{Dan Roberge}
Today we're going to be talking to Christopher Lehman from a company called Elephys Care that uses radar and AI technology to detect patterns in residence, including fall prevention.
Hi, Chris. How are you?
{Christopher Lehman}
Good, Dan. Yourself?
{Dan Roberge}
Doing well, doing well. Um, so tell me a little bit about your technology and what you offer to, uh, the senior care marketplace.
{Christopher Lehman}
Sure. Our company, uh, Gold Sentinel, Inc. Has created an AI-based platform called Elephas Care. It is a non-camera, non-wearable resident monitoring, uh, resident activity monitoring system, uh, specifically designed for long-term care, uh, residents that are living with dementia or other form of cognitive decline.
{Dan Roberge}
Oh, wow. Okay, so now, how did this idea come up? Come about? Like, what led to, uh, your team developing this type of technology, which is quite advanced?
{Christopher Lehman}
Yeah, it is. Um, we originally started, uh, looking at fall detection, uh, as falls are a critical, um, issue in long-term care. Uh, well documented, well researched. And as we started down the path to meet stakeholders in long-term care and discuss with them our ideas to validate the business opportunity, we learned, uh, very early on that while fall detection is great, really, uh, what stakeholders, what operators of long term care are looking for is a way to prevent falls before they happen.
Uh, so we kind of switched gears, and the system, um, that we developed arose, uh, out of asking questions. And when somebody said to us that fall prevention was more important, we asked them, we said, well, what does that look like for you? How can we help prevent falls?
And, uh, an early partner of ours that became our pilot site and later, uh, became our first customer said, uh, well, there are two ways that they identified right off the bat to be able to prevent falls. One was detecting movement in bed for non-ambulatory residents with dementia, uh, and being able to alert the closest caregiver to that room.
Um, in long-term care, uh, they don't restrain residents. They used to, uh, it used to be that somebody with dementia, uh, that was non-ambulance because they've forgotten or they don't realize that, uh, they've lost the ability to walk. They'll still attempt to exit their bed, and that will result in a fall 100% of the time.
Those residents used to be restrained, physically restrained, using, uh, restraints to the bed. Uh, they realized that it caused more, more damage, both physically and, um, mentally, mental trauma to the resident by restraining them than it did by not restraining them. However, uh, you still need to be able to alert somebody who can get there fast enough for a positive intervention.
So one way to prevent falls would be able to be to be able to detect movement in bed. That could indicate that that resident was about to attempt a bed exit, uh, and let somebody know who was close enough to intervene. So that was something we worked on and something that's an important feature, uh, of our system.
The second way that was, uh, identified was to be able to track and report on, uh, toileting activity and specifically changes in toileting activity. So when, uh, average or baseline toileting activity changes for any resident, it could be an indication, uh, of a secondary and underlying condition like a UTI and, uh, or urinary tract infection.
In seniors, uh, an undetected and untreated UTI, uh, an untreated infection of any time can lead to sepsis and death. Um, however, with UTIs , you know, before it gets to that stage, it can still lead to symptoms, uh, that can lead to a fall. So, uh, it can lead to loss, uh, of balance, dizziness, uh, you know, for any healthy individual, if, uh, we have a UTI, we recognize that there are some symptoms, we would go see our doctor, our GP, describe the symptoms.
They would, they would usually prescribe, uh, an antibiotic. We would take that. It would clear up and go away in a couple of days. Um, seniors living with dementia don't recognize that there's been a change. They don't necessarily recognize, uh, that there's discomfort. So if it's unobserved, if the change is unobserved, uh, it can go undetected and it can lead to a, uh, fall or, uh, a cascading effect of just undetected illness.
So those two waves primarily were identified to us as being ways that, uh, through monitoring, through detection, we could help to prevent falls. So we built a system around, uh, more around detection of, uh, changes in activity and allowing or helping a, uh, facility to be able to prevent falls, rather than just detect falls.
We do detect falls instantly, an alert for it, but it really became, um, it became the impetus for forming a clinical advisory group then to say, okay, let's really deep dive into, uh, the daily workflows of the clinicians and see what else we can do. To help, uh, them with their care planning and assessments.
{Dan Roberge}
Yeah. And I find it interesting that just like any other great, uh, business out there, when you start off, you might have a plan and then you talk to your customers and they redirect you in a certain, uh, avenue that becomes your core business model when you listen to people.
So how far had you gone along the path of a certain area and you switched over? How was that kind of okay, redirection, uh, handled within your company?
{Christopher Lehman}
We had done a lot of research on the technology side. So we had built out the non-visual, uh, part of our technology is radar sensors. So there's no implication for privacy or dignity. We can cover areas that other technologies can't, other sensitive areas like sleeping areas and washrooms, uh, because there's absolutely no, uh, imagery captured.
It's radar signals. We're really, uh, bouncing signals off of or into the room and they're reflecting back and we're interpreting, uh, through AI, uh, what those signals are telling us. Um, so we had done a fair bit of work on the fall detection when the request was put into, uh, to say, you know, could we detect movement in bed and alert for that?
And could we detect and report on toileting activity? So we said, sure we could. And we started down that road and we started building, uh, those features into the product. And we really at that time said, let's ask more questions. Let's get a. We formed a clinical advisory group and we said, you know, based on that early feedback, let's see if we can uncover, uh, potentially other ways that we can help reduce harm, uh, improve outcomes, improve quality of life in long term care.
And what we discovered was that there were a number of, uh, what's called activities of daily living, or Adl's that our technology could track. We don't track all of them, but there's a number of them that we could track and report on and through working with our clinical advisory group, they said, so it's impossible for us to review and look at every resident every day for every activity.
Uh, however, what would be beneficial would be if your technology could baseline everybody and then just flag those residents where there's a change. So what we came up with that we measure, in addition to toileting activities. So we measure toileting frequency and duration, the number of visits and how long a resident spends in the washroom.
We also track, uh, um, active versus sedentary time, uh, you know, how long they're still, whether they're in bed or sitting in a chair. Um, we track movement how fast they're moving, what their gait is, or how fast they move. Uh, we also track how long they take to get out of bed.
We look for, uh, so we baseline everybody on those measurements, and then we look for variances, uh, or deviations to the baseline. And we report only on those residents because. And we do this every shift. So at, uh, every shift change, the frontline care workers can review those residents in the prior 8 hours whose activity has changed from their baseline.
So it may only be a handful of them, but these residents often are, uh, again, because there's a very high percentage in the demographic that has dementia or suffering from some other form of cognitive decline caused by stroke or other reason, um, aren't able to communicate. Uh, when there's a change, we let their activity, and their actions speak for them.
So if they're suddenly taking longer to get out of bed or they're getting out of bed faster, both of those things could indicate, uh, for instance, again, that may be indicative of a, uh, utI or an incontinence issue. Um, uh, any change in condition caught early leads, uh, to a better outcome and less suffering for the resident.
So that's what we do. We have a mobile app that we can present that data in a report that can be reviewed in seconds. And, uh, that was also really, really important for long-term care. One of the things they told us really early on, Dan, was that if we're gonna build any technology for long-term care, it can't be anything that takes time.
It has to be something that saves time, because they said, frankly, we don't have any time. And in North America, there is a crisis, a staffing crisis, uh, in nursing and personal support workers, or, uh, in the US, they're referred to as nurses aides. Um, everybody's short-staffed, so we said, okay, great, let us come up with something then that can actually save you time on a shift by shift.
Uh, every care worker can actually recover time in their shift and put that back into direct care time. Caring for the residents.
{Dan Roberge}
That's incredible. And it's like every good AI these days, as we're seeing, is all about taking big data and making it, you know, um, usable for people to read on a daily basis. When it's too much information, you can't do much with it. But what you're doing is you're bringing it down, like you said to a report that allows the nurse's aide and the staff to be able to go through that.
Now, are you feeding that into anything else? Like, are you integrated with other tools where the actual resident care is being registered for compliance purposes and so on?
{Christopher Lehman}
Yeah, absolutely. So there's two different, uh, uh, other um, first of all, there's a dashboard for data visualization. The mobile report is really an exception report where, uh, deviations to those baseline measurements. When those happen, we report them. But for every resident, any other, any other practitioner. The continuum of care and long-term care, uh, can view long-term trends for all the same measurements for residents.
So if a physiotherapist or a recreation coordinator wants to take a look at the long-term trends, for instance, for active versus sedentary time over the past three months, that data is visualized by day for the past week, by week for the past month, or by month for the past year.
So, uh, MDs, physiotherapists, rye MDs coordinators, recreation coordinators can all go in and look at any resident and see how they're trending over time. Uh, because some of these things, the daily changes, uh, maybe within the thresholds where they're not being reported as an exception because they're in an allowable range to say, look, if they're, if they go to the washroom once more a day or once more per shift, we don't need to see that.
Because it's not meaningful enough. Um, but if they're going once more per shift and it's not being captured, but it's continuing to trend upwards, that may be a concern over a week or over a month. Secondly, we do also integrate with any, um, ehr electronic health record system. Um, we're presently, uh, integrated with Pint Click Care.
That's where our first customer utilizes Point Click Care for their system. So we receive all the, uh, resident census data, demographic data from point click care that is beneficial for the facility because they don't have to maintain two systems. Um, changes to the resident census data that, uh, occurs within point click care are automatically reflected in our system.
So if a resident changes rooms or moves or somebody comes in or goes out, they don't have to maintain two systems. Um, we're also working with the facility now to find, uh, anywhere that the data, uh, can be uploaded into. For instance, things like progress, uh, notes, and, uh, different areas within point click care.
Uh, on the clinical side, for instance, if there was a fall or a change in toileting activity, where does the facility or where could the facility benefit from having that information inserted? Where somebody that's reviewing that, um, outside of that frontline care team for, uh, the exception reports. So we're trying to look for, uh, every opportunity we can to try and save time and increase that direct care time where clinicians and frontline care workers aren't occupied with busy work and searching for data, um, we make it easy to find and easy to act on.
{Dan Roberge}
Yeah, you want to automate as much as you can in that process.
{Christopher Lehman}
There's a fear with a lot of people, with AI that, um. Oh, you're introducing an AI tool. Uh, some workers aren't. Um, they think, well, they're coming for. Our jobs, it's going to be in the age of AI. They're not going to need me anymore. Um, we're really not doing anything like that. We're just presenting data to allow, uh, clinicians, frontline care workers to do their jobs better.
Um, we don't diagnose, uh, or predict anything. We present the data and we still require those frontline caregivers, they're the ones that are the experts on the residence to be able to, uh, interpret that data. Um, we're just collecting and providing a view of data that is unobserved and would otherwise be unavailable
So to try and help them as a tool to improve their care planning and provide better care for the residents, ultimately improving quality of life.
{Dan Roberge}
Yeah. And you're doing it effectively, and I think that's the difference. You know, I think, um, if you think about, you have competitors out there and most come in the way of having a wearable on, um, you know, on the resident. And so you are maybe doing that without that, that need for, you know, having something to put on every day or forgetting to put it on or it gets wet or whatever it is.
Maybe we can talk a little bit about the technology itself and the hardware and so on. So what's the physical environment? How do you set this up? And, uh, what's the process for getting that up and running?
{Christopher Lehman}
Sure. It was something that, again, was very deliberately designed. From the beginning. We wanted to make a system that was really zero effort for the resident. It had to be zero effort for the resident because of the nature of the demographics that we're dealing with. Um, wearables just, there's a lot of studies.
It's not that it's bad technology or that it doesn't work. It certainly works. And there's a lot of great wearable technology out there. The important thing to understand with wearables is that number one, it has to be worn to work. And in long-term care settings, specifically, residents that are living with, uh, dementia or other forms of cognitive decline, um, they may not want to wear it, they may take it off, they may damage it.
Um, this becomes a management nightmare for the facility. Now they've got an inventory of devices that they have to make sure are functioning, and charged in good working order. Otherwise, what's the point? Um, so we knew we didn't want to go in that direction, that wearables weren't really suitable. And again, this is backed up by other research, not just by us.
Um, secondly, we knew it couldn't be camera-based, uh, for obvious reasons. In sleeping and washroom areas. And there were a couple of other considerations that, uh, weren't so obvious initially. But again, we learned from asking questions from the stakeholders. Uh, a lot of our competitors are building, uh, Wi-Fi-based devices.
Um, that's problematic in today's world in long-term care, uh, primarily because. Well, there are two main reasons why Wi-Fi can be problematic. One is signal propagation. It's dependent on the construction of the facility. And there are wide, wide variables in how those, uh, buildings are constructed that could negatively impact wi Fi signal propagation.
Secondly is that in today's world, everybody's buying mom, dad, grandpa, grandma, their uncle or aunt, or their loved one in long-term care. Um, wireless devices, they're buying them tablets so that they can Facetime. Um, they're buying them tablets so that they can stream Netflix or other streaming services.
Uh, there's a finite amount of bandwidth that's available, and the bandwidth contention becomes an issue as well. There isn't infinite bandwidth available. And for a mission-critical, um, solution platform like, um, fall detection and measuring, uh, activities of daily living to prevent falls, uh, if you're reliant on WiFi, that's a problem.
Uh, and there's another way that we wanted to make sure as well, was that powering the devices. So our devices, it's a really simple, uh, wiring a cabling. They're powered over Ethernet. Ethernet, for anyone that's not familiar, is a, uh, networking technology that's been around for years and years.
Uh, there's a lot of companies that can, uh, put the cabling in. It's inexpensive cabling. Much less expensive than having to cable, uh, an electrical outlet for a device. So the sensors are powered through a single Ethernet connection that goes back to a power over Ethernet or Poe switch that's all networked together.
We've got a server on-site that does the, uh, uh, signal processing and the AI, uh, and sends the alerts back to, uh, the phones and the dashboards. Uh, so it becomes very cost-effective for, even for a retrofit. Um, again, one of the less obvious, uh, considerations that seem to be lacking in some of our competitors that are AC-powered is that in any setting, um, the AC is down near the baseboards, which is fine.
You can put a sensor or a camera high up on a wall and you can run a cable that's long enough to plug in. It's not the length or getting down to that baseboard again, it's having something that, where you've considered the demographic and residents living with dementia or cognitive decline are going to unplug things.
If they're plugged in, they shouldn't be reachable. Um, now that's a very expensive proposition. If you have to install AC at the top of the wall or in the ceiling where it's not already present, uh, it's much more expensive running AC than it is, you know, cabling a single ethernet cable.
So those were all of them. Very deliberate design decisions, uh, to power it over Ethernet. It gives us a physical network connection that's much more secure than WiFi, it's less expensive for powering than uh, putting in distinct AC power for each one, or batteries for that matter.
Batteries, again, that's a management. You're putting a burden on the facility to uh, check that devices are charged or that the batteries are good if they need to be changed. Um, in our implementation, in a 150-room facility, there are 420 sensors. Um, we didn't want to put that burden on, we wanted to make it zero effort for the staff as well.
So those were our hardware decisions where really zero effort for residents and zero effort for staff.
{Dan Roberge}
Yeah, and I could see that because there are definitely less uh, building code regulations around uh, wiring Ethernet compared to, like you said, uh, just regular uh, voltage wire. Um, so now the other element though is the signal itself, which is, you mentioned a radar signal. Does that interfere with anything else in the building?
Is there a consideration around how that may interact with the nurse call system and so on?
{Christopher Lehman}
No, it's um, the devices are all uh, tested. They're certified through uh, the regulatory body, uh, the FCC certified for indoor use, which um, industry Canada adopts the same uh, certification. So they're designed not to interfere and they're at a frequency that's much different than any medical uh, devices or uh, in fact, WiFi or anything like that.
So, and they're very, very low power. They're, they're less power, much less power than the average cell phone.
{Dan Roberge}
Wow. And so now in general, um, you know, you've now had a lot of experience. What's your like in the industry? What's your um, how did you get to this point in your life to be able to uh, talk, um, and have so much knowledge around senior care, but also technology, and now AI within the technology world, which is fairly new.
How did you get to this point?
{Christopher Lehman}
Well, my background personally was not uh, it wasn't in long-term care, uh, but it was in technology. So I'd worked for a couple of different uh, technology companies in mostly telecommunications, uh, but also doing some data center work and um, cloud computing. So I had a good grasp on a lot of the components of the solution kind of before coming in, and uh, had really kind of, you know, spent some years growing businesses from startup to commercialization and scale.
So my co-founder actually came up with the idea and uh, it was post uh, an event that happened in Ontario, uh, Canada back in 2015, and was really kind of unrelated to what we eventually built. But there was an incident where a nurse was, um, uh, murdering, actually was killing residents in long-term care by switching their medications.
And while we didn't develop anything to address any of those concerns, others have since, and there's certainly been a, uh, move to address how that happened. It really shone a light on what's happening in long-term care. And, uh, we started looking at, or my partner started looking at some of the issues in long-term care, and that's where he came across the huge incidence of falls, not only among seniors in general but how acute it is.
One in four seniors are going to fall, uh, every year, and that number, uh, two to five times more in long term care. And the severity, uh, of the injury is much higher in long-term care. So that really kind of provided the genesis of it. My co-founder was a guy that I had worked with over the years, and, uh, he kind of came to me and said, are, uh, you interested?
Can you help? I said, sure. And, uh, as it happened in 2017, which was around the time he had the initial idea, um, my dad fell, and he fell in a. He wasn't living in long-term care. He was living in a retirement building, though. That was a 55 and over seniors building.
We're seeing lots of these buildings going up. It was brand new, it was state of the art at the time. Um, and state of the art for retirement living, uh, to this day, is pull cords on walls, um, which is wonderful if you happen to fall next to one and you're not incapacitated by the fall.
Um, my dad fell and unfortunately was, uh, undiscovered for, uh, almost 12 hours. Um, and, uh, he was brought to hospital and died shortly thereafter. Uh, I had a good motivation after that when he had this idea, I had a personal interest in saying not only for long-term care, but our roadmap is to bring this into retirement living as well.
Um, because we know that, uh, residents that are living in retirement living, they also have similar concerns. They have the same kind of privacy concerns. They don't want cameras. Um, there's a stigma around the wearable devices, uh, even for independent living seniors, uh, wearing the pendants or the fall detection devices.
Um, so we saw a real opportunity, and obviously, everybody's aware of the aging demographic globally. Uh, this is, uh, uh, a problem that's not going to go away, and it's an issue that's only going to grow because of the aging population.
{Dan Roberge}
Yeah, absolutely. And you found a way to use technology to give you the information that you need, but at the same time, keep, uh, the privacy for that resident. And I think that that's one of the key differentiators for. Your product that I uh, really like about it. Now, the return on investment, um, we talked about obviously the residents' health and safety.
That is probably the number one return on investment. But dollar value wise, um, you also talked about saving time, but with um, the nurses' aides and so on, getting information, not being able to track that information before. But what kind of other rois would you say your product brings that a facility would say this is worth it for me to put in here outside of let's say the resident care aspect of it?
{Christopher Lehman}
Yeah. So there's uh, there is a cost, every facility, regardless of the jurisdiction, the costs differ a little between Canada and the US. Um, but there is a cost of falls, there's a cost to hospital um, readmissions, there's a cost to the facility, you know, for resident, uh, transfers and emergency department, um, ah, admissions.
Um, there's a cost to the facility for ah, vacancies. And you know, if they, if a resident has to leave to uh, go to a hospital long term because of injury, you know, now they have to get another resident in. So there are measurables that our system, if you're, if you're reducing the causes of or identifying potential causes of illness through changes in activity, um, early intervention results in better outcomes and better health for the residents.
So you're not only reducing trauma for the residents, you're having a measurable impact on costs for the facility. And uh, for retirement facilities, we're providing a competitive benefit for them by they can go out and market their facilities as uh, uh, while still being privacy compliant, a much safer facility for seniors to live in than their competitors.
Uh, so that can increase uh, their revenue and increase their bottom line. Uh, the increase or the cost savings shouldn't uh, be underestimated for the time savings, though. That is a big component where if we can uh, allow facilities to increase direct care time without adding headcount. So now if staff are not, if they can reduce their care planning or their assessment time across the continuum of care, if they're not responding all the time to for instance, bed alarms because we can detect those movements in bed and bed alarms aren't that good at it because uh, again they're pressure based.
They're typically going to go off when the resident is on the floor. So again it's something that they're only responding to. All of those things are measurable. Um, when the staff can be redeployed to be doing uh, direct care rather than busy work, rather than administrative work rather than.
Or we can reduce the time for care planning and assessments. Um, those are all measurable. That's all important. And it. Provides the facility the ability to provide a higher level of care, a higher quality of care without, uh, adding headcount. So our measured ROI, uh, again, and we've tested this with our facility operators, is roughly 16 months.
Um, after the capital costs of the hardware for the solution and the installation, um, they're uh, ahead of the ROI. Uh, the cost-benefit comes out around 16 months.
{Dan Roberge}
That's incredible. So is there anything else you want to add about your product or the story behind, uh, uh, the evolution of the product and where it's going in the future?
{Christopher Lehman}
Uh, no. I mean, we're incredibly excited. The interesting thing with AI, so our device today is nothing considered, uh, a uh, medical device by either Health Canada or uh, the FDA in the US. Um, because at this point we're not claiming to predict, prevent, or diagnose anything. Um, when we talk about fall prevention, we're not directly preventing falls.
We're assisting the facility, uh, for them to be able to potentially prevent a fall. So if we can, again, get back full circle to where we started with something like detecting movement in bed, well, that's a detection that we're doing and we're alerting a staff member to that so that they can use that information to potentially intervene and hopefully prevent a fall.
But we don't claim that that alone is going to prevent a fall. However, uh, AI, the wonderful thing with AI is that, um, as we grow, as we gain customers and we're uh, covering more and more residents, um, the interesting thing is what happens when we've got, uh, we're covering 500 residents or 1000 or 10,000 in multiple different facilities.
Maybe we've detected five falls on the same day. Now we have the ability with looking at the data that we've collected to say, even though they're across multiple different facilities, what did we detect uh, an hour before that fall happened? What did we detect 6 hours before or a day before or 24 hours before, uh, or a week before?
And is there any correlation? These are things that are really hard for humans to look at, that data. And again, the facilities only know what happened in their own facility. Um, nobody's really got this data today. They don't have, uh, uh, the activity data for seniors that proceed a fall, um, when we could correlate that there's uh, a real hope that we can find a correlation and see something in the activity data that we can actually say with a high degree of confidence when something happens that we can measure.
This looks like a marker that precedes a full. And you know, let's elevate this, this resident to uh, you know, a higher observation level for the facility because you know, we believe that they're at a higher risk for ah, ah, an imminent fall. Um, you know, that comes later down the road.
There's a lot of different work we're doing on the R and D side and we're really committed to innovation and continuing to look for ways that we can um, help push the envelope for detection, uh, in terms of activity detection and reporting, uh, that we will kind of continue to do.
And on the AI side, it's a "green fields, blue sky" opportunity, uh, whatever euphemism you want to use. Um, uh, we don't know, I mean we may be able to look at um, uh, detected activity prior to a stroke, uh, in addition to falls. These aren't things that are, and excuse the pun, on the radar today necessarily.
But as we can integrate closer with um, the EHR platforms as well, we can start working with those groups and with the homes to say while we don't detect strokes and that's not in our data set today, if that data is available to us through the EHR record, we can start looking backwards now and saying, were there any unique changes in activity prior to any number of uh, health events, uh, for seniors?
So it's exciting.
{Dan Roberge}
Yeah, absolutely. And that's probably one of the most interesting things about AI, is that we're going to all realize we're all creatures of patterns. And um, once there's enough information in there, we're going to be able to predict a lot of things. So where can people find more information about uh, your product and your company?
{Christopher Lehman}
So uh, they can visit our website. That's the best place to go. It's www.elephascare.com. So it's elephascare.com and then uh, if they want to email infolificare.com if they have any questions or they'd be interested in seeing a demo for their uh, for their long term care or retirement community, um, we'd be happy to uh, meet with them and provide a demo and show them how we can help them.
{Dan Roberge}
Well, I think it's great technology, very interesting, uh, application of AI and uh, really thank you uh, for talking to us about that today and uh, we'll talk again in the future and see how it's evolved.
{Christopher Lehman}
Thanks Dan, I really appreciate it and uh, love what you guys are doing as well. Just shining a light on all of the different uh, technology and tools that are coming to market to help improve the quality of life for seniors.
{Dan Roberge}
Excellent. Thanks, Chris.
{Christopher Lehman}
Okay. Thanks, Daniel.
.svg){kind=small}