saw a doctor at, you know, Rush, and the doctor said you had high blood pressure. now we're able to sort of say you are one person in 2019 who had high blood pressure in that geographical zone.
speaker-1 (09:09.964)
And we were able to then take that address and map it to the census track level, which is a really small unit of analysis. That's like 1,500 people often, 2,000 people in the city sometimes, and sort of say how many people in that little circle of the world have high blood pressure. so that way we would get a better sense of what we're sort of thinking about as sort of hot zones or blue zones. So areas that are doing really well and what can we learn from them and areas that are having more disease for a variety of reasons, and sort of figure out if we can deploy resources with.
local partners to those areas to then sort of say, okay, can we start seeing some signal of improving, say, the high blood pressure, the diabetes, the high cholesterol in that area? We all the reason that this is important is that historically, we in on the public health side, the way that we used to historically get that data was going into the community saying, hey, you know, this is in Lincoln Park, Chicago, we're gonna just do a survey, a random survey of people that are, you know, walking down the street, if you will,
Hey, did your doctor tell you you have high blood pressure? And the answer would be yes or no. And we would use those estimates to get a sense of how many people in that local area have this condition. What we decided to do was the the reverse. how many times did the doctor tell you tell you you have high blood pressure, whether or not you remember it or not? and so it's a di a different, a different pro and con. Neither data is right or wrong, it's just different. And so what we wanted to do was to sort of take that condition data that we generate as health systems.
And then represent that in a map where we could better understand the health of our community.
speaker-0 (10:40.748)
So what you're describing is a situation where a lot of it was self kind of self reported and now you're getting like kind of cleaner cleaner data. social determinants of health has been a hot topic for many years. do you feel that better data is like a missing piece for us to address that?
speaker-1 (11:02.09)
Yeah. I mean, I think it's one big part. I mean, we do know, you know, from studies that now have gone back ten, fifteen, twenty years that a large probably the biggest driver of healthcare outcomes are actually your your social needs and less your genetics or your biology. And so, those often intersect with each other for different reasons, but often those are a big driver. And so to your point, I think one of the things that we haven't done well in people who are thinking about population level health.
in urban areas or rural areas is really f get a sense of where are the diseases centralized and where are where are people really the most sick that could benefit from more help. We know that we have limited resources and we can't deploy a blood pressure van in every corner of the city to screen for high blood pressure. But if we knew the areas that had the highest rates of blood pressure, maybe there are other people in that community have no idea they have high blood pressure. And maybe we can diagnose them and and then mitigate that the risk from high blood pressure in the future. And so
I think this is one way, it's it's not the only way where you can say, How do we take data, have different partners, whether that be a researcher, older men here in the city of Chicago, who we know for those of us who live in Chicago, have a fair amount of power, if you will, and o elected representatives, you know, mayor, other people like that, who can take a look at the data that is available now and sort of make may help make more informed decisions so that way we can deploy resources or make partnerships with local resources to sort of
hopefully take this data and then drive it into action in some way. So that way, some of the big drivers we think for the life expectancy gap are really both heart conditions, metabolic conditions like obesity and diabetes. And so can we sort of use this data to then partner with people to then sort of start working in the community to sort of mitigate that on a really hyper local level, right? Like the census track level is really unique and we don't, you know, historically if you
If look at data like this, even if it was available, it was on like the zip code level. And and for those of you who are listening who live in Chicago, you know that in one zip code, it could be very different. Like the top part of the zip code or the bottom part of the zip code could be totally different communities. and so while it's good for mail, it may not be good for health.
speaker-0 (13:13.772)
So you're describing a lot of different stakeholder groups that could use this tool on the serving clinicians, researchers, policymakers, et cetera. Was it difficult building something that creates value for so many different audiences?
speaker-1 (13:28.438)
Yeah, I wouldn't say it was straightforward, certainly. I think the biggest thing that we were fortunate enough to do is two things. One, we are fortunate at Rush to receive a large philanthropic award from the Srill family, Chicago, you Chicago Humanity Trust by the Srill family. And so that that is in part funding this this mission around having a studio, having both an internal and external facing arm of the studio, with this health map project being really a a center stone of the the studio itself.
And so we were lucky to have funding, which is you know, for better or for worse, really important these days for really doing anything. and then number two, we were really fortunate to have the Capricorn Network all and all the hard work that they have done.
speaker-0 (14:12.302)
So you've been working in this project for some time and now that it's live, is there anything that has surprised you since all the data's come together? Were there any findings that maybe challenged your own assumptions?
speaker-1 (14:25.122)
I would just say I've been surprised that, you know, the reception has been really positive. You know, it's always nice to get feedback from people saying that this will be helpful and useful. I was at a meeting where we were doing kind of a pre-launch with a couple a a group of pastors here in the Chicago led area who have kind of a health faith based network. And one of them actually said, Hey, I you know, I can actually, you know, use this to help figure out where I'm gonna put some of the resources for our that we have in the church to sort of
deploy locally. and so that was really helpful that you know, I heard that feedback that someone could really take that information, you know, make that sort of exactly what we were hoping to do, which is sort of data to action.
speaker-0 (15:08.51)
And how do you know when you've actually made a difference? Like what will tell you that you're creating a real impact?
speaker-1 (15:14.95)
I you know, I think ultimately for us it will be really how do we feel like in the next four or five years as we have this project slated to at least continue through two thousand twenty nine? sorry, twenty eight. can we sort of see longitudinally some changes as we help work with community partners to sort of deploy some resources locally?
speaker-0 (15:36.024)
Okay. If another city wanted to build something similar, where would you advise them to start? what were like some of the key some of the key learnings, like things that maybe you wish you knew on the
speaker-1 (15:45.518)
I I think we're what I would say is if you have any ideas, you know, like I w I was just on a call a couple of days ago with some people from the New York area and they have a similar research network with a lot of the hospitals there. and so, you know, I think you don't know what is possible till you ask. So I think, you know, what I would say is be persistent and and resourceful. and don't assume you have to build from the ground up. And sometimes people have already done work that are really helpful and and for our in my case, Capricorn already was a
a well formed research network that had data, had partners. and so I didn't have to sort of build from the bottom up. And I think when I first wrote this proposal many years ago, I thought I would have to. And so the lesson learned is sometimes the easier way is better. So
speaker-0 (16:32.064)
And now you can be a resource for some of these people.
speaker-1 (16:35.606)
Yeah, yeah, that would be the hope. I mean, I think I I think our hope is that that this could be really a a resource not just for the city of Chicago and the communities around it, but also for those other people that work in at this intersection of sort of public health and data, to sort of think through how can health systems and research networks better serve the community that they are kind of studying and impacting, right? And so this is just one way of doing it.
speaker-0 (17:05.358)
And there's a lot potential because we kinda touched on this earlier, but healthcare spent years collecting troves of data. And so to be able to t to to operationalize it the way that you the way that you are doing, I think it'll have a huge impact in a lot of places.
speaker-1 (17:25.198)
Well that's what we hope. That's what we hope. That's the that's the dream and I think we're looking forward to it.
speaker-0 (17:29.804)
If this succeeds the way that you hope, what is what does Chicago look like five years from now?
speaker-1 (17:35.242)
I mean, I don't I don't think I think this is one piece of the pie, right? I don't think that any one check can can solve all the problems we have in any mer major urban area and as it pertains to healthcare outcomes. But I just think that this is an example that I'm hoping this is that this is a a lighting rod, if you will, for people to think, hey, like if someone can do this, maybe I can do something very similar for a different type of disease or a different problem that
is related to healthcare outcomes, maybe it's like building a a map. I mean, we have put some of this on there, but like, you know, a map of what are all the resources out there for people to use that are already out, you know, like a a food pantry, a clinic, et cetera. We have some of that on the map, but we are, you know, that's I wouldn't say the main feature of the map, right? So could someone else sort of take what they learn to sort of build a repo or a resource for people that is like in my area, what is available? Yeah, we have some of that on our our product, but I would love for someone else to to take that you know,
baton and be hey, I'm really going to make a wonderful, curated, thoughtful resource that, you know, things open, things close, and and make it really res resource rich, but also correct. those are all things that I'm hoping in the next five years that this potentially can be a starting point for all these little sub projects that become a coalesce around improving the healthcare outcomes of our patients at in it both at Rush but also more broadly across the city.
And then obviously the the big dream, which I think will take longer than just five years, is can we use resources like this, this sort of idea of data to action to create local and and then also citywide initiatives and more region wide initiatives that are really aimed at decreasing those life expectancy issues that we see. And then, you know, in a dream state, if this was just one small piece of the pie that
10 years from now, we're able to say that the life expectancy has shrunk to 10 years from 19 to 22 or something like that, right where it is now in certain communities, that would be a huge win. You know, if I look back 10 years from now or five years from now, that would be a huge win. but really just sort of providing value day to day, right? I mean, the, you know, we've seen some good engagement with the site already, trying to get the word out through mechanisms like this podcast and others to sort of get people to, you know, play with the data, see or sort of see it, experience it. I think
speaker-1 (19:53.538)
You know, what I've learned is that people I can say something in words, but when they can see it in a map, it's more real. and so I think seeing is believing, I think. and so I'm I'm hoping that people go visit the site and and sort of experiment with it and you know, we have a little contact us. So if you have questions, concerns, ideas for making it better, you know, we're we're always trying to improve it over time.
speaker-0 (20:18.998)
had a chance to to play around in there and I feel like the user interface is really good. Like it's it like you said, seeing is believing.
speaker-1 (20:29.846)
Right there. Yeah, no thank you. We pre we've spent a lot of time thinking about that user experience. So
speaker-0 (20:34.678)
So, let's switch gears for a quick round of rapid fire questions, given your AI background. just please the first thing that comes to mind. So what's one AI tool that you use personally that has nothing to do with medicine?
speaker-1 (20:50.494)
there's too many to name. well I guess what I would say is, you know, I have really gone into this idea unless probably last year really. I think in December of twenty twenty five, when the Opus models came out, I think there is, you know, this whole idea or this joke in Silicon Valley that like all of a sudden all these developers went home and they had a holiday break and they just made projects. And I was like I was just like that same person 'cause I was a doctor doing at the same time.
And that's the first time I realized that like, you know, with Claude Code as a harness and then a model like Opus and now even better, Fable, you can just do things that were no not possible for me to do before. You know, like if I wanted to, I'll give you a real life example. I did I work I work in a research lab. I wearing some of the hats I didn't talk about today. And I was unhappy with our project management tool. So I just built one over a weekend. You know, like I never would have done that before. There's no
reason I have any business making a SaaS service for lab project management, but I did. You know, so so what I would say is for those of you listening who are maybe not feel like they're not technical, I may have been, yes, in a pro you know before all the AI boom, I did know how to code and touch the computer, but it wasn't that wasn't what sort of got me over the hump is really this realizing that right now, even where we have today the AI and that's the change more and more, if you have an idea, now
you you still need to sort of think about the right idea and who's gonna use it, but AI can help you get there. It's a vehicle. and so, that could be this health map project that we just talked about, or it could be a personal project like, hey, I just want to ha organize how I work out better. You know, before you would have you'd have to go, I don't know, Google something, but now you can just work with AI to make something that works for you. So what I would say is, you know, one AI product is probably not enough. I think it's really just I'm just excited about this idea of having
I don't know, citizen developers, for lack of a better word, where people can sort of build things and maybe solve a really small problem for you. Maybe it's not gonna solve a million problems for everyone else, but it it just works.
speaker-0 (22:55.788)
What's one AI habit you wish every clinician had?
speaker-1 (23:00.982)
I would say, you know, just be curious. I feel like, you know, healthcare in a lot of industries are sometimes for a lot of legitimate reasons scared of kind of moving fast, right? and I have a lot of colleagues who think I'm cra you know, who think when I type into a terminal and it talks back to me that it's, you know, it's it's scary. But I guess how people like try it one time, come up with something that's a pain point for you. And I would say
you know, I've had sessions one on one with people like, hey, I what what is a problem that you wish you could solve? And I I would spend 15, 20 minutes with them with whatever local AI friend they have, and we're able to sort of come up with at least a starting point. And then my my homework for them is now next time you have a problem, do it yourself. and so for me it's really curiosity and sort of just being curious and and excited about how this could potentially change your
workflow for all the things you don't like doing in medicine. There are so many things we do in health in healthcare that people just don't like doing. they want to spend more time with patients. They want to spend less time typing on the computer. Guess what? What not the only way, but one of the ways that that could happen is with AI, right? So you know my advice to clinicians is to just experiment and think about how you can use this. And then more importantly, advocate to people like me who have leadership roles at health health systems to provide them with tools, right?
You can only do this in a safe sandbox. You don't want to have shadow AI where you're using your own AI at at work. And so if you're like, hey, I really think this tool is great, phone your local AI champion and and let them know, hey, this is something that might really be a great boon for other people, not just me. And then hopefully get that approved.
speaker-0 (24:42.478)
Last question. What's one healthcare AI use case that you feel is maybe underhyped? I know right now the ambient scribe is like probably the most common use case, but is there something that you feel is under under s underhyped?
speaker-1 (24:59.224)
There's a lot of underhyed use cases, but I well I'll tell you one that's actually happening that we're really excited about. we're in the middle to late phases of kind of an early pilot for something we're calling a healthcare tumor board orchestrator agent. and essentially in healthcare, especially in cancer, doctors meet usually on a weekly or bi weekly basis to discuss what they think is best as far as treatment options for patients who have certain types of cancer. And it's a multidisciplinary group.
at different a surgeon, a medical doctor, nurses, pharmacists, et cetera, who get together either in person or on Zoom these days and discuss what's the best approach. What we realized, which is not unique to Rush, but is unique to having these tumor boards, is that it takes a lot of time for humans to prepare for set tumor boards, right? So for any one patient, if if I'm a patient in a tumor board that they're discussing that day, it could take a, you know, one of the people in the in the staff
somewhere between fifteen and twenty minutes to sort of collate all the information from the computer to to put together a story that people could digest in a relatively click way to help make it an informed decision for that patient. And so that to me, when I heard that, I was like, you know, light bulb went moment. I was like, I think AI can help, right? And so, we're, you know, in that process where we're sort of taking data from the computer, translating it into something that people are
can see and iterate and and you know and still have a human in the loop. They can take a look at it, make sure there's no hallucinations, no issues. But a lot of that initial fact finding that people do that ends up being a lot of copy and paste in healthcare now is solved or not solved, but at least assisted with AI helping you, right? And then so if we can make that that nurse who's putting together a a a PowerPoint deck, a Word document for that screen that is going to be shared for everyone to see, can we make that
take her four or five minutes, you know, as opposed to 15, right? that's what we're really excited about. So stuff like that. I I think a lot of those use cases are to me like are equally as important. but also a lot of use cases that people are worrying about now are things like, you know, revenue cycle and call center stuff. and I have to say that that isn't super important too to make healthcare more efficient in all the other ways that we are inefficient. But I I guess as a clinician, I always have sort of a soft spar soft
speaker-1 (27:18.808)
part of my, you know, my being that I really want to help the other frontline people doing the work 'cause I know that they're spending so much time doing this work and they're trying to do the best job, but can we have AI help them do it? Right. And I think it's not to replace that that nurse that's doing that job, but just to free loan free free their time to do something else that they wouldn't have been able to do without having this sort of AI helper help them put this report together.
speaker-0 (27:42.252)
Yeah, the back office stuff is always the safest kind of, especially in healthcare. So it's great to hear from folks like you who are looking at it from the clinical side as well. And I definitely think we should check in again sooner than later. I'd love to hear about more of this stuff. So Dr. Rojas, thank you so much for joining us today and for sharing how your team is helping turn healthcare data into decisions that can improve communities.
speaker-1 (28:07.022)
Thank you so much, James, for having
speaker-0 (28:09.388)
And thanks to everyone for listening. If you enjoyed today's conversation, be sure to subscribe to the Scope World Podcast wherever you get your podcasts. We'll see you next time.