Innovation That Changes Lives: AI in Home-Based Healthcare

Artificial intelligence is changing healthcare—but what does that actually look like in a patient's home? In this episode of Home Is Where the Health Is, host Ashton Jones sits down with Evan Kramer, Senior Vice President of Innovation and Operational Excellence at Compassus, to explore how AI is helping clinicians spend less time on paperwork and more time caring for patients. From ambient clinical documentation to AI-powered intake and referral workflows, discover how technology is reducing administrative burden, improving access to care and enhancing the caregiver and patient experience.

Transcript

Ashton Jones  0:00  
Hi everyone, I'm Ashton Jones, and this is Home Is Where the Health Is. This podcast looks at the latest progress and innovation in the field of home-based healthcare in association with Compasses. In today's episode, we'll explore how Compassus is using artificial intelligence to enhance patient care and improve the caregiver experience. Our guest today is Evan Kramer, senior vice president of innovation and operational excellence at Compasses. In his role, he's responsible for driving innovative initiatives at the company. We'll discuss the ways AI is alleviating burdens for caregivers while also improving patient outcomes. He'll give his thoughts on why home health companies should embrace AI. What's next in the tech space for Compasses, and how this transformation is blending innovation with mission-driven care. There's all that and more coming up on Home Is Where the Health Is. hello, and welcome back to Home Is Where the Health Is. Today, we're diving into the topic of artificial intelligence and its role in advancing patient care. Today, we have our guest, Evan Kramer, who's the Senior Vice President of Innovation and Operational Excellence at Compasses. Evan, thank you so much for being here. We are really grateful for you attending and sharing your expertise today.

Evan Kramer  1:27  
Absolutely, it's a pleasure. Looking forward to chatting about how we're deploying AI at Compasses.

Ashton Jones  1:33  
Do you want to start with telling our audience a little bit more about yourself?

Evan Kramer  1:37  
Yeah, absolutely. I've been at Compasses for about four years now, I come from a long background of management consulting, where I use technology to solve problems and streamline processes, mostly for healthcare companies. As far as education goes, got my MBA from University of Texas, and then this is kind of a fun fact, but my undergrad was at Berkeley College of Music, so I was a guitar player there, still play guitar, and it may sound kind of random with the rest of my background, and in some ways it is, but I really find that I use the same creative muscles solving business problems as I do playing music.

Ashton Jones  2:15  
Tell me a little bit more about your personal life.

Evan Kramer  2:18  
Yeah, well, we've lived in Nashville for almost 10 years now. I married a woman that I met at Berkeley, actually. So she's a songwriter here in Nashville. Well, there's the tie-in. Yep, so she's.. she keeps me very close to music still. And then we've got a four-year-old boy who's also a budding rock star as well. He's listening in Nirvana and Ozzy Osbourne, and all that stuff.

Ashton Jones  2:45  
So, going back to your professional background, was there a moment in time that really interested you in AI, or did you just kind of stumble upon it?

Evan Kramer  2:54  
It's a little bit of both, to be honest. When I was in consulting, I did a lot of technology-focused work to optimize processes and companies and things like that, so I've been close to this idea of intelligent automation for a long time, but AI is really, I think, kind of what cracked the code on a lot of it, and so I'll never forget where I was when someone showed me Chat GPT for the first time, but it was like this just kind of light bulb unlock moment, and you know, immediately took it back to my work here at Compasses to start exploring how we might be able to leverage it here.

Ashton Jones  3:28  
Let's talk a little bit about what AI actually is. How would you describe AI in its simplest terms?

Evan Kramer  3:35  
It is very broad at this point, but for purposes of this discussion today, I think about it in two categories, so there's the generative AI, which is like the Chat GPT, for example, where you ask it a question and rather than it querying the internet and giving you an, like, an exact link to something, it's actually generating something that's totally unique to the question that you asked, and you can apply that to audio and to visual, like you said, you can create pictures and videos with it as well. So, generative is kind of this whole category over here. Then the other category, which we're going to talk more about today, is agentic AI, and that's AI that uses generative AI in some ways, but then it's also pairs it with automation to actually take action on behalf of someone, and that's where things get really interesting and really powerful from a process improvement standpoint.

Ashton Jones  4:28  
So, now diving into how artificial intelligence is being used in the home-based healthcare space, what are you seeing across the industry?

Evan Kramer  4:38  
Well, from what I can tell, home-based care companies are just now starting to meaningfully look into the possibilities of AI, and I'm really proud to say that Compassus, I think, is a leader in this space. As I mentioned, I mean, Chad GPT came out, and we almost immediately started running down the path with it. AI, particularly the generative and the agentic AI that I talked about earlier, here is totally. Really transformative, and it can be applied in all kinds of different ways. And here at Compasses, we are rapidly adopting it, because we see the many, many opportunities to improve nurses' job satisfaction, support our partners, and most importantly, elevate the quality and the experience that we provide to our patients overall. So the technology has become really reliable, and it continues to get better and better every day.

Ashton Jones  5:26  
What are some ways that AI can be used to promote wellbeing and honor quality of life?

Evan Kramer  5:32  
I just mentioned our clinicians, partners, and patients. I like to say this a lot. Our goal with AI at Compasses is to profoundly differentiate the services that we provide to all three of those groups, so when I think about our clinicians, AI is already supporting their well-being and quality of life by drafting their clinical documentation for them. This is a highly burdensome task, and AI is enabling them to spend a lot more focused time with their patients and is supporting them in achieving healthier work-life balances too.

Ashton Jones  6:04  
Yeah,

Evan Kramer  6:05  
for our patients, our documentation tools are actually enabling a better care experience, because the clinician is actually able to focus on them more in the encounter versus actually taking notes and writing their documentation at the same time.

Ashton Jones  6:19  
Right,

Evan Kramer  6:19  
it's also producing a lot more data that we can now use to ensure that we're providing the best care to that patient too, so as we continue to progress, applying AI to other processes like scheduling and engagement, and eventually care delivery will really continue to enhance the experience of both our caregivers and our patients.

Ashton Jones  6:40  
When you say that it's helping to extract some more data, what types of information is it showing you?

Evan Kramer  6:45  
Well, and we can talk a lot more about how these tools actually work, but essentially the documentation tools are recording the entire visit with the patient, and using that transcript, AI is able to extract all kinds of details from it, not just to fill out the documentation, but to also add more robustness to that documentation too. So we're, we're seeing our narrative notes, for example, go from maybe one paragraph to three paragraphs, and there's just a lot more meat on the bone there that we can use to assess the patient's condition and evolve a plan of care, etc.

Ashton Jones  7:21  
So when you're talking about some of the AI, where there's ambient listening going on, and it's helping the clinician gather some information from that visit, were they excited about that, or was it difficult to get people on board?

Evan Kramer  7:36  
Generally speaking, they were very excited about it, and to give you an example, as we were rolling this out, I probably got two or three emails a week from different nurses across the company, just asking, "Hey, when am I going to get this in my location? So, there was a lot of demand for it. It takes 30% of their time to do it. It's really burdensome. It's not something that they enjoy spending their time on right, so the idea of being able to get help with the automation was very attractive to them, so you know there's always like the 5% of people or so that are a little bit more hesitant, but I'd say by and large this was honestly the easiest change management journey that I've ever been on at Compasses, or even in all of my consulting days, and I think that the technology is there at this point. It works well. It's been proven out really well in the acute setting before I got to home health, and of course, there's a whole compliant consent process too that we go through with all the patients that I've made our nursing team feel very comfortable using the tools,

Ashton Jones  8:40  
so with the AI alleviating some of the burden for clinicians, I'm sure there's also some accountability to it. So, what does the accuracy rate look like, and what's their part in making sure that the information is correct?

Evan Kramer  8:54  
Yeah, it's a great question. The clinician is totally responsible for what goes into the medical record, so the way that we explain the tool to them is it's not an automation of documentation, it is a draft of the documentation, so we're saving them a ton of time by not having to start from scratch, but it's their responsibility to go through and read and make sure that everything okay reflects their clinical judgment, and you know that should be actually placed into the record, and that you know on the accuracy point, we all know about how AI can hallucinate. I will say that we don't see huge hallucinations with these tools, because it's pulling just from a transcript of the visit itself.

Ashton Jones  9:36  
Gotcha,

Evan Kramer  9:36  
but there are little minor things that a clinician would want to edit here and there, and so that's we really emphasize their responsibility to do that

Ashton Jones  9:45  
well. Of course, there's another elephant in the room when it comes to AI, and that's around how people feel in the security of their jobs as a leader. How do you address those concerns?

Evan Kramer  9:57  
Actually, I think healthcare is one of the safest. Industries to be in, in that regard, it's inherently a people-centric business. Patients want that human to human connection when they're receiving care, especially at a really sensitive time like the end of their life. And on top of that, our industry actually struggles with a shortage of nurses right now. There's there's just not enough of them to meet all of the care needs within our communities to begin with, so for me the question isn't will AI eliminate jobs, it's rather how will AI enable the nurses and the support staff that we actually do have to spend as much of their time as possible with patients working at the top of their licenses, so through that framing I think AI may actually be one of the best things to happen to healthcare in decades, and I think it'll benefit everything, everyone from clinicians to patients and their communities, as well as our teammates here at Compasses.

Ashton Jones  10:55  
Yeah, it really is about driving them back to what they're best at, which is taking care of the patient and allowing them to focus on what's most meaningful.

Speaker 1  11:04  
It's a really significant time commitment of the documentation. I mean, we estimate that most of our nurses spend about 30% of their week on that, so if you can even give half of that time back, there's whole patients in the community that now get care because they now have that extra capacity.

Ashton Jones  11:21  
Are there any specific stories or instances from caregivers that come to mind for you as you were going through this change management process?

Evan Kramer  11:29  
Yeah, absolutely. I mentioned this is the easiest change management journey. It's also been one of the most fun and most gratifying, because we have gotten so much positive feedback from so many different people. I've lost count of the number of nurses that said that Compasses has given them their nights and weekends back because we adopted these tools, so I know that we're making a really positive life impact on all these clinicians, and another more specific example would be that we serve patients in all types of different communities where English might not be their first language, so one really cool feature of these tools is that they can translate like they've got about 30 different languages that we can translate into, and so I hear stories of out on the West Coast, where we have a large Mandarin population, they're able to speak in Mandarin, and these tools can translate back into English. Yeah,

Ashton Jones  12:24  
so from what I've heard, there's also another key innovation going on, which is using AI to improve the intake process, and so it seems like it's really streamlining patient onboarding. Can you share how that technology works and how it's impacting workflows.

Evan Kramer  12:42  
Yes, and this one I think gets me really excited from the process standpoint. As impactful as AI documentation has been, it was actually very straightforward from a people and process point of view, because the process stayed exactly the same, like the nurse continued to do the visit, you know, and the only difference is that they had a tool that drafted the documentation for them, but there was really no difference to their day to day, if that makes sense. Intake, on the other hand, is totally different. The process was the problem to begin with, and no amount of technology would have added much value to it without fixing that process first. So, to give you a little bit more context, there our patients come to us through referrals from doctors and hospitals, etc. and we have to act really fast because each of them is a real person waiting to go home to continue their care journey. We get these referrals through all sorts of different channels, they can be like 60 pages long and they're often incomplete, and so we'll have dedicated teammates whose full day is just focused on watching for these referrals to come in, reading those 60 pages manually, taking data out of there, and manually keying it into the EMR, and then coordinating with our team to complete those referrals, so it's a very inefficient, frustrating process for people, and the worst part is that we actually can't admit every patient that comes to us. So now they're like, for instance, if they're not in our service area or if they're not on our insurance, we actually can't admit them, and so you have these teammates that will spend sometimes 40 minutes going through a referral only to hit this hard stop that realized, like, we couldn't have even taken that patient in the first place. So, to fix the problem, we took a clean sheet approach. We basically redid the entire order of operations, we rewrote the job descriptions and the reporting lines. It's a completely different process now than it was before,

Ashton Jones  14:43  
yeah.

Speaker 1  14:43  
But then what we did is that we stacked AI agents on top of it, so now you've got this optimal process, and then you've got AI agents that are really kind of turbo charging that process. If that makes sense,

Ashton Jones  14:56  
yeah. When you say an AI agent, can you explain what that is? Yes,

Evan Kramer  15:00  
going back to the definitions that we talked about up front, you got the generative AI and then the agentic AI. The agents are the ones that can actually take real actions. For instance, when the referral comes in, rather than having a human read that 60 page document, we can now have an AI agent do it, and it knows what it's looking for, so it can pull out all the little pieces of information that a human might have pulled out, or another example would be all those checks to make sure that we can accept the patient before a human had to check the zip code, check the insurance, check the acuity level, and all that, and they did that one after the other. We now have an agent that does each one of those, and they do it all at the same time. So, what used to take a person 45 minutes to an hour now happens in about eight to 10 minutes.

Ashton Jones  15:50  
That's really interesting. So, what sort of impact are you seeing from the agents and from all these tools that you're implementing?

Evan Kramer  15:58  
So, on this intake process, in particular, we're live now in three states, and we're continuing the rollout process from there, but we're actually seeing a lot of really measurable results, which is exciting. So, I mentioned the time savings already, that's been really, really impactful. One of the other big impacts that we're seeing is, is that because we're able to move so much faster, and because we're so much more confident that we can actually accept a patient, we're actually missing out on fewer referrals than we used to. So, for context, hospitals in particular will typically blast out referrals to a lot of different providers, and whoever takes it first gets the business. We would oftentimes miss out on a lot of those because we just took us too long, but in the three states that we're currently live, we've seen that rate drop by about 20% so we're capturing more referrals, and that's actually directly showing up in ADC growth, which is really exciting. And then the last thing, and this is the most important, our patients are getting care faster than they were before, too. So we measure one of our quality measures is timely initiation of care, and we've seen that just in the last two months jump up about seven percentage points. It's

Ashton Jones  17:09  
amazing. Could you share a moment where the results of this technology really came to life, something that reinforces the why behind what you and your team do

Evan Kramer  17:22  
absolutely, and I've got a lot of these stories, but the one that's really stuck with me was that I flew out to Ohio to do one of our training launches for the documentation tools I was telling you about, and after that training was over, a nurse approached her manager, and she actually had tears in her eyes, saying that the training after the training she had called her father, who she was caring for at home, and when she told him about the tools, his response was, Does this mean that we'll get to spend more time together? And that nurse was actually about ready to leave Compasses, she was ready to submit her resignation, and because we were able to make the job less burdensome, we were able to free up more time for her to spend time with her father and care for her father, and she's still employed at Compasses today.

Ashton Jones  18:11  
Well, thank you for sharing something so personal and inspiring that really ties back to why the work matters, and it really illustrates the importance of focusing on the human element that goes with the artificial intelligence. Looking ahead, I want to talk about what's next for Compasses in the AI space, and how you see AI further enhancing hospice and home-based care.

Evan Kramer  18:38  
Well, as I mentioned up front, our goal with AI at Compasses is to profoundly differentiate the services that we provide in the eyes of our caregivers, our partners, and most importantly, our patients. So, I've got an AI roadmap that goes out for the next two years with all kinds of initiatives mapped out, and it includes things like scheduling and optimizing the patient experience and engagement process, our care delivery process, and then there's a lot of back office things too, so we're lots of opportunities to use the AI agents that I told you about to optimize things like revenue cycle and finance processes, and things like that, so ultimately I think AI will have a very positive impact on healthcare overall. As I mentioned up front, it will always be human centric. The clinician judgment will always guide care decisions, but AI can really provide a lot of support there, and it can remove the vast majority of the administrative burden and process friction that we all experience today. And as things get further along I think it's going to be really interesting to see how tools like remote patient monitoring and virtual care and digital patient engagement get layered in with all of this and I could really envision an ecosystem of all these different tools where AI is ingesting all this data from all these different points and. To help us drive smarter, more informed, higher quality care.

Ashton Jones  20:05  
Where do you see compasses in the AI ecosystem right now as compared to some of our peers?

Evan Kramer  20:12  
I think we're leading the way. Honestly, you're starting to hear a lot of discussion in our space about clinical documentation, I think there's folks are really starting to look into that technology. We've got that fully rolled out already, and it's not just one or two tools, it's all documentation and all disciplines. So definitely feel like we're leading the way there. And then things like the intake tool that we talked about,

Ashton Jones  20:39  
right?

Evan Kramer  20:40  
That's one that I'm very excited about, because it's not even a commercially available tool. It's something that we built in house with a developer, and I think that's something that's really differentiating Compassus amongst the rest of our peer set.

Ashton Jones  20:54  
Before we wrap up today, I want to ask you, what's that final takeaway that you want the audience to have coming out of today's discussion,

Evan Kramer  21:02  
overall, I think the main theme for me is that AI is a force for good in home-based care. It already is, and it will continue to help us achieve our mission here. That said, as amazing as that technology is, it's also not magic, so there's no shortcutting the hard work that's required to prepare our processes and our teammates to realize its full potential, but if we do that and if we're really effective at that, I'm really excited for what the next couple of years at Compasses look like, and our ability to really positively impact our clinicians, our partners, and our patients,

Ashton Jones  21:42  
that brings us to the end of today's episode of Home Is Where the Health Is. A big thank you to Evan for such an insightful conversation. In today's episode, we explored how Compassus is blending artificial intelligence with mission-driven care to redefine what's possible for patients, caregivers, and home-based healthcare. If you enjoyed learning about AI's role in caregiving and found this discussion insightful, please like, share, and subscribe to us on your favorite streaming platform. And don't forget to join us next time for another in-depth discussion on innovation in home-based care. Until then, this is Ashton Jones with Home Is Where the Help Is. Thanks for listening.