Today on This Week Health.

The culture in your organization that you create around creative problem solving and encouraging the leaders in the organization to not be thinking like they thought even a year ago, let alone five years ago about how you're going to solve problems.

Allright today, we do a solution showcase from the HIMSS floor and we interview Andrew Gostine with Artisight, and Stephanie Lahr with Monument Health. And we're talking about the application of AI and cameras in the health system. And we specifically talk about platforms. It's really fascinating what Stephanie has been able to do with this Artisight platform. Not only around clinical automation, but also automation within the supply chain automation just across the board. This is what we see with true platforms. That you can apply this artificial site, this Artisight solution. And because of the way it's designed and architected that you can then start to dream up what additional ways you can apply this, this simple camera with pan, tilt and zoom anywhere within the health system with AI in the background, looking for specific things. And when it sees those things, it generates an alert, generates a message, or even puts things into the EHR. And Stephanie's incredibly creative. And I appreciate just the really fun and robust conversation we have here. So here's a solution showcase with Artisight. Hope you 📍 enjoy.

To borrow a phrase from Stephanie where we're really trying to eliminate the friction in healthcare on the clinical side. So what we provide our solutions to automate a lot of the frustrating parts of providing clinical care, whether it's in clinics, the patient rooms, the operating rooms, even now the loading docks, parking garages and sterile supply. So it's looking for the low hanging fruit to remove friction through automation.

And you are a clinican?

I am a clinician still practicing.

Yeah, it was, it was interesting last night because I heard you, the two of you present and it was interesting to hear you tell the story of you go into practice straight out of med school and you're all excited. And you're looking at this thing, you're like, wow, this could be better. There, there has to be a way to reduce this friction. And you took a process that was taking interns multiple hours every night, and you took it down to like 15 minutes every day. It's that kind of thinking that you're bringing to us.

There's really not a lot that is changed seemingly from when my dad was starting his residency to when I started mine. And so it was helping to see better ways of doing things without the friction so that we can actually get back to patient care.

Alright, I'm going to come back to the architecture. You're using cameras, artificial intelligence, processing this information but I found it best to just go straight to the source. You're using this technology. Stephanie, how are you using this technology?

So for us, it's been a journey that we're on. We go live here actually in just a couple of weeks with our first use case, but to play back a little bit to kind of what brought us in and how we decided. So this is because it's clinically facing having the clinical team early on, involved in helping identify the workflows was super important. So my chief nursing officer and I had an opportunity to go to the Sirius innovation center down in California and Andrew was presenting. And honestly, at the time I knew nothing about the solution had not really even heard anything yet.

And this one just in June, I think of this last year. So really not that long ago. And by the end of the presentation, when he had gone through the use cases and the perioperative setting in the patient hospital room, fall reduction stuff and then added on a few other things like parking lot management and whatever. We actually cleared him out of the room and there were six CIOs in the room and we said, we all want this. This could be transformational to the way we deliver care. These are the tools that we're going to have to start utilizing, to figure out. I mean, again, if you look at our staffing shortages that we're dealing with right now, no better time for a solution like this to come along because those friction points add people that add human capital to the work that we're doing, not to mention making it less enjoyable.

We started with three primary areas. One is hand hygiene monitoring, not super exciting, but a requirement for all of us that are in healthcare organizations, to be able to prove that we are doing decent work in hand hygiene, and maybe where we're not, we are working to make it better. That comes down to surveillance, capturing information, understanding it, and doing something about it.

We do it not surprisingly, manually today. We have people walking around. So many of the things that Artisight can do. And again, this is because of the approach of artificially being a person to an extent, right? You see a little, you hear a little, you can feel and or know where things are in space.

If we can take those senses and leverage the information that's coming off of it, where we have typically employed people to do it, now we can use the technology. So again, with handwriting. We can leverage the cameras to monitor and help us understand. Is there appropriate hand hygiene practices happening or not?

One of the other things I really appreciate is all of that's happening in an anonymous environment. So as we started talking about putting cameras all over the hospital for hand hygiene, the peri-operative and the patient room situations, which we can talk more about, obviously some of the first questions that come up are w what do you mean you're going to put cameras up everywhere? And how are you going to potentially use that against me? And so I love that that is not a part of the solution at all in hand hygiene. I don't know that it was you Bill or me washing hands. What I know is, was a person and they did or didn't comply. And then we can leverage that information that was against low hanging fruit. We were ready to buy a point solution and we said, Nope, we're not going to do that.

And that's the point I want to get to. So people are gonna hear this and go hand hygiene. Is that what Artisight does. Yeah. But it's, it, it starts there and you go, okay. That's, one of the things it's another set of eyes and with automation, really another set of hands that's putting stuff onto the keyboard, potentially in the EMR, potentially putting notes in and those kinds of things. That's, that's what it is. I'm going to come back to you for some more of the use cases. Give us an idea of the setup. What does it look like?

So the set up in full capacity. Right? We've made digital equivalence of many human senses, and we have sensors that go beyond human senses. So it starts with those sensors, that camera, the speaker, a microphone, RTLS technology, radar accelerometers. We take all of those feeds across the network into a data center or the cloud doesn't matter. We're completely hardware agnostic.

I was just going to, because you just rattled off a whole bunch of things. Is there special equipment?

So with the servers, we're really dependent on NVIDIA GPU's. We've standardized on to a lot of the NVIDIA software stack, deep stream, Triton, MBAE, to make sure that we can deliver that enterprise grade solution. But when it comes to the individual sensors, the cameras were very much agnostic.

So as long as it has those sensors that you were talking about, that will work within your platform.

Exactly.

That's, that's pretty interesting. All right. So we put these cameras in each one of these rooms. It's bringing that feed in what what's going on with that feed once it comes in.

So the feed will typically route across the network into a server and we'll deploy these servers and duplicate for redundancy so that we can deliver that high availability. The proof system, once it actually gets to the server, we stream that video straight into an Nvidia GPU where algorithms will start processing it. So we take what is typically just this digital stream of video. And we process it to extract events that clinicians would care about, like a patient getting out of bed or someone washing their hands as they enter a room so that we don't actually have to have someone watch this. And I think that is one of the things we've been able to overcome as a kind of camera based technology is that fear that this is big brother, because people are not watching our video feeds in most cases, we have algorithms that are doing it.

And nor nor is it being recorded. Right. And so I think those are things that people are thinking about are do you know, it's me. How do you know what to me, what are you maybe going to do with that? And are you archiving it somewhere to use against me or analyze later? And the answer to all of those things is no. And in fact in my own situation, my general counsel had lots of questions, right? I mean, this was, this is new territory for healthcare. And I said, yeah, I get it. Probably the best way to address that is let's get on an airplane and go to Northwestern and look at it and see it. And she was able to work through all of those concerns that she had so that she could get on board and being a supporter cause she wanted to support it. But those questions had to be answered.

Yeah. And you had shared with me a story of essentially a general council coming and saying, we need the video feed. When people see a camera, they assume it's being recorded. That's not the nature.

People see a camera and they think security camera system. They think surveillance, they think storage. We're not any of that. I literally think of a camera as a light sensor that feeds light and pixel data to a GPU. It's just another sensor. It is not the security system.

So traditionally we would train those things by taking millions of pictures. We put them in front of people who would say, yes, that's a unclean room. That's a clean room. That's a patient that's turned or patients getting out of bed and we would train it over a very long period of time with humans, but you're doing it a little different aren't you?

So by definition, if you take that approach, you have to save video, you have to save images. And that works if you're trying to make an elder them to identify CATS like Google might do. It doesn't work in healthcare. I can't put a thousand video cameras inside a hospital and save a bunch of images of patients and send it somewhere to have it annotated. What we do is very different. We have our systems understand and process the video feeds in real time to generate a general understanding of the hospital environment. And then clinicians through interactions with our software are also serving the double duty of training the system for the things that we care about. Hand-washing events, patients getting out of bed, patients turning to prevent pressure ulcers as examples. It's learning from board certified us clinicians in real time.

So it sends an alert. This person's getting out of bed that actually comes back and trains your system. That that's what it looks like for a person to get out of bed. Exactly. How does it learn though upfront? Sorry, I'm getting a little geeky. I'm gonna come back to the use cases.

The real breakthrough technologies kind of been in the area of self supervised learning. Okay.

So that's, that makes sense. So you're, you are processing those images, but the GPU is actually processing it and it's, the algorithm is looking for certain things. Correct. When it sees those things, it can then extrapolate what's actually going on in that image. It teaches itself.

Correct. So that the algorithms that self supervised learning approach, what it offers us is a, it's kind of a foundational understanding of what happens in the hospital. And then with just a little bit of guidance towards the events that we care about, they can get very good, very quickly at recognizing the things that clinicians are telling it are important.

Well, and that's why, I mean, one of the other super compelling things for us as a health system who has, we've got room for improvement in the OR we've got room for improvement in the hospital rooms, we've got room for improvement in the clinics was a platform for reducing that friction in all of those areas. And so that basic understanding that's already there is great, but it also meant that we weren't just buying hand hygiene and fall reduction and turn monitoring. What we were buying was an opportunity to sit at the table and say, here are our high friction problems. They may or may not be totally similar or dissimilar to another organization. How can we leverage the system to fix that? So there are some core things, some central things they had already done, most of which were deploying. But the reality is, is as I share with the rest of the executive team across the organization, what the system is capable of and how it works, they're coming to me now and saying, Hey, so could we use Artisight in the loading dock?

Could we use Artisight in a variety of different places because it's not a box of ready-made solutions that I decide, which ones I want. It's really a platform for me to be able to buy and say, great, I have the platform. Now, what problems do I want to solve?

As a CIO, you appreciate a platform because it saves you from a a hundred points solutions, which are going to really tax your organization from a complexity standpoint and a cost standpoint. Right. You're gonna repeat things over and over. And it's interesting. You rattle off three clinical solutions, then you rattle off a loading dock. Describe, I mean, how does that process start? Where it's like, you show it to people and they go, Hey, wait a minute. That thing's going to be able to recognize when something shows up on the loading dock.

And it's going to be able to recognize like when a bin of inventory has just gone empty it's I mean, because you can teach it to see things just the way we see things and go, that's bins empty. But it's a, it's a platform that's been put in place and now you can use it anywhere. You can really imagine. It's really kind of amazing.

It is amazing, which is why to be honest for us, it started at the highest level. My chief nursing officer and I, who saw it in California, went back and immediately took it to our senior executive team so that our most senior leaders could mull over what this meant that got our legal counsel able to start thinking about how we crossed them and those bridges, but it also meant them that our CFO, where the supply chain team lives and our market presidents, and they're much more familiar with their day to day problems.

It could go, oh yeah, this is something that we are going to need to support organizationally. And then as we started looking at the use cases for the ROI, then we went to the leaders over surgery and said, what do you think about this? And when the leaders over the clinic areas and said, what do you think about this?

And then word just started to spread by itself at work about a week or two from going live with hand hygiene. So without even being live yet with our first use case, the requests just keep coming in. But I, I think some of it too, is the culture in your organization that you create around creative problem solving and encouraging the leaders in the organization to not be thinking like they thought even a year ago, let alone five years ago about how you're going to solve problems.

And certainly the way we're not going to solve problems, which we are notorious for doing in healthcare is to bump up FTEs and hire people. It might be a silver lining in this crisis that we have right now around human capital. Because again, it was very easy in the past, just for the loading dock to say, well, I'm hiring three more people because of the way things are coming in. There's not three people to hire. So we're forced to be creative in our thought processes. And so that is the reason that then the leaders of supply chain are coming and saying, well, I can't hire a person, even if I want to, how do I replicate some of what the people are doing? And with technology so that I can let the humans that are going to be there, do the highest level work possible.

Yeah, it's interesting because normally I would go through this and I'd go, okay, how am I going to sell it to the clinician to sound? And you just answered that. How am I going to sell it to the CFO? You answered that. How am I going to sell, sell to administration? You answered that. The thing I want to talk to you about is I want to go to Northwestern. So you've practiced at Northwestern. Correct. How many times do you present this and people go and you go well, and here's what you can do with it. And they go, yeah here's what I'd like to do with it. And you go, we can do that. I mean, does that happen pretty often?

For the people that the platform approach really clicks that's how I know when it clicks. When they start asking me and doing the ideation for us about where they want to take the platform.

Is there a benefit for them to do the ideation?

I think so. I think we're starting from a lot of buy-in if they're the ones driving the clinical use cases, and that was what we saw very early on, even being a practicing physician. I don't see every specialty. I don't see it from the perspective of a nurse or a pharmacist or the people on the loading ducks. There's going to be so many problems. We can apply this technology to, we need to open source and crowdsource. So that ideas can come from anywhere inside the institution.

Where do you envision this will go next. Do you have any ideas?

Well, I mean, obviously there's a whole bunch of use cases across the health system, but I, what I'm starting to think about are ways to integrate this. I mean, again, given that Andrew doesn't care what technology it is, what hardware it is, and in general, that we can have a relationship of cross-pollination.

I'm thinking about things like, well I have another product that does ambient clinical intelligence in the room for documentation for providers. Is there a way for us to parlay things together so that I can leverage some of the algorithms and some of the documentation capabilities? And that's just one example, but I'm thinking about how do I take the platform to the next level and integrate it with some of the other things we're already doing? Make and the EHR is another great example. We're talking about use cases where we can teach the algorithm to potentially recognize certain kinds of clinical care that's being delivered. And instead of asking the nurse to do the work and then go to the computer and document the work, maybe we could teach the algorithm to recognize the work and then it becomes a matter of just confirming that's the work that you did and the documentation is complete, right? So I think it's taking it to that next level of how do we take what's happening in the platform and maybe integrate it with some of our other core systems and leverage them together.

As a founder this has to be pretty satisfying to have somebody who's gets the vision where this could potentially go. As people are listening to this and they want more information on this, where, where would they go?

So I would send them to our website, www.Artisight.com. And we'll take that. We'll bring in some of our partners like CDW and Sirius Healthcare to make sure that we can explain and then subsequently deliver that enterprise grade solution.

This is fantastic. Stephanie, thanks for your time.

What a great conversation with Stephanie Lahr and Andrew Gostine with Artisight. I love talking about this artificial site and artificial intelligence and what we're doing there. Self-directed learning and just amazing amount of opportunities in this space. And I love the creativity that health organizations like Stephanie's are applying to this, and really looking across the board and saying, how can we become more efficient as a health organization and provide better care and augment the clinicians who are in the field with this kind of technology, very exciting stuff. We appreciate them. We appreciate CDW healthcare for making their booth available for this conversation as well. If you're looking for some more conversations like this, this is the conference channel. We have another channel over there. It's called this week health newsroom and all the interviews we did at ViVE and at HIMSS are out there on that. We're getting great feedback. There's gonna be 40 plus of them. I'm gonna keep releasing them over the next couple of weeks. So go ahead and check that out if you haven't done so already. And that's it. That's a wrap. Thanks for listening. That's all for now.