How AI, Data & Health Tech Are Transforming Federal Healthcare

Healthcare is evolving fast—and AI, data, and health technology are driving the next wave of change. But how do massive systems like the VA and Department of Defense actually modernize healthcare? 

Now leading public sector strategy at RGP Federal, physician and health tech leader Dr. Roshni Ghosh breaks down what it really takes to transform healthcare from the inside out. With 20+ years leading federal health IT strategy, electronic health record (EHR) modernization, and healthcare analytics, she shares how technology is improving care for Veterans and millions of patients nationwide. 

We talk about the real-world impact of AI in healthcare, interoperability challenges, health data analytics, and digital transformation in government health systems—plus what future physician and tech leaders need to know about the next era of healthcare innovation. 

If you're curious about health tech, healthcare leadership, digital health innovation, or how data is reshaping patient care, this conversation pulls back the curtain on the future of medicine. 

#HealthTech #DigitalHealth #HealthcareInnovation #AIinHealthcare #HealthIT #FutureOfHealthcare #VeteransHealth #RGPFederal #HealthcareLeadership #HealthData #FederalHealthcare #FORUMPowerPodcast

FORUM POWER Podcast delivers insider conversations on GovCon and federal IT leadership, hosted by Mary Ann Brown, President of FORUM.
Follow FORUM on LinkedIn for updates. New episodes weekly. 

Transcript

SPEAKER_02 0:01

The choices leaders make today shape how government works tomorrow. Featuring candid conversations with leaders at the forefront of federal innovation and govcon, foreign power examines modernization, procurement, cybersecurity, AI, and more. This is Forum Power. And now, here's your host, Mary Ann Brown.

SPEAKER_00 0:24

I am being joined today by Roshni Ghosh, public sector lead at RGP Federal. Roshni, thank you for hopping on me today. Thank you for having me. You have this fascinating background as both a physician and a technology leader. So, what led you from clinical medicine into the world of health IT and federal technology, and how does your medical training sort of inform the work you do today?

SPEAKER_01 0:46

It was definitely an interesting path and one that I probably would not have predicted when I left clinical medicine. I started medical school at a really young age, younger than most. And throughout my school and my residency year, I kept being focused on like the system issues. Like, for example, like one of my rotations as a student, I was working at the Newark VA, and that was my first interaction with Vista. And I'm like, this is so neat. I can put order entry in, there's note-taking. And I don't know if everyone else was as interested in that part. And so the system side became really important to me. And even while I was doing my interim year at Penn, I ended up working on a research project at Wharton around the financials and business side of healthcare. So it was just one of those things that I knew that it probably clinical medicine probably wasn't where I was gonna land. And so then when I finished my year, I went to Columbia to get a master's in public health. And I was there to like look at hospital administration. But what it really did was provide me with a more comprehensive view of healthcare, including Medicare and Medicaid. And I started understanding more about the financials of medicine. And it was still a while ago where health technology was coming in, just it's like in the in the foray of healthcare. And while I was at Columbia, I started working at a healthcare startup, and I wrote an investigational new drug application for the FDA and was able to present it to the government. And that was also kind of my first interaction with the federal government and realizing there's a whole healthcare world there. So after that, I went into consulting, and that really kind of set me on my path now. Um my first project, well, my second project, was um doing the Epic deployment at Kaiser Mid-Atlantic and being the physician on the project, I started looking at clinical adoption and training and all the different pieces that are really still relevant to doing IT deployments. And then I um just kind of stated it it was just been a windy path and just kind of taken every opportunity. So it's been great.

SPEAKER_00 2:47

It sounds like you're a super overachiever. Like you go into everything and use it just as a stepping stone for the next thing, which is really smart. That's a really smart way to like get ahead. So now you've spent significant time at the Department of Veterans Affairs as the executive director advancing analytics and innovation to improve veteran care. What did that experience teach you about the unique challenges and opportunities in serving our nation's veterans through technology?

SPEAKER_01 3:14

I want to start by saying I love that job and the potential that it had to really change the landscape of healthcare delivery through innovation and understanding the process and stuff around how veteran care is delivered. So we were able to think through so many different challenges, draw from commercial best practices, and then even customize the models so that it could succeed in the VA environment. But really, the great lesson that I learned was that everyone wants to support the veterans. And I was especially proud, like we started a pilot called Vet Smile, and it was set up to be a system where leading dental schools all signed on and who are willing to see veterans on a sliding scale to help kind of alleviate that gap in benefits that currently exists. This is really important for veterans that were in need of oral surgery for or cancer care or any kind of head and neck surgery. And so it really just started tying these pieces together because, you know, obviously dental care is really important for health care and for just kind of the holistic piece of reducing chronic conditions. I didn't really know much about dentistry and it's a dental world, and there was just a lot of innovation there that was very aligned to what the veteran needs were, especially those with PTSD. And I started seeing the impact of VR and noise cancellation evolving like lighting and then having special rooms to help the veteran want to come in and get care and then also feel very comfortable and seen. And so the unique challenge that I faced was that it's difficult to implement innovation quickly and meaningfully, and also like foundational data was hard to find data that was validated, that was clean, and able to like be in a format to support these initiatives and validate the problem and measure the success and also determine whatever gaps there may be. And governance was also an issue, um, just because it's such a massive organization and clear ownership is always hard.

SPEAKER_00 5:10

Yeah. Healthcare data and analytics have enormous potential, but also enormous complexity, especially in government. What are the biggest barriers you've seen to actually turning data into actionable insights that improve patient care?

SPEAKER_01 5:23

That's a big challenge. And um, I think it's a large challenge, particularly because in government it's complicated and the data comes from multiple sources, but data quality, data interoperability, data sharing, as well as data consistency and comprehensiveness are all needed to drive actionable insights. And that exists in pockets, but it's not standard across the full healthcare realm. And there also needs to be like data insights that are coming in from the states, from other federal agencies, and even the commercial market to really make it truly actionable for the veterans.

SPEAKER_00 5:59

Until I started doing this podcast, I never quite realized how new really government VA IT sort of is. I blindly thought that if a veteran were to go into an office in, say, Alaska and then get transferred to say Mississippi and go into the VA, they would know all of his info. And the fact that all of that isn't just transferred, that's that's crazy to me a little bit.

SPEAKER_01 6:24

My assumption was that it could be looked at as like a community care referral and it was going to dentistry, and they would have some way of letting the VA know that this dental care happened. And it wasn't that simple. And we had to actually have a PDF come back and go into the system. So it's still validated that the like, you know, we still knew that the dental care happened, but it wasn't easily integrated, and we had to ask for that.

SPEAKER_00 6:48

That's crazy. So now you've held chief medical information officer roles in both public and private sectors. How do those environments differ when it comes to implementing health technology? And what can each sector learn from each other?

SPEAKER_01 7:02

Honestly, the challenges are all the same. Spending more time understanding the problem, finding the correct solution, not chasing after shiny objects, spending the time to truly kind of incorporate the innovation or tech into a workflow that works for the institution, training and planning for the change, and then building an accountability. I think being able to measure success, understanding the return on investment, holding the vendors and the leaders accountable, those are all big pieces that it's in federal and commercial. But in order to do all of that, you need to find the success metrics up front. So it comes back to what is the problem, what are you trying to solve? Will the technology really solve it? And do you have the right contracts in place? And then do you have the tools in place to vet that this is actually solving the problem? I do think that there's a speed that exists in commercial that the government could obviously benefit from, but that also requires a lot of changes to happen within the government acquisition cycle. So I'm very mindful of the fact that, you know, there are larger contracts, there's a lot more complexity involved in trying to do that. But that's why my approach has always been to start small and then test a concept, prove that it works, work out the kinks before you let a big contract. And that's what my job at the VA was try a little pilot to see if we can see what sticks or what can work. And if it doesn't work, then we can stop it, right? Before you put a large investment into it. And I think that being able to stop something is also as important as implementing it.

SPEAKER_00 8:40

And you also work at the intersection of healthcare, technology, and data. And we're seeing the explosive growth in AI, predictive analytics, precision medicine. What excites you most about these emerging technologies and what concerns you?

SPEAKER_01 8:55

Well, I think it's all super exciting. And of course, there's so much that needs to be figured out around policy, performance, patient impact, provider impact, security, um, data privacy, and the list just keeps going on and on, right? I think that one of the ways, though, to approach any of this innovation is to really tie it to foundational elements, right? Do you have an existing platform in place that it can tie into? Do you have an existing governance structure that it can fall into? It just can't be something new that just gets thrown at the institution. One of the things that I'm excited about, and I'm gonna give a little plug for your disruptive tech event, is um we're putting together a pilot with agentic AI and ServiceNow and our kind of full capabilities looking at like, you know, helping with room readiness and cardiac care, because it's a very complicated place where there's a lot of equipment. But really the premise of that is we want to use existing technology that's already been paid for and then build on the agentic capability of tools that have already exist, right? So you're adding in innovation, but you're not changing the entire infrastructure to navigate this new change.

SPEAKER_00 10:13

And you're gonna be demoing that at the Disruptive Tech Awards. This may or may not come out before the event. So if it comes out after, you guys missed out, but be sure to check it out. So interoperability remains one of healthcare's most persistent challenges. From your perspective, working across VA, private sector, and now federal health solutions, what's it really gonna take to achieve seamless data exchange that serves the patients?

SPEAKER_01 10:38

That's like the billion-dollar question. I honestly don't have a magic answer for that. I've been in this space now long enough to see how they try to do it through policy, they tried to bring in private sector, they've tried many ways.

SPEAKER_00 10:56

Right.

SPEAKER_01 10:56

I think the challenge is the incentives. Data is power, data is revenue, data has a lot tied to it. So until they really figure out what types of data should sit where and who should have access to it and what needs to be opened up for everyone. And I don't know who that they is, right? I think that's part of the problem, the governance model around this. I think it's gonna continue to be a challenge. But I think that as you start building in programs and look at the problem and figure out what the solution is versus coming at it from what the technology is, some of those challenges around interoperability start being alleviated just because the institution or the government is aligned in solving that problem. So then maybe it's not like a grander policy, but it's case by case, like maybe this is where we share the data for this because we're all aligned to get to this um outcome versus everyone must share everything, everything must be open, everyone should use the same data standards.

SPEAKER_00 12:00

That's just harder. Yeah. So you've spent over two decades in this space. How has the conversation around health IT and innovation changed from when you started to where you are today? And what's remained stubbornly the same and you really wish would change? Well, it's been pretty cyclical.

SPEAKER_01 12:18

Despite all the innovation, despite it all changing and evolving, it's really still interoperability, it's adoption, it's change management, it's training, it's business process engineering, it's privacy, all of which has been called something different along the way. And all of them have had different levels of importance over the last two decades, but they're all just as important as the tech. And it really comes down to, which is just mind-boggling to me because it's been around forever, is it is people, it's process, and then it's technology. Around that is also just making sure you have the foundational data to support that. So I think people looked at it individually. Like, I'm gonna look at the people part of it, I'm gonna look at the process part of it, I'm gonna look at the technology part of it. But you really need something to tie all of that together. They can't be separate initiatives, and they all require the same amount of focus. There was all these statistics. I remember when we were doing the military deployment, you know, all this work went into change management, but um, I think Gartner had put out where like I think 80% of IT implementations failed because they didn't do change management. But then when you looked at budgets, that was always the first area to be cut. Yeah. And I think we're still in that space of not making that a true pillar of any kind of technology implementation.

SPEAKER_00 13:40

So veterans' health needs, they're complex, often involving multiple systems, long-term care, mental health, unique services that are related conditions. How do you ensure technology solutions actually meet those nuanced needs of this population rather than just checking the boxes?

SPEAKER_01 13:57

From my time at the VA, I learned that obviously these health needs are there and they're also truly complex and it's not one size fits all. It's also different from the general population just because of the added elements of the past experiences that our veterans have had doing the work that they did and serving our country. And I also think that before choosing any innovation, it really has to come back to understanding the problem that you're trying to solve. I know it sounds like a little of a broken record, but it really is the problem. And this is not a problem that can be identified in a vacuum. You need to speak to the veterans, you need to talk to their caretakers, you need to talk to the providers, the veteran support groups, and then you need to get a full understanding of the technology in its current state, also where it's going to be in the near term, and then also where it's gonna be in the long term, and understand that full comprehensive set of capabilities that you are potentially purchasing. And then finally, I think just as important is the infrastructure, right? So is there a governance model? Is there an ability to deploy and manage and sustain and measure this implementation, right? In addition to this already complex environment that you're trying to potentially revolutionalize, but it's not that simple. It's not like a plug and play.

SPEAKER_00 15:20

And leading innovation in government healthcare requires navigating regulations, privacy concern, procurement processors, and stakeholder management. What's your approach in moving innovation forward when there's so many potential roadblocks?

SPEAKER_01 15:34

You know, I think this still comes back to the pilot concept, you know, testing big ideas in a simple way before scaling. And I think if we change the focus from selling a tool to solving a problem, I think a lot of those items move away from being potential roadblocks, right? It's getting everyone aligned, getting all the stakeholders aligned all around a problem that can be solved. Those pieces I think become a little bit easier to resolve.

SPEAKER_00 16:01

And you have the MD from Rutgers and the MPH from Columbia, clinical expertise plus public health perspective. How does that dual lens shape how you think about technology solutions for population health versus individual patient care?

SPEAKER_01 16:16

When I did that program, my focus is on hospital administration. So I learned a lot about how hospitals are run and managed. But then one of like the best aspects of the program at Columbia was the fact that we learned the history of public health and the policies and the programs that helped to kind of shape how programs are put into place in this country. And then I had the opportunity to kind of align that with like the business side of healthcare by taking some of the more healthcare financial classes at the business school and then put that together with a clinical understanding. I think what ends up happening is when I look at technology and the solutions that are out there, I try to look at it in a more holistic way. And I probably ask a lot more detailed questions around its value and utility while also trying to keep in mind that there are stakeholders, potentially providers or patients, or even like other individuals that we don't even know of, like caretakers at home, that are going to be impacted. So what are we who are we looking to serve and how is it gonna impact them?

SPEAKER_00 17:16

And federal health agencies are under enormous pressure to modernize while maintaining operations and serving beneficiaries. How do you help agencies balance that urgency of transformation with the reality of the limited resources and mission continuity?

SPEAKER_01 17:31

Honestly, I don't have a good answer for this question. I think that for my time in government, as well as serving as a contractor over a very long time, it just shows me how complicated it is to modernize while not breaking the current system. From my perspective, I think looking at problems again and solutions and more discrete packages helps minimize kind of that initial impact and cost. We're focused on helping to more like balance the skills in terms of like maintaining operations, serving beneficiaries, but then also adding in innovation where applicable. And I do think it's fair to set the expectation that government may not need to be at the forefront of innovation in all areas. And I think that we could probably benefit from incorporating innovation, tech, health IT, whatever it is that's been tried, tested in the commercial marketplace, so that there are best practices and lesson learned that we can incorporate. Whatever happens on the commercial space, it has to be customized for the government space. So it's never just a pickup here, it worked here, and it's gonna work here.

SPEAKER_00 18:41

And you've worked with some of the most vulnerable, underserved populations through VA and federal health programs. How do you ensure that technology solutions promote health equity rather than widening existing persparities?

SPEAKER_01 18:53

There was a constant challenge in the work that we were trying to do. I think it goes back to always identifying and really understanding what the veteran is experiencing, what is the problem that they are saying that they are facing, and then determining if the technology is the needed solution or can it be addressed through a policy, a workflow, or a data adjustment? I just think that for RGP federal, I you know, I've been thinking about this a lot because a lot of the work that I wanted to do has always been around the disparate populations. I was working on the homeless side at the VA also after I left CCPI, and it really came down to do we understand the data systems and are they aligned to technology? And then also do we have like this human-centered design aspect? So any program, any pilot, anything that we're going to do has to incorporate all those three. And then are we looking at existing systems that have shown not to break healthcare? And are we building around that so that we have more of a 360 view about what we're trying to accomplish? It's hard to know what the downstream impact is going to be, but with those populations, especially those that are in rural areas, innovation could be incredibly helpful. But again, it just has to be rolled out the right way to kind of alleviate some of these disparities that might come from having immense amounts of innovation in cities, but then not having it in areas that probably need it most.

SPEAKER_00 20:21

And looking at the federal health landscape, VA, HHS, DOD health systems, and beyond, where do you see the greatest opportunity for technology to fundamentally improve how we deliver care and serve beneficiaries?

SPEAKER_01 20:34

I mean, it has to be AI, right? Unfortunately. But the challenge is that healthcare is so complex and with so many interdependencies, I think it just has to be used judiciously. I like learning from other industries outside of health to see like what kind of intersection they have with health and get a different perspective around some of the innovations in the last 20 years. I'm just gonna call out something that we just did recently, which I thought was really helpful. At RGP, they surveyed, I think, 200 US CFOs from different industries to kind of better understand the gap between AI expectations and the operational readiness required to get measurable value. And most of the organizations had established some sort of AI governance, right, to guide responsible use, but they also found that fragmentation was abundant, the risk maturity, you know, varied across industries. And that in itself creates vulnerabilities, right, that can intensify as you start embedding AI because we really truly don't understand the full capabilities of something like this. And I thought the most interesting thing that one of the most interesting things in terms of relating to healthcare was that 20% of CFOs and tech and financial services say their AI governance frameworks are advanced, compare with just 4% of CPG and retail CFOs, and say. Of healthcare CFOs. So that's a big area, right? To like kind of think about. And so I think as different industries potentially have the ability to incorporate AI without the patient safety piece, without these big outcomes that could really, really hurt individuals, it's good to learn from that, right? And see what kind of models can we translate over into healthcare. And again, it's that same thing: like try it and test it and then see how it can be customized to work in healthcare or the federal government or whichever industry.

SPEAKER_00 22:30

For physician leaders or clinical professionals, considering a movement of health IT leadership like you've done, what advice would you give them about making that transition and what skills do they really need to develop?

SPEAKER_01 22:40

That's a really tough question. 20 years ago, it might have been like, I don't know, just go for it. But there's like so many different aspects of health IT, and I think many of those would benefit from physician leaders and clinical professionals just because of their insight. But I will say that my main advice would be if you want to work outside of the clinical setting, it's not enough that you only worked in the clinical setting. I think that if you're serious about considering a move into health IT, you have to expand your skill set. So you have to understand technology, you have to understand policy impact, financials to some extent, process, and just the general strategic operational vision that goes around these things. So you can definitely come at it from like how it would impact you as a physician, but the broader picture is really important.

SPEAKER_00 23:29

We're gonna do a rapid fire round and wrap it up with that. So it's gonna be the first thing that comes to your mind. Are you ready? Okay, yep. Okay. So one technology would transform healthcare in the next five years.

SPEAKER_01 23:41

AI, but I'm gonna say AI to reduce administrative burden.

SPEAKER_00 23:45

What's the most important quality in a health IT leader? The ability to problem solve. Coffee or tea to fuel those strategic planning sessions.

SPEAKER_01 23:52

I'm unfortunately gonna say water because I have one cup of coffee a day and I'm very particular about it. And after that, it's just water.

SPEAKER_00 23:59

Well, that's very healthy of you. So that's that's gonna be the best piece of advice from your medical training that applies to technology leadership. Always keep the patient or the end user in mind with whatever you do. And if you could fix one thing in federal healthcare technology tomorrow, what would it be?

SPEAKER_01 24:15

Interoperability of data, but meaning like health, benefits, state programs, exchanges, other agencies, all of it.

SPEAKER_00 24:22

Excellent answers. This has been a great conversation. Thank you so much for hopping on and speaking with us today. Oh, thank you so much for having me. I really enjoyed it.

SPEAKER_02 24:30

You've been listening to Forum Power. Stay connected by following Forum on LinkedIn or visit govforum.io. Thank you for listening. Be sure to join us again for a new episode and a deeper look at leadership and action.