Today in health, it,

Another one of our interviews and action.

This comes from the healthcare to healthcare event, which I was a guest

at from the serious health care team.

It was in Montana.

And I was able to sit down with a handful of CEOs.

And I'm going to share those with you here shortly.

My name is bill Russell.

I'm a former CIO for a 16 hospital system and creator of this week in health.

It.

A channel dedicated to keeping health it staff current and engaged.

I hope you're enjoying these interviews and action.

We were able to do these interviews at the health conference, the

chime conference, and now the healthcare to healthcare event.

I've really enjoyed doing them.

, just a reminder.

We're going to get back to our normal programming where I take

a new story, break it down.

And talk about why it matters to health.

It.

We're going to be doing that as soon as the interviews are done we have

done 10 from the chime conference eight from the health conference and

we have five from the healthcare to health care conference so i hope you

enjoy another one of these interviews

All right.

We're doing an interview from the healthcare to healthcare event.

It's a serious invite only event.

And I'm joined today by Dr.

Andrew Gustine, our CEO, CEO,

founder

CEO, founder of artist site.

And we're going to talk a clinical automation essentially

is what we're gonna do.

So tell us a little bit about artist's site

before we get going.

Doug artist's site is it's really an IOT sensor-based automation.

To bring automation into the clinical space.

You know, as a practicing physician, I was looking out at the different automation

tools that were available and I realized most of them were focused on back office

types of automation, but the biggest problems we have in healthcare often,

not in the back office or at the bedside.

And I saw a huge gap in the marketplace.

And so developed this platform as a solution to a lot of the problems

I was seeing at the bedside.

Th

the reason we do it on the administrative side and not the

clinical side is the quality of data.

And so I think it's fascinating that you're getting the data raw

data directly from a camera that's embedded in that clinical setting.

And then that obviously for our listeners is going to create a whole

bunch of questions, but talk about

the solution.

So you talk about the first part of your statement, you know,

the back office automation.

Just the first evolution of automation because of all the discreet data

that was available, uh, you know, the amount of charges we generated,

the, um, different insurance broker that was handling the payments.

All of that information was quantifiable and available.

Just brief data easily to work off of.

Okay.

All of the information I have available at the bedside to

me is coming from a clinician.

It's a clinician that is seeing.

Is talking to the patient is feeling the patients and then putting

that information into the EMR.

If you don't have the clinician that is there to capture the data,

you don't have any discrete data.

And so what we saw in the marketplace, uh, really at the

bedside was that there was a gap.

We didn't have access to the discrete data.

So we developed these IOT sensors to structured data for us, putting

a camera in the room and developing.

To look for events or pieces of information that we can capture

in real time and build automation around so that we could remove

the human component of data entry.

No,

we only remove the component, but actually assist them

because now you take that error away there, you know, they forgot

to put in information about whatever whatever's going on in that room.

That's missing information.

You can't really act on it, but the camera is there.

The camera's always there.

It's a lot cheaper than a clinician.

It doesn't need a bathroom break.

It can always be watching, looking for things that we tell it to, to see,

and kind of to the other side of that, won't look for things that I don't

want it to see, because I don't want to put a system in a hospital that

clinicians feel like his big brother.

And that was interesting.

So looking at some of the video clips that you were showing

yesterday, everything's blacked out.

The patient's face is blacked out.

Blacked out, but you're still picking up that data.

You're still processing that data.

So what are you doing around privacy?

Obviously that's one step, but you're doing some other

things around privacy as well.

So I think people have to change how they think about a camera.

When people think about an IP or security camera, they think of it

as the closed circuit TV system.

That's recording video.

That was a lot of the initial use cases.

When we talk about an IP camera, we really think of it as just a light switch.

That if you look at those pixels makes a picture, but we don't have to record it.

We have elder rhythms that just watch the video feeds in real time.

And we'll look for different combinations of pixels that represent

objects or events of interest.

They will structure some type of data from the event that it's witnessing, but

we can then just save the discrete data.

We don't have to save the video.

So we can just retain the parts of that video feed that are interested

to the hospital, but not any of the risks or any of the concern

that a clinician might have.

If we don't train the camera to see something, it is totally blind to it.

Yeah.

And it was interesting.

One of your users was here and they said their legal team came back and said,

Hey, can we get access to the video?

And the answer was, it doesn't exist.

It goes into the AI engine.

It gets.

It generates those insights, if you will.

And then it's,

it's, it's somewhat like time where we live

in the present and I don't really have a recording of the past.

It's the same thing with our video feeds.

We run inferencing on them in real time, but we don't save videos.

There are some exceptions for other use cases where surgeons want records of their

surgical videos for education purposes.

But for the vast majority of our use cases, we don't say that.

So let's talk about some of the use cases.

So people are probably thinking right now, if I had a camera in there,

obviously we're doing tele-health, uh, fall detection, hand-washing clean rooms.

I mean, it really is limitless.

Really?

What if it's really limited by people's creativity of what they can imagine?

Computer vision is very powerful, uh, in many

ways, a lot better than humans in some ways it's not as powerful.

That's the best way to think about computer vision is it's

a much more cost efficient way that can capture the same data.

As if I put an educated person in every part of the hospital, just

collecting data with their eyes.

Computer vision can see the same way that humans can see.

And if a human can witness something with their eyes and tell me about it,

I can most likely train a camera to do the same thing for a much cheaper.

It's really interesting.

So you showed up fall detection.

So I assumed, is there training that needs to go on at each

institution or do you have, are the algorithms already trained to detect?

So

it's again, similar to humans.

You know, if I teach you at one hospital, how to monitor what it looks like when

the janitors are cleaning a room and then I take you to another hospital and

you witnessed janitors cleaning a room, you're going to be pretty accurate.

They might use.

They might have different scrub colors on, but you're going to be able to

tell me with reasonable certainty that you're pretty sure a janitors cleaning

an operating room or a room or whatever, same thing with computer vision.

If I train it at one institution and take it to a second out of the box,

it's going to be pretty advanced, but we will do some retraining over a week

or two to make sure that it lends the peculiarities of that new institution.

What's interesting.

So I want to talk to you about deployment, but.

The, you talked a little bit about the highest, um, use case, the highest best

use for the first couple of, of, of things that you're gonna use it in your

health system, because the platform can really return almost a complete

return on investment with one or two use cases, but you still have the platform.

And I find that interesting in of itself.

What have you found to be some of the, some of the things

that people have deployed?

The initial.

So

there's a, there's a lot of business strategy

that goes around technology.

And there's a lot of very interesting use cases that, you know, as a physician

really interests me, but healthcare is still a business and we have to make

sure that we're justifying the things that we're doing in the hospital.

So we build business, use cases around some of what I would

consider the first best use cases where hospitals are collecting.

And we can show this is your baseline performance.

And after you put in our fall algorithms, this is by how much you've reduced false.

Well, the problem with healthcare is we don't collect data on most things.

So if I could build an algorithm that reduced the amount of infections, but

you have no idea what your baseline infection rate is, then I can't show you.

You made X.

You saved this many dollars and it will justify its existence.

So all of the first use cases that we recommend to hospitals like reducing

falls, reducing pressure ulcers, uh, bringing them up to compliance with new

regulations for hand-washing is around things that we know they're measuring

where we know they might be sustaining penalties, or they have a large expense

to maintain some level of compliance.

We go after those first to help them show to their board and decisions.

That this is the quantifiable value.

And then once the platform is in there returning that value from those first

use cases, you can now use it for things that might be more interesting,

might provide a greater impact to society, but are harder or more

nebulous to quantify in terms of the

value.

So I thought the other thing that was interesting is innovation partners really.

I mean, when I think about the academic medical centers could really use this.

In a lot of different ways that we probably couldn't even imagine ourselves.

And, uh, you talked about sort of incenting that, that environment

of innovators to, can you talk about that a little bit?

Yeah.

So it kind of a two part answer.

So I, I may be the world's best physician, most likely I'm not, but even the world's

best physician is not going to think of all of the use cases for this technical.

And the goal for this company is not to become the richest company

in the world it's to fix healthcare.

And the only way we're going to be able to solve as many problems as this platform

can solve is if we get everyone to help us, if we crowdsource the ideas and

the opportunities, and co-develop the solutions now for the institutions, most

often academic or very large medical.

That want to help with that and are very interested in helping us bring new

solutions on the platform to market.

They should totally be able to share and the financial gains from those solutions.

And so we have revenue sharing agreements.

We have cost reduction terms and all of the contracts where if they help us

develop a solution, we will start spitting back revenue from other clients that are

using the solutions they co-developed to chip away at the contracts.

These.

As kind of a way of saying you participated, you helped provide this

value to the healthcare ecosystem.

The overused phrase of app store comes to mind, essentially,

you're an app store for computer vision solutions in the hospital.

And I think we will ultimately evolve into something that

is even beyond the applications that we develop, where people who research.

Develop a very interesting algorithm to solve a problem, but have no

way of managing a camera network of 10,000 cameras securely behind the

health system, firewall with security compliance and firmware upgrades.

We can be that portal for them to deploy their algorithms on

the infrastructure we maintain.

The nursing shortage is written about a lot at this point.

And I think we're, we're looking at potentially half a million

in the next three years, right?

How does this help in that with regard to that?

So we're going after things that nurses have to do

now, but shouldn't have to do tomorrow.

So in some of the literature I've presented here, we see that nurses spend

about a third of their time documenting.

So eliminating some of the things that don't really require a nurse to do,

like documenting that he or she turned to patient, eliminating that from their

workflow is going to make them more.

It's also going to burn them out less because they didn't go

to nursing school so that they become a stenographer in the EMR.

They went to nursing school so they could be at the bedside

taking care of the patients.

So we're trying to bring the joy back to nursing, eliminate the

nonproductive non-patient care aspects of their workflow so that we can help

them see more patients, more timely and take better care of patients.

The

lastly, I want to talk to you about.

It's deployment because you're not talking massive, expensive care bros, and you guys

even have a mobile solution to move in.

It seems like you could actually ramp this up pretty quickly at a health system.

So we, we did that even in the middle of COVID

where there was a lot of challenges of getting into the patient rooms.

In six weeks.

So with the mobile systems, we can do that in a few minutes of bringing

them into a hospital and adding them to the secure wifi network.

When we talk about server infrastructure, for those that have

a hybrid or a cloud presence, we can turn on new virtual machines

for AI processing and a few hours.

So the rate limiting step is typically camera installation.

The rest of it is hours to minutes of the bringing new systems.

So

people are going to ask me about security, talked about privacy.

They're going to ask about security.

You're streaming this information.

I'm not sure that this information itself has value to someone who's going

to hack it and that kind of stuff.

But if it did what, what's the, uh, what's the information around security.

You know, just the, the first things we're going to do to

keep this secure is always deployed behind the health system firewall.

So the cameras for obvious reasons being in the patient rooms are by definition

behind the hospital firewall, but even the server infrastructure or the

cloud, we enter and deploy these on virtual machines in our client's cloud

tenant, so that we're not streaming this outside of their ecosystem.

So the first and second line of defenses are always the.

Security measures the VPNs, that directly route things up to the cloud.

If that's how they deploy us in terms of the data that we collect by not saving

any video or in cases where we do save video, but de-identify, or anonymize

the video, we're preventing that risk of any Phi leaving the health system with

the data that we generate and structure for them from those unstructured.

We send that into their enterprise data warehouses or their EMR is we're never

holding our own version of the data.

So we're trying to make sure that we're putting this data in a place that is

super safe and controlled by the client.

Andrew, thanks for your time.

We're really excited to solution.

I'm looking forward to seeing what you guys

do.

Yeah.

Thank

you.

Don't forget to check back as we have more of these interviews

coming to you, that's all for today.

If you know of someone that might benefit from our channel,

please forward them a note.

They can subscribe on our website this week, health.com or wherever you listen

to podcasts, apple, Google, overcast, Spotify, Stitcher, you get the picture.

We are everywhere.

We want to thank our channel sponsors who are investing in our mission to

develop the next generation of health.

VMware Hill-Rom Starbridge advisors, McAfee and Aruba networks.

Thanks for listening.