← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Sudipto Srivastava, Chief Data and Analytics Officer at Montefiore Health System
00:26:21
Loading summary
Transcript20 lines
- [00:01]
A
Welcome everyone to the Beckers Healthcare Podcast. I'm Laura Dearda, Vice President Editor in Chief here at Becker's Healthcare and I'm thrilled today to be joined by Sudipto Srivastava, the Chief Data and Analytics Officer at Montefiore Health System. Sudipto, it's a pleasure to have you on the podcast today.
- [00:18]
B
Thanks, it's great to be here. Big fan of Becker. So excited for this opportunity, Laura.
- [00:24]
A
Absolutely. And we're big fans of yours. And everything that you're doing at Montefiore Health System as well I think is such an interesting time in healthcare and we certainly excited for our conversation.
- [00:34]
B
Indeed. Yeah, it is very exciting.
- [00:37]
A
Absolutely. Well, I know we talk about a good many things in thinking about technology and data and AI and more. But before we do, can you introduce yourself and tell us a little bit more about the health system?
- [00:48]
B
Absolutely. And actually even before I do that, you know, because this is an audience that's near and dear to me, my other colleagues in healthcare and tech who are trying to solve big problems. I want to maybe start with a couple of hypothesis that as we're going through our discussion today, Laura, I just want the audience to keep in the back of their minds. Number one, do you think that we have solved all of our problems in health care? And number two, given all the focus on AI and things like that. But how would you change work if you had access to Several really smart PhDs who are working for you? Of course, albeit with a little contact, with little context for operational and cultural context that you work in, like these AI tools can do that. But just how would you change? And I want the audience to keep that as a frame in the back of their minds because we'll be talking a lot about it and I think as we have the conversation, it might prompt some thoughts and ideas in their own heads, which is of course the goal of this discussion anyway. So with that said, just to let me tell a little bit about our organization and myself, Montefiore Health System, Montefiore Einstein is a 13 hospital health system. Covers the New York City, Bronx, Westchester, Hudson valley area. About 7.5 million encounters in a year. About 36, 38, 7,000 employees. We also are an academic medical center in the Albert Einstein School of Medicine. And the interesting fact that some of your audience may or may not know is that we have a $0 Maryland student tuition because this was made possible by Dr. Ruth Gotsman, who announced in February 2024 that all current and future medical students will have their tuition waived in perpetuity through a billion dollar donation. So that's something that is fascinating and we love sharing that fact. As for me, my title is BP Data Analytics. My role includes oversight for data analytics, organizational reporting, AI governance and also research it partnering with our School of Medicine. And then as we sort of think of roles, titles and what we like to pivot on within our organization is that we have made a conscious decision not to create a separate, say czar for AI or anything else. And again, there's nothing wrong with organizations that are doing it, but our philosophy is that AI and tech is everyone's responsibility and we find that this frees up our staff and docs to explore things from their unique vantage point. Now of course we have guardrails, but we want to be deliberate in not taking and making AI an individual's response.
- [03:50]
A
That makes a ton of sense and thank you so much for that explanation. It's so fascinating to see the different perspectives that health systems are taking as technology and AI move so quickly. I love the idea that it's everyone's responsibility that you want everyone from their seat to take that unique vantage point and really figure out how they can apply it within their own roles and workflows and get the best results possible. That's really cool to hear about. When you think about the last year or so, what was the most important initiative that you led? What did you do and what were the results?
- [04:25]
B
Yeah, well, you know, plenty with like 37,000 employees thinking and working every day. Smart bunch of clinicians and staff and nurses that we have. So lots I guess, you know, in terms of this conversation, I'll pick a few to talk about. Of course, as with every self respecting health system, you know, we launched ambient voice AI solutions. So we have over 550 clinicians enrolled in that covering areas such as primary care, medical and surgical specialties. And we have an ongoing demand of additional departments that want to use Ambient Scribes. We have a very successful implementation there. We have over 48% high utilizers. These are folks who are using scribes more than 50% of the times in their sort of visit and over 50% of their visits. So we've had about like over 300,000 encounters and you know, this thing sort of continues to grow. And you know, back to the question that we sort of prompted the audience to think of is, you know, think of, you know, have you solved all problems in health? And I think Ambient Scribe becomes a very interesting one because this was a problem that we like over the last few decades. We kind of created for our doctors and nurses, where documentation burden sort of kept growing. And here we have a technology solution that allows them to solve that problem, that classical pajama time that they were sort of all giving. And you've done many podcasts with Becker's and articles around that, so I want to bore the audience with it. But here we have this huge problem that now that we can try to solve with tech. Continuing with our other initiatives, we also launched our AI Governance framework and our AI policy for the entire health system to guide how things go. We've had our evolution in that last year when we started, we launched what we call Governance 1.0. We learned from our lessons. The starting of this year, we launched Governance 2.0 and we're already thinking about Governance 3.0. If your audience is saying, well, what's the difference between each of these? I think as we went through our phases, we started thinking of AI governance in terms of quick decision making because we were facing a lot of volume of requests that are coming in. And then as we were able to stabilize that, we realized that we needed significant Input in Governance 2.0 from all walks of our ecosystem, of course, the clinical teams, the nursing teams, as well as people in finance and legal and privacy and cyber and so on. So that is the culmination of our Governance 2.0. And then we're seeing how even this process can be improved given the new volume of requests that we're getting on AI. And we're starting to think of governance sort of 3.0. So it's a very deliberate, iterative process and we're really sort of proud of what we've done. We've looked at over like 80, like we've launched over 80 subsolutions. We have several, close to 100 more in the pipeline and covers the entire gamut of like radiology and acute care, cardiology, revenue cycle, patient access, GI it, security, nursing, just to sort of name a few. Now, with that sort of said, we also feel that the most important initiatives are also the ones where we have deep partnerships with our business in this area. We've done a lot of work with our population health teams within our care management organization when it comes to having data on care gap closures, having data on how they're doing with their hedis measures. On the clinical side, we've started auto tagging off cancer types using genai tools like the bioclinical birch capabilities. On the research side, we actually entered last year into a partnership with Dandelion Health and there's details written up about it. But essentially we will partner. We are part of a consortium of four other health systems around the country which are offering de identified data to test out AI tools, AI algorithms, advanced research, all in a safety identified sort of framework. So we're very proud of that sort of partnership. And that's just to name a few. You know, I'm of course not talking about, you know, things that we did with our pharmacy. We have a strong one Amazon, one medical partnership and so on. But I'll shut up for now and hopefully that gives you a little bit of a perspective of what we've been doing.
- [09:17]
A
Absolutely. That's incredible. I mean, you know, to have so many different AI solutions in the pipeline, close to 100, is really, really a huge accomplishment right now. And I appreciate you talking, you know, whether the governance process, that structure, I know a lot of health systems and leaders that I talk to, that's a sticking point for them trying to figure out how they're incorporating AI smartly into that governance process to make sure they're not missing anything, but also not reinventing the wheel and keeping the aspects of governance in place that have worked across the board. So it seems like, you know, being able to move quickly is important, but obviously you want to mitigate the risks and make sure you're not moving too quickly. And finding that right balance is critical, it seems like.
- [10:01]
B
Indeed.
- [10:03]
A
Absolutely. Well, I'm curious, you know, kind of building upon all of these things that you've been talking about, these different AI applications and really incorporating that into the daily workflows for your team members, what do you see as being some of the big priorities and headwinds that you're focused on for the rest of this year?
- [10:21]
B
Oh yeah, Well, I think from a. When it comes to priorities to the point that I made earlier, we really want to align with our business and clinical teams in exploring the problems that they face day to day and how it can be solved through tech, through AI and other tools. In that there are certain areas that have popped up big in that space. Many of our colleagues here who are listening in have been thinking about rev cycle and we see a huge opportunity in revenue cycle, whether it's in terms of denials, management, prior authorization, other functional areas where we have a huge opportunity to work better, to have better throughput, get better outputs as we partner with our health payers as well. Supply chain is another huge area. If you think about most health systems, supply chain is a significant, significant, probably number two or number three when it comes to the amount of cost and revenue that they have to handle and they look at contracts, there's a lot of information that they have to extract. They have so many different vendors looking across the different vendor landscape, I kind of always say it's they're buying pencils to MRI machines. So using AI in a thoughtful manner there is huge. HR is another area from a business perspective because we get a lot of internal questions around basic things like what's the vacation policy, how it's happening, what are the rules around X, Y and Z that we get a lot of staffing requests on. On the clinical side, there's a ton that we're thinking of in diabetes management outreach and population healthcare gap closures in our surgical areas, we do so much in our periops space with that. All those are active exploration and discovery and sometimes even some deeper that's happening. Finally, from a tech perspective, from a priorities point of view, you know, we're looking at conversational AI because, you know, that allows our nurses and sort of front desk staff to, you know, focus on the higher acuity, sort of, you know, patients. And so we're looking into sort of implementing tools in that space. And we're also, you know, starting to look deeper into the partnerships that we already have with our sort of vendors. I mean, you know, we, of course, it's the. We use EPIC as our emr and EPIC has been pretty amazing in terms of exposing tools in the space of AI and productivity. And we're actively looking at those. We have about 10 or 12 in our pipeline that are going. Many of them are live in use right now, many are about to go live, and we have a longer roadmap of how we switch those capabilities on. It doesn't just stop with our EMR systems. We're looking at our operational Systems. We use ServiceNow. What can we use there to increase efficiency and respond better to our customer requests? Workday, we went through that implementation and Workday itself has tools. As you look at our application landscape, every vendor is trying to build tools in for added efficiency, cost savings and so on. Those would be a little bit of a smattering of the priorities that we're looking at. I think you asked about headwinds as well, so. Well, there are plenty. And your audience is no stranger to those. I like to say that healthcare likes to go through. Not likes to, it just goes through feast and famine cycles when it comes to funding and so on. So as we look at 2026 and 2027, the macro environments will kind of force us to focus a lot on the cost savings. The other part will be filtering signal from the noise. Healthcare is embracing, unlike many or similar to other industry, but unique to healthcare now is we are embracing AI at a faster pace. The interest is there, both the leadership level, the clinical level, the nursing level to do that. And then how do you filter out the signal from the noise? What is worth investing in, what is not worth investing in it? The other headwinds are around protecting our data and our information. You know, because of all the big bold ideas that we have, we have to be mindful that, you know, we are, we are guard, you know, we have to protect our patients set of data. You know, we there, there's always a potential for breaches. So we partner very closely with our, you know, cybersecurity teams. You know, especially with AI, there's new aspects of data come that come in, which is what is the vendor doing with their data, how are they training this partner with the large LLMs? What is going on in terms of the safety of our data, what kind of agreements we need to have there? Those are going to be. They've always been challenges for us, but I think they will continue to be challenges. Finally, just the other headwind is we have to go be steady in the storm. I think there's so much news media articles that's coming out, all the things about the glorious things that AI can do and all the mistakes that it can do and all the doom that it can cause or all the amazing things it can do. So staying true to our mission and aligning with the needs of our staff, doctors and nurses, that will be a big sort of headwind that we have to manage as well. Anyway, I said a lot, but that's what we're sort of tackling right now.
- [16:25]
A
Yeah, absolutely. I mean, you know, it's a huge space and certainly a lot of challenges out there across the board. And so, you know, when you're looking at the opportunities out there, I think so many possible partnerships you could have new companies coming up every day as well as, like you mentioned, you know, existing partners coming out with new AI applications in ways to ideally increase efficiencies and become more critical within your daily workflows. So I'm curious, when you're evaluating all the opportunities out there and trying to bring your teams along and especially in health care, you know, there can be some resistance to change in certain ways. How do you kind of bridge that gap and really make sure you're one, selecting the right things and then two, putting Your team in the position to actually use it. So, you know, these new applications that you're bringing in or working through aren't left dormant, you know, a year from now.
- [17:26]
B
Yeah, no, I think, you know, and this is where all 37, 38,000 employees, you know, have a role to play. And, you know, there are many aspects that go in, you know, one, we don't want to force AI down any sort of particular area. There has to be a conversation and understanding of what it does, what it doesn't do. There's a lot of training and coaching that goes in there in terms of what is the value of that. Talking and being actually honest about some of the skepticism that people have. And that is a journey. And sometimes you take two steps forward and one step backward when it comes to people in teams who may be genuinely, you know, nervous and anxious about the quality of the tools, the stability of these tools, the accuracy of these tools. So, you know, so we have to definitely sort of do that. We also, you know, have certain measurements. So we haven't talked a lot about the control towers and the control centers that we have to overall monitor the use of AI. And not just monitoring from a performance drift, other perspective, but also from a, hey, did it meet the hypothesis that you had going into it? Because if you think about it, many large organizations like ours, and especially in healthcare, I like to say that we're really good at addition but not good at subtraction. Let me tell you what I mean by that. What I mean is we're really good at adding more and more solutions and tech to it, but we're not as deliberate about subtracting it and taking them away. So we've been very conscious in terms of our approach to tech and AI is as we work through our governance process, we will go back to solutions and say you had, you know, when you launched the solution and when AI governance approved it, the hypothesis was it was going to do X, you know, X number of patient encounters and improve in productivity or maybe some dollar threshold or numbers or volume thresholds or something, something, something. Is it doing that? And let's be very honest to measure if it's doing that and if it isn't, let's take it out. Let's take it out of our ecosystem and do that. And trust me, these are not easy conversations to have. But that is a framework that is needed so that two years, three years down the line, we're not taking on so much tech debt that we have to have a separate exercise to unwind all of that. Does that make sense?
- [20:02]
A
Yeah. That's so helpful to understand how you're thinking about these things and you know, really truly trying to avoid that technical debt, for sure. I'm curious, you know, in looking ahead, what do you think the hardest thing you'll have to do in the coming year will be?
- [20:16]
B
Yeah, I think. And again, this is something maybe your audience can appreciate as well, especially if it's in a large health system like ours, with an academic medical center as well. Is just picking the right problems to solve. I like to joke with my. Not joke, but at least have a conversation with my stakeholder team. And my leadership is we have 100 problems to solve, but the funding and resources to do 17 of them. And then how do we do that? And it is very hard because when someone comes in and says this is going to help me detect a particular disease three days earlier, five days earlier, or this will have a clinical impact on someone who is working with diabetes, or here's a cost saving or it is very hard to make those decisions to pick those 17 or 20 out of the hundred. But we have to because we don't have unlimited resources. So allocating our funds wisely in that area, investing in the right things where we see the right quote, unquote. I'm doing air quotes right now. Roi because ROI is defined very differently for every organization. Some look at it from a purely dollars perspective, some identify other factors. We of course have our framework for that, but it really will be how do we pick the right problems to solve? To the question that I asked your audience to think through is do we feel that we have solved all our problems? So as we look at our health systems, what is the biggest one is access. BIG is reaching out to a certain patient population. BIG is looking at the high acuity patients who may have a lung nodule that goes undiagnosed for a period of time to get a referral appointment back. So there's so many of them, but just picking the right problems, I think that has been always a challenge. But I think especially with the supercharged environment of tech that we're in right now, that I feel will be a pretty hard thing for this year and many years to come.
- [22:19]
A
That makes a lot of sense. And certainly, you know, having that strong process, as you mentioned earlier, can make a big difference in keeping things, you know, relatively focused in on what's most important for the health system. Tying back to the mission and vision of where you want to go. Before we wrap up here, I want to talk about Growth too. Where do you see some of the best opportunities for organizational growth in the future?
- [22:45]
B
Yeah, well, that's a fascinating question. If I had to pick maybe one, I would say I see a unique opportunity in redefining the way we work to those in the audience who are either rolling their eyes because every consultant walking to the door right now is using these big words, but I think there is a huge opportunity there. Maybe just let me share a story, a small one. I was speaking recently to someone who is in er, radiology and also teaches radiology residents what they've been observing. They said was that first year residents in radiology are catching up quite fast with their third year or even fourth year colleagues. And their hypothesis, again not proven, was that this is due to AI because a lot of first year residents are adept at AI and are using it different ways. And then the discussion evolved into what can be done to further expand the scope and scale of what is taught to residents using now perhaps AI as a tool. So now I'm not clinical, so I'm paraphrasing some of what they said. But can AI be used to expose radiology residents to other business fields that they interact with many earlier, much earlier, so that instead of looking at certain things from a very silo lens of radiology, they can look at it from a lens that a cardiologist would look at or a neurologist would look at. So now doctors in the audience may say, huh, well that's a cockamamie idea. But which is why I'm, you know, paraphrasing someone else, you know, who is way more intelligent than I am in the clinical sphere. But I give that as an example to say, now imagine your processes, your workflows, the things that you do, whether it's operations or clinical or nursing or call centers, supply chain or population health. Pick an area, look upstream and downstream. Are there ways that we can redefine things a little bit? I think we have a once in a long time opportunity to be able to do that because we now have the tech tools, the AI tools, the capability, the desire, the capital required for it, the funding needed for it, the leadership and the board level interest in that, that it'll be a missed opportunity if we don't do that from an organizational growth perspective. And this won't be a one year effort, this might be multi year efforts and so on, but it allows us to focus on that.
- [25:24]
A
Yeah, yeah, definitely. I think that makes a lot of sense and you know, it's really helpful to have that type of focus in and then understanding where you know you're going from here. So I appreciate it. Sudipto thank you so much for joining us on the podcast today. This has been a fun conversation. I'm really kind of focused and I appreciate your time and energy on it and look forward to continuing the conversation soon.
- [25:45]
B
This was fascinating. Laura, I love your podcast and I'm glad that we had an opportunity to discuss it.