Leading Quality
Welcome to Leading Quality, the show that dives into the real-world stories and strategies of healthcare quality improvement leaders at all levels, from Frontline Champions to C-Suite Executives. Each episode uncovers how these dedicated professionals tackle complex topics in real healthcare environments. Discussion range from QI fundamentals, to leadership, technology, AI, and beyond. If you’re passionate about elevating patient care and want practical insights that go beyond the buzzwords, this podcast is for you. Tune in for inspirational conversations, innovative frameworks, and the behind-the-scenes details you won’t hear anywhere else, and discover how you, too, can lead quality improvement from wherever you stand in healthcare.
Leading Quality
Healthcare Needs a GPS for Life Outside the Hospital with Dr. Joshua Liu
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Why This Episode Matters
Patients spend most of their lives outside the hospital, yet healthcare teams often have limited visibility into what happens after discharge, between visits, or during long-term recovery. In this episode, Dr. Joshua Liu discusses how SeamlessMD was built to close that gap by helping patients navigate clinical journeys, giving care teams better insight into patient progress, and showing why successful health technology depends as much on workflow, incentives, and leadership alignment as it does on the product itself.
Key Ideas Explored
- The “black hole” after discharge and why patients need more support outside the hospital
- SeamlessMD as a healthcare GPS for surgery, chronic disease, cancer care, pregnancy, and recovery
- Why surgical pathways were easier to scale than complex chronic disease management
- How workflow fit determines whether health technology helps or burdens care teams
- Why strong outcome data may still fail to drive adoption without strategic and financial alignment
Takeaways for Quality Leaders
- Look beyond the hospital walls when designing quality and safety interventions.
- Do not assume better data or better outcomes will automatically create executive buy-in.
- Evaluate technology by how well it fits real clinical workflows, not just by its features.
- Engage both frontline teams and senior leaders early if pilots are expected to scale.
- Use patient questions and after-hours concerns as signals for improving education, navigation, and care design.
Continue the Conversation
Resources & Frameworks Referenced
Leading Quality is a podcast for healthcare leaders committed to improving systems, culture, and outcomes.
If you found this episode valuable, follow the show, rate and review the podcast, or share it with a colleague working to improve care.
Connect with Jason Meadows on LinkedIn for more insights on healthcare quality and leadership.
New episodes published every other Thursday at 7AM Eastern Time.
Credits:
Host, Writer, and Executive Producer
Jason Meadows, MD
Produced by
Thrive Healthcare Improvement
Edited by
Milan Milosavljevic
AI That Answers The Right Instructions
SPEAKER_00Now you basically have like a Chat GPT, but only for your instructions as a patient. Imagine you were to go to ChatGPT and you've just had your near place and you'd ask it, when can I take my bandage or my dressing off? ChatGPT would actually say, oh, like you could take it off in one to two days. But for a lot of like hospitals and surgical teams, patients get a special bandage now where you shouldn't take it off for maybe seven or ten days. And if you go into CM Less and you ask that question, we'll tell you exactly what your surgeon wants you to know, which is don't take it off for seven or ten days because you have a special bandage here.
Show Setup And Why Seamless Matters
SPEAKER_01Welcome to Leading Quality, the podcast spotlighting the people moving healthcare forward from the front lines to the C suite. I'm your host, Jason Meadows.
unknownDr.
SPEAKER_01Joshua Liu is a physician and founder of Seamless MD, a Toronto-based health technology company focused on helping patients navigate the parts of care that happen outside the direct view of the clinical team. His path into this work began in medical school, where research on avoidable readmissions pulled him toward quality, safety, and the persistent gaps between hospital care and what happens after patients go home. This conversation focuses on Seamless MD's work to support patients across clinical journeys, from surgery to cancer care to chronic disease management, pregnancy, and recovery after discharge. Joshua describes the platform as a kind of digital GPS for healthcare, guiding patients through each step of care, helping them understand what to do next, and giving care teams better visibility when patients are starting to drift off course. We also discuss a deeper issue about technology and healthcare. The issue is not that healthcare needs more technology, it is whether technology can actually fit into clinical workflows, respect the reality of busy care teams and limited resources, improve patient outcomes, and still scale in systems where incentives do not always reward better care. Josh shares hard-earned lessons about failed pilots, executive buy-in, workflow design, patient engagement, and why strong outcome data is sometimes still not enough. What makes Josh especially valuable in this conversation is that he speaks both languages. The clinical language of safety, quality, and patient care, and the operational language of building and scaling technology inside real health systems. This is a grounded conversation about what health tech has to get right if it is going to make care safer, more connected, and more humane.
unknownDr.
SPEAKER_01Joshua Liu, welcome to the show.
SPEAKER_00Okay, Jason, great to be here.
SPEAKER_01Good to see you again. We've uh our relationship has been kind of an intermittent LinkedIn messaging relationship mostly, and we've had a few virtual encounters, but I'm excited to get further into your backstory and and the work that you do with Seamless MD uh today.
From Med School To Startup Founder
SPEAKER_01So I you know wanted to just start with a zoomed-out view of who you are, where you're based, your clinical background, and the the path that led you into building Seamless MD.
SPEAKER_00Yeah, so I'm born and raised uh in Toronto. I actually never fell in love with a certain type of medicine the way most people do. So my well, I thought I was gonna be a pediatric neurologist when I started in med school. And then I started shadowing some folks in urology and realized it wasn't for me, and so I was lost again. And I kind of just bubbled my way around for the first couple of years. But I had some really great mentors who took me under their wing and said, hey Josh, why don't you spend a summer with us doing non-clinical things? So I did an internship at uh it used to be called the Center for Innovation and Complex Care. It's now called Open Lab based out of UHN. And what they do is they bring together designers, engineers, and clinicians to kind of redesign or re-engineer uh care delivery for the most complex, costly patients in the healthcare system. And I had some fantastic mentors, um, Howie Abrams, Dante Mora, who brought me under their wing and said, Hey, Josh, you know, we're really interested in preventing readmissions back to hospital. Why don't you go and study that with us? So I spent a summer doing that, became quite obsessed with that problem area. You know, this idea that patients can leave hospital, but wait, they're not truly back to baseline yet. So there's some real chance they bounce back to hospital. And, you know, why can't we catch and prevent those issues earlier? So that was kind of around my um in my mind halfway through med school. And around the same time, I got kind of bit by the tech bug. I uh, you know, it was around the time that I think startups started becoming a thing, and uh mobile and cloud computing were relatively new. The iPhone had come out a few years before. And I had a couple buddies in med school who were equally interested in in tech and healthcare. And so we actually tried starting a completely different healthcare tech venture, and we quitslash gave up pretty quickly. And one of my challenges was that we were all clinical people. None of us knew how to write code or or build a company. And so um that kind of died on its own. I did my third year of clerkship, and then my fourth year of clerkship, I still had kind of an itch to do something. And I came across a local startup incubator called the Next36. And on a whim, I applied for it, um, got in and started the incubator actually in parallel with my last year of med school. Met a couple of brilliant engineers and brought the readmissions problem to the conversation. And we basically started SeamlessMD through this incubator. I graduated, was supposed to start my uh family medicine residency in Toronto, never did it because I said, Hey, can you give me a year to just explore the startup thing? It's probably gonna die in a year, and then I'll come and you know beg you to let me back and start residency. And as you know, Jason, like healthcare innovation is very slow. So after you're trying to vote seamless uh full-time, we didn't get very far. And I said, Hey, can I have another year to see where this goes? And long story short, uh, I never did residency, and it it's been my full-time gig for about 14 years now.
SPEAKER_01That's great. Thanks for sharing it. And I I think I remember you telling me that they actually held your residency spot for uh for a total of six years.
SPEAKER_00Yeah, that's right. I would go back uh once a year to meet my program director, and he'd asked me, Josh, are you are you gonna come back this year? And I'd say, uh, no, not this year. And then after five years, uh there was a change in program directors, and understandably, the new program director said, like, look, like uh uh I'm new, I'm trying to like, you know, uh have a fresh slate here. And basically, if I didn't come back after six years, uh that would be it. And I said, Oh my gosh, I'm I'm surprised you let me do this for this long. I'm grateful for it. But I I knew by that point I was so committed, it just wasn't gonna happen.
SPEAKER_01Yeah. You've described the period after hospital discharge as kind of a black hole for clinicians. I think that resonates with anyone who's cared for patients in the hospital and then had very little visibility into what happens once they leave. How did that metaphor emerge for you and how did it shape the problem that you wanted to solve?
SPEAKER_00So it's interesting. Well, we we started um exploring the idea of seamless. I interviewed a lot of my like past mentors and peers of my mentors that I met
The Black Hole After Discharge
SPEAKER_00during my research on readmissions. And a very common concept was hey, we feel like we're sending the patient into a black hole after discharge. And my research initially was actually on preventing readmissions for exacerbations of chronic disease, so CHF, COPD, etc. Um, what was interesting though is most of those mentors and peers weren't interested in the idea of using technology today to monitor patients for different reasons. And so we actually had a lot of trouble finding early adopters for this vision. And then off-handedly one day, uh, one of those uh folks said, Hey, why don't you go talk to surgeons? Um, they they're more tech savvy, they like toys, they also have readmission issues. And we said, Okay, that's kind of disappointing, but sure, we have nothing to lose. Started code outreaching to surgeons, and they said, Yeah, we have readmission problems too, and no one's talking to us about tech like this. You know, let's go try it out. And so for me, like a big lesson actually was you've kind of got to like throw your idea out into the universe and and see where it sticks. And often it sticks you know on a part of the wall that you didn't expect. You know, they they also viewed it as a black hole after discharge. Um obviously um uh you know our vision has grown beyond just discharge um 14 years later. But I'm when I think back, I'm kind of surprised at how much this concept was persistent across many parts of healthcare, and then and also I'm surprised at you know not being able to predict where it would stick. I think that's true for a lot of innovation in healthcare.
SPEAKER_01Yeah, and and so I'm also curious, as a tech founder with a clinical background, how has your clinical training helped you build in healthcare? Where has it been most valuable? You know, understanding the problem or earning trust, designing the product, building the team.
SPEAKER_00It made such a huge difference, especially at the beginning, that if there wasn't someone clinical on the team, especially back then when technology was not as pervasive, or at least digital technology wasn't as pervasive as it is now. I'm not sure if we could have gotten it off the ground. Um, even in those first conversations with those um, you know, clinicians. I mean, you Jason, you know this as a clinician, like you're you're you can speak very quickly and fluently and using all kinds of clinical language that someone who's not from that world just won't get. And and to for them to get it, you'd probably have to really like step back and simplify your language and and really break it down. And most clinicians probably wouldn't have had the time for that. So the fact that I could just speak the same language as the clinicians and understand the problems they were describing uh I think really made a difference for us um to learn more quickly. And then I was able to kind of you know help educate my my co-founders on that language and bring them up to speed over time. But secondly, you know, you brought up trust. And I think even if we haven't worked together before, the fact that we've both gone through clinical experiences helps bridge that trust much more quickly to not only even get the meaning in the first place, but being a clinician adds a lot of credibility in those conversations because they know you've you've been in that environment with them. Even now, actually, in our company, about a quarter of our team comes from the clinical world. They may be you know from the pharmacy world or uh dietitian world, the nursing world and so forth. And so we actually have those folks be critical parts of the implementation and the partnership with an organization because what we found is we're you know getting so involved in the clinical workflows and the clinical protocols that when they know someone on the other side of that table speaks the same language, has been in the trenches, it not only makes implementation smoother, but like you said, there's that trust that gets instantly built. And I think that also leads to not just a better relationship with health systems, but even the product is much better because we have so much like internal clinical DNA that we have a lot of really good natural inclinations about where the product is good, where it's bad, where it's safe, where it may not be safe, and course corrected quickly. And I think you'd be missing that if you didn't have substantial clinical DNA in the team.
SPEAKER_01Yeah. So is that a is that a conscious choice when you add new members to your team? Are you selecting for that? And are you are
Clinical DNA Builds Trust And Safety
SPEAKER_01you seeing other health tech companies also selecting for clinical DNA like that?
SPEAKER_00So I would say early on it was an unconscious decision, and now it's a very much an intentional decision where there was a specific role in the company that that works with healthcare systems for implementation and maintenance of the clinical side where that person has to have come from the clinical world. So that that's kind of baked in at this point to the way the role is designed. But early on, it was more accidental, we just kind of uh you know brought in people we thought were great, and it turned out that a quarter of them in them need to be clinical. I think a mistake a lot of health tech companies make is they kind of conflate having real clinical DN in the company with simply having clinical advisors they call up like once a quarter. And I think what we found is that unless someone is truly in the weeds of your product and your healthcare partners day-to-day, they just won't have the context. Like you can have the smartest clutch in the world as an advisor, but if they're on an hour call once every quarter or once every month, like they're only going to be at the surface level of what you do. You have to be truly embedded inside the organization, I think, to make a really substantial impact clinically.
SPEAKER_01Yeah, it sounds like you feel you use the phrase being in the weeds clinically, uh, or being in the trenches, maybe you said. And maybe being in the trenches on both fronts is maybe some part of the magic ingredient there.
SPEAKER_00It is. I mean, I mean, I'm sure imagine like you're because you've been in the hospital environment for many years. If you think about, let's say, your electronic medical record implementation, um, the way one organization does it is different than another, right? So you have specific workflows that you've implemented, or you know, whether it's orders or something else, that's only done a certain way. And unless you live in that, it's hard for you to have the full context for like what's going on in your organization and how you might want to improve it. And so it's the same thing here, where unless you're working in a product every day, unless you're working with healthcare organizations and seeing all the outlier edge cases and the really unique workflows and things, you just want to have enough context to make a difference. And so I think a mistake health tech companies make is that they think just if they can just say they're working with great advisors, like that's gonna build trust. But actually, like I think what I found is having real subject matter expertise about the problem you're solving and the workflows and your product, like that's what builds trust with healthcare systems, so that when they ask you a question, you sound like you actually know what you're talking about because you do. But if you don't actually live in it, you can't even answer those questions as well.
What SeamlessMD Looks Like In Practice
SPEAKER_01I'm aware that we've mentioned your your company, Seamless MD, a couple of times at this point. Uh, but for listeners who haven't seen the product or haven't encountered it before, what does the product actually look like and and what does it do?
SPEAKER_00I'll start with an analogy to help explain it. Um, imagine you're traveling from point A to point B in a car. In the past we had paper maps, and now we have GPS. It's faster, safer, it's better. And my favorite thing about GPS is if you go off track when you're driving, it alerts you and nudges you back on track. And what struck us many years ago was there was no GPS for navigating patients through a complex journey. Maybe it's you know, pre- and post-surgery, a cancer journey, a chronic disease journey. It was still mostly paper and verbal. And patients would tell us they'd forget things, they'd lose things, those instructions were hard to follow. And then as you and I discussed, the care team felt like patients were in a black hole when they weren't in front of you in the clinic or the hospital. And so CBSMD is basically a digital GPS to guide patients through these journeys. So um, let's say you're going through a heart surgery, we automate your pre- and post-surgery instructions through text, email, and an application on your phone or your desktop. Um, so that way, night, you know, a week before surgery, you stop your blood thinners on time, the night before, stop eating on time. Um, we deliver education automatically. So when you go home, here are the rehab exercises, the videos that you should follow. And then very importantly, we help your healthcare team track data uh about your symptoms and your health status when you're at home so they can catch things earlier. So let's say after a heart surgery, patients would be tracking their pain levels, they could take a photo of their surgical incision, and then that data is available on dashboards that can be embedded right inside the electronic health record. So your team can get alerts, they can monitor your incision remotely, catch a complication earlier. And by keeping patients on track and helping the care team catch problems earlier, we've helped healthcare organizations reduce length of stay, readmissions, visits back to the ED, and then ultimately cost for the healthcare system.
SPEAKER_01And then you mentioned that your interest was initially in kind of what I'm guessing was the internal medicine and chronic disease space. And as you put those ideas out into the world and realized that those ideas fit nicely with surgery, you eventually shifted back into chronic disease. I can imagine how the surgery use case would have fit really well initially as you're describing, you know, this defined event with specific preparation, specific follow-up, you know, an incision that you can take a picture of. But what does it look like now with trying to monitor chronic disease and and what kind of diseases are you uh looking to monitor?
SPEAKER_00Yeah, I mean, what's interesting is a lot of the building blocks are the same, but to your point, the way you arrange the blocks that can be very different. So surgery was nice as a starting point because it's a very well-defined pathway. You know, there's a set of pre-surgery instructions that the patient has to follow. There's a set of post-surgery instructions they have to follow and symptoms to monitor. And it's a very clearly defined window, like maybe two to four weeks pre-op, and then you know, two to four weeks post-op, and then and then quote unquote it's over for the most part. You're hopefully back to baseline mostly by by you know a month or two afterwards, and then the patient's back on the next thing of their journey. And card disease management, as you know, is very different because it can be very it can be a lifetime journey. There's a lot more ongoing management as well. But the building blocks are are um very, very common. So, you know, let's say it's uh you know, hypertension, you know, we are tracking, let's say, vital signs and and symptoms. We are educating them about lifestyle and dietary management and medication adherence and and so forth. So a lot of the building blocks are the same, but there were some things about the the workflow or the platform design that had to evolve a little bit to support something that wasn't as episodic as a surgery. Um, similarly, when we branched out into um like women's health for pregnancy, we had to figure out how do we extend this to you know nine months of pregnancy and then you know months postpartum for oncology treatment. We had to figure out how to support um you know chemo radiation monitoring and then survivorship over a much longer time horizon, often beyond a year. And so the billion blocks are the same, but we had to kind of implement different configurations of that. Um I'm kind of oversimplifying it, but but I think it's largely true. Um but now we've we've done heart failures, COPD, hypertension, diabetes, stroke, cancer treatment. I mentioned women's health, but but we've kind of even gone into like the start and end of um the life cycle. So we started doing work and pediatrics in the past year, palliative care as well. And so I would say at this point we've supported over 150 different conditions, procedures, or treatments across the continuum. And there's still some things to the edges we haven't done yet. So, you know, transplant care, for example, is something that that where it's on the list, but we haven't done quite yet. Um, but I would say almost every major condition folks want to support patients on, we we've done, and now we're just kind of working at the edges. Um, the one other
Scaling From Surgery To Whole Journeys
SPEAKER_00thing we have that we're still trying to do more of is how do we grow this across more of the continuum of care. So we've done a little work of home care, um, we've partnered a little bit with primary care, but there's so much more to do beyond the kind of acute care hospital setting that we're we're still working towards I wanted to pause for a moment to invite you into something I'm starting with this podcast.
Share What You Are Seeing
SPEAKER_01If something from this episode connected with your own experience, where you've seen it work or not work, I'd love to hear about it. There's a short link in the show notes where you can share what you're seeing in your own work. It takes about a minute. I read every response, and over time I'll be sharing what we're learning together in future episodes and in other ways, and giving shout-outs during future episodes to people in the community who share ideas that really move the conversation forward. If you're up for it, I'd really value your perspective. Thanks. So I'm imagining some of these longer term journeys as you describe the heart failures and hypertensions.
Keeping Patients Engaged Long Term
SPEAKER_01How do you keep those people engaged and what what kind of additional value are they getting from the seamless experience?
SPEAKER_00Yeah, so I mean to your point, um, it's harder to engage folks like that long, long term. So the nice thing I'll I'll back up as a to contrast, so with surgery, what makes it a bit easier is that you have a life-changing event. Oh my gosh, I'm gonna have like major heart surgery. And so patients are a lot more primed to wanting to prepare and have the best outcome and make sure they recover properly and rehab and all that. And so it's almost many ways much easier to do this for surgery. That's maybe that's probably why it's a good thing we started there. Um, for chronic care management, you know, it can go on for months or years, and to your point, it can feel repetitive. It can feel like, well, I've I've had I've heard this education before, it can be very um repetitive in that sense. Um, I think one of the most important things that we found is when patients are when they know that what they're doing is being monitored um by. Someone back to the healthcare team, that's one of the most important ways to kind of keep them engaged. There are a lot of um, you know, there's a lot of consumer applications out there where you can, you know, track your blood pressure or track some other things where you might start off doing it, but if there's no accountability partner, you know, it's easy to kind of fall off track. Or if you feel like, well, like um, I'm taking my medications, my numbers look good now. Guess I don't need them anymore. I think we've all seen that, right? Um, yeah. But it's been very powerful when someone on the other side, maybe it could be a NP, a PA or physician who's actually looking at the data, and then you and then at the visit, they're commenting on the data that you've trapped uh remotely, and they're seeing it inside the EMR. That closed loop on the data, I think, has been the biggest reason um that we've seen engagement. Um and by the way, this is true both for chronic care, cancer, but also you know, surgery, even in the surgery world. Um, it seems obvious now, but when when the patient knows, oh my gosh, like I put a symptom in this program, and then 10 minutes later the nurse calls me to talk about it, that's a very strong, like you know, positive reinforcement to like, okay, like I should track my data because next time something goes goes wrong, like the nurse will call me and talk about it. We still have like pretty good engagement, even if no one's monitoring it, but it's it's kind of obvious. But like if you know your care team's looking at it, you're just way more motivated to to care because you know someone else is looking at it. And there's expectation that like, oh, they're expecting me to do this. Right? My care team wants me to do this. It changes the dynamic.
SPEAKER_01Yeah, it it it's almost sounds like some kind of almost social media experience of you know putting something about my myself out there and then getting getting some kind of feedback from the world, in this case from my from my clinicians. So I'm also curious about the the education part of this. I think you've you've mentioned before that the education, maybe to your surprise, was something that was durable and valuable to patients on the chronic care journey. And I will say, from years of internal medicine experience, the there's lots of conversations I've initiated about dietary education or you know, blood pressure management or or something else that um patients would look at me and say, wow, that's the first time I didn't know that these foods were gonna cause my, you know, contained a lot of sodium and were
Education That Patients Actually Use
SPEAKER_01gonna cause my heart failure to get out of control. So yeah, I'm curious how the the education part of it has played out with seamless MD.
SPEAKER_00Yeah, um, a lot of really interesting learnings along the way. So so one is, you know, to your point, we often I think we often overload patients in education, so we'll either kind of like uh verbal vomit, because we're trying to get everything out there to them. And and we know from the literature that they only retain maybe like 40% of what they've heard, or or or and they've heard that, I think I've read that they forget up towards 80% of the education that they that they you know hear. And then the second thing that we do is we give them these like really large booklets, which either it's just too much all at once, sometimes it's too high of a reading level, keeping the wrong language. And so one of the things that we really worked hard to do on the digital side is not just like digitize the education and just give it like a PDF version, which I think unfortunately is too often the case still in healthcare, it's how do we actually transform that education into something that's digestible and accessible to patients? So, what does that mean? So for us, it's let's break it down. So instead of saying the patient like 50 pages on day one, it's okay, maybe day one you get this bite-sized thing about you know, um your sodium intake, right? And then maybe on you know day three, it's something about learning about your medications. But the idea is that we break it down to bite-sized pieces so that way it's spread out, and then we actually repeat it every now and then to kind of positively reinforce education that that you you would have learned before, but maybe you've forgotten. Um, I think the second thing that's important is, as I mentioned, like making the literacy level right. So we bring it down to the you know, grade six, sometimes grade five reading level for all the education. And so we actually have um health literacy folks on the team, and everyone's trained on health literacy on the clinical side here, so that way we can ensure that even as we customize the programs to match the vision of the healthcare organizations, we're still investing in ensuring the right literacy level is there for patients, so that way that's not a barrier to digesting the content. Um and then the third thing that I think would be really interesting is just um the language piece. I mean, so often you'll find you walk into a hospital and you have these diverse communities, especially in Canada, and yet most of the education is only in English. And so, you know, we have some partners here where they may have seamless in like almost like 10 different languages, and the patient can just choose their language and they get it in the one that they can understand. And then we've started actually a couple years ago using AI to accelerate that because one of the challenges is no one can safely trust, let's say, Google Translate in real time because medical translation is still not perfect. And so folks would actually spend you know lots of money like hiring professional translators to turn our content into different languages in our program. And that's just not very scalable for organizations with very thin margins. We actually um began implementing AI translations in our back end where we tested a bunch of the LLMs to figure out okay, which LLMs are are good for clinical information for patients. And depending on the LLMs, some were better at Spanish, some were better at French and Chinese and Arabic, et cetera. And so we figured out which ones were the best for each language. And then what we've done is we've actually drafted translations for organizations in different languages so that way they wouldn't have to spend all these dollars on professional translators. They still may do that to like proofread and double check the translations and correct them, but it's it's brought the cost down significantly. And so it's not the coolest you know AI application that we've done, but it's been very impactful. And so there's, I think we found that a lot of these AI use cases are, you know, almost like under the hood, but can actually increase accessibility for patients. It's nothing to do with the technology, right? It's just the content. But if you don't make the content accessible, it's like it doesn't even exist, right?
SPEAKER_01Yeah. I can imagine a lot of people in our audience, you know, immediately seeing the value of that because it's so challenging, as you say, to, you know, it's the day of discharge or the day before discharge, and you're having to give all this information and you're trying to give, you know, even if you have some translated materials, you know, do the papers, where do the papers go after we give them to them? Do they get, you know, shoved under a uh, you know, into a bookshelf at home or something and never seen again? And if they keep them, you know, are they accessible? Do they understand them? It's it's a it's a big, big challenge. I also understand that a recurring problem in health tech is that technology gets introduced without fully understanding the workflow that it's entering.
Workflow Fit And Reducing Friction
SPEAKER_01Then the tool may be good in theory, but poorly aligned with how care actually happens. Um, what's your approach to understanding workflow before implementation?
SPEAKER_00So I'll tell you kind of what really mattered at the beginning and then how it's evolved over time. So, you know, one of the nice things when you first start is that you're completely naive, you don't know anything, and you're willing, and then you're willing to be able to do a lot of things that maybe are not as scalable. And so when we first started, we actually would like plant ourselves in the hospital, and we would just like sit in the clinic, watch how things were done, we'd we'd follow the patient through the journey and understand what the workflows were like in the in the actual complex clinical environment. And I think that was really important because it helped us actually figure out some of the workflows from the ground up. And there's a lot of things that are like non-intuitive, non-intuitive if you're not from the clinical world, like you don't actually realize how busy people are. And when clinicians say, Hey, like if I have to do it, I have to have to add something that takes another minute, as an outsider, it doesn't sound like a big thing, but for a clinician, you know, like a minute's actually a lot, right? Um, so be able to actually live and see that and figure out um how to embed what we do into the workflow. Um, that was really helpful. And it's not just the technology, it's about even things like, well, you know, watching a clinician educate a patient on what seamless is and why they should use it. We realized, oh, actually, there's a good and bad way to do that, right? And so we started building out like talk tracks and scripts um so that people wouldn't have to come up with something to say. They could actually have a best practice way that we found to work well in reaching patients. Um, and then realizing, oh, hey, actually, like it's important to have even basics, like it's important to have posters and brochures up on walls and and placards for staff to use, and even have like the talk tracks printed on a sheet that they can go back to, like all these little things that have nothing to do with technology, but you have to actually figure out how do you make this work in the workflow and make the initiative like organized and comprehensive and and and and and reduce the friction to doing things. So if you have to, if a clinician has to figure out how to, you know, the script for this, that's like one more layer of friction for them to make this work. And so I think a lot of it was almost like understanding how can we reduce the friction in the workflow. And later on that turned into okay, we could have integrations with the EHR and automate a lot of things. And so now, like we've we've done this in so many different clinical environments, each with their own unique aspects to it, that now when we work with the new organization, almost every new organization we can probably map to someone else that we work with already, and how we did it there made it work. That we now have like a playbook for like, okay, if you're like this organization, this is what we recommend. If you're like this other one, it's what we recommend. But it took like 14 years of just like working in so many different settings to have a playbook. And so it's almost like if you were starting out, I almost don't want to hand you a playbook. So I feel like we because like every product is so unique that you almost need to figure out yourself and have a playbook that works
Why Data Alone Does Not Drive Adoption
SPEAKER_00for your product, but you have to willing to do the hard work of like being literally in the trenches and like watching and learning and taking notes.
SPEAKER_01Yeah, 100%. Uh in healthcare, we often assume that if the clinical data are strong, uh, that people will adopt the solution. Uh I know that you guys have generated some really strong clinical data, but the buyer might also be thinking about return on investment or other kinds of incentives that are driving their own decision making. What's been your experience with the clinical data, the ROI data, uh as it relates to seamless?
SPEAKER_00So, you know, I remember when I first when we first started, I had this belief that I think a lot of clinician innovators do that if you just prove it in the data, uh, adoption magically happens. We retained this world of evidence-based medicine. And so it seems it seems like intuitive that if the evidence is good, then everyone just adopts it. I think we now know that that's not as simple as that. Um we had some tough lessons though. Like we we had you know several times early on where we'd meet some great clinical champions at hospitals, and they'd say, hey, if we just prove that this reduces readmissions and ED visits, trust us, the executive team will invest in this long term. And we're like, great, let's prove it together. And we'd spend six, 12 months doing these pilots and cut readmissions in half, and then we'd co-present the data to the hospital C-suite, and um they'd say, Hey, great outcomes, great for patients, but uh given the way that we're uh you know we're paid and funded in Ontario healthcare, um, if this actually saves us money, the government will think we need less money next year. And and so actually it just doesn't really save us money because we'll just get a smaller budget next year. And it shocked me that our incentives were so perverse that if there was a great way to improve an outcome, that there wouldn't be aligned incentives with the funding model. Um, but I think the other wake-up call was, well, you know, making a lot of assumptions about what actually mattered to the organization this year. And so now what we've realized is like, look, like we have to trust yet, verify. So even if we're solving the right clinical problem, does this align with the quote-unquote business problem right now, a business priority for the organization? And and you only get that by asking. So now, you know, we have to go to the executive sponsor before even like implementing this and getting alignment on, hey, you know, does this actually address a priority you care about? Does reducing readmissions or reducing length of stay actually make a meaningful dent on your business goals this year? And if the answer is yes, then there's alignment. And if we implement this and show value, it can scale. But the answer is like, no, actually, uh actually not, but maybe there's something else that we could solve instead that we align with and find alignment, but maybe there isn't. The point is that by by skipping that step, um you you can you know run into a ditch basically, despite great data and great outcomes. I think the second thing we've learned too is that adopting technology is very much an emotional uh decision, and often data is used to justify an emotional decision afterwards. So um, you know, if people are bought in, like they they want the data to be good. And so they look they look for reasons for the data to support that emotional decision to invest. Whereas if folks were not involved in the decision and maybe were were excluded in the decision making, it doesn't matter how good the data is, they they often have a bad taste in their mouth and and and they'll they'll feel like they weren't part of it. And so there's this really interesting human dynamic we found with adoption where like you have to engage all the right stakeholders, you have to get buy-in, people have to be emotionally bought in so that when the impact is good, they were part of it, they're happy with it, and they'll want to continue. Um, it shouldn't be the case, you would think, but humans are a fickle species. And healthcare is very human.
SPEAKER_01Yeah. Um, it is. And it sounds like you've learned your way into um this sophisticated change management strategy. I don't know if you call it that, but that's kind of what you're describing. And I think you've written about your general approach to having kind of three key champions that you uh need in order for something to succeed.
Three Champions You Must Have
SPEAKER_01Can you tell me about that?
SPEAKER_00Yeah, so the one that most people think about is the clinical champion. You know, is there someone in the frontline clinical world who's willing to champion it with their fellow physicians, nurses, Allen Health, and so forth? So that way you'll get buy-in um at the front lines to put your technology in the clinical workflow. Um, then you have, I think we we think of them as like an administrative champion, someone who like often it's like a director or manager of a service line or a clinical department who has to actually make sure that everyone there is aware and supports it and promotes it and they keep an eye on the data. Um, they make sure that the day-to-day operations runs well. And then there's the executive champion who's really that sponsor, usually in the C-suite or the VP role, who directly holds the budget to fund this and is responsible for some organizational level KPIs that hopefully your technology is helping to solve so that way they can justify the investment. And if any one of those three were to disappear tomorrow, your initiative could be in trouble, right? So, like obviously, if you don't have your executive sponsor, it doesn't matter how good the data is, no one's gonna pay for it long term. But if you have a great executive sponsor, but you don't you but your frontline like champions leave because they moved organizations, maybe they got got promoted and feel like they have other priorities now, and you lose frontline staff believing in this, you lose them wanting to invest in promoting it to patients, um, it can fall there too. And so I think we've learned that even if you get this into an organization, we are constantly having to educate, meet new stakeholders, navigate stakeholder change, and and continue to evangelize what we do. Um, because people ultimately are the drivers of innovation in the organization. And if people are not consistently engaged, um the innovation will eventually just die in most cases. And so it you almost have to like re-evangelize this constantly, more than we realized, actually. It's very much a people engagement um partnership beyond just the tech now.
Getting Past The Pilot Into Partnership
SPEAKER_01I I've also understood that health tech can kind of get stuck at or or maybe shortly after the pilot. And I know you've talked about how important it is to create durable relationships with with healthcare partners and not just have impressive pilots. Um, what can you tell me about that? What have you learned about kind of making durable partnerships?
SPEAKER_00Yeah, I I think a key thing is um not assuming that everyone knows what's going on in the organization. I'm always surprised at how like we're doing this innovative thing, and and even if some executives actually were part of it and know about it and have helped support it, that doesn't mean everyone else in the organization knows about it. And so one of the very powerful things that we've done is just work to reduce friction to folks educating others internally and their communities about the initiative. So what does that look like? So, for example, that could mean we say, hey, like, you know, why don't you do a hospital like announcement about this initiative on your website um or in your local community so that way your patients know about it, so that way your team knows about it, because often people are reading their own hospital newsletters. Let's do a case study about the results that we can celebrate that you can then share with your organization so that they um feel great about the success. Um Data. We produce you know monthly and quarterly data reports that we you know share with the team. So even though we have these dashboards where a team could go in and look at the data and create their own reports, people are busy, right? So we put in the effort of like, hey, here's a quarterly report, here's an annual report of the data, the milestones, what you've achieved with us. And actually, I'm surprised at how often those reports end up being used to inform like annual reports for hospitals. And so because we've reduced the friction to make the data available, people actually use it and share it more often. Whereas if you say, oh, hey, can you on your own prove the business case and ROI to your organization like completely on your own? It's a lot of work, right? So the more that we can just streamline getting data in the hands of people to make it easy to share, that spreads the word. And to us as a company, we think it's obvious that like, oh, why wouldn't they want to just put in the effort to share it? But I mean, you know this, Jason, being the hospital, like you have a million things going on. Like the top of your list is not, hey, how do I make it easy to share my seamless innovation data with my hospital? Like, that's not near the top of your list. It's probably not even on the list, to be honest, right? So how do we make it easy for you to celebrate success and share data? It's a very, I think, under under um explored thing in health tech companies because we just assume things will magically spread on their own.
SPEAKER_01Yeah, that I mean this is pure gold learning for for anyone who hasn't gone through the the journey that you have. You know, they can speed right past a lot of these hurdles.
Conversational AI With Tight Guardrails
SPEAKER_01I think in 2026, no health tech uh conversation is complete without talking about AI. I know SeamlessMD has developed conversational AI within the platform. When did that become part of the product and how's it going?
SPEAKER_00So, you know, like a lot of companies, we've we've spent a couple of years thinking about well, what's what's the way in which we can use AI to like really accelerate what we currently do. It started off with the translation piece I mentioned earlier, but over the past year, and then this finally became public because we uh went live with healthcare organizations a couple months ago, we we came up with, I think, a very safe and effective way to embed AI into the patient experience itself. And so what we launched a couple months ago, and now it's on live in about, I think, 10 healthcare organizations across Canada, the US, is what we call seamless answers. And basically what we did was apply a retrieval augmented generation AI model on top of the healthcare-approved education already in our platform for organizations. So that way a patient essentially can interact with a conversational AI chatbot where they can ask any question they want about their healthcare journey, and it only responds to patients based on the education that's already been approved and customized by their healthcare team. So, what does that mean in practice? So, you know, I'll go back to our example about um like heart surgery. In the past, before we had AI and Seamless, if you wanted to learn something, you'd have to go through like, you know, our entire educational library digitally, which again was still better than paper. You could search it and stuff like that, but you couldn't ask a natural language question and get a specific answer. Now you basically have like a Chat GPT, but only for your instructions as a patient. So why this matters is imagine you were to go into ChatGPT and you've just had your near place and you'd ask it, when can I take my bandage or my dressing off? ChatGPT will actually say, Oh, like you could take it off in one to two days. But for a lot of like hospitals and surgical teams, patients get a special bandage now, where you shouldn't take it off for maybe seven or ten days. And if you go into CMLS and you ask that question, we'll tell you exactly what your surgeon wants you to. To know, which is don't take it off for seven or ten days because you have a special bandage here. And so the idea is that we're giving patients the appropriate safe answer that matches the instructions that their care team is giving them. And what's really interesting is that this is being used in times of day where you can't normally reach your doctor or nurse. So about 36% of all the questions patients are asking is coming in after hours, you know, between 6 p.m. and 8 a.m. And so we know that we're filling a critical gap for patient education and making just easier for them to access the content that they need.
SPEAKER_01So my experience and maybe some of our audience uh experience with AI, even if you give some kind of custom guardrails, let's say you know create your own custom GPT within ChatGPT, it's really easy to make that go off, you know, outside of the instructions. In fact, it can be hard to keep it in the within the instructions you've given it. I imagine that's different for for seamless answers, but I'm curious how you guys have developed that so that it is safe and reliable to stick to the script.
SPEAKER_00Yeah, it's one of those things where we end up having to be almost more strict than a chat GPT would be. So I think we uh we will end up having more cases where the AI says, I can't answer that question because we have such tight guard rules, unless unless the model is very confident that there's specific content in the knowledge base that relates to that question. And so this the good thing is that it's very safe, it's not gonna it's very unlikely to say something that's very inappropriate, but it also means there's more cases where it doesn't feel confident enough to give an answer. And so one of the things that we've been doing is monitoring, okay, like what are the questions that we're getting where it doesn't feel confident enough to answer? And then we go back to the healthcare team and say, hey, like these are topics that people are asking about where the AI doesn't feel confident enough to answer. Can we add more content to the knowledge base so that way we can answer it next time? And so I think that the really neat insight is we're now like learning about the topics and questions patients are actually asking, which the healthcare team sometimes didn't even realize was important to the patient. And so whether or not they put that content in seamless, it gives insight to the healthcare team about the questions patients really care about, and that's gonna change education in some way, whether it's through seamless or just how the clinic or the organization does things. I think that's gonna add some really neat data that we that they just didn't normally have before. And I think they're gonna be surprised at what patients actually care about, and it's often not what we think.
SPEAKER_01Yeah, I'm sure, I'm sure it's not what we think. Um what have you what have you learned on that front so far?
SPEAKER_00Yeah, well, it's interesting. Patients will often ask like very, very, very specific things that we didn't realize was important. So, for example, I'm trying to think. Um like there's a lot of like operational things. I mean, now this is not an issue, but like in the past, like very specific very we think are simple things like, oh, like where can I park? How do I get there? It wasn't clear in the instructions exactly how I got there. Um, or where do I find this thing in the hospital, like wayfinding and stuff like that? So like there are all these like non-clinical things that patients like care about or worry about that should be in the knowledge base, but as a clinician, it's not really top of mind for us. Or sometimes we find some very unique like symptoms that maybe it turns out they weren't the most common thing, but now that we have enough data, it's common enough that we should probably have something in there about that symptom or side effect. But we wouldn't know unless it was surfaced in the questions.
SPEAKER_01And uh are these questions coming out in the middle of the day? Are they getting asked after hours when you know people are just going to bed? When when do you getting when are you getting these questions?
SPEAKER_00So, surprising or not surprising, 36% of questions are coming after hours, which for us is defined as between 6 p.m. and 8 a.m. Which which shows I think there's a real gap in the healthcare system where you know your care team has to go home and see their family and sleep too, but you still have questions. And I think what's neat is if the patients couldn't ask that question seamless, they probably would be going to ChatGPT, which probably is fine for some questions, but maybe not fine for other questions.
SPEAKER_01I'm really excited about where this all goes and and the work that you're doing towards fixing the black hole that we talked about.
The Five-Year Vision For Seamless Care
SPEAKER_01If this work achieves its highest aspiration five years from now, what is SeamlessMD doing that it's not doing today?
SPEAKER_00A couple of things. I think one is just um well, I think it's really it's impact, but it's on a few different dimensions. So one is growing across many more uh conditions and treatments at the edges that we haven't done. So we talked a bit about, I think, transplant earlier. We're looking to do work in inflammatory bowel disease, more pediatrics, and so just being able to reach all kinds of patients and families who we currently um don't support yet today. Uh, the second thing is just um geographical impact. So right now we have a footprint in Canada and the US, but hopefully in the next five years, we're doing this work outside North America, um, hopefully in Europe and Asia and some other places too. Um and then third is just continue to do more across the whole continuum. So, you know, we're we're still doing a lot of work at the hospital, a little bit of primary care, a little bit of home care. But my my dream would be that um as the patient moves between primary care and acute care and home care, that they can stay on the same seamless experience. And then maybe the the clinical parts are different, maybe the care team is different, but they're all using the same platform and the patient is using the same platform. Um and so that only happens if we're able to grow across all these different um parts of the market. But if we can do that successfully, then hopefully we can have truly a seamless experience for the patient digitally, uh, no matter where they move in the healthcare system.
SPEAKER_01What a what a great vision to have that GPS, as you put it, uh be able to navigate them through all those aspects of the medical experience. Josh, this has been a great conversation. Thanks for taking the time, especially. I know you're a new father, and congratulations again. Really appreciate learning more about Seamless MD and everything that you and your team are building. So thanks so much.
SPEAKER_00Thanks, Jason. Appreciate the opportunity to be here with you.
Where To Follow And Final Thanks
SPEAKER_01If anyone wants to uh follow your work more or get in touch with you, how would they do that?
SPEAKER_00Uh Joshua Liu on LinkedIn and Joshua P. Liu on I still call it Twitter, but you can call it X.
SPEAKER_01I still call it Twitter too. Um very good. So I'll uh we'll link to those and any other uh resources that we've mentioned in the course of the conversation in the show notes. Again, thank you so much. Great speaking with you, and I hope uh we get to do this again sometime. You got it. Thanks, Jason. Thanks so much for listening to today's episode of Leading Quality. If you enjoyed the show, please take a moment to like, subscribe, and share it with someone who might find it useful. You can find all our episodes at leadingquality.budsprout.com or in your favorite podcast app. The show is written and hosted by me, Jason Meadows, edited by Milan Milostavievich, and produced by Thrive Healthcare Improvement. See you next time.
People on this episode
Podcasts we love
Check out these other fine podcasts recommended by us, not an algorithm.
