Transcript
Marla: Hi everyone, and welcome to the Intelligent Clinic battling Burnout with AI. We are so glad to have you here today and really appreciate your time joining us for this really important conversation that we need to be having in rehab therapy right now. I think we've all been combining together talking about how administrative burden has been skyrocketing.
Staffing shortages are worsening, increasing presage pressure on margins are occurring every day, and at the same time, we're getting clinicians really burnt out because they just feel like documentation, paperwork, and all of the pressures along with all of that, is taking them away from what they love most, which is treating patients and providing excellent quality care.
So that's why we're here today, is to talk about how ai, when used in the right way. Can help clinics reclaim time, retain staff, and really refocus on patient care. And I know there's some hesitations around ai, so we wanna talk to the leaders who are using this, who are adopting it and who are even creating it, and ask a lot of questions and get a lot of feedback.
Before we dive in though, just some quick housekeeping notes. We really encourage you to ask questions. We want you to be part of this conversation as well. So please use the q and a feature at the bottom of the screen. You'll see it right there, and you can drop questions at any point in time. We'll get to as many as we can during today's webinar, and if we don't, we will absolutely reply to you afterwards and respond via email.
And this session will be recorded. We will send out a replay link afterwards and it'll be emailed to all registrants. So please feel free to share it with your colleagues. So we do wanna hear from you as I said. So we're gonna start with a quick poll. You are gonna see it pop up on your screen and really just asking you what your AI experience in the clinic is right now.
I'm gonna fill it out as well. So is it, you're just starting, you're actively researching, you're already using it or you're using it extensively and you're a leader in the space. Great. So as you guys are filling that out, that's great to see. And honestly, it looks like a lot more people are experienced with ai, which I love.
Since the last webinar we didn't see as many. So that's great that everyone is starting to adopt it and really starting to say, wanna learn more. So now let's meet our amazing panelists who are here with us today. We have Will Ro, who is an operations leader at Physical Chicago, overseeing 18 clinics and 64 clinicians.
Thank you so much for joining us. Will. Pleasure to have you. And we also have Dr. Steve Ryland, who is a board member for the United Physical Therapy Association and CEO at Align Physical Therapy, which opened three clinical locations in under nine months. Steve, such a pleasure to have you as well.
Steve: Thanks, Marla.
Glad to be here today.
Marla: Great. And really, you too are leaders that have gone beyond theory and are really putting AI to work in actual clinics with real results. So can't wait to dive in and ask you some questions. But before we get to your feedback and your responses, we also have Adam Baliatico, who's the co-founder of Prompt Health.
Adam, great to have you.
Adam: Thanks Marla. Looking forward to being here,
Marla: and Pedro Teixeira, co-founder of PredictionHealth. Pedro, thanks so much for joining.
Pedro: Hi, happy to be here.
Marla: And I'm gonna turn it over to both of you to just share the story behind Prompt’s merger with PredictionHealth, and just about how it's combining missions and the reason we're doing that.
And I'll let you guys take the stage.
Adam: Awesome. Thanks Marla. I can get us started and then pass it off the Pedro, but yeah, Mike and I, we started Prompt around the same time that Pedro and Ravi actually started PredictionHealth and we've known them and Adam and a lot of the other early PredictionHealth team members for, it has to be more than five years now.
And whenever we think about products and acquisitions and things like that, first foremost, and, probably first, second, third, fourth on the list are the people in the culture. And, it's always been amazing to work with the PredictionHealth team as partners. The visions were so closely aligned that the mission was around helping practice owners, clinicians, and patients achieve better outcomes.
Allowing our technological solutions to make the rehab therapy space better and putting our customers first. And that was something that we always shared between Prompt and PredictionHealth. And as we've both grown over the years we were really able to accomplish a ton separately as partners.
And we had this great integration and we were both building a ton of AI and automation to help empower the rehab therapy practices that we serve to be more efficient and deliver better experience. And ultimately, what kind of became obvious in the past few months was that by combining forces we could build something like truly special that over the coming months and years, we'll actually revolutionize how the rehab therapy practice operates.
Both for prompt EMR clients, and we'll talk about this more, but also for non-pro EMR clients, there's so much that by combining our two AI teams. We can accomplish regardless of the EMR you use. And we're really looking forward to talking more about that today as well. Pass it off, Pedro.
Maybe you have more to
Pedro: add there. Yeah, so when we started right back in 2017, right around when you all started, it was this mission that we had. We wanted to build useful tools. We thought that AI was a really powerful one into the electronic systems that clinicians use to take care of patients.
And, it's great to meet another company, right? That isn't just, saying that, but that actually has shown over the years that they really do prioritize that. And so that alignment I think was really exciting. But then. As we've gone through the years, we've done really well by listening to our clients, right?
What can we do to help you? What do you wish the AI could do to make your life easier and help you take even better care of patients? And just time after time, especially more recently, there were just these things that we could build a lot better if we were working together, right? To just build it straight in and have, certain data sets available to make certain features really seamless and enjoyable and easy to use.
And then to expand beyond what we do for the clinician to other folks, right? The administrators that you know, et cetera. Like a lot of people, basically everybody that uses the system to take care of patients would benefit from some AI help. We could all use a little sidekick as we like to joke.
So really great alignment and really excited about what we're gonna be able to build together. It's just another way for us to make our clients happy, and helping people that help patients is a very rewarding thing to do. So we're super excited and we hopefully hopefully you all are as well.
Marla: Yeah, we're really exciting for this dream team, which is combining missions where both organizations are just really tackling the same core issues of how do we get back to therapist, reduce burnout, enable smarter clinical and operational decisions. Excited about all of that, but we wanna hear from the clinicians and owners and operators that are doing that and that are actively using it and find out some real results and outcomes.
For our two guests I'd love to just ask you guys, what are you using AI to do in your clinic, and what results are you seeing? Steve, why don't we start with you and tell us a little bit about how you're using I AI and what's happening with it?
Steve: Yeah, so we're using both Prompt and PredictionHealth, which is really good 'cause the two of 'em work very well together.
But we're running it for everything, basically we're doing it for documentation, we're doing it for chart audits. 'Cause let's be real. These insurance carriers these days, they're getting, for some reason, physical therapists are getting picked on. We just went through a big Medicare targeted, proven educate audit.
And if you've ever been through one of those, I'd rather have all my fingernails pulled out. But we passed it with flying colors and I think I wrote you guys an email. I said, man I was stressed. And it was right around Christmas time where they pulled 40 charts and I'm like, oh my God. Being new.
I've been doing this for many years. Private practice. I've sold my first practices in 2015 and then I went and worked for corporate America, ran hospitals, did all that stuff, really missed private practice, came back in, scaled three clinics super fast, eight months, three clinics up and operational.
We didn't have time to go in and do all the audits and, 'cause I don't sleep well at night. The two things that stress me the most is where's our next payroll gonna be coming from? And are our charts clean? And I used to have somebody full-time, my mom used to do it for me when I ran my previous practice, and she would go through charts every day.
We're finding 20 to 25 mistakes a day, not only on improper CPT codes, but improper documentation is medical necessary. And when we run these reports, I can look at my charts and say, we've gone from an 80 low, 80% compliance rate since using PredictionHealth to upper nineties, which is unheard of.
And because it tells you not only as a practice what are you doing correctly or what could be improved on, or each therapist, you can go to each therapist. I got some therapists that are really good at putting the keywords in and what the insurance companies are looking for. But you can go into each therapist and monthly I can give them a report and say, Hey, this is the areas you need to focus on.
This is what PredictionHealth is telling you need help on. Whether it's flow sheets, whether it's, and I'm telling you, when we sent those notes and I got the phone call, I was scared to death when the lady called me from our carriers First Coast Service options up in Maryland, Virginia, I believe.
But she goes, I wanna be the first to tell you. She goes, I've been doing this for 15 years. You were the first one that I have not had to call and ask for more information. Your notes were so I, we passed the first round. And that gives you a clean bill of health for 12 months. But it was scary.
And knowing that behind the scenes that a, we were using PredictionHealth to, to be able to go in and look at all that stuff and knowing. That you can put your head on a pillow at night and have to not have to worry, what are my therapists putting in their nose? What are they doing? It basically, it writes it out for you.
You run those reports, you can send those reports to your therapist, they can correct what they need to. Correct. And it's awesome. It's really, as a practice owner, putting me more at ease knowing that I have that running behind the scenes on every day, during every node.
Marla: And as a clinician, having that training, almost as we say, like that assistant, that mentor right there through ai, it's actually teaching them what they didn't document correctly and they're learning from it.
So you're not investing in all these training resources and courses to everybody. It's just targeting the people for what they want.
Steve: And basically I say it makes 'em a better clinician. We're not taught how to document in school. Maybe we get 15 minutes of it, but you're just not, and then you go on your clinicals.
You may be taught how to document properly, but. The way we're going to protect our profession because physical therapy, we all know this. I'm, I joined LinkedIn groups. I follow all the private practice, physical therapy. I've run hospital outpatient departments before. We're not, they're not coming to, even though we save billions of dollars every year to these health insurance carriers, they're still picking on us.
And they're not gonna stop until we give them a reason to stop. And that's why where AI comes in, because when they start auditing our notes and they can't find anything, they're gonna move on to the next target. But we need to, we've been really bad for years on documentation and they know it.
And these are easy target. It's low hanging fruit for these insurance carriers to go in and just pick on us.
Marla: Yeah. And will, I know you, you're seeing the exact same thing here, but you have 18 clinics, so would love to hear from you a little bit about how operationally and those dashboards and analytics have also helped your practice.
And any metrics maybe that you've seen change from that.
Will: Yeah. I don't have a clinical background, right? We got into this business more from the business mindset, and for me, what sidekick does, I always joke around, it's like almost like you're in your high school project and you got paired with the smartest person in class, or you even got paired with the teacher, right?
And so it's just this resource that you can rely on. And it, for me, it reduces the ambiguity in everything that you do, right? So when you are sitting down with your therapist and you're coaching them on, Hey, these are the codes that you messed up, this is the justification of coding that you messed up.
It's just, there's no debate on is it right or is it wrong? And particularly, I've been using AI across the board in many different aspects of my life, but prompt and PredictionHealth work so well together. That you really almost don't even notice that it's there. And it's just this constant thing that's really following all of our therapists in terms of, Hey, are we doing everything right?
What can we do better? And this is how you do better. In terms of auditing, like Steve said it's world's different than what we were doing before as opposed to, here, let's take a couple charts here every month and review. This is going after every single note that all of my therapists are doing.
And for me, it saves a lot of time, right? We've got 18 clinics, 64 clinicians, for us to implement something, train everyone across the board that could really honestly take a month or multiple months with this feature. It's like the snap of a finger. And so one thing that I am trying to move towards in the physical therapy realm is.
I feel like we are slow to get information across the board. Okay. A month later, two months later, we're finally figuring out a problem was happening with prediction elephant prompt. It's instant and you can fix problems in a week of them starting as opposed to 2, 3, 4 months down the line when you have a real issue.
And obviously we all know reimbursements are declining and this is the major function that we use to help protect ourself from sliding reimbursements.
Marla: And either of you, would you say you've seen any increase in revenue per visit or any way that, you've seen that ROI come back with the training and with the coding
Will: almost instantly for us, like the same
Steve: week.
Yeah. We've seen a $4 increase in our average reimbursement come back. 'cause I'm a big statistics guy. I love to run statistics and I know pretty much within a dollar, you know what our reimbursement, we've seen a $4 increase since using PredictionHealth and Prompt together because basically.
It'll tell you, the therapist, the AI technology is telling you, Hey, that code that you build really should be built therapeutic activity or neuromuscular reeducation. Instead of thera therapeutic exercise, the therapists tend to go lean towards that therapeutic exercise, which is really a less skilled code to use.
But if your documentation backs up neuro or therapeutic activity, it's a higher reimbursement code. And basically the AI is doing that for you, where it's making sure and protecting that if you do get audited, you've got the documentation, the backup to show no, that was actually functional activity, or it was more balanced and coordination training versus a just a straight line exercise.
Marla: Yeah. A absolutely. And I think, like you said, it's just helping clinicians be better at what they're doing and Bill for the time that they're giving. And wanted to ask you guys a little bit more about that too. How are you using AI for your clinicians and what they're doing in their day-to-day work?
Will: On that first, go ahead. Go ahead.
Steve: Sorry.
Will: Go ahead. Yeah for us, so while I'm not a pt right, I'm not actually writing notes, but I, the The interesting part about PredictionHealth Sidekick prompt is for me as like a business owner the secondary things that it allows you to do that you could not do before.
And I can give you a prime example, we. We really believe that we want our clinical directors and our clinicians to be out in the field meeting doctors helping us pull in those referrals. And with prompt PredictionHealth the whole shebang, we have actually reduced the amount of time that we have to set aside every month for them to do chart audits for them to do coaching and whatnot.
And we actually had a 7% increase in referrals because we've allowed our clinical directors and our clinicians to actually go out and see doctors a lot more. And so there, there's just like a, almost a never ending possibility with what you can also do because you're running this stuff. And as Steve mentioned, it's not, there's no questions in the numbers, right?
It gives me the peace of mind every day that, that, that is running all of my audits, it's running all my coding. And if we can document correctly, then we justify all the care that we got. And like Steve we call 'em top codes. Our top codes increased 9% since we implemented you guys.
Marla: It's a big difference.
As we said, those reimbursements are going down. So getting any lift you can is important. And Steve, you were saying the same thing with that is now that clinicians could use the other part, which is sidekick, how are you using that AI piece, whether it's the AI intakes, the Sidekick for actually treating your patients?
Steve: Yeah, it's cut down on documentation time significantly. I, we've seen a 50% decrease in documentation time. It's nice when my therapists we work shifts, we're open seven to six. Some come in at seven, some leave at four, some come in at nine, leave at six. But when they're walking out the door and they're not having to go home and do documentation, or they're having to stay late to do documentation, or we're having to track 'em down saying, Hey, you can use that data too.
If somebody doesn't finish a note, it's gonna tell you that therapist didn't finish a note. It's also helping you to get that money out faster to insurance carriers. We bill on a daily basis, so it's helped, it's really helped increase our revenue flow and getting money in faster. It's, we've seen denials drop by 23% because we got the correct coding, we've got the correct CPTs going out.
It's making sure it's telling you, Hey, these two codes or these two diagnosises, they're Medicare's kicking out diagnoses now. So those two diagnoses don't go together. So it's smart intelligence where you're not having to bill those codes out and get the rejection and then have to wait another couple weeks for the insurance to reprocess it.
It's saving you, especially as a business owner. But the, I'd love seeing my therapist because it stressed me out watching my therapist, just because I'm a nit picker, as a private practice owner, it's my business. I've got to protect my business. It's bigger than just one person, but knowing that they can use sidekick.
I'll give you per example. I did two evals this morning, and by the time I walked out of both evals, my evals were done. And not only were they done, they showed function, they had functional goals. I cannot tell you how many times I've had to look at therapist goals and go, there's no function in that goal.
I don't, it's great that they can lift their arm to 140 degrees, but what does that mean? They're able to screw in a light bulb. They're able to put on their shirt, the AI does that for you. It writes the functional goals. It writes the functional, how did they respond to their therapy session last time?
How did they respond to today's session? It writes that for you, and it's doing it, it's crazy. It's almost like having a robot standing over your shoulder that's listening in on your conversation and it, it writes exactly what you were going to write. It's almost it's it's cool.
I never thought that we would see it. This quick in the physical therapy industry. But again, I can't tell you, we have to protect our industry. I'm a big proponent of that, and I'm telling AI has just been, it's been a game changer. It really is. I don't know how we did it before this.
Will: Yeah. I can give you a real life example too.
For us, we have a so obviously we have 18 clinics, people off sick vacation, whatnot. And so we actually have internal travel therapists that don't really have a clinic. They're bopping around all over the place. And because of this one, their life is, a lot of it is spent in traffic. And not only that, but when you're in the clinic.
They're pretty much spending all their downtime just catching up on the notes and learning about the patients. One of them was actually about to, was like, Hey, this isn't working for me. I just don't have a work life balance. Because she was having to do all her notes at home.
We implemented sidekick and actually got to retain an employee that I did not wanna lose because she got her work life balance back. So just based off that, it's invaluable.
Marla: Yeah. And especially with, as you said, those people covering when someone's out the AI insights right at the top, it's it's telling you exactly when went on in that patient, the past history, what the trends have been, and any compliance issues too.
As a therapist, I really value that. If you're picking up somebody else's patient, for sure. Yeah.
Steve: Flow sheets kill us. It'll tell you up there, Hey, your flow sheet hasn't been changed in two visits. You gotta change your reps, you gotta change what you're doing. Again, medical necessity, it's all about medical necessity and having those, right when you pop on that chart, it's saying, Hey.
The last couple visits you haven't changed, and that's nice to have right at your fingertips.
Marla: Yeah, absolutely. And I know Steve, you're very passionate as I am about PTs becoming more into that primary care provider space and they really do need the assistance, again, assistance of AI to be able to see more evals and not take as much paperwork home to be able to do those higher value tasks.
So we'd love to hear from you. How much has changed in terms of how much paperwork is going home at the end of the day? How much burnout do you have? Any real life examples or even just retention of employees? Like the example, but in numbers?
Steve: Yeah. We have not lost any therapists, which is great, but again, we've seen a big reduction in documentation time, 40 to 50% reduction in documentation time.
Notes are done in the morning, notes are done in the afternoon. People are walking out and going home. It's, again I don't, I used to spend, especially scaling three clinics in eight months. I was treating patients as a full-time clinician and then going home at night and doing my admin stuff and finishing up.
It was like my wife was like, I don't have a husband anymore. She goes, I get it. I know this is your passion and you love private practice. But since we've, we went with PredictionHealth about a year ago, that was cleaning up a lot of it. And then when we went with Prompt and merging the two together, it's like my documentation.
She's don't you have documentation to do tonight? I'm like, Nope, not tonight. Don't have it. And being able to run my statistics and be able to see which my therapist, what codes are they billing, is there any variance? Is there anything that we can do to improve? 'Cause we're fighting for every dollar now.
My, it's nice seeing my therapist. I've seen the mood in my clinics, like we are very upbeat, we're very happy. There were days where I could see, somebody called out sick and they had to bring on a couple extra patients and move it in. They don't mind so much now anymore. They do what they need to do because they know with sidekick running and with the AI kicking in and helping them with their notes.
It's, I used to walk in and I'd look around the room and I see my therapist with their faces buried in their computers and I'm like, this is not healthcare. This is not what I signed up for physical therapy. Now it's nice. Their computers aren't, their faces aren't buried in their computers.
They're talking to their patients. They're educating their patients. They're doing more billable time because we know documentation, it's not reimbursable. So if we can do less, re less documentation and more hands-on and more education, that alone increases your revenue per visit.
Marla: Yeah. Yeah, exactly that.
That's spot on. And will, what about you? Any unexpected wins in terms of recruiting, retention, or burnout that you've seen in all of your, I said, I think you said 60 plus therapists, clinicians.
Will: So when we started, we had probably 80% of the p my therapist, they actually wanted a sidekick. The other 20 just said, no, I don't even want it.
I can tell you right now, we have a hundred percent of our therapists on it. And so for me, an unexpected thing was again the, I can't stress enough, like the integration between PredictionHealth and Prompt is so seamless that some of my therapists who have been in the industry for a long time, who, it's a struggle to even get 'em on a Zoom with with their technology and whatnot, they actually adopted it and use it probably better than anyone else does because it.
Takes away from the actual computer work that you have to do. And so employee retention employee engagement, efficiency across the workforce, it's, it, they all loop into this seamless process that is prediction, health and prompt. And so yeah, there the possibilities are endless, right?
And I think it really, for me, it saves so much time for my clinicians that, I've got 18 clinics, each one has its own problem. And so my clinic directors, my area directors, they're able actually to hone in and focus on those issues. And one, it has made training training new therapists a breeze or much easier than it was before.
And as you said, it is a recruitment tool, right? It's a competitive industry. I think during COVID, about 20,000 therapists left the industry and we have not recovered from that. And everybody's therapists, they're getting offers from everyone, right? And so this is just one added way that one, we can get therapists, and two, we can keep those therapists.
Marla: Fantastic. Thank you. And I have some questions coming in. I know we were gonna answer them at the end, but I think this is a really good one, just saying, taking a 30 foot level, what is prompt and prediction? It seems like an EMR and steroids. But just wanna know a quick overview. Adam Pedro, if you wanna answer that for us.
Adam: Yeah, I could start EMR on steroids is probably a good, three words set, description of it, but at a high level, maybe like AI powered operating system that's inclusive of the EMR. Patient engagement, practice management, billing software, obviously analytics like Steve, Steve has been talking a lot about on this.
And then the AI components of, a, sidekick and practice intel. And then, we have a lot of AI around filling the schedule keeping patients engaged with their care. And just really an all-in-one. Everyone says it's an all-in-one platform, but I think really like encompassing the ability to run the large majority of your business with a single tech stack that's really optimized for delivering great outcomes for the business and for the patients.
Pedro: Yeah. And then to dive in a little bit more on the sidekick and the practice intel piece. So Sidekick is just within your EMR, so you know, you're impro, you don't have to go to a different place or anything. There's just a little widget there in the bottom corner. And when you're walking to the room or when you get back to your office, you're just hitting the, you.
It'll record button, and then either you have a conversation with a patient and let it just run in the background. And when you hit stop, it'll just draft that note in a few seconds so you can review it quickly and easily. And then accept the entire sections, as you're going through which, is probably why folks are getting that big efficiency jump.
Or again, sometimes you may decide you want to quickly just tell it. Or, in your office you're catching up on a note maybe or something. You can also do a dictation, but not like a comma period, whatever. More if you had a really competent human scribe with you and you, and they maybe just weren't in the room, you'd be like, oh yeah, I just saw them.
They're doing great from last time we did this, and that. Here's how we progressed today. And then it just goes and writes all that and puts it into the right spots so you can just quickly click to accept. Perfect. That's the sidekick documentation assistant. And then the practice intel is AI models that are reading the notes.
So for all this compliance stuff and the CCPT coding that folks were mentioning, these models that we built to be able to read the sentences or the notes, et cetera, and then give you back analytics that incorporate that information that it read from all the charts either about your compliance or your CPT coding.
Marla: Great. And that's where you get those insights. Yeah,
Steve: And to his point, one of the reports I ran when we first started using PredictionHealth was that 73% of our therapeutic exercise was actually more neuromuscular reeducation. So just making that one change and now we're down to 10%, but I've watched my revenue for dollar increase, which is amazing.
And you didn't even talk about the intake part. Yes. So both of my patients this morning that came in had already had, you can push out. The objective testing, the lower extremity functional scale, so they get all that stuff ahead of time, so you're not having to do that.
You're not having to sit there and type all that stuff out. Both of my patients had filled out their intake form. My eval was 80% done before I even started speaking to the patient.
Marla: Actually, I do wanna touch a little more on that because we have a question about it is how do you use both Prompt AI and PredictionHealth and how do they compliment each other?
And I don't know, Adam, if you wanna talk about the intake and how it goes into the sidekick or any of our
Adam: yeah, I can start on that and then maybe Steve can actually talk about how it played out today with his patients. Histor historically, like the intake paperwork, 10 years ago it was paper.
It was obviously a static piece of paper. It can't be dynamically changing the form as the patient's answering it. But then even as the world shifted to like digital intake it was digital but still static, right? Like you send a form to a patient and the set of questions on that form are gonna be the questions they answer.
And yeah, there was some like basic if then logic, if they answer this, ask this, but. The problem we always solve when we talk to clinicians is I had my patients spend 20 to 25 minutes filling out a digital intake, but they come in and I still have to ask 'em a ton of questions to actually learn about their condition.
What we've built at Prompt, and this part is specific to prompt EMRA. A lot of the stuff we've talked about so far works, whether you use prompt or not is AI powered intake questionnaire where based on how the patient is answering questions it'll ask them different, more relevant questions to learn about, their prior level of function, their current level of function.
And what it allows you to do is, like Steve said, when the patient shows up, you've actually asked them less questions than you used to ask them. But with the information they provide, you can draft, some, between 50 and 80% of the eval, their current level of function, prior level of function an as assessments comorbidities.
It will recommend long-term goals based on what the patient's told you. Then when you combine that with sidekick, I think that's why Steve was saying like most of the time his eval's done before the patient leaves because one, already half of it's written when the patient shows up, and now you're just talking to the scribe as you go.
And by the time the patient walks out, you've got very little or nothing to do,
Steve: really just your objective. That's the only thing that you need to do. It picks up on the subjective, the assessment, the plan, the goals, functional goals, which are, anybody out there who is a private practice owner or clinician knows we are horrible at writing goals.
Sorry, everybody, but it's true.
Adam: I think the objective is actually as we start talking about the roadmap that's something Pedro and I and our teams have been working on is being able to take your conversation where you could just say, their shoulder flexion was 35 degrees, like populating that field in the EMR, whether that's prompt or another EMR or like we did, three by 10 squats, like actually creating that as an exercise on the flow, the digital flow sheet.
So that's a major part of our kind of near term roadmap with sidekick, both for prompt customers and non-pro customers alike.
Marla: We actually have a question here are you allowing new graduates to use the AI documentation? And just to let you guys know, we are giving this to universities, but we're telling them not to turn on the AI until their final third year.
But would love to know from you guys, are you letting new grads use Scribe and all of these intake features.
Will: Yeah, we do. We make them wait a month, so let's make sure that they know how to do it. Without it, but yes a hundred percent it's almost really mandatory for us. It just doesn't make sense for us to not use it.
Steve: Yeah, same here. We, what I'll do is, whenever I get a new grad, I want them to come in and shadow with the senior therapist first, whether that be a week, two weeks, whatever they need to get comfortable. 'cause you just came from school and you don't really have, yes, you've had your rotations, but you're still walking in that first day and you still very uncomfortable, et cetera.
So that's what we'll do is we'll have 'em shadow with the senior therapist until they're comfortable and then start gradually building up their schedule so we can mentor them a little bit more. But it's really helped us in attracting new grads to come to us because. Not a lot of clinicians out there are using the AI technology and the documentation, and it helps us because it's like, Hey, that's one of the things that your hearing therapists say the most is, oh my gosh, the documentation component of physical therapy is horrible, and this just really helps Big time.
Will: Yeah. Especially in our busier clinics, one thing that I think isn't set enough is that you can actually use sidekick for two patients at the same time. And so it's funny. All and all of our busier clinics, when that month hits that, that new hire, I almost get the email like exactly one month after.
Because everybody else in the clinic has it and they're obviously pushing to get it a lot, but especially when you're seeing multiple patients at the same time. It is it just changes everything.
Marla: Great. Great. I like that. That process that you're doing, that makes sense. Then they're eager to have it.
'cause they know the frustrations. You're not just giving it right to them. And what about Pedro, this one's a little bit more for you, is can you discuss how the AI has been trained and who validates the things like functional goals or indeed what we want? So we'd love for you to dive into that.
Pedro: Yeah. It's a really good point is that you can't just call an AI model and expect it to be just like magically, perfect. So we spend a lot of time having either subject matter experts. So for like the compliance model, we labeled huge numbers of examples of here's a compliance statement, here's not a compliance statement, et cetera.
And then having those subject matter experts. Then also review notes and then like confirming. So then same thing for the documentation, like what makes it a compliant, thorough note? Here's like the ideal version. And like then comparing what the human, taking their time.
Writing a perfect note is versus what the ai, what sidekick would generate. And making sure that their functional goals that they are accurate, that it's comprehensive, that it's factually complete. And then doing analyses on okay, let's look at all of the facts that they mentioned in the visit and the facts that the therapist documented and where they documented it to.
'cause again if it's subjective information, you want it in the right section, right? You don't want it in the wrong section. And then confirming that what a, what an experienced human would do that the models are mirroring that. And then continuing to quantify and confirm. And then if anyone ever sees something.
That they think is not as a good an output. They can always give us a thumbs up or thumbs down on all of our outputs. And we review those every week to see okay, it's never good enough, basically for us. So a lot of processes to make sure that it's great and that it continues to get better.
Marla: Great. And the therapist is the final say. They choose to put it in or not, they review it and they can add and edit it. So it doesn't just go in there, you actually do insert it at you. You click and say, yes, I wanna add this.
Steve: And it's been, Medicare is very, if you go to the Medicare policy manual, they're very specific about what they want in evals, what they want in goals, what they want in prior level history.
All of that stuff is built in. So you don't have to worry about is my eval or is my daily note? Is my progress note? Does it have the components that Medicare requires? And every insurance company's different. This puts you at ease knowing that all the components that Medicare policy manual requires of a physical therapist is in there.
And I can tell you the goals I've had to change very few little goals. I'm a, there's four components to a goal. Who is the goal entail? Is it the patient, is it the caregiver? What is the function behind that goal? What is the deficit? A lot of people are really good about writing deficits, like I said before, 140 degrees of shoulder flexion.
What does that mean functionally? And then there has to be that therapy qualifier. Why does that goal require the skill level of a therapist? It puts that in the goal, proper scapula, thoracic rhythm, all of that stuff. Yeah, it's really it's more than what I expected when we first started.
I thought I was gonna have to go in and start changing some of this. I really don't have to do that.
Marla: And we've got a lot of questions coming in. Now they excited to hear all about the sidekick and the CPT coding and the AI insights, but they wanna know what else are we doing or are you using for administrative burnout or patient engagement?
What else are we using AI for in your clinics? And will Steve, I'll let you guys jump on that and then Adam you can chat on it as well.
Adam: Yeah,
Will: go ahead Steve.
Adam: I was gonna say maybe, sorry, like with how PROMPT thinks about it and then, will and Steve, you could talk about with how it plays out in practice.
So I'm glad you asked that question. It's actually when Mike and I started Prompt and we sat, we spent like a year and a half sitting in PT clinics and, we all obviously saw the therapist struggle, but one of the things that blew us away in a negative way was. How much work, especially the admin and the billing staff had to do to keep the clinic running and like mountains of paper and they were making a thousand phone calls a day to schedule patients and there was just so much stuff that it felt like technology could solve.
And we really try and the clinic is so interconnected, right? Like a disconnect. A patient who's not engaged doesn't just hurt the therapist, it means you're making less money or not getting an off doesn't just, mean it's an admin issue. There was clinical billing impacts, patient impacts.
So we've really built this interconnected system that streamlines the administrative workflows, automates a ton of scheduling, eligibility, billing interactions. And allows everyone to focus on, like delivering that exceptional customer experience and less on the tasks that, automation and AI can handle for the practice.
So I, I think we've really done a great job on, reducing that administrative and billing burden and keeping the patients more engaged in their care. But, we'd love to hear from Will and Steve, how that's played out in practice.
Will: Yeah. Admin wise for us there's several different ways that just the package deal with PredictionHealth and Prompt helps.
I would say the first list, or the first and most major thing is the wait list function, right? So I think the worst thing for any practice owner is those blank spots on the schedule, right? And prompts wait list feature. You can actually go and every time we check in a patient or for their initial eval, we say, okay, this patient wants to call 'em in the morning or in the afternoon or know what specific times they want to, and immediately when we get a cancel, we can shoot a blast out to all of our patients.
So immediately what that does is now my we call 'em CCCs isn't physical, but my receptionist does not have to call 50 people to fill that slot. And just off that time, save alone, that's an hour to two hours every single day. We also use the kiosks, or we're testing the kiosks now. Especially in our busier clinics where you've got one receptionist and you've got three or four people checking in, checking out, they're able to actually reduce the amount of time that they have to spend with the patient because somebody can check in.
If someone's here for their 13th visit, they know the drill, right? And they don't really have to talk to our front desk at all. And so in terms of getting the patient in, getting them seen, getting information out to them, and also adjusting our schedule, which I personally believe is, one of our top five things that you really have to get right in a PT clinic.
It assists them in all that. And then what that allows is now my front desk has a little more free time to maybe wipe down a table in between in between patients and whatnot, which then. Allows my clinicians to have more time to do their notes. And I think Adam hit it right on the head. But everything is so interconnected that just these little chunks of time that you can save throughout the day, like I, I really think it probably saves me, like a whole employee across the board.
And it just makes everything more efficient. It's efficiency on steroids really.
Marla: And I do wanna point out that AI wait list feature. And with the online scheduling, it actually does look for the highest value patients, meaning ones that their off has already been approved, that they are maybe a post-op or a certain insurance that is, is right to put in there.
So it does go through and use AI thinking to, to ask the patients that fit that time first, which is really nice.
Steve: Yeah, and from my standpoint, yeah, 10th visit Progress Notes. It lets you know the 10th visit Progress notes due today on a Medicare patient. Plan of care is expired, and it gives you, you can set reminders, whether it's two visits before, three visits before, it allows you, but it's a big red banner that pops up and says your progress note is about to expire.
You need to do a new progress note. I have a friend who just went through a targeted probe up in the Northeast and his progress notes weren't done on 10th visit. That's one of the biggest thing Medicare's looking for right now. Are you doing your 10th visit Progress notes? He just had to pay $45,000 back to his Medicare carrier because of non-compliance.
And that's a game stopper right there. You gotta write the check. Now. He's probably, I told him, I said, you're gonna have to go hire an attorney. You're gonna have to go to an administrative law judge. You're probably gonna get the money back if you can prove, you know why those were late. But we shouldn't even be in that situation where we're having to write a check back to the carrier for services provided.
So I really like the fact that it's telling our therapist, Hey, your progress note's due. It's due either in two visits or it's due tomorrow, or it's due with the next visit. But at least they know, Hey, I have to go in and do my 10th visit progress note. My plan of care is about to expire. I need it. I need it.
If I'm gonna keep that patient, I have to extend a new plan of care, a new recert, send it off to the doctor. That's the other thing, the fax feature, it automatically faxes it to the doctor. It tells you when it's back. It'll tell you if it's not faxed back. Are you being? Are you getting that plan of care?
I know the law's changed for Medicare in January, we no longer have to have a signed plan of care, at least for the first round if we have a referral. But that second plan of care still needs to go out. So all of those features as a business owner, I can go in and I can run those reports. I've got a plan of care compliance report.
I know what plan of cares are about to expire. I know KX modifiers, how many of my patients are approaching. It'll tell me, A, how many patients have passed the KX modifier and B, how many are getting ready to approach the KX modifier? So you can go to those therapists and say, Hey, what are we doing with this patient?
Because we knew we now have that 3000 target threshold. When you go over $3,000, doesn't mean they're gonna ask for the notes, but it means there, there's a very good likelihood they're going to ask for the notes. Knowing that your documentation, your progress notes have been done, your plan of care research, the doctors, if you need a sign plan of care, which is one thing we're working on getting rid of.
Hang tight on that one, but knowing that all of that is done behind the scenes, I'm not having to worry as a business owner, because the last thing we wanna do right now is if we're struggling for every dollar we get in the door, having to write a check back to the insurance companies and Blue Cross Blue Shield's doing it too.
We're getting audits from Blue Cross and Blue Shield. I'm more comfortable knowing that I'm sending those notes out, that there's a very good likelihood that nothing's gonna come out of that audit.
Will: Yeah, and even expounding like a little more on what Steve said, I think that AI has allowed I think a common theme that I've noticed in the physical therapy industry is it's mostly backwards facing, right?
Okay, this happened. Now we realize now we change. AI has allowed me to look forward in certain things. And he mentioned the plan of care compliance. It's very obvious and I can give you an example. Hey. We've got a lot of people who are not getting in as often as they should be. Why is that?
And so for us, we've got 18 clinics over the last six, seven years. We are constantly in a ramp phase for all of our clinics and we have actually been able to reduce the amount of ramp because we know when to hire that next therapist much more efficiently than just, Hey, I think this is the right time.
I think we all know that the PT industry is somewhat seasonal, right? And so when you hire that, that employee and you're about to go on a down on the downswing of the seasonality it really can affect a younger clinic that only has one or two clinicians. So we've been able to hire better at the right time and reduce the amount of time that it takes to ramp a clinic considerably, like by six or seven months.
Marla: That's great. And Adam, I'd love if you tell us a little bit more about some of the AI features that just came out, like the AI no-show and AI discharge like you said, will, it's very predictive. So those are some of the AI features that we're doing for predicting outcomes.
Adam: Yeah, absolutely. And some of these are in early release, so if there's prompt customers on here, there's a chance you may not particularly have access to them yet, but we'll be rolling them out more broadly soon.
One is around no-show prediction. Obviously a no-show is a very expensive cost to the clinic. Whatever you make 90, a hundred bucks a visit every no-show, you have no opportunity to fill that spot. So we built a AI model that can predict, reasonably well actually. If the visit is a high likelihood to no show.
And then we have two kind of pieces of functionality baked into that. One is we send an automatic additional reminder to that patient. It doesn't say Hey, we think you're likely to no show, so we're sending you a reminder. It just sends them an extra reminder. And then also a list and a dashboard tool for the admins and the clinicians to be able to see so that maybe, they're in the clinic today, but their visit tomorrow's flagged as a no-show risk.
Like you can have those conversations and make sure you're following up. We're also gonna be extending that to drop off. The patient's just gonna entirely stop coming to care and, we'll hopefully be launching that soon as, as well. And then building on the clinician experience drafting, like no, no show discharge summaries we'll be rolling out shortly.
So using all the context of the case and actually some of, the sidekick functionality combined with some of our AI functionality to draft a discharge summary that's compliant. That will be rolling out soon as well, since, that usually takes five or 10 minutes for a therapist and it's for a patient that stops showing up.
So it's not a revenue producing task, it's purely a compliance task. And now with this AI model we'll draft it, therapists and spend 30, 40 seconds reviewing it, maybe tweak what they need to tweak and then finalize it and keep the team compliant without having to spend all this time on a non-revenue producing task.
Marla: And I always wanna bring it back to the clinician who that is such a matter of quality care because I know I would, if somebody was about to drop off, you'd be calling them, why aren't you here? Weren't you showing up? This allows you to predict that and talk to them when they're in the clinic and really change your plan of care to help them continue and complete and get better.
Steve: Yeah, and the email functions too, where I can now send out an email to my patients, like once they've fill out their intake paperwork, we know they're now a scheduled patient. What to expect for their first therapy visit. What to where, what to bring. It's a nice introductory email. After the third visit, we send an email, you've now started your physical therapy, is there any questions we can answer?
Et cetera, et cetera. If they've fallen off the schedule, it's gone 90 days, I send 'em an email just checking in on you, how you doing? That's all. Run through the ai, it's all through the system, which is so nice. I don't have to have a person that's picking up the phone, calling 'em, or stuffing envelopes and mailing 'em or whatever, like what we used to do.
So that's a big feature too that I love.
Marla: Will any more insight from you about that and also just about how to even adopt these and the how either hard or easy it is in your clinic to do that?
Will: Yeah, I think just the general theme is we're moving towards ai across the board and from what I've seen, every high functioning company is getting involved with AI to move forward.
And so I just think that if you're not. Onboard, you just will start lagging behind a little bit because of all the functionality that it does and really I hate to be vague, but it helps everything. It's helped me hire, it's helped me keep employees. It's helped me. It's helped them become more efficient, have more time at home.
It's just, it's really been, I would say probably the most influential thing that we have done at Physical since we started. So you just, it's just something that you have to leaned on
Marla: and Adam, Pedro, any, just what's on the roadmap, what's coming out? I know we haven't even touched all of the pieces of AI in the platform, but would love to turn it over to you guys to give us some future of what is coming out and maybe what we didn't discuss yet today.
Adam: Yeah. Pedro, why don't you,
Pedro: Go ahead and take this. Yeah obviously as one entity, there's a lot of things that we can do in a much more seamless way, so we're really excited about that. You'll see things like the objective findings or the objective exam items that people ask Hey, in prompter structure it'd be great if sidekick nicely connected with that and enter that in.
Same thing with the intervention and stuff too. So that's obviously have a really high priority getting it baked straight in so there's no extra steps or anything. They're turning it on within Prompt is another really exciting bit, and you'll also see better integration of some of the native AI features, right?
Like the intake with Sidekick, et cetera. So that, all of that just seems more and more like one service. And then broadly we also have lots of other improvements that we're doing around making the results more custom in some thoughtful ways for clinicians. And then this also opens up the door for everybody, for us to continue to really improve the quality of the models.
Obviously the more data and the more training and opportunities we have with the system. The better it gets for everybody which is really exciting to do too. And then on the analytics side that's another thing where you should be seeing that become, again, more seamless, more integrated across the products.
So for the compliance model, for the CPT insights model, for scanning your data, working to make that easier and easier for folks so that's something that, doesn't require, almost any lift at all from folks as we're able to just build that right into the data processing at prompt.
Marla: And I think that's so important, not just to have ai, but have AI integrated into your systems because nobody wants to sign into multiple platforms or use it here and then jump over and none of it's connecting. So love how we are continuing to make it more seamless and usable in your daily day-to-day tasks without adding operational hurdles.
And Adam, what about you?
Adam: I think Pedro hit on a lot of things and multidisciplinary support and especially prompt has really grown a lot in the pediatric space and, obviously it's such a different care model. You can't just, take your eyes off the kid and start writing a note.
So expanding sidekick as quickly as we can to cover, as much of the pediatric use case as possible is a big priority. And just continuing to look for ways to automate, things that can be automated so that you know the parts that can't be right, like the actual treating of the patient and building that relationship and.
Can be done better and more effective by happier people that are less burnt out. What that will be like, three years ago, I don't think anyone would've thought that, like building a scribe would've been the thing, because scribes had been around for a long time and no one ever found them valuable.
But, the technology changes quickly. And now with our combined team, we have such an ability to adapt and move quickly and build as the technology changes. So I'm sure there'll be things six months from now that we launch that we're not even thinking about today.
Marla: And I did get a lot of questions in here about pediatric OT speech happy to hear that we are in that area and advancing, like you said, with the scribe and the intakes.
Pedro: Yeah. And if I can't just it's nice to just take a moment to encourage people, reach out, tell us. 'cause again, like listening to the folks that use our software and hearing what sort of problems or improvements would help you the most is such a key thing to us. You can do just about, anything it seems with ai.
And so just hearing what really valuable and helpful to folks like, better covering speeds, et cetera please, do reach out, do tell us, give us the feedback. We wanna make sure that what we're building is really useful for you all.
Marla: And I do wanna ask this question as well, 'cause I know it's something that we're working out and should be out pretty soon, is just about the objective section and how much it'll go into it versus clicking.
Pedro: So that's definitely something we're working on. We just, we're trying to make sure there's so many different fields that you can put data into and, having a number, the office is not really an acceptable thing. So we're just making sure that we're gonna hit a very high bar of quality and accuracy when we do that into the many, apparently thousands of fields that can potentially get filled in within the EMR.
Marla: Absolutely. And I know we have tons more questions, but not a lot of time left, so I do want, you'll see a popup on your screen. If anybody is interested in learning more about prompt, more about PredictionHealth or about both, please feel free to indicate there. We will answer all your questions and reach out to you and make sure that some of these other questions that you're asking we get to in deeper depth and make sure that you guys are using AI in your clinics as well.
All right, so as that's popping up, let me see if we can get one more question in here. Oh, lot of good feedback. Let's see. I guess a great one just to end it out is if you can give us an example, will or Steve, about how it is implementing this, how long it took and the change management to put AI in your clinics.
That would be great. In closing.
Will: Yeah, relatively seamless. Obviously it's a new system, right? So you gotta learn how to use it. But one, I think that the support that both of you guys, your guys' company and now together have offered I have had multiple different EMRs and you guys get back actually quickly and you do the things that you say you're gonna do.
So one, the support that you get when you're training is really stellar. And two, it's it. It really isn't that difficult to implement, right? Like we're stressing the seamless part because that's really what it is. It's not a bolt on, it's not something that you have to go to a different window. It is really just turning it on and then AI learns you better, right?
So you just have to know going into it, Hey, the first time I'm gonna use it is gonna be a lot different than the 15th or a hundredth time that I use it. And so as it learns you it just becomes really efficient. And like I said, like I've had, clinic directors or clinicians that have been in the industry a long time, even back to when they were handwriting notes, and they actually adopted it, I think, quicker and easier than some others because it really reduces the amount of time that you're doing other things on the computer.
So very quickly, yeah. You're,
Steve: yeah. Your onboarding team, by the way. When we made the switch from another EMR system to prompt your onboarding team, what a different experience that was. It was just you guys were there with us basically holding our hands through the whole way. It was seamless.
One of the recommendations you guys had was, Hey, come Monday when you guys go live with this, you might wanna scale back your schedules because we wanna give therapists a time. Guess what? I did not scale back my schedules. I'm like, wait a second. I'm not scaling back my schedules. My therapist walked in.
I would say that was Monday. By Wednesday. Everybody had a handle on it. There was no like, oh my god, their documentation was done. There was a few questions here and a few questions there, and we sent you guys and you guys were really good about jumping on the phone and it was, it really wasn't that painful.
It wasn't at all. And if I could just make a couple comments. I've had therapists who call me that say, Hey, I'm using chat GPT. There was a big warning that came out this morning. I did happen to catch it on one of the private practice websites. Chat GBT is not HIPAA compliant. People are using chat GPT to write their notes for 'em.
I do not suggest you do that. If you're wanting to get yourself in a lot of trouble, be very careful because that is not something I would suggest that you do. Cutting corners, et cetera. The other thing is stuff I could tell therapists out there, being a private practice owner, it's scary.
Change is not fun. But I say, the industry is changing. We've been getting beat up. I keep, I'm trying to keep it positive. I think the future of physical therapy is gonna be extremely good. Insurance companies are waking up, physical therapists are needed. We just need to get paid like we're needed.
But right now, with using the AI technology, using the combination of the AI with the documentation system, you have all the reports at your fingertips. It makes it so much easier because I'm still in the process of scaling offices. I've got two more offices coming this year and it just makes it seamless and it really is helping me right now to recruit physical therapists to come in and, hey, listen, we're using ai, we're using documentation.
The documentation component, if it's gonna be really simple, you need to just come in and treat patients, et cetera. It's scary. And a lot of people are sitting on the fence and wondering, is this something I should do? I'm a hundred percent. I'm like, this is something, if you're not doing it now, you're gonna eventually be doing it.
You're gonna be forced to do it.
Will: Yeah. And not only that too, it's like you can use it as needed, right? So if you are the kind of person who just, wanted to change a little something here or there, you can use it a little bit. You can use it for every single note that you take. You can use it for only evals.
So it, it's not something that is like always gonna be pushing you to go towards it. You use it when you feel fit.
Marla: I can't thank you guys enough, honestly, so much for your insight, your tips, and for inspiring everyone and helping them see what's out there and what's possible. But please, for any of you we're available, please reach out.
We did put the demo and trial links in the chat. We'll keep an eye out for the replay email that'll be coming in your inbox and reach out to us or reach out to Steve or Will about any questions. Pleasure to have you guys here and we look forward to help bettering the industry. Thank you all.
Thanks everyone.
Pedro: Thanks all.