← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Rethinking Hospital Operations with a Central Nervous System Approach
00:24:30
Loading summary
Transcript22 lines
- [00:00]
A
Hello and welcome to the Becker's Healthcare Podcast. I'm Molly Gamble. Thanks for joining me. Few questions are more pressing for hospital leaders right now than this one. It's not just how to manage through today's environment, but how to fundamentally rethink the way a hospital runs. My guest today has spent the last decade and a half on exactly that. Mohan Girodaridas is the founder and CEO of leantas, a healthcare AI and analytics company that works with many of the country's leading health systems on capacity, staffing and patient flow. Mohan, thank you so much for joining me today.
- [00:31]
B
Thanks, Molly. Always good to chat with you.
- [00:34]
A
It's great to be with you in this conversation. Today we're going to be digging into what a modern hospital operating model actually looks like and why Mohan believes the next leap forward isn't about automating existing workflows, but building something closer to a central nervous system, as he describes it, for hospital operations. We're going to get to that in a moment. But Mohan, maybe first we can establish the stakes here. Obviously, health care leaders are dealing with rising demand, staffing constraints, policy changes, more and more financial pressures all at once. From your vantage point working with 200 systems, what has fundamentally changed in how hospitals operate today? What feels different now even compared to five years ago?
- [01:14]
B
Molly, as I think about the last year, most of last year, there was a bit of a freeze with all the new regulations coming out of Washington, I think, you know, with the tariffs and with research grants and with Medicaid policy changes, et cetera, I think we've settled into a new normal. But the new normal is not an easy new normal. I think health system leaders have fully absorbed that. The demand continues to be high, labor continues to be constrained, and reimbursement pressure will continue. None of those feel temporary. And so that's the new normal that health systems need to operate against now. In the past, health systems could build their way out of problems by adding beds, adding ORs, hiring staff, putting in new hospitals. And what capacity does is it masks the underlying problem because as long as you've got plenty of capacity, a dashboard showing you how yesterday did was good enough to run the operations. But now that the slack has gone, what we are seeing is persistent gridlock. If you talk to any health system leader, they wake up every day to this pretty much the same set of messages. The ED boarding is high, the census is going to be high. It's taking three or four hours to get a bed, its discharges are getting delayed, and that's the Problem. Now, if you go back to what you said, what is different over the last five or ten years? The reality is clinical complexity today is massively higher than it was five years ago, 10 years ago, 15 years ago, just massively different. Meanwhile, the operational sophistication has not really changed that much. The same dashboards from the same EHR companies, plus spreadsheets and human middleware is how most health systems operate. That's why we're seeing what we're seeing, because health systems are running a very complex, capacity constrained systems using tools and methods designed for a different era.
- [03:15]
A
So it sounds like the clinical complexity line has just really skyrocketed in terms of being massively different and the operational sophistication has remained plateaued, that has not budged. As you described it, Mohan, given that shift over the last five years, what do you think a modern hospital operating model should actually look like?
- [03:38]
B
When I think about what is the starting point, the starting point is recognizing one fundamental truth. Healthcare is an asset intensive service delivery. Of all the kinds of service operations, asset intensive service businesses are the hardest to optimize. Why? Because you've got very expensive assets that typically are purpose built. They have a very specific purpose and they have very specially trained people to operate it. Think about an imaging machine or an or not. Anybody can come in off the street and learn how to operate those things. And so delivering any service requires that you line up the asset, line up the right staff with the right skills, in the right room, with the right patient and the right supplies. Otherwise the service can't happen if any one of those things falls apart. That's what makes asset intensive service operations particularly complex. That's why calendar gazing and appointment making simply doesn't do it. Now when I think about what else is an example of an asset intensive service business, it would be airlines, Delta, United, American or logistics companies, UPS, FedEx, et cetera. Now when you think about those industries, airlines or logistics, they've invested billions of dollars and several decades building sophisticated capacity, flow, yield optimization, asset turns, asset velocities, those sorts of things to their operations. Healthcare has not done those things. So when you look at a very simple metric operating margin divided by the value of the operating assets, healthcare is an order of magnitude lower than airlines or logistics. Now in fairness, planes can fly all night, healthcare can't do surgeries all night. So by definition it's a bit of a one arm tied behind your back problem. But healthcare is also focused on clinical excellence, which by the way is the right thing to focus on. No patient is going to say, you know, the surgery wasn't effective, but the operational experience was better than a Disney cruise. So I'm good. They're going to want the clinical outcomes to be correct first. And so that's the challenge we've got. If you go back to what you said, what does a modern operation look like? It requires continuous prediction. Not a dashboard about yesterday, but a prediction about the rest of today, tomorrow, this week, next week. It requires continuous optimization because you got to do the asset, the staff, the room, et cetera. It requires continuous adjustment. It cannot be done through manual interventions or frontline heroics. So what we lead into is what we call a central nervous system, meaning a system that can sense what's happening, understand it, and interpret what's happening, take actions, and then learn to get better over time. That's kind of what we are describing.
- [06:25]
A
I appreciate you defining what that is, because I think to your point, Mohan, you walked us through. Whether it's the cruises or the airlines or logistics companies, a lot of parallels or comparisons for health care only go so far. So it sounds like the central nervous system is really a more transferable way to think about how a health system in modern times operates.
- [06:47]
B
Yeah, that's. That's exactly right. So describing it as a system that senses, understands, acts and learns. So let's make a. Take a simple example. Imagine you touch a hot stove. What happens, Your fingertips sense the heat. You understand instantaneously that that was probably not a good thing to do. Before your brain even processes it, your spinal cord issues the reflex motion to pull your hand away. And then a second or two later, your brain processes it and says, that really wasn't great. I'm going to learn never to touch a hot stove again. So if you think about what happened, you sensed, you understood, you acted, and you learned, all in the space of a second. So if you took each of those concepts and applied it to healthcare. Let's start at sense. Sense is gathering input. The fingertips sense that it was hot. Hospitals generate thousands and thousands of sensory signals every day. Right? But they throw away 99% of it. It's written on post, it notes, it's device details that nobody recorded. It's stuff on a spreadsheet. It's stuff on an EHR dashboard that no one looked at. So all these signals are fragmented, delayed, not captured or thrown away. That is kind of step one. It's like going through the world with your eyes closed and your ears shut because you're shutting off sensory signals that you should have Captured. So what do we need to do? We need to capture those signals, filter out the noise. We can't overreact to everything that happens. Your fingertips would not have freaked out if the stove hadn't been on and it had been cold. So you've got to filter out the noise and then you've got to understand it. So understanding has operational understanding, meaning all the queuing, theory flows, bottlenecks, et cetera. So it's mathematical sophistication on operations and then clinical sophistication on acuities, what's happening, what tests, et cetera. And you need both of those. And then you've got to act, take the necessary steps, whether they're operational steps or clinical steps, and then learn so you get better. So what does this look like in practice? A patient presents at the ed. Imagine, based on that initial encounter, you could have predicted this patient's gonna need a chest scan and is gonna need a bed in the pulmonary unit. That's the sensed it, understood it. Now you start acting, open a slot on the CT on the imaging, send messages to the right unit, get stuff started. By the time the patient flows through it, that's happened. Now, if we guessed wrong and it wasn't a ct, it needed to be something else and it wasn't a pulmonary bed, it was a bed somewhere else. Then over time, the algorithm learns and the next time it predicts better and learns better. So this is how a system would adjust and absorb and learn data. And that's the notion of how you take a central nervous system concept and make it into an operational reality.
- [09:46]
A
Let's add AI now into the fold. You mentioned health systems asset intensive services, sometimes most often the hardest to optimize. As you had said, Mohan, I'm curious now about how this is playing out with the automation options available at Health Systems through technology, you're hearing so much about AI, but in your expertise, what really distinguishes a system that is functioning like the central nervous system. So it's sensing, adapting, learning versus a system that's just automating pieces of existing workflows.
- [10:22]
B
So a couple of things. I think what makes healthcare complicated is that the last mile is incredibly hard. If the Scheduler doesn't put Dr. Smith with the patient John Doe in OR 2 at 9:30am, with the right team of nurses, anesthesia techs, the robot and the implant vendor delivering the artificial knee, that surgery won't happen. You can't vibe code your way to that. And expecting AI to solve thousands and thousands of these last mile problems that happen everywhere. So that's not going to happen if you just blindly automate, you just make a bad process faster. That's all that automation accomplishes. So part of what the central nervous system needs to do is make the workflows more intelligent. Automate, where you can offer assistance by surfacing things. For instance, nurses are expected to have read a 20 page medical record before a patient walks in and they had 30 seconds of time before they picked up the patient. There's no chance they're reading all of that. So intelligent assistants imagine if they could read all of that and surface the two or three things you need to do. So there is the workflow, there is the assistive support, there is agentic support, there are some tasks that are low risk that can happen automatically. At the same time, you need the human in the loop to make the last decision. And so all of this forms a very tight control system that can be governed with intelligent AI to do the last mile. Let me give you another example that can bring this to life. Way back when, before 9 11, I was on a Lufthansa flight, Singapore to Frankfurt, 14 hours, middle of the night. And I got upgraded, so I was happy with that. Turns out this pilot liked to welcome passengers from the upper deck into the cockpit one at a time. And so the moment I saw that happening, I put my hand up and told the flight attendant I'd love to be invited back whenever the pilot is up for it. And sure enough, an hour or two later they pulled me in. And I wasn't sure what to expect. I'd never sat in a cockpit of a 747 in the middle of the night. And it was fascinating. Still completely dark. The pilots were bored out of their minds. That was the first reaction I had. They were just so happy to have someone to talk to. So both of them are turned around, looking at me and chatting. And I was fascinated, looking at all the dials and looking out the windows, etc. And then I finally said, how come you guys don't fly the plane? Wouldn't it be a lot of fun? And what he said has stuck with me. The pilot said, sure I could, but the autopilot can fly it better than I can. So I said, really? Why is that? He said, it's taking a thousand signals every minute. Wind speed, ailerons, control surfaces, flaps, tail position, etc. And it's making many, many adjustments. Because the objective is to keep the cabin stable for the passengers and not let them get bounced around and to keep us on track for where we are headed, and that's hard to do manually. So what he is implicitly, what he was implicitly telling me is that from the days of a Cessna, when the pilot could feel the plane on the controls and manage it to go to a 747 with a level of complexity and 400 passengers on board, the number of signals exceeded his ability to react correctly. We have the same problem in healthcare. The number of signals has exceeded the ability of anyone to truly get it in their head. The ways you could arrange a day in an MRI machine is a number with more than 100 zeros. Yet we're expecting people to decide in what order patients should be seen. So that's why we've got a unique opportunity here.
- [14:15]
A
Very much so. What year was that again? That happened?
- [14:17]
B
Oh, it was 2000. Before 9. 11. So a long time ago. Nowadays there's no chance on a commercial plane you're getting let into a plane.
- [14:26]
A
How relevant, how relevant 25 years later to still be thinking about that in terms of healthcare and what systems are trying to do and build.
- [14:34]
B
Honestly, Molly, it was a bucket list thing to sit in a cockpit of a 747 in the middle of the night. I still think about that very much so.
- [14:42]
A
I'm going to add another caveat if you'll let me, Mohan, which is that everything you've described thus far, the adapting, the learning, the sensing, making sense of all of these signals that are too overwhelming for any one person, this also needs to be connected in the health system. You can't have the ED and the or the inpatient units, imaging, all the other parts, going about this with operational intelligence, if there's still silos. So each of those individual units can't go alone in this. Where do you see the limits of trying to do that with the EHR alone when we're talking about connecting operational intelligence?
- [15:22]
B
That's an excellent question, Molly, because here's the thing. You can think of the hospital as a living system where each department is like a vital organ, right? So you've got imaging, you've got the ORs, you've got EDs, et cetera. So you've got vital organs. And they're all interconnected. And they're interconnected because there's so many support services. Imaging touches everything, patient flow touches everything. Evs, physiotherapy, lots of things touch everything. So if you thought of it as organs that are connected to the circulatory system, then it starts to get make sense that they're all interconnected. So what happens today? Health system leaders in each unit there are two things that are just happening continuously. One, they're making decisions based on data that was either yesterday or right this moment. There is very little decision making with precise, forward oriented predictions. Okay, that's, that's kind of thing. One thing. Two, many decisions are being made based. Made based on a purview of the immediate environment, meaning their local optimization. Not saying they're making bad decisions. They make the best decisions for their field of vision and that sometimes has adverse consequences downstream. So the imaging registrar who decides to do patient A and then patient B, because the two of them had the same setup on the machine. Yes, it's more efficient. Well, but if you'd done patient C before B, yes, you'd have had two setups and you'd have wasted five minutes. But patient C would have got discharged two hours earlier had you done that. That. But they don't have the visibility to do that. So it's not like they're doing something wrong, they just don't have the visibility to do that. EVS will optimize the rooms to clean based on their crews traveling the least. So they will do room 1, room 2, room 3 instead of room 1, room 10, room 3. Well, if 10 needed it faster, they don't have the macro view, so they're optimizing for what they've got. Everywhere you go in health care there are those two problems, which is they're working with current data or old data, and they're working with a local model, not a global model. That's why a nervous system that connects it, so you know when you sense danger to one part of your body, your legs get into motion to run from it. So. So you've got to be able to connect everything. Now the EHR is a fantastic system of record, but it was never built to be the system of intelligence because firstly, it doesn't have 80% of the data you need to do all of this. You need stuff from workflows, management systems, you need data from devices, you need environmental data, you just need all kinds of other data. So you can't expect the EHR to solve it any more than you could have expected an orthopedic surgeon to do cardiac surgery, brilliant surgeon, no doubt, but it's a different thing. And so when we think about this scale of technology transformations, it'll take 1,000 companies to do it. So when you think about the Internet was not delivered by IBM, AI will not be delivered by Google or Microsoft alone. It's a mosaic of 1,000 companies and 900 of them have probably not been born as yet.
- [18:36]
A
We've been talking so far about the operating model being rewired by operational intelligence and how that just will fire differently. But beyond the model itself, let's talk about some of the effects and the results that systems are trying to achieve through this work. Can you talk about Mohan? If a health system begins to operate more how we're describing connected, it's learning what changes should leaders really expect to see? Whether that's in patient flow, staff experience, financial performance, what are some results? Most likely here the benefit and the
- [19:14]
B
beauty of operational excellence is it's very tangible and it's quite quick. Meaning you don't have to wait years to see the results. Right? You can start to see it in weeks and months. You'll see improvements in flow and flow will manifest as shorter wait times, shorter cycle times between discharge order and actual discharge, bed placement request and actual bed placement. You'll see fewer delays at every step, fewer bottlenecks emerging and a better utilization of existing capacity. That'll also improve access because the moment you get higher velocity and higher movement, it enables you to see more patients. The shortest cycle time to unlock new capacity is to move an existing patient onwards because that frees up the resource that they were just using. And as a result, financial performance will improve because you're getting more value out of the assets you already have. Staff spending less time on manual stuff and chasing and more time with patients will make them happier, will have less burnout. Nobody went to nursing school or medical school to become a data entry operator into an EHR or to become human middleware to patch things from here to there and cut and paste from one spreadsheet to another. Letting them do what inspired them to seek this career in the first place is very helpful. Leaders will spend less time firefighting on day to day issues and more time on strategic things that they need to. And over time, what happens is the system will become more predictable, more coordinated, more resilient. Think about this. You have probably taken 2,000 flights in your life, maybe more. Many of them were connecting flights. I bet you've missed less than 5 connections in your life, right? And think of the uncertainty. It's a plane, it's a complex piece of equipment flying six miles high at 600 miles an hour. There's weather, there's crews, there's delays, there's, you know, all kinds of things. But they managed to hit 90, 95% reliability on making you make the connection and making your bag make the connection. Because the number of times you lost your bag is probably once or twice in your lifetime, not, not 20 times. Right. And so that is also very helpful for health systems.
- [21:32]
A
The predictive and resilient piece and comparing it to flying is such a great point and that that is what's possible as these systems continue to learn from their own operations over time. Mohan, you mentioned improved access, the financial performance will improve staff, less time, chart diving, chasing down things, more time with patients. If you step back holistically. Any other reason that this matters most for hospital operations to modernize and actually get things right at this level for patients and access to care specifically?
- [22:07]
B
I might be a bit biased, but I actually think it's one of the highest value things that a health system leader can focus on. Because today's operating model shows up as inefficiencies in delays. Whether it's waiting for beds, waiting for a scan, waiting for treatment. And the consequences of the delays are actually hard to quantify. If you don't squeeze a patient in for an imaging scan and the earliest you offer them is four weeks out and you lose the scan. It's not just the scan, you lost all the treatment that followed because where you went the scan probably goes the treatment. So the consequences are quite high. So improving access allows more patients to be treated, which is on point and on mission. When operations improve, patients move through the system as we talked about, more predictably, with less frustration. The staff are less frustrated. So we've talked to a bunch of CFOs. We just published a report recently on what CFO said and 70, 72% of them, some number of that order, thought that operating margins are in the 2 to 3% range, which is really, really tight margins for running such a complex, high risk, high value Service for Humanity. 2 and 3% is a very tight margin to operate on. But what that does is it points to why this is important. Say a health system wanted to improve its operating margin by $20 million. If it tried to do it through just regular growth, it would need to find a billion dollars of net patient revenue for the 2% of a billion to translate into 20 million, right? That's a billion of new net patient revenue is hard. 20 million in operating improvements by getting sophisticated on this stuff is a lot easier, is a lot quicker and the ability to deliver it within six to nine months, we've shown it 100 times all over the place.
- [24:01]
A
Mohan, I want to thank you so much for this conversation today. It's a lot to think about in terms of where health system operations have been, where you see them going, and ultimately why it matters so much. So I want to thank you so much for your expertise and thank you to our listeners for joining us. For more conversations like this one, please visit Beckershospitalreview.com or subscribe wherever you get your podcasts. Mohan thank you again.
- [24:27]
B
Molly thank you.