#143 Home Health Therapy Documentation with Rohit Shetty

We all know that documentation is a heavy burden for therapists. But home health therapists arguably face one of the heaviest loads, with mandated OASIS reporting layered on top of standard clinical documentation — a dual requirement unique to this setting.

They also face challenges that clinic-based therapists rarely encounter: coordinating care across disciplines — physicians, nurses, and other therapists — often without shared documentation systems, as well as navigating privacy considerations and internet instability when documenting in patients’ homes.

But, as we’ve already discussed around outpatient EHRs, there are changes on the horizon. In this discussion, we’ll zoom in on what’s becoming possible in this complex environment, from clinical decision support to practical automation. We’ll be joined by Rohit Shetty of AutoMynd — Rohit is excellent at explaining exactly what’s possible, while staying grounded in reality.

See full course details here:
https://otpotential.com/ceu-podcast-courses/home-health-therapy-documentation

See all OT CEU courses here:
https://otpotential.com/ceu-podcast-courses

Check our our live webinar schedule here:
https://otpotential.com/live-ot-ceu-webinars

Support the show by using the OTPOTENTIAL Medbridge Code:
https://otpotential.com/blog/promo-code-for-medbridge

Try 2 free OT Potential courses here:
https://otpotential.com/free-ot-ceus

Transcript

SPEAKER_00 0:00

We all know that documentation is a heavy burden for therapists, but home health therapists arguably face one of the heaviest loads, with mandated OASIS reporting layered on top of standard clinical documentation. Home health therapists also face unique challenges coordinating care across multiple disciplines with physicians, nurses, and other therapists, often without shared documentation systems. And on top of all this, they navigate privacy considerations and internet instability documenting in patients' homes. But there are changes on the horizon. In this discussion, we'll zoom in on what's becoming possible in this complex environment, from clinical decision support to practice automation. We'll be joined by Rohit Chetty of Automind. Rohit is excellent at explaining what's possible while staying grounded in reality. So let's dive in. Welcome to the OT Potential Podcast. I'm your host, Sarah Lyon, OTRL, and I wanted to let you know that this podcast may qualify as continuing education for you. You are probably listening to this podcast on a free podcast platform, but to gain CEU credit, you will need to be a member of the OT Potential Club, our OT Continuing Education platform. You can go to OTPotential.com to learn more. Okay, here we go. As I mentioned at the top, we are joined today by Rohit Chetty. Rohit is the founder and CEO of Automind, a health tech startup that builds AI-powered documentation products specifically for home health clinicians. As a financial disclosure, I just want to highlight that Rohit clearly has a financial stake in this topic, being that he is the founder of Automind. But that being said, I just truly believe Rohit is unique in his deep and extensive home health therapy, documentation expertise, and in his ability to communicate with therapists, just making him one of the best possible guests for our topic today. So without further ado, I will patch Rohit into our live studio. Rohit, welcome to OT Potential. It's so great to have you.

SPEAKER_01 2:15

Thank you for having me, Sarah. It's great to be here and uh excited to chat today.

SPEAKER_00 2:20

I am so excited to chat. My mind is just swimming with home health like numbers and information after reading some of the journal articles that we have attached to this episode. Uh I've always said on the podcast that the future of healthcare is in the home. I think we're going to see so much care transitioning to patients' home. But even though I say that all the time, I did not realize that home health really is one of the fastest growing healthcare segments. Right now, annually, it serves around 3.6 million Medicare beneficiaries. So a lot of people. And it's just a very complex setting. And number, another number that stood out to me was around 20% of home health patients face an unplanned hospitalization or emergency department visit during their stay. So that to me just speaks to the complexity of this caseload. It feels like one of our goals is for that ER hospital stay to not happen, but there's clearly a lot of ways that we can improve our care. And we're really hoping that our documentation systems are going to be an important step in that. So I'm so excited to talk about where the industry is today, what the challenges are, um, and what's on the horizon and what's happening in real time for some of the agencies across the United States. Before we talk about all those awesome things, I'm so excited to learn a little bit more about you and your journey. Of all the healthcare founders I talked to, um, I'm just going to say or disclose, like you are definitely one of my favorites because you have such a deep background in home health. And I feel like you've been so focused on just advancing the space for so long. Can you share with us your story of how you got involved in like the IT side of home health and then where you're at today?

SPEAKER_02 4:25

Absolutely. Uh thanks again for having me here. And uh, you know, this is such an important topic as so many home health clinicians are trying to consider, you know, advancements and use of AI in their day-to-day. And even for folks who are uh currently working in outpatient or primary care, trying to, you know, look for some flexibility that home health offers. So it's a pretty attractive space. But a little background on myself, been in technology for 15 years. I'm based out of Northern Virginia. Uh, you know, I uh I've spent a whole lot of my career just listening to people because one of my primary roles was to solve complex problems, break it down into simple concepts that can be then applied to technology. So I've been very fortunate that way. But I used to work for a large home health agency uh that serves like millions of patients a year. And uh one of the things I was brought on to do was do an inventory of uh all the different things our clinicians and back office staff do that takes a lot of manual effort. So that kind of gave me the opportunity to be, you know, alongside those operators and clinicians to truly understand what a day-to-day looks like and what can we do to improve it. So, yeah, worked for that agency for several years and until, you know, uh until the last couple of years, I realized that I have so much of knowledge in my head and I need an outlet to actually execute on. And that's where I started my company, Automine, that is focused towards uh, you know, clinical documentation, back office administration for home health and hospice agencies. And uh we are one of the leading platforms today that provides uh AI-powered documentation to AI-powered intake and uh several other workflows that home health agencies use for. Yeah.

SPEAKER_00 6:24

I agree with you that home health is such an attractive setting for us as therapists. We love being able to get into the home, see the patients in their natural environment. Um, I hope some of the people who listen to this uh episode are considering home health as like a future setting. But I do want to say the challenges of documentation, just out loud, kind of uh uh summarize them for people. One of the articles that I read for this episode is called the match rates between home health assessment and Medicare Claims Data. Um, this came out in JAMA. I think this year it came out this year. It's some great background reading to just get some of the numbers around home health. One of the challenges I want to lead us with is um what the article was looking at was match rates between an OASIS assessment, which is the main home health assessments that's done, and then matching that with the actual beneficiary, which is in a different Medicare database. And the match rates have been going down over the years. I think it went from like uh 98%, it was working well, down to 76%. So just right there, that shows you like you're working in these big complex with these big complex databases, the systems. Uh I don't want to comment on these huge systems, but it seems like uh instead of getting better, they're kind of struggling. Um, so you're working like with in this background with all this data. That's what I took away from this article. I know there's so many other challenges to home health documentation. Um, how would you summarize them? And also, can you just explain Oasis to us for people who don't know what it is?

SPEAKER_02 8:17

Absolutely. Yeah, so let's start with that. So Oasis is type of uh CMS required documentation that is specific to home health. Uh, essentially what it is uh doing is like making sure, unlike your primary care, where you can take soap notes and visit summary, Oasis assessment is an objective scoring of patients' functional and cognitive status, you know, across different uh segments of your patient chart. So that said, when we think about uh the complexity around Oasis, it's the fact like there is uh a requirement to understand the logic behind how do you score a patient's you know, uh functional status or a cognitive status? Are they able to uh get in and out of bed effectively? So that you have a range of you know from zero to five or different models. And we can dive deeper into that, what that looks like. Uh, but in a sense, it's an extensive set of questions that you need to fill out when you first see the patient for their home health visit, and then uh across the episode for uh different OS type of visits. But maybe it will be useful to explain how the home health episode works uh in a nutshell, so we can uh you know lay a proper foundation for this conversation. So in home health, uh, we go by the context of episode. What that truly means is when a patient is discharged from the hospital, you know, they are referred to a home health agency if they require home health services and if they qualify for it. Uh, once the patient is accepted in a home health agency, a clinician, a PT, or a nurse or OT in some cases will go and do their start-of-care Oasis assessment and create the clinical picture for the patient and generate a plan of care that will essentially last for 30 to 60 days. You know, uh, so you are basically planning for the next 30, 60 days of how many visits per week this patient would require to be able to function appropriately or make progress towards goal. Uh, and that's what it's called episode. You know, typically an episode, you know, ranges from 30 to 60 days, and it can go on, like you said, if a patient is readmitted to the hospital during this episode, they would be extended back uh, you know, uh into the system. The agency that was taking care of them uh will be uh extending their certification period. Uh, but that in a nutshell, uh, what Oasis and episode is. And I totally agree with the uh the statistic which you mentioned, and I can probably highlight why I think that is happening. I mean, healthcare in general always had the data problem and more so of like an interoperability problem. Systems don't talk to each other. So, what has happened with home healthcare is over the period of last 10 to 12 years, with the rise in technology, there's also been rise in fragmentation. So agencies in search of efficiencies have onboarded more tools and more systems. And your EMR is not always the system of record anymore. You know, you have several different data sets in different places. So fragmentation being one, but uh the complexity of episodes and multiple admits in the same for the same patient, I also feel that has led to multiple claims and not able to map it to uh the actual you know assessment and the medication, Medicare beneficiaries. Uh so there are several different reasons on why that is happening, but we are seeing trends, you know, uh positively that will eventually you know fix that gap as well.

SPEAKER_00 12:11

So uh I'm hearing there's it feels like there's some cool things going on in home health, actually. Like with the the standardized OASIS assessment, you're gathering so many data points, which is actually a problem to me in some of our outpatient and even acute care settings, where in that soap note, we're getting all this unstructured data. So then it's hard to compare episodes. But home health, I guess, feels like a front runner in that they're really organized in what data that you're trying to collect. Just so I understand, if I'm a home health therapist, I'm walking into a patient's house, I'm guessing with like a tablet or a computer, and I'm trying to fill out this four oasis uh intake with 70 questions, or is that a good guess of how many data points I'm gathering?

SPEAKER_02 13:07

Oh probably uh on the lower side of it, yeah. But uh on average, yeah, it's uh on average you're looking at roughly 150 to 200 questions. And I don't want to scare anybody who's trying to get in home health, by the way. You know, things are changing. And uh uh I think uh that's where technology is playing a huge role today. Uh, but you're right, yeah. So if a clinician typically, if they're not using appropriate technology, they would probably take notes on a piece of paper uh or probably use their device, company-issued device, to capture the patient's clinical picture across these different Oasis assessments. Uh and just to kind of caveat that Oasis assessments are not for every single visit, they are only on your primary main visits during the episode. So it happens roughly about four to five times during a 60-day period or a 30-day period. Yeah.

SPEAKER_00 14:06

And I can imagine as I'm in the patient's home, I'm picturing that like internet's unreliable, which to me sounds like a huge challenge for scribes, and you're just hitting all these challenges that are different than an outpatient therapist or an acute care therapist would hit, then uh with all these variables. When I think of the advances in technology, I'm I in documentation, I do tend to think of outpatient therapists. I think because their uh workflow is and setting is more structured. But when I think about home health, I'm like, these are the people that need these advances the most because they have this documentation burden and challenge uh that is unique. And if we can solve these problems here, it feels like we can solve them anywhere. Can you tell me a little bit about just what is becoming possible in home health documentation? Like what's out there that wasn't there five years ago that's helping us meet some of these challenges?

SPEAKER_02 15:14

Absolutely. So, you know, over the last five years, uh we all have seen, including, you know, primary care and outpatient, uh the evolution of how we use AI in our day-to-day and especially in clinical setting. But you hit some really interesting points which I want to kind of emphasize is the variability in a home health setting is quite different than your traditional outpatient, outpatient setting or primary care, where you are sort of in a very controlled environment, right? You are uh you're walking into someone else's house, almost a stranger initially, and don't know what to expect. You have a parking dog, you have TV with a lot, you know, TV on the highest volume, family, you know, walking around the house. So there are a lot of different variables that a home health clinician uh anticipates for their visit. And uh identifying a technology that could meet uh all the needs for documentation in that setting was one of the most complex parts for majority of the health record uh systems or EMR systems. Uh but one of the most popular use of AI we are seeing in home health today is around clinical documentation scribes. So, you know, home health went through the same journey of going from paper uh to like an electronic uh medical record. Up until like last three, four years, AI scribes were introduced uh that essentially would capture the conversation between the clinician and the patient and generate that Oasis documentation for you. So that has been one of the most popular offerings that AI had to contribute to the home health community. But as time passes, it is also like now used for medication reconciliation, where you could snap a picture of medications and generate a med profile. I I don't know if you're aware, but in home health, on average, clinicians document roughly 30 to 40 medications in their first visit. And that is the name of the medication, the frequency, the route, and on top of that, they do drug interaction verification to see uh you know risks associated with same drugs uh or multiple drugs, you know, uh are properly communicated to the patient. So there's a whole bunch of different things they do. So AI and visual models are used to document medication today, which is a huge hit across home health agencies. So uh all that to say, I think home health is finally sort of catching up on technology, which was typically behind your primary care and outpatient setting. But with access to AI applications um easier than ever, we are seeing that adoption curve slowly improving.

unknown 18:23

Yeah.

SPEAKER_00 18:24

I love the benefits of scribes to us. I think across settings, we are seeing how much time they can save, how they can help our accuracy. Um, two other like interrelated parts of that I wanted to like just linger on just a little bit were or our clinical decision support. I read this really great article that I'll link to called Exploring Home Health Care Clinicians Need for Using Clinical Decision Support System for Early Risk Warning. One thing that they pointed out in here was that there's already been 20 studies developed looking at clinical risk prediction models that leverage ERs to predict adverse outcomes. Um, what a great tool for us in home health, where again, we're just working so hard to reduce those hospitalizations, those emergency department visits. Um, it feels like one of our main focuses as home health is to home health therapists is to stop that from happening and how cool to have tools that are coming into our documentation systems to help with that. Something I wanted to say out loud from this article, too, is that they talk about the five rights of clinical decision support, which I had heard in pieces but had never heard together. Um they talk about the challenge of getting the right information to the right person in the right format in the right channel at the right time. Yeah, let's linger on that clinical decision support just for a moment. And anything else you can add to like what does that look like logistically for therapists? What would an actual like early warning system look like in a modern documentation system?

SPEAKER_02 20:09

Absolutely. Yeah, that's a great point. I think because that's where the real value is. Uh, you know, while we shoot for efficiency gains and reduce time on documentation, but how do we empower clinicians with intelligence uh support uh that could help with clinical decisions? One of the things which we are seeing a lot of success, and this trend is across the industry, is ability to proactively create a risk profile for the patients. And that can typically come through the referral paperwork and the discharge summaries, right? When a patient is discharged from the hospital, the hospital systems generate a whole bunch of paperwork of what was the reason for admission and what are the clinical notes that are associated with the patient. What uh what we have seen is uh using models that can process a bunch of pages across discharge summary, create a clinical picture for the patient that can be surfaced to the therapist, you know, even before the visit gets started or the case manager. What that lets them do is like appropriately think how can we ensure within the 30, 60 days time we have with this patient, you know, to make the best decisions as associated with the plan of care, uh like the goals and the interventions. You know, what are other uh risk uh risk flags that we should be aware of? Like, hey, is the patient at the risk of fall? Do they have trouble getting in and out of bed? Things of that nature uh can be easily identified. And when you are in the patient's home, the AI assistant can actually help you. To say you might want to educate the patient on these different factors, or even educate the family, you know, on what they need to be doing to make sure that the patient does not uh fall while you know going up and down the stairs, uh, things of that nature. So there are several different tools that are being uh implemented today, but one of the very simplest ways to do it is essentially proactively identifying risks associated with the patients because you do have the data. It's a matter of you know running it through clinical intelligence layer to surface them in a way that is can that can be consumed by physical therapists or occupational therapists uh in their plat of care.

SPEAKER_00 22:45

Yeah, when I think of working in the hospital, um, we literally have like signs up like fall risk uh and different signs on the doors and in their rooms. And with the EMR systems I've seen in the hospital, I can't imagine an interface that would be helpful to show me those things because they're so cluttered and so confusing. So what an important advancement to um be alerting us of risks and like that five rights framework to do it in a way like right time, right patient, in a way that's actually helpful uh for us to consume. The other thing I wanted to linger on was I actually did an episode on therapy documentation this winter. It was pretty focused on outpatient therapy. Something that feels like has changed even since this winter is seeing more AI agents coming into our documentation system. I honestly don't understand what these agents are able to do currently, like what is hype and what is actually happening. What can you tell me about AI agents in our documentation? I guess like both broadly what's happening and specific in your world.

SPEAKER_02 24:00

Absolutely. So when we think about agents or AI agents, you know, essentially that term is a representation uh of an automation layer that works with large language models or in simple words, chat GPD, open AI, all of these uh large enterprise AI shops. So what agents are able to do, they are able to browse the internet and perform actions similar to a human, right? So let's just say if you want to book dinner uh at a restaurant, you would simply go in and say, like, I want to book restaurant at this, uh, I want to book dinner for this restaurant, and my preferred times are this. So you would typically go and browse the internet, what are top five restaurants in this area, and where where do I book, what time slot, etc. What this agent can do, it's can replicate the same steps and having access to your computer or your browser, able to perform that same task. And now what that looks like for healthcare is you can empower these agents with a knowledge base that is specific to your organization, and then generate and retrieve information that is custom tailored for your needs rather than when you go to Chat GPT. It's like a very broad information you get, you know, but in a true healthcare setting, you need very specific domain-focused outputs. Uh, so AI agents, in a way, does a pretty good job of that. They are also being used in a lot of different places for insurance and claim submissions where they can go to the payer portal and submit the claim. As far as use for a therapist or a clinician, the practical use that I have come to find out like it's meaningful, is around surfacing guidance and assistance uh on how to document a certain thing or how to handle a patient's plan of care in a certain way based on the condition, but also based on the enterprise knowledge, that's where it is truly useful, you know. Uh especially if you're using an AI scribe already, I think agents can complement that by providing more custom-focused knowledge. But broadly, I think agents, although very hyped and very popular uh in a healthcare setting, like I think the maturity curve is still not there, in my opinion, not at least in home health setting, just because of the complexity and variability around different workflows. Because it's never a linear workflow uh in home health when you think about you know accepting referrals, scanning referral paperwork, or doing a home health visit. So it could be used, but it needs to be used with human or clinician in the loop, uh, you know, throughout the workflow. Yeah. So, in my opinion, we uh, you know, I think there's value to it, but it it needs to be integrated with uh clinical oversight uh to provide true value. Yeah.

SPEAKER_00 27:26

So hypothetically, during the session, I can say in the future, I can say, like, hey, Mrs. Patient, I want to see you again next Thursda on Thursday. Let's get you on the calendar at two o'clock. And the agent goes in and books that and then I'm like, Would you like a reminder? And then it can enable a reminder. Is that kind of stuff reality anywhere yet?

SPEAKER_02 27:56

Yeah. It is absolutely a reality. Yeah. Um, because that's one of the low risk uh yes, yeah.

SPEAKER_00 28:04

Scheduling feels like low risk.

SPEAKER_02 28:07

Low risk use case where you're not necessarily going to, you know, toy around with the clinical picture of the patient. Uh, so so I think like that, uh, those types of use cases are prime use cases for agents or even just standard workflow automations, right? Uh, for example, even scribes today, when they hear a conversation and talking about schedule or talking about goals and intervention, can automatically generate and plot that visit for you for the next few weeks, or plot that plan of care for you, uh, or a draft version of the plan of care that you can go and approve, and it automatically flows into the workflow. Uh, a lot has to do with what system you're using, right? If you're using a legacy EMR, you know, some of these workflows, even after the use of AI, cannot be supported. Uh, and it feels like a copy-paste ordeal. So that's something to take it with a grain of salt, like is useful as long as you have a system that can support it uh from a medical uh EMR perspective. Yeah.

SPEAKER_00 29:17

Yeah, it feels like these agents are part of why we're seeing, I want to call it like industry consolidation, where if your scribes just like a standalone tool, it won't be able to go in and do these automations um for you. So the more things can be integrated, the better. That kind of leads me to we're talking about like these cool things that documentation hypothetically can do for us. And um in Outpatient, when we did that episode, I was very much like, go out and try these tools because when you're on a soap note, you can just tack on a scribe and copy paste in four blocks. And it's just easy to try what's out there. In home health, it feels much harder to try what's out there because you're not really saving time if you're copying and pasting in these hundred like data points into your OASIS documentation. Which leads me to my question of just like, what do you think the current adoption rate is among home health agencies? How many therapists are actually seeing this like new era of tools in their documentation? And how many are still on these legacy platforms? Like a year ago, I will uh on the podcast, probably a year and a half ago, I was like, within a year, 70% of us will all be on these new modern platforms. It hasn't been that fast when we did the outpatient episode. We guessed between like 25% and 50% of therapists were on a more modern platform now that have some of these tools. Uh, what's your guess for home health of what adoption is looking like? Is it lagging? Is it ahead? What's like the real world looking like for people?

SPEAKER_02 31:17

Yeah, let me start with saying uh what surprised us the most, and no no offense to nurses here, but therapists were the first uh category of uh clinicians to adopt uh scribes uh and these AI co-pilots faster than nurses. And uh, you know, the common theme we heard across agencies that we used to work with uh is like our therapists hate documentation. I mean, everybody hates documentation, but our therapists hate documentation. So that kind of helped for a lot of technology companies to kind of you know uh trigger that revolution, so to speak, as how we can change the way we think about documentation. But as far as adoption rate goes, right, it is probably in the same ballpark as outpatient. Yeah, I would say it's about 30 to 40 percent of therapists today are on some kind of scribe tool. And even they, if they are not officially on a scribe tool, they're using AI in some capacity provided by their organization. You know, we have seen almost 50% of the large enterprises, you know, which is not a whole lot, probably there are 100 agencies that are called enterprises, and 50% of them are already using some sort of like documentation platform to help their nurses and therapists. So I would say comparatively to last decade or the last five years, the growth curve or adoption curve is a lot faster than home health has ever seen before, you know, uh because this is a true shift for home health community. Primary care typically uh or outpatient, it operates almost like a consumer-style market. So you have much more awareness and access, while in home health, unless your agency is forward thinking and tracking innovation in what's happening, that information or access to new technology or new tools often goes through a lot of red tape within the enterprise, right? So uh sometimes that adoption kind of slow down. Uh, but we are seeing a drastic change over the last year, especially. You know, every now and then we see a new enterprise adopt uh a documentation tool for the clinicians. Yeah.

SPEAKER_00 33:49

I was so surprised looking at the numbers. I read that there's 12,000 home health agencies in the United States. That is way more than I ever would have thought. I would have guessed like a hundred. Like I was so off in my thinking, which tells me there must be lots of small ones. And I'm just like brainstorming here. I'm like, are you more likely to get on a new documentation system if you're like a small one or if you have to convince a big player? Anything you've observed as far as like the agility of the agency or just trends around this, I guess, highly fragmented industry?

SPEAKER_02 34:32

Yeah, uh, so that's a great metric, right? So we have roughly about 12,000 Medicare agencies uh for home health. And what's interesting about this is like roughly 70 to 75, 80 percent are mom and pop shops, small agencies that to 5200 or 200 patients. And uh they are also the most innovative of them all, right? Uh, because yeah, they always run in a startup style, uh, even though you know they've been in the industry for a long time. So we did we do see like they are a lot more agile and nimble when it comes to adopting new technologies, and uh they are actually the most uh they're actually the segment that benefits the most by use of new technology, partly because they operate under like 50 clinicians and 10 back office staff compared to an enterprise who may have 500, 600 you know, users in their back office and 10,000 clinicians on the front line. Uh so the the impact of change is uh lot more controlled uh in a small to mid-size agencies compared to enterprises that do millions of patients a year. So we have seen uh we have seen a trend that you know majority of the requests we get these days are actually small to mid-size. How do we apply AI to this workflow for documentation or for the intake or for billing? You know, though uh those are the play, uh those are the agencies who are already thinking what their operating model will look like, you know, in the next three to four years with the evolution of AI and the use cases of AI that are going in production. Yeah.

SPEAKER_00 36:24

You mentioned at the top that one of your main skill sets is just listening and hearing from home health therapists. I'm so curious the stories that you've heard over the past years of um what the upsides for therapists have been. And I'm equally curious about like downsides or unintended consequences. I think when you think of AI, whenever we bring up AI on the podcast, we definitely get um comments and concerns about environmental impact, about energy consumption, about um just where our regulations are at. So I just want to say those out loud as just a big picture um thing that feels important to say. But I'm just particularly curious about the upsides and downsides for that therapy session and what you're hearing from therapists on the ground.

SPEAKER_02 37:22

Yeah. Uh so if you think about like uh a typical therapy evaluation assessment in home health spans anywhere between 60 to 90 minutes in home, right? And then roughly about an hour for that patient documentation after hours, right? And uh throw in the complexity of Oasis and the regulation requiring you to finish documentation within or submit documentation within 48 hours. Uh, all of that kind of makes the workflow for any clinician a lot more not just complex, but also uh administratively intense, you know, for that 48 hours. And you're not seeing one patient every single day. You're typically seeing five to six patients a day in your community or wherever the zip code you are serving. So the upsides we have started to hear from clinicians, the biggest, of course, uh is the time savings and uh not having to chart after hours, or even if they have to chart, it's less than probably 30 minutes to 45 minutes after hours charting, right? Uh, but that's most that was one of the biggest uh benefit that some of these products are bringing to the table is when you are in home, you are fully locked in and having a meaningful conversation with the patients, and you are not juggling your device or handwritten notes while speaking to them. So that overall interaction with your patient becomes more meaningful, you are able to actually build a relationship off of that and you know execute your plan of care effectively. But that those are some of the upsides, and we can go on about how comprehensive the documentation looks like because it is catching all these uh important things which otherwise will get missed out. But when we started, you know, tinkering around the idea of ambient AI, AI for clinical documentation, the question I had to ask myself, and a lot of my peers also had to ask ourselves is like, hey, how do we convey to the community that this is going to assist and not going to replace the clinical thinking? Right. This is there to take off the clinical work and not necessarily the clinical judgment that is required for delivering care. Uh, so that was one of the important pieces we needed to kind of navigate. The second one was like, how do we ensure the patient feels comfortable? Right. Uh, we were worried, like in in, you know, in the world where we live with so many tight regulations, how do we ensure the patient feels comfortable to use AI for their plan of care? So uh it was important to articulate the message and convey the upsides in a way because at the end of the day, if used effectively with the clinician in the loop, the patients are going to reap the benefits as well. That their therapist is actually truly listening and having an engaging conversation. And I mean, broadly speaking, like you mentioned, right? I think consumption of energy, water to run these data centers, they remain a challenge that you know the big players in this space are also trying to solve. But as healthcare, home healthcare tech founders, which the work we do, we need to kind of share that responsibility on how efficient our systems are to prevent overconsumption or prevent prevent waste, essentially, while delivering these tools, you know, because how often we actually, you know, go to chat GPT today or Claude for our questions instead of going to Google, right? Um, and even it's like the most simplest question, and that kind of adds up when you think at a macro level. So identifying our shared responsibility is going to play a key role uh as we think about the big picture around energy consumption and uh using AI responsibly. Yeah.

SPEAKER_00 42:10

I have two more concerns that came up in the comments that um I want to say out loud. I'm excited to hear your take on them because I know you're a um broad thinker. One is around our therapist's composition skills and those just quickly deteriorating. That with the scribe and uh they're writing a lot for us. And I couldn't I could make two arguments about this. I would say how you write is how you think we have to preserve our writing skills. Are we creating this unintended consequence where our critical thinking skills are actually deteriorating because our writing skills are deteriorating? Um, I've noticed that about myself, just in I'm so quick to plug my writing into like a chat to help me polish it. Like, I can tell my writing skills are deteriorating, like they're getting worse. But I would also like that's one I would be like, I'm worried about our critical thinking. But then I would also say our composition skills aren't our main therapy skill. They're something that we've had to grow in because of um insurance. I don't know. How do you think about yeah? I I haven't seen it documented anywhere, but it just has, I totally believe it, it has to be real. I think we're all going to get worse at writing. How do you think about that? Or do you agree with that? And how do you think about it?

SPEAKER_02 43:34

I do, yeah. I think if you think about it, like if you go on LinkedIn, right? It feels like entirety of LinkedIn or social media for that matter is AI generated content. And in fact, I have this discussion with my team and peers often that is there a scenario that five years from now, the entirety of health record, you know, is To be an AI generated content, right? And uh we are just feeding AI generated content in loop over loop every time that patient comes back in the system. So there is an argument to be made if it's a good thing or a bad thing. I think uh we can go down that route of kind of deciphering what that would mean for the next 50 years, and uh you know what does that mean for the human and clinical skills of writing and reading and our attention span? So there's a right uh argument to be made on that aspect, however, yeah, I think I think uh in the near term, the way I always encourage people is like if you're using AI or any scribe function, try to opt for a technology that allows you to personalize as close to as how you would do things. So when you are reviewing and editing, you can relate to it. And I think it's not a writing problem so much, but I think it's an attention problem, right? In my opinion, when you keep using AI for different solutions, oftentimes you are desensitized to the information generated by AI, uh, that you don't review it enough. So I think repurposing the attention or the time we have gained by the efficiency into reviewing it is going to play a more critical role for maintaining the integrity of the patient record. Uh and yeah, uh, I think as far as our personal skills of writing as a clinician or as an operator, that's the work we are we'll have to do internally and find that balance. And that's a trade-off in the world of technology. You could easily get comfortable and stop writing effectively, but then an argument can be made. Do you want to write? You know, do we see a world 10 years later? Maybe it's all voice, you know, it's all voice, or this whole administrative work which we used to do is pretty much an invisible layer while we are simply focused on delivering care. But uh where we where we stand right now, where I stand right now, I don't feel comfortable to completely rely on an AI scribe for my clinical narrative. I want to have my, you know, uh clinical lens before I can submit and make it part of the patient record, even if it means doing spot checks, because 99% of the time it's gonna do a great job. Will it be exactly the way you would do the note? Maybe not. Uh, but if you opt for a technology vendor that can personalize to the way you like, it it would, it would probably a good first step.

SPEAKER_00 47:04

It's so interesting to just think about how as therapists, as human beings right now, we are part of this huge experiment experiment that's going on uh where everything's changing so fast. I almost wish all of us would like write down the different parts of our jobs and then write down what parts, like check which parts we want to be replaced, um, and really like hold on to and hone the parts that we don't want to be replaced. Like for us in therapists, I'm like maintaining that patient therapist relationship interaction. That to me is the prime thing that we need to hold on to, uh, that we don't want to get replaced. There is something truly special. And there's definitely going to be technologies that are trying to replace that, and that's the thing we need to hold on to. Um, and then just to be aware about how these things are impacting us and maybe we need to be, I don't know, journaling to like maintain our own voice. So our voice doesn't become the voice of the algorithm. Um, these are very existential. I just want to say out loud uh as far as comments, that this is just an annoyance. I feel like a lot of therapists are being asked to use their phone as a hot spot, even in hospitals, and they're like draining their phone's data and it's unreliable and the technology is awesome, but it's reliant on that internet connection and feels shaky at the time. And I can imagine that's high in home health. Any quick note on just like internet connection? And I'm just like, oh, headache is what I see there.

SPEAKER_02 48:40

Yeah, that's a great point. Uh that's a very prominent problem in home health. And uh especially if you're not part of uh if you're a technology company trying to insert yourself into home health uh and not aware of these nuances, it can quickly become like a you know liability because internet connectivity is like standard, it's like default sometimes. So good news is although legacy system did not have an effective way to support offline work, the newer systems, most of them today consider offline more, where you can still leverage you know AI models and scribes and even the standard systems without the internet in rural areas. And that like for Automind, my company, I think 30 to 40 percent of our usage is offline, uh, which tells us like how what are the rural parts of our country where people still need home health, you know? And uh that's a really interesting insight we got. And that said, like, yeah, I think with newer technologies, that's rarely a problem. You always have an option or a workaround. It's with the legacy systems where you struggle where you have to sync and things of that nature. But there are so many workarounds, even like people uh basically were take handwritten notes and dump that to AI models after their visit is done and they come back to an internet sufficient area. So there are several different uh workarounds that people have, you know, invented on their own uh over the last couple of years uh to tackle that problem.

SPEAKER_00 50:25

We're headed into our last seven minutes. I want to squeeze in two questions because I'm so curious to hear both of them. The first is around um our like analytics dashboards that a lot of these new systems come with. I love dashboards, I love market like visualizing progress. Some of the therapy documentation dashboards feel like big brother punitive of therapists. I don't know, I just don't like what they prioritize to show. I'm curious when you're thinking of like the KPI, the key performance indicators, the KPIs that should be on a therapy documentation dashboard for the therapists, what are like the big picture data points that you think we should be seeing about ourselves in our practice?

SPEAKER_02 51:15

Yeah, so as uh as we think about uh you know moving the use of AI technology from efficiency to patient-focused outcomes, that leads us to how do we deliver value-based care to our patients and what does that look like in practice? So a lot of the KPIs that uh myself and my team tends to focus is around value-based care. What does that attribute to uh a patient's outcome? And that could look like have you educated the patient on the medication or patient's safety? Did we treat the patient with gentle uh gentleness and you know, describe the purpose of our services, things of that nature? Uh I think play a huge role in ensuring that you are representing your agency as well as your discipline in the right way. And you can track that part of the dashboard, uh, you know, even after the visit is done. And what that also unlocks is because now we have this nuanced data point through scribes and through these different technologies, the clinicians also have access to educations and CEUs through these clinical education platforms. And that there, that's also another space we see evolving of micro, bite microeducation, bite-sized education, uh, to help clinicians and therapists get acclimated to what what it means to deliver value-based care in home health setting. Home health setting also follows a really interesting metric uh and some other settings also, but home health specifically, it's called star ratings or HH caps ratings. It's essentially a series of questions that Medicare tracks to see how likely it is the patient would recommend the agency to another uh someone else, or how likely it is that patient is not going to be readmitted again. So uh some of those data points are certainly uh useful to track part of the dashboards and KPIs.

SPEAKER_00 53:33

That gives me the language, I think, to articulate what I don't like about some therapy dashboards. Like if you're just in the fee for service mindset, uh then these dashboards look like did you squeeze in as many units as possible? Did you always build the highest code for that time? Um, which is the game we play, I understand it. But if we just use AI to like automate this old way of doing things, like yes, that will lead to skill deterioration. It's going to lead to unethical therapists. Uh, this is all why we need to shift towards a value-based care model where we can actually use AI to help us improve our care. And I love seeing the early glimpses of that. That's actually my big picture takeaway from this whole conversation, like excited for we need these new payment models. I think we'll have the technology. And if done right, it will make us better therapists. Our skills won't deteriorate, they'll increase. And if we're using a tool where our skills are decreasing, we need to be vocal about that and be telling come just be loud that it's the wrong thing. Like these tools should be making us better. What's your takeaway from this conversation? What you and just like where you hope things are going uh in the next couple of years?

SPEAKER_02 54:56

Yeah, uh in my opinion, I think uh folks who are building technologies and working for home health or healthcare, we need to quickly start expanding our focus from documentation focus time savings versus how does documentation lead to uh a great patient outcome, uh, you know, or an improved patient outcome. I think that's where my focus is, and that's where I think the industry will go as well, because we will have to start rethinking about what is the role of a clinician in the new world as we think about taking the administrative work from outside of their workflows and having them fully dialed in with their patients and being responsible for outcomes. I think that's where I think industry will go. And more importantly, hope would be like we can do more for our patients in the time we have. And a lot of our patients uh come from communities that are underserved and don't have access to the standard uh, you know, facilities that uh folks in the cities have access to. And they are also uh in most cases come from a very uh you know difficult uh you know uh demography. So providing them education and educating them on access of to Medicare, to the benefits they have, those are the principles that I think will get more emphasis in the next two to three years as we think about use of technology or use of any kind of AI application in our setting. Yeah.

SPEAKER_00 56:50

Well, Rohe, I know we were focused on home health documentation, but this has just helped me think about how we are in this rapid change cycle. It's so important to be vocal about uh making sure this change leads to better access, leads to better care. Thank you so much for your work on this front, for uh pouring over these home health details and for just being here today to share with us what you've been learning. I so appreciate your time today and the work that you're doing.

SPEAKER_02 57:20

I appreciate the opportunity and uh this was a great discussion. I hope the listeners, you know, have had something to gain from this. And uh yeah, feel free to uh reach out to me if I could provide any assistance or answer any questions as far as home health and uh therapy goes. Yep. Thank you for having me, Sarah.

SPEAKER_00 57:40

Yeah, thank you, Rohit. Thank you for joining us on the OT Potential podcast. To earn one hour of AOTA approved continuing education for your time today, you will need to sign in or sign up at OTPotential.com. Once you're in the OT Potential Club, you will find a five question post-course quiz connected to this episode. When you pass the quiz with a score of 75% or higher, you will be able to download a PDF certificate that certifies your completion of this course. Okay, I want to thank you for joining us today, and we'll see you next time.