Skip to content

Episode 24

Tackling home care staffing challenges with AI technology

Jeff Howell (00:01):

Welcome to home health, 360 a podcast presented by AlayaCare. I’m your host, Jeff Howell. And this is the show about learning from the best in home healthcare from around the globe.

Jeff Howell (00:18):

Welcome to another edition of home health 360, where we speak with leaders in home care and home health from across the globe. Everyone I speak with in home health loves to talk about data and analytics and predicting negative health events. And everyone’s favorite buzzwords. These days seem to be machine learning and artificial intelligence. Well today I have a certified expert. Naomi Goldapple is the head of a division within AlayaCare, known as AlayaLabs. Alli labs supports the corporate innovation strategy of Alia care through applied research in machine learning and optimization applications. So in this episode, we will be talking about how to solve the staffing crisis with technology. What is the future of home health scheduling and exactly how can we make predictions to help keep people in their homes out of hospitals and be able to improve outcomes? So, Naomi, this is very long overdue. Welcome to the show.

Naomi Goldapple (01:23):

Thank you so much. I’m really happy to be here.

Jeff Howell (01:25):

Give us a little bit of background before joining AlayaCare. You were at a company called element AI, which was a world class company at the time that launched an incubated advanced AI solutions and partnership with large organizations. Can you give us a little bit more background about what that company was all about and the work that you did there?

Naomi Goldapple (01:45):

Yeah, sure. So element, I was quite revolutionary in 2016. The idea was really to bring the best and brightest researchers and together with, you know, business professionals to really turn all of their research in artificial intelligence, into operational code and actually apply it so that companies big and small could use it to improve their operations. So what, what we did was we really attracted, you know, people from all over the world, which was, which was wonderful. And we focused on a few sectors and I was in charge of transport and logistics. So that meant using all kinds of technologies you know, machine learning and optimization technologies natural language processing, all that to tackle some of the problems in the world of transportation and logistics. So we worked with, you know, big ports and train companies and airplanes and all kinds of trains, planes, and automobiles, and you know, leveraged all those technologies to bring solutions into those industries.

Jeff Howell (02:51):

Yeah, I would imagine there’s all the big three PL companies and anyone who’s in eCommerce that that would be the type of company that would be coming to you, you guys, with the problems. And then I would imagine your research teams would then, you know, go about with their data sciencey ways to come up with solutions for these real world business problems then.

Naomi Goldapple (03:09):

Absolutely, absolutely. So we would even, you know, partner with with some of these logistics companies, because they would want to know how they can help their clients. So we would help them you know, with their own optimization algorithms and how they could use all of that data that they collect to actually deliver better service. So, you know, we were mostly B2B so that they could serve their markets even better.

Jeff Howell (03:32):

I remember a few years ago at the Eli care better outcomes conference, there was a speaker from element AI. And I was really struck by the comment that he made that there was something like only around 25 people in the whole world. And I’ll, I’ll probably butcher this a little bit, but could really be this like end to end project manager that could design and implement and oversee an AI project. And I’m curious, so this would’ve been, I guess in late 2019, and I was wondering why there are so few people, if it’s just the complexity of the work or a reflection of the maturity of the industry, or maybe a little bit of both.

Naomi Goldapple (04:10):

Yeah. So 25 people, I’m wondering if

Jeff Howell (04:14):

2,500, sorry. Yeah. Did I say 25? I told you I was gonna butcher it.

Naomi Goldapple (04:22):

Yeah, so definitely, you know, looking back a few years ago it is complex. There’s a lot of moving parts you know, from actual data extraction and cleaning and data wrangling and pre-processing building pipelines you know, training models then putting into production and retraining models and monitoring. There are definitely a lot of moving parts to getting it right. From a to Z, I would say that since then the, the industry, you know, itself has really upskilled and there’s many more kind of ML tools available to help democratize access to these technologies and put them into, put them into production. You know, there, there there’s a real difference between putting regular, you know, code deterministic code into production versus machine learning because the data, the data evolves. And so the results of the models that you trained on certain data will evolve as well.

Naomi Goldapple (05:19):

And so that does require monitoring and, and retraining. And then another important aspect of this, which maybe is why he was referring to, you know, few people that really understand how to put it into production from a to Z is you have to think about the end consumer, who’s actually consuming these classifications or predictions in their daily workflow and using it as decision support and how are they consuming it. And also, how are you kind of creating a feedback loop so that, you know, if your predictions are correct, you know, Netflix you know, will always ask you, you know, was this a good choice? Did you like this? Mm. You know, they’re, they’re getting you to give them them feedback on their models. So then they take that back and that helps with the retraining of the models. So kind of having this living and breathing system is, is complex to design and to maintain so that it keeps delivering consistent value. So yeah, I can I can agree with that.

Jeff Howell (06:12):

I think I’m gonna get this right. Maybe, you know, this when you’re doing an e-commerce checkout and you’re asked to click on the pictures that have a traffic sign in them.

Naomi Goldapple (06:22):

Yes.

Jeff Howell (06:23):

That is powering Google’s feedback loop for their self driving car technology. Is that not right? Have you ever heard of that?

Naomi Goldapple (06:30):

Well, I don’t know if it’s for their car technology, but definitely it’s for computer vision. So yes. Everything, everything that you do is helping to train their models when, you know Instagram or, or Facebook will, you know, automatically tag certain people it will do facial recognition and if indeed it tags them correctly, then it will know you’ll, you’ll, you’ll be helping them. So, yeah, lots of, you know, even when you’re composing your texts and it gives you predictive texts on what you should be. Sure. Yeah. Writing next, if you say, yep. I agree with that. I’m gonna swipe then you’re, you’re helping to train all those models.

Jeff Howell (07:03):

Right. Right. And when you say democratization, I’m assuming what you’re referring to is taking it from an early immature standpoint and making it more available to the masses. And I’ll give you an example with GTP three. You now can get access to these like AI writing bots. So say if you wanna write your own Instagram ads or emails or , you could even use it if you’re a student. Although I don’t know if I would advise about to do that, but there are all kinds of like these content writers that are really bots, where you just put in the the topic that you wanna write on, and it’s scraped, let’s call about 10% of the internet and it’s learned, and it can actually string together paragraphs on a particular topic. So that, that’s what you’re getting at with the democratization really right. Where you, you start to the learnings so that the common person can have a tool that they could just use without having to be an expert.

Naomi Goldapple (08:04):

Yeah, definitely. That’s part of it. The part of democratization also is, you know, you want other industries also to be able to use and leverage these technologies without having to, you know, hire a expensive, like full data science team as well. So, you know, it shouldn’t be that the only companies that can really put AI into production are like the Facebooks and the Amazons and the Googles, because they have so much data and they have, you know, deep pockets that they can do all the research and then they can hire lots of people to be able to do this. So there’s a lot of tools that will help automate the processes so that even, you know, if you have one data scientist, they can have a lot of auto ML tools at their fingertips that they can then put stuff into production, much easier than having, you know, an entire data engineering team and data science team to support them. So there’s, mm-hmm, affording, you know, the, the actual practitioners much better, so they don’t have to create things from scratch. And then there’s also allowing other industries, and as you said, individuals, to be able to take advantage of this technology.

Jeff Howell (09:12):

So for your work at Aly care would you be more focused clinically, like predicting negative health events or operationally like reducing drive time or mileage or predicting caregiver turnover?

Naomi Goldapple (09:25):

So do I have to choose

Jeff Howell (09:27):

Yeah. All of the above, yeah.

Naomi Goldapple (09:29):

All of the above. Okay. And, and, and it, it is all of the above. You know, I always say we’re sort of tackling three of the biggest problems that our customers face. So unless you’ve been, you know, not paying attention the past couple of years, there’s definitely a huge shortage in the healthcare industry. And home care has been hit pretty hard in terms of caregivers and nurses. Not only is it a challenge to kind of attract, but it’s also a challenge to retain and reduce the turnover of, of these sort of precious resources. So we work on, you know predictive models to predict who is at risk of quitting and what can we do? How can we mitigate those circumstances? And very closely coupled with that is schedule and route optimization because in our research we’ve seen that the main reason why people quit this industry, why the caregivers quit is because of the scheduling. So that could be, you know, schedules that don’t give them enough hours, which seems crazy when there’s a shortage

Jeff Howell (10:33):

Sure. Yeah.

Naomi Goldapple (10:34):

Or schedules that really don’t suit them. So, you know, getting weekends or evenings when they can’t, they have young families. And then, you know, one of the other pieces that came out of our research was you know, if you can reduce the lag time between when they’re hired and when they get their first shift, you’re gonna have a lot more success keeping them around for longer because they can get disenchanted very quickly. They don’t wanna be hired, not be given hours and not be earning money right away and kind of be part of the fold. So those are, those are two of the areas. And then on the clinical side, as you’re saying, we’re definitely working to, you know, overall the, the whole goal is to improve patient outcomes and to minimize negative outcomes. And when I say negative outcomes you know, we’re talking about reduce falls, reduce ER, visits, reduce you know, readmission into hospitals.

Naomi Goldapple (11:23):

And the way that we do that is the home care workers are actually in a very privileged position where, you know, every visit, they are collecting a lot of information, a lot of status, a lot of observations on how the the patient is doing. And so if we can collect all that information and kind of have a constant narrative on how they’re doing, we can see anomalies a lot easier, and we can really pinpoint when risks are brewing and, and we can alert. And so they can act on these right away, like make a trigger and, and mitigate these situations to avoid these these negative outcomes.

Jeff Howell (12:01):

So there was a lot of great stuff in there. Let’s go back to the very first part of what you said. Sure. I understand you worked on a schedule reshuffle project and going back to the route and schedule optimization and trying to make everything so efficient so that people can ramp up at their hours. The project where you did this big schedule, reshuffle involved field supervisors, can you give us some insight as to how that project unfolded?

Naomi Goldapple (12:31):

Yeah, sure. So we one of our clients came to us with a challenge, which was that their field supervisors spend a lot of time on the road mm-hmm and only managed to do, you know, two or three visits a day where they would prefer if that’s more like five or six visits a day, and that they’re spending a lot of money also in you know, paying for mileage and paying for what they call windshield time, which is just time spent, not with the client, but time spent actually getting to the, to the appointments. We help them to look at the, the overall schedules for a month and put all of the fixed visits and even like last minute visits, cuz sometimes things happen where they have to drop everything and, and actually go to for, for if it’s an emergency or the beginning of care or some other reasons. So we left some slack for, you know, anything that could be ad hoc. And we, we re rejigged the schedules so that the routes were a lot cleaner and they were able to do, you know, five or six visits a day and we’ve reduced the travel time by about 46%. And this allowed them to, you know, spend more quality time with the clients and spend less time on the road. And that’s even more important today where the, you know, gas prices have gone through the roof. So that’s very much complicated as well.

Jeff Howell (13:52):

I’m a very lazy consumer, I’m the last person to talk about gas prices, but I’m now talking about it. So I know it’s bad. and, and so a schedule reshuffle then would sort of be like, if I’m a field supervisor and I’ve got six visits in that particular case, am I going from scheduling my own visits to basically clustering? So I’ll have a, a visit in the field day and, and that’s how that’s sort of the end result.

Naomi Goldapple (14:17):

Yeah. So, I mean, we actually make it kind of flexible so that they can decide, but the idea is that yes, it is kind of on demand reshuffling and it’s it can be self, self shuffling. So a, you know, we can put in a schedule for the for the month and say, this is a, the optimal schedule for the month with the information that we have right now, knowing that you have certain clients that you have to see, you know, once a year, some, you have to see them twice a year. Some you have a special case while you have to see them and then leaving enough slack for, you know, anything that comes up up. So you could, you could use that monthly schedule, but then as soon as, you know, a visit comes in that you have to insert into that schedule, you can then hit the, you know, reshuffle button. Yeah. And it will then allow you to insert those visits in and reschedule so that you maintain the optimality that you, that you had in the first place.

Jeff Howell (15:12):

Okay. So you’re designing the goal post to start with. Yeah. And it really starts with this mindset of the way we used to do it was you would just schedule according to what the client’s preferences were. And now we’re really trying to make sure we minimize the windshield time. So we’re giving the client a different set of parameters, but it, it is just a bit of a change of a culture within the office that this is one of the objectives we need to achieve is the, the reduction of miles. And so the clients probably don’t know anything O other than here are the times that we’re available. And, you know, so it’s not much of a change in an experience for them.

Naomi Goldapple (15:50):

Oh, for the clients. No, for the field supervisors, it is because the way it worked before was here’s your visits for the next six months, you figure it out, right? Like you, you cluster things, however you want you. And, you know, I remember there’s a lot of different color highlighters going on and, you know, just trying to fix that up. And then if somebody had to cancel or something have to be inserted, there’s, you know, some crossing things out and moving things around. And that’s very you know, reactive as opposed to trying to you know, proactively hit some metrics that will make sure that things are being efficiently, efficiently delivered.

Jeff Howell (16:25):

Okay. Well, let’s move on to another case study that where you built a satisfaction dashboard. And I believe this was for an agency of maybe a hundred or 200 clients or something like that. It’s not, it’s not a, a massive agency. It’s, it’s reflective of a lot of home care agencies out there. And my understanding was that it was a very sort of bespoke kind of consulting assignment, where you sat down with the client and you jointly came up with the five or six elements. That was the hypothesis of, this is what happy looks like for a caregiver. And that if these caregivers, if we can improve satisfaction, even through COVID, that it will lead to reduced turnover. So tell us about that project.

Naomi Goldapple (17:13):

Yeah, absolutely. So this one actually was a very proactive owner agency owner who had a very clear idea of what they thought drove satisfaction in their caregivers. So indeed we had sort of a handful of variables that were thought to drive satisfaction. So what we did is we automated the extraction of this data every day, coming right out of the all Eli care system and put into a model. So a mathematical equation basically that would then come up with a score. So this was done every, every day. So every day these scores are refreshed. So you can see you know, and we would display them in descending order. So you would see, okay, your happiest people are on top. And then as you go down, you’re like, you know what, let me look at the bottom 10% and see why their scores are so low.

Naomi Goldapple (18:08):

And you know, if you see, you can also see the trending and see, well, this, these people used to be happy and now they seem to be, you know, trending downwards or others seem to be trending upwards. So this really gives you know, a good kind of pulse on how the caregiver satisfaction is on a daily basis that the actual scheduler can look at and they can, you know, pick up the phone and say, Hey, you know, I see your score’s gone down. What’s going on? Can I give you some more shifts? Are these shifts satisfactory? Is there anything going on? And they could actually try and try and improve their situation to avoid them becoming disenchanted and, and, and finally leaving. So one of the, one of the interesting things that we found right at the beginning was when we did these scores, we found there, there was a whole bunch who had zero scores.

Naomi Goldapple (18:53):

Like it was just a score of zero. And we’re like, well, how could that be? And then we see that, you know, there’s some caregivers that hadn’t had shifts in a while. So if there’s no shifts there’s, there’s no scores in home care a lot. The, the caregiver doesn’t actually quit, they just sort of fade away. So you keep them on your roster, but they’re not actually, you know, employed. They’re not actually doing any, any activities. So this was really good just to actually clean up the roster and see, huh, look at all these zeros, let me pick up the phone and see if I can even resurrect some of these resurrect their interest and rejoining into the fold or actually just clean up the roster and know exactly what I’m dealing with and how many, you know, I have to go higher and, and of what types.

Naomi Goldapple (19:35):

So just bubbling up all that data at their fingertips was hugely beneficial. And, you know, COVID hit at the beginning of this. When we were, when we put this into production and we saw that the satisfaction was increasingly was increasing by about 30% after six weeks. And we’re like, why is it increasing so much? And a, the workforce was getting smaller because not everybody could work because some of them, you know, they didn’t have childcare. They had to, they had to stay home or maybe they were afraid to work, but those who were working were getting full utilization, they were getting full schedules. They were kind of, you know, getting to see the people that they wanted to see there, all, all the factors that drove satisfaction were very high. And so you had a very satisfied workforce, you know, at the, at the beginning of COVID

Jeff Howell (20:23):

Oh, interesting. And it sort of weeded out the people that were gonna be less likely to be happy and that the more loyal employees were left behind. So let’s, so let’s recap them. What, what were the inputs? It was getting the shifts that you want getting a high percentage of the number of hours that you want. What were some of the other metrics that were the inputs,

Naomi Goldapple (20:42):

There was so utilization, which is really the Delta between what are the hours that you’re getting versus what are the hours that you want? So what’s your capacity. And that one was highly, highly weighted. That really drives a very, very high percentage of the satisfaction is just, you know, can I have a full schedule? Sure. Which, which makes sense. Then there were things like, you know how often am I seeing the same clients? Cause you know, a lot of the satisfaction that comes from this work is building relationships with your, with your clients. So if every day you’re seeing a whole bunch of different people and you don’t get to revisit the same ones, then it’s a little less, less satisfying. You know, if you’re doing something called quality shifts which, you know, depends on, on the agency, but some like longer shifts versus smaller shifts, you know, there’s other things that other agencies find important. Like how much do they have to travel during the day? You know, and surprisingly pay is not, not a huge driver, but you know, it, it, it can be as a, as part of it. So we’ve this model can be reused for, for any agency. And if an agency has a very specific metric that they feel is very important for their situation, then we can add that in to the to the equation and then include that in the in the actual score.

Jeff Howell (21:57):

Yeah. I mean, and it all makes sense. And I think post-study, you went back and looked at the, sort of a, more of a longitudinal basis just to see how things worked out thereafter. And there was a strong correlation between the caregivers that were happiest, ended up sticking around longer.

Naomi Goldapple (22:15):

Yeah, yeah. And yeah, we almost, I suppose it was a, more of a pessimistic way to look, but we almost looked at it the other way to see yes, those who had poor scores and declining scores did in fact leave you know, within three months. So it does show that as their scores go down if not mitigated, then, then they will, it, it is definitely a, a predictor of of churn.

Jeff Howell (22:37):

Yeah. Well, that’s an interesting way to see if you can intervene and, and take early action for the ones that you’re identifying as going down that slippery slope and see what you can do to make a difference for them to stay for longer.

Naomi Goldapple (22:52):

Well, and I think that’s a that’s like an actionable thing that the schedulers or the HR can do every single day, right? Because that data is refreshed every single day mm-hmm . However we also have like another suite of metrics that we call retention metrics, which can also help you to plan. So when we look at things like, you know, let’s look at the age category, so we’ve noticed definitely that, you know, caregivers 30 and under definitely don’t stay as long as caregivers who are 45 and up, you know, they tend to keep their jobs for a lot longer. We look at the referral sources, like where did we get them from which job boards or which, which colleges did we get them from? Which ones seem to drive more loyalty. And so looking at all those type of things, looking at you know, are they starting to clock in late, is clocking in late a sign of maybe becoming a little disenchanted. So we take a look at, at all those and try to kind of spot those early warning signs that can be, that can be acted on, and that can also help in planning your recruitment strategies.

Jeff Howell (23:55):

Got it. So you’ve also worked on something on the clinical side, which I just call like a, a natural language processor where it scans these unstructured nursing notes and progress notes, and comes up with a score based on, I presume what you and the client jointly agree to. So for example, loss of appetite might have a certain score, but swollen foot might have another score. Can you bring us into what a, what a li labs has done to come up with something that will just kind of scan the system and alert coordinators to some patients that might require more attention?

Naomi Goldapple (24:36):

Yeah. Yeah. And it’s funny because just piggybacking on what we were just talking about, when we first started looking into notes, we had a hypothesis that caregiver notes might contain information that will let us know that they’re becoming disenchanted or frustrated and that they might be at risk of leaving. So we started looking at the notes to see, you know, are they starting to get sloppy? Are they starting to use language that shows that they’re frustrated? And when we started looking into these notes, we’re like, geez, these notes are chalk full of clinical information.

Jeff Howell (25:06):

Ah,

Naomi Goldapple (25:07):

And so we’re like this, this, this should be bubbled up somewhere, right? Like they should know this patient fell, this, you know, they lost appetite refusing to use their walk, you know, things, things like that that were kind of embedded in some unstructured notes. So what we started doing is we started we kind of extracted and collected notes that were left. These could be, you know, overview notes, progress notes, visit notes activity of daily living notes. And we worked with some clinical managers from our, our clients to basically classify what would be a note that was really important that you would wanna know about right away. And we call that a red flag. And what is a note that is kind of something you wanna know about, but maybe it’s more, you just wanna keep an eye on it as opposed to, you know, be it being actionable right away.

Naomi Goldapple (25:54):

And we call those yellow flags. So the red flags were things like, you know, a fall anything cardiac, anything, you know, skin breakage anything about COVID you know, things, anything about neglect or abuse or things that had to be dealt with right away. And then, you know, the yellow flags could have been anything like, like, you know, canceled visits could be, you know, loss of appetite could be, you know, change of mood could be, you know, other things that they should be that they should be aware of. And so what we did is we developed a a machine learning model, a deep learning model actually to automatically read these notes and every hour post them on a dashboard divided into yellow and red flags. And the clinical manager can look at it at any time and see if these CC what’s going on.

Naomi Goldapple (26:42):

And we did pilots with a few of our, a few of our customers and they found this was extremely useful. They was able to, they were able to act on things that they probably wouldn’t have found out about for, for a while. Some just do not have the personnel and the time to read through these notes and some read through, but they only read through a percentage of them. So just by bubbling up, the important ones can really help to understand the status of a patient and dictate whether, you know, action should be taken to improve those outcomes.

Jeff Howell (27:13):

And the great thing is, is that there’s no additional training required. You’re just implementing a system on the back end. That’s kind of scanning all these notes that are going in anyways.

Speaker 3 (27:22):

Yeah. I mean, the training is just we, we ask them to, you know, we were talking about before it labeled them, so we wanna know, did we predict this? Right? And then they tell us, yes, you did thumbs up or no, you didn’t thumbs down.

Jeff Howell (27:34):

Right. There’s that feedback loop we were talking about.

Naomi Goldapple (27:36):

Yeah. Like you want, you want your, your model to always get smarter and always get better and learn from, learn from what what the clinical manager says.

Jeff Howell (27:44):

What are the common misconceptions about the work that you do in, in data science for like the people of home care, you know, I, everyone I speak to, you know, we have this sort of perception that technology is gonna save us all. And I’m curious your thoughts on may maybe some limitations or, you know, what the future may look like with all this stuff.

Naomi Goldapple (28:05):

Yeah. It’s a good it’s a good questions, man. I hope there aren’t too many misconceptions cuz we try to try to be pretty clear on what we’re doing. But I think when people hear, you know, AI or machine learning, they definitely start thinking about, you know, things like robots or they think about maybe all these like cool devices that are out there. I mean, there’s things like, you know, sensors and the souls of shoes that can, you know, detect changes in gait or, you know, potentials for falling and things like that. Those are fun. And I, and I think they’re, they’re very innovative. But that’s not, it’s not where we’re focusing right now. Like the adoption of these devices is tough right now. And we’re, we’re definitely open to collecting all of this internet of thing, data that, that is available to help measure the status of the client.

Naomi Goldapple (28:52):

But we’ve been focusing a bit like on the bigger operational problems. And then we can focus on some of the, on the fun stuff in terms of you know, what’s, what’s coming next. And as our sort of aging population progresses, and the, the older people are more and more comfortable with with technology, then that’ll be the, the adoption’s gonna gonna be much easier, but I’d say less of a misconception, but more of like an eligibility requirement is, you know, everything that we’re doing depends on good quality of data. And before clients can start to take advantage of predictive analytics, we need to be sure that they’re capturing clean, consistent data so that they have, you know, a good grasp of the basics. You know, we were talking about churn before like employee turnover and, you know, in speaking with, with a lot of our clients, like some don’t have a good grasp on, you know, how many active employees do you have?

Naomi Goldapple (29:47):

What is your current turnover rate? You know, those things are much more important as a base before you start predicting on, you know, who’s gonna churn, you have to know who do I even have? What am I dealing with here? And making sure that like the data at the point of capture is an easy process and can be used later on. So I think maybe a misconception is that it’s easy to just start using these technologies, but unless you have a really good, you have kind of your data house in order, you can’t really take advantage of these technologies. So it requires a little bit of grunt work sometimes to, to get there and to start using the data in your workflows and really to try to think, think about being a data driven company to be able to take advantage of this.

Jeff Howell (30:32):

Yeah. Your data’s only as clean as the people that are on the front end, getting it in there. How, what would you recommend for agencies that are trying to, you know, take steps to clean up their data?

Naomi Goldapple (30:42):

Yeah, I mean you know, having, using like some sort of business intelligence tools is very useful to be able to, you know, easily visualize and consume the data. So you can actually spot it. We’ve done some work building what we call like missing data dashboards. So we can actually help you to say look, here’s all of your employees and your system we’re missing half of their home addresses. And if we’re missing their home addresses, we’re not gonna be able to put them into a route optimizer, cuz we don’t know where they’re starting from. Right? Like basic stuff like that. Sometimes if they just understand why the data is important, then they can start to, you know, come up with processes to include it.

Naomi Goldapple (31:20):

We were, we were, you know, just working with one of our one of our clients and you know, if we wanna be able to predict who is at risk of being hospitalized, who is at risk of falling, we need to understand when people were hospitalized in the past and when did they fall? What were the circumstances leading up to that so that you can actually teach, teach a model when some of the risks are present, but if they’re not capturing hospitalization dates, how long they stayed in the hospital when did they come out of hospital, then it’s hard to know right. It’s hard to have something to train it on.

Jeff Howell (31:53):

Yeah. It’s hard to to just say how, how much have we improved by if you weren’t really measuring how good you were before kind of thing. Right. So

Naomi Goldapple (31:59):

Exactly, so baselines are super, super important and you know, sometimes not it’s annoying, but because you know, you wanna get to the fun stuff, but sometimes you just have to get your baselines set and then you could be able to compare off of it.

Jeff Howell (32:12):

Mm-Hmm well for anyone who’s listening, if you wanna learn a little bit more about Alia labs you can just Google a L a Y a L a B S Alli labs. It’ll bring you to the Alli care.com domain. And then it’s just slash Alli labs. We’re bumping up against our time here, Naomi, but I will get you out of here on this. Give us a reason for hope that the future of care in the home is bright.

Naomi Goldapple (32:41):

Okay. I will try. So I think one thing if I, if I have to be a half a glass, half full silver lining kind of person, which I, which I try to be this whole pandemic, one thing, it has definitely shown the, the, the spotlight on the deficiencies of the home care industry. And so I think that has really put the industry at a place where, you know, politicians and people are very motivated to fix this because they realize how important this is for our society and how important it is for people to age at home and for that to be possible for them. So that’s I, I think more and more people are gonna be wanting to age in place and that the infrastructure is being set up so that the, the systems will be in place.

Naomi Goldapple (33:30):

So that can be possible. So I think that’s that, that is one positive thing and the technology is going in the right direction. Right? So another thing with the pandemic, the fact that we were able to quickly you know, turn to virtual care, I don’t think that would’ve happened so quickly without the pandemic, but we saw that we could actually accomplish quite a bit through virtual care, even when I say virtual, that could even just be a phone call. Right, right. To make sure that, you know, seniors are okay, the more and more people will become comfortable with using the technology, including seniors. And so we’ll be able to kind of offload some of the tasks and some of the, the visibility to the technology, as opposed to needing, you know, an army of an army of new people to be able to come into this industry.

Jeff Howell (34:14):

Well, I’m always happy when I speak with you cause I’m ex inspired and the stuff you’re working on is just so cool. Don’t be surprised if you see my resume, come across your desk at some stage, I’d love to be working on the stuff that you guys are working on. So thanks for doing the show today and I’m sure to have you on at some future stage as well. Thanks, Naomi.

Naomi Goldapple (34:35):

Thank you so much. Byebye

Jeff Howell (34:38):

Home health 360 is presented by Aly care. First off, I wanna thank our amazing guests and listeners to get more episodes. You can go to Aly care.com/home health 360 that’s spelled home health 360, or search home health 360 on any of your favorite podcasting platforms. The easiest way to stay up to date on our new shows is to subscribe on apple podcasts, Spotify, or wherever you get your podcasts. We also have a newsletter you can sign up for on Alia care.com/home health, 360 to get alerts for new shows and more valuable content from Alia care right into your inbox. Thanks for listening. And we’ll see you next time.

Listen to episodes on your favourite platform:

HomeHealth360Podcast-Ep24-UF

Episode Description

A.I. technology in home care back-office software is increasingly guiding organizations to better outcomes by helping find solutions to challenges such as caregiver recruitment and retention, optimal scheduling, and more.

AlayaCare’s VP of AlayaLabs, Naomi Goldapple is our guest in this week’s podcast to talk about solving the current caregiver staffing crisis with artificial intelligence technology. In this episode, she will dive into the future of home health scheduling, how to make predications to help keep people in their homes and out of hospitals, and improve overall outcomes for home care businesses.

Episode Resources