BoneChat #49 - Bringing Smart Implants into Orthopedics - Bill Hunter, Founder/CEO at Canary Medical

So, uh, welcome to Bone Chat discussion number 49. Uh, today's extra special because we're joined with Dr. Bill Hunter, founder, CEO of Canary Medical. Uh, Bill and I talked on my podcast, I, what was it, 3 years ago, I think.

Um, at least, yeah.

At least, and a light bulb went off in my head when you said two things. You said, uh, cardiology is all math, and you said, why do orthopedic implants have to be passive? And I thought, wow, what a great question. Um, so fast forward to today, today, I can't wait to kind of catch up with your progress and, uh, handing the mic over to you, Bill.

Thanks.

Well, well, thank you, Tiger. I— Tiger was really the first one to Discover Us. I lost my way, spent 25 years in interventional cardiology making stents and balloons and all kinds of things like that. So, go easy on me on the orthopedic side here. But I thought what I'd do is a bit of a semi-random walk, just some quick slides that will give you an idea of not just what we do, but the challenges we face and ideally, you know, stimulate some discussion. So, You know, when I first started doing this and Tiger Buford interviewed me and said, this is kind of interesting stuff, but, you know, what does it mean? And I don't know that I really had a good answer to that question, but this is part of what I— the way I've been looking at it is we all talk about digital surgery. We're doing all this stuff with robotics and VR, AR, and all these types of things. I'm fond of saying that we're putting in implants right now with a micrometer, but we're measuring them with a sextant. And, and what I mean by that is that the implant itself provides absolutely no feedback. And the feedback that we do get from the patient is, is word of mouth. It's, it's subjective. It's, it's, you know, patient-reported outcomes. And I believe that one of the reasons that robotics has not demonstrated clear benefit over existing surgeries is that we're just not measuring it accurately enough to be able to tell. And so I think a big part of things is you can't fix anything you can't measure. And the first step of doing that, in my opinion, was to have a device that provided some sort of feedback and some sort of digital interface that was at least at the level of resolution of the devices that were used to put it in. And what we're seeing as we get the outcomes is that we're starting to have objective outcomes. And I can tell you right now, the objective outcomes don't always overlap perfectly well with subjective outcomes, but it does at least give you something granular to hang your hat on as, as you go back and look at the data. And I'm going to show you some stuff that's really, really brand new. So just for those— one slide for those who haven't seen it before. Basically, this is a glorified pacemaker. So it has a pacemaker battery in here for power. We have an antenna on the top. We use the telemetry that pacemakers use to transmit, and that's relayed through a base station in the patient's home. And then the next day, the doc and patient are looking at their data. If we transmitted in real time like your iWatch, we'd be out of power in a couple of days. But because we just transmit once a day, we can make this thing last for 20 years. Just a little, you know, advancement, Tiger, since we first talked, there's almost like Moore's Law is happening in batteries. About every 3 years, the power of batteries doubles. And so we launched this, which is the stubby version of the Persona Stem, last week. And this actually has more power than this one does. And we can get 20 years of patient monitoring off of this.

And the bulk of the stem contents is the battery.

It, it really is. So this is actually where we are technologically now. So this is now approved by the FDA. So this is 3-year-old technology. So this is where we are now. And so what we'll be able to do in the future is just drill a hole in the implant. And so we'll just put it in like that. So we don't even change the outside dimensions. And that's how we'll get into hip and shoulder and other things like that. So we'll go to kind of a cartridge design. To, to get into more indications. Took us about 2 years to really build what I would call the digital underpinnings. So the first thing we, we had to do was create what we called recovery curves. And for the docs online, I think of these almost like the pediatric growth curves, 95th percentile at the top, 5th percentile at the bottom. We had no idea what normal was. So, how many steps does a patient take on day 21? No idea. What's normal range of motion on day 28? No idea. So, it took us several thousand patients before we could do this by sex and age. We'll soon be able to do it by BMI and diabetes status, but there's basically a line that you're looking at. And for any one of the parameters we collect, and we collect a bunch of activity parameters, so step count, cadence, stride length, distance traveled, walking speed. Is my patient on the couch or are they moving around? And then we have very, very precise gait kinematics. So we're looking at the device movement in 3-space. We're doing it at 40 times the visual scale, hence my first comment about robotics. So we're able to see, is it loose? Is it vibrating? Is it medial lateral, anterior posterior stable? Or is it, is it not? So we built that and that allowed us with a single glance to say, hey, my patient's average, above average, or below. That was a big first step. Um, my biggest user has put in about 1,000 of these now. And so you can imagine, no matter how simple we made those curves, no one's going to look at 1,000 curves every day to see how their patients are doing. So we built an algorithm that does that. Um, and that with a single click of a button, you can go through and within a fraction of a second, see the 5 or 6 patients that are falling behind or are not meeting your clinical expectations. We can push that to your phone so that you're looking at that on a Monday morning as to who you might want to book in office that week. And most importantly, we've now integrated it into the billing system so docs can get paid $1,000 per year, uh, to do remote patient monitoring. And if you're doing 700, 800, 900 joints a year, it's probably worth having somebody dedicated, uh, to look at that. So this is kind of just the single clinical example I'll give. This is a patient who doesn't have adequate flexion postoperatively. The patient is completely oblivious to this. They don't know what to expect. They've never had their knee replaced before. They just know that they're moving, but they don't know that they're not moving enough. A signal goes off to, to let the practice know. The practice contacts the patient, says, hey, I'd like to modify your extra— your exercises, you know, maybe some compression stockings, wherever the case may be. And you can see that within, you know, you know, 2, 3 weeks, you have the patient pretty much back to normal. If that patient didn't see their doc until the scheduled 3-month follow-up, you might have a fixed deformity. You might have something that, you know, you weren't able to just fix with PT. So the object here is to identify these people who are falling behind, identify why they're falling behind. Is it an extension problem? Is it a flexion problem? Is it a stability problem? And then allow, you know, for a non-invasive intervention to get them back on track. That's really the principle of the whole thing. So this is brand new data. It's based on many thousand patients, and it's meant to be a little bit cheeky and a little bit evocative. So what is the real failure rate of TKI? So we've been saying for, I don't know, 40 years that 20% of patients are unhappy with their outcomes. And that hasn't really been touched by biomaterials or robotics or anything else. And that's what we see. The data would suggest that all patients are not created equal. So what I'm trying to show in the middle here is that it looks like the mechanical failure rate— and these are my words, not anybody else's— but the mechanical rate of failure is about 5 or 6%. And by that, I mean, it doesn't seem to matter what you do to this patient. You give them great care, you, you monitor them every single day, you change their exercises. It just doesn't seem to matter. So I suspect that these are the ones that either just weren't good candidates for TKI, or maybe it was a situation where, you know, the implant didn't go in quite right or wasn't the right anatomic match for the patient. And for whatever reason, again, I'm not a surgeon, but they seem to be really resistant to treatment. Incidentally, and this for you, Mark, 55% of people do really great. They can sit on the couch, eat Krispy Kremes all day long, do absolutely nothing, and they get a fantastic outcome. We spend $15,000 on rehab on those patients. I'm not sure we're getting much bang for our buck for those patients in that cohort. And then about 40% of the people you can move around on the chessboard. Some people start off really well and end up poorly. In fact, if you start off too fast, you have twice the failure rate as you do if you start off too slow. And on the other hand, some people, you know, look terrible, but, but you're able to, with the right interventions, get them into the green. And so my belief is twofold. One is that if we can manage those yellow patients, which I think we are looking at soft tissue healing there, we're looking at the patients that can actually be maneuvered and optimizing that soft tissue healing, you could take the failure rate of TKA down to 5 or 6%. I believe that if you had enough data on that 5 or 6%, you could do feedback with robotics to figure out if the anatomy was different in each one of those and start to tease out different surgical techniques in order to even take a crack at the 5 or 6%. Again, all speculation, but, you know, we're now at 4 billion data points. And the data is starting to tell us a bit of a story.

So, so, Bill, the data is in the cloud somewhere and there's a collective amount of data, but there's also hospital-specific, surgeon-specific data. So who can see what?

Right now, we— everybody can see whatever, you know, they want within reason. So the doctors see everything. So we show them all the stuff. The doctors don't necessarily want us being too explicit with the patients. They want us to let them know whether they're doing okay or not. They, you know, if somebody's in the 32nd percentile, that's actually fine. They're probably going to recover just fine, but they don't want their patients freaking out that that means something's going wrong. So the user interface for the patient's a little bit different than the user interface for the doc. The doc gets the numbers. The patient kind of almost like a Peloton, you know, keep up the good work, you're improving, you know, your arrow's in the right direction, that type of thing. So Tiger, you asked me to write a piece probably 3 years ago about data and what it was. And at that time I said that data companies in healthcare start as monitoring companies, they migrate into becoming diagnostic companies, and they end up as patient management companies. And so this is where we are at. So we, we started with the monitoring. We can now do the screening digitally. Think about it. That's taking what is historically a very labor-intensive process. Patient comes into the doctor's office, is, you know, watch the patient walk, do range of motion, provide feedback to the patient. That's now all done in the wild every day, if you will. Without any input. Again, back to the one practice I was referring to, they think they're going to save 2,400 office visits this year by not doing some of the follow-up on that 80% of patients who are cruising right through to the finish line. Obviously, if you can't get paid, no one's going to do it. If you get paid for, for seeing a patient in person but not remotely, you know, that's going to skew. So we had to get that done. We've already got some diagnostics. We can tell what a loose implant's like. We can do some interesting things with that. We've actually diagnosed a couple of strokes. So patients started walking abnormally, and they thought it was their knee, and we got a video and said, no, that's a hemiplegia. You know, you need to go to the emergency room. So we're starting to match all these gait abnormalities to personalized care. When I, when I went to medical school, they taught me that there were 31 gait abnormalities, you know, Parkinson's and MS and congestive heart failure and all kinds of things. We're actually starting to be able to see it. So this is actually a health monitor as well. And then ultimately you want to get into patient management, and I'm going to show you some cutting-edge stuff that's not available yet, but, but that we're working on. So ChatGPT-2, if I'm not mistaken, was actually trained on only 2 billion data points. Um, this is 4 billion data points plugged into ChatGPT-4. Uh, and when you do that, you can train the algorithm how to read those recovery curves. So once we were able to tell whether a patient was above average or below average, that was great, but all we did was shift the bottleneck. So now somebody in the practice had to look at the data and make a clinical decision. Otherwise, this was just a giant waste of time. And so there was a lot of time spent teaching somebody in the practice how to read these algorithms. Well, that can all be done now by ChatGPT-4, just like a pap smear or, you know, an X-ray or an ECG is, is read by that. If you want to look at them, you can, but we can give you an interpretation of, of what the likely outcome is. So that eliminates kind of the practice bottleneck. And then of course the last step is to throw all this stuff into a large language model. And, you know, in that situation, we're, you know, not quite there yet, but Mrs. Jones can pick up her phone in the morning and say, hey Siri, how's my knee doing? And they can say, all right, here's where you're up to. I'm happy with your activity levels. Not so, you know, looks like we need to work on your range of motion a little bit. And, you know, here's the exercises I'd like you to do based on what the data is telling you. And so if we do that, We've digitalized the entire clinical process, obviously with oversight of the clinician, but we all know that we don't have enough docs. We all know that we don't have enough people to do all this stuff like that. This is a way for a fairly significant proportion of the people to do this without, you know, having to be in their doctor all the time. And this is where we're getting to. And the last slide here is—

Bill? Yeah. Is that interface a subscription or is it just part of the implant service?

It's just part of the implant service. Got it. So this is for Mark. So we start to look at things and I used to, when I started doing this, I'd say, all right, people would say to me, well, Bill, if you don't, if you can't detect infection, this is kind of a big waste of time. And so we went and looked at the numbers to see if that's actually true. So this is 1.4 million Medicare patients. This is actually getting published because to my, my astonishment, nobody had ever actually done this before. And so Medicare spends $5 billion per year creating knee patients. So that's what they pay their doctors. That's what they pay their implant companies to put in total knees. They spend $15 billion looking after them. So to no clinician's surprise, creating a patient is nowhere near as expensive as looking after a patient. Very intuitive, but not often collected. And the reason is, is that the aftercare is a grab bag of things. It's X-rays, it's MRIs, it's home health, it's, you know, readmissions and the like. But if you— I don't know if you can see my cursor or not, but the complications, the cost of complications, actually relatively small, right? Your big expenses are ER visits and readmissions. Your home health, 70% of Medicare patients get home care visits. X-rays and MRIs are, I think, all of those are at least double the cost of all the complications combined. So it really, it's patient management where the money is if you're looking at it. And being able to do this is very important. If you look at how Medicare spends their money, they spend $5 billion. So the equivalent of the entire knee market in days 1 to days 90. They spend the knee market again, $5 billion more, day 91 to day 365. And the data I didn't believe, which is year 2, they spend another $5 billion in year 2. I thought that all of those— all of that cost would have worked its way through the system by the end of year 1, but that's not the case. And so I'll leave it there. And hopefully there's some points of discussion here, but I think we've already reached the killer app, and the killer app is looking after the patient. And I call this continuous glucose monitoring. I call it CGM for orthopedic surgeons. And the key with diabetics is to keep them out of the hospital, right? The key with diabetics is to keep them— their glucose within normal bandwidth. And if we can keep our patients within normal recovery limits, that's actually where the money is. You know, I would love to get to the complications. We're working hard on it. But I don't think that's actually where this goes. I think this goes to skewing those yellow patients into becoming green patients and cutting the cost of care associated with it and the number of human beings required to actually do this, you know, amount of work. Because total knees and hips combined are the largest surgical procedure in the United States. They're even more common than C-section. So it's, it's not a trivial cost for the system.

Great. That's where we're at, Tiger.

Thanks for that update. Wow, it's been— you've done a lot in 3 years. So Elizabeth's up first and then Mark Maloney.

So first of all, I have to tell you that the Canary Assistant is already pinging on my phone. I'm in physical therapy and this morning I only did half of my exercises. So the app would be going, hey Elizabeth, get going.

Exactly.

I'm blown away by this. This is very impressive. I seem to recall hearing about it a couple of years ago. I didn't dig into it. I'm not sure why. And it's, that's my main point. And I'm curious, what does the next 6 to 12 months look like for you, for the company?

Yeah. So, you know, we've kind of got 3 major efforts underway. So the first is to basically do what I call, as corny as it sounds, the knee CG, right? So we have all these curves, right? We have all these, you know, how the knee's moving. We have 6 dimensions of knee movement. And what we would like to do is to be able to say, see this squiggle right here? That's a fractured patella. See this squiggle right here? That's an evolving infection. See that squiggle right here? You know, that's a patient with a DVT. And, you know, again, my background, I spent 25 years in interventional cardiology. You know, we use an electrical wave, an ECG, to diagnose myocarditis, ischemia, valvular heart disease. None of those are, are electrical events, but an abnormal heart beats differently than a normal heart. And we see that in the tracing of the ECG. We are starting to see that in how a human being walks. If you've got a hot red knee, you walk differently. If you've got a DVT, you walk differently, right? If you are developing Alzheimer's disease, you walk differently. Differently. If you're in CHF, you'll walk differently and your activity metrics change. So that's where the big data comes in, where I'm hoping that in addition to what I just showed you, there'll be a little box on the right-hand side saying this patient's at high risk for a DBT or this patient's at high risk for loosening and so on and so forth. Because really gait is a biomarker on steroids, right? It requires an intact respiratory system, cardiovascular system, musculoskeletal system, neurological system. And if you're monitoring somebody's gait from 7 in the morning till 10 o'clock at night every single day, even subtle changes you can pick up. So that's really important. The second thing for my orthopedic surgeon colleagues, and I'm way out over my skis here, but everybody's interested in, okay, let's look at preoperative gait. Let's look at intraoperative data. Let's look at postoperative gait and see if we can tease some of this stuff out. You know, we've got all these different surgical approaches. We've got all these different ideas of what the right way to do, and we want to personalize the surgery to a patient. Is kinematics a way that we can use to really personalize that surgery? So those are the big research efforts that we've got going right now.

Very exciting. Congratulations.

Thank you. Mark.

Yeah, that was, that was an amazing presentation. I don't know how many epiphanies I can handle these days., but that was pretty, pretty phenomenal information. In the last couple years, I've been working really hard as an advocate for independent medical practices and bringing them interesting ideas on how to create revenue. Yep. And I've really hammered home things like remote patient care, remote patient monitoring. There was a, what, Canadian company I work with that had a, an app that was monitored balance.

Right.

Because the number one reason for readmission after total joint is, is falling down.

Yep.

Yep. Your 31% really captured that very well.

Yep.

Are you familiar with a Canadian company out of Montreal called Imovi?

Yeah. They, they have a, a wearable, do they not? Or, or brace, brace-based?

Well, it's a, it may be temporary. It's, they, they, they throw some infrared markers on your knee and then they, you walk 5 minutes in treadmill and they call it the knee KG.

Yeah.

Yes, yes, yes, yes. I, I know exactly what you're talking about.

Yes.

Okay.

Yeah, just see, it just seems like, you know, their 14 years of work in this might in some way help you as a, especially as a fellow Canadian.

Yes, exactly.

So is any of what you're doing powered by any of those CPT codes I mentioned, like remote patient monitoring and so forth? You taking advantage of any of that?

Yeah. So as you know, the CPT codes are very prescriptive. You know, you have to have 16 days worth of data. It has to be physiologic in nature. So our device is actually programmed specifically to deliver that. So one of the ways we cheat on power is that we monitor every single day in the first year, but then we only monitor 4 months in years 2 to 5, and then we only monitor 2 months in years 6 to 20, because at that point you're just looking for pace, you know, changes from baseline. And so, but the way we monitor still would allow the doctor to bill that annual follow-up because they get 16 days worth of data within the 30-day period in order to do that. So we had the advantage of designing the device to fit the codes, not the other way around. And that, you know, that allowed us to, to keep up.

Okay.

My last question is, would you like a fresh introduction to those guys? I know that, that your Zimmer connection would It makes them super excited. I know they're kind of in, they're in between rounds and that kind of stuff. They could really probably use, uh, what I'm feeling, an injection of enthusiasm.

Absolutely.

Okay, thanks.

Please do.

Thank you.

All right, Kyle, and then Joe.

Thanks Tiger. So Bill, you touched on what I was gonna ask you about a little bit. You know, you show that you gather data and one of the data pieces that you get is say range of motion and you, have a patient monitor that shows they're only at the 10th percentile and that says, okay, we're gonna suggest something as action. But that dataset, it's all part of your post-TKA patients and you've looked at the range there. Well, someone in the 10th percentile post-surgery, that may actually be great for them because they were in the 1st percentile before.

Yeah.

And it's actually a big improvement.

Okay.

So I, I'm curious about your plans for some sort of preoperative thing where as your device gets smaller and smaller, do you put it in a patient for a year before their operation to have a baseline, or do you do something contralaterally? What's the plan? What are you looking into next for that?

Yeah, sure. So we built a wearable version of this, but we, we really don't want to be in the wearable field, but we, we built it for the docs where it's, it's the same IMU, right? So you have at least an apples-to-apples comparison pre and post-op. And the idea would be that you put this on the patient a little bit like a Holter monitor, send them home with it for a day or two so that we could get an idea of their preoperative gait function. So we at least have a way to compare pre- and postoperative function because you're, you're absolutely right. I mean, for some people, you know, that would be a 10x improvement even if it wasn't, you know, relative to their peers. So that's something we, we want to bring in place. For the reimbursement folks, that's also an interesting dataset in terms of pre-qualifying patients for surgery as well. So using kinematics as a pre-qualification metric. So that, that's in the works. And I just, I'm waiting for the docs to get enthusiastic about doing gait assessment preoperatively.

May not happen. Yeah, one, I mean, we, when Bill and I talked 3 years ago, one of the aha moments was this solves the compliance problem with wearables because people don't, they don't, after they take a bath or shower, they don't, you know, or sleeping, they take their thing off. Well, well, when you put the implant inside, that solves the compliance problem for 20 years.

Yeah, that, that's a really good point. A professor of mine used to say that the problem with patient compliance is the patient. And, you know, the reality is, and if you look at the wearable data, it doesn't matter how good the device is, you know, you've lost about 75% of your patients by the time you get to about 3 months. So there's a good uptake in the first, you know, 30 days, but then it tends to taper off. The other thing that I also didn't realize when we last spoke, Tiger, is the benefit of having those sensors cemented into an anatomically consistent position. So with a wearable, somebody might put it on slightly differently every time, even if they're compliant. And so you don't get that real apples-to-apples comparison with time because ours, you know, anatomically is fixed. Even if we're looking— and we have patients out 3 years now— even if we're looking at somebody at 3 years and we see a change, we're likely to believe that that change is real as opposed to, you know, potentially just an artifact.

It's like stereotactic beads that they used to put in bones.

Yeah, yep, yep.

Go ahead, Joe.

Yeah, hey, can you guys hear me all right?

Yes.

First of all, Bill, I think that what you, what you're working on there at Canary is incredible. I think it's super fascinating, and, uh, that's the future. I think the data collection and getting that stuff down pat is going to make a lot of decisions for both physicians and patients. So It's impressive work.

Uh, thank you.

Exciting presentation. And then similar to, uh, what Mark had touched on and maybe a little bit less significant, I was wondering if you had worked or if, if you had heard of any universities who are doing something similar because I know that one of my friends is in a PhD program at the University of Florida and he's studying gait mechanics and fall risk in patients.

Yep.

Yeah. And I wonder if you had tried to partner with some So, you know, do a project with the university and have them collect that information and compare it with your product at all?

No, we're, we are not working with the University of Florida, but we are working with more and more folks. It, it's a very fascinating data set. It's, it's now many thousands of patients. And although there are bigger, bigger registries, the Australians have a big registry, the Canadians have a big registry, but our data is unique in that it's linear. So we have day-by-day, um, on these patients for long periods of time. So there's, you know, the problem with the registries is that you might have 2-week data, 3-month data, and 1-year data, but you don't know what, what happens necessarily in between because it's, it's based on clinical notes or, or EMR. So yeah, we're— our— this is probably also controversial to maybe people in the industry, but we are planning and do run an open platform. We share our data openly, and that's for very pragmatic business reasons. If you're interested in using our data, you're probably buying our device because you, you know, you want that data. So we hook into everybody who wants to use it to try and get this as, you know, a way for them to augment their part. So any research introductions you have, please pass them on. I'll, I'll put my email in the chat.

Yeah, that would be great. I'd be happy to connect you guys and see if there's anything there. I know that for a fact he's looking at fall risk and trying to find, you know, he's looking at the graphs and determining what's going to show a higher fall risk or not.

Yeah, it's funny that somebody mentioned this earlier, but if you look at the hospital readmission rates, uh, actually musculoskeletal causes are really low. So the most common reasons for a TKA patient to be back in hospital are good old-fashioned cardiac decompensation, atelectasis, and pneumonia. So the complications of surgery, and then fall risk and stuff like that. You know, people aren't coming back in because they've got a rip-roaring infection. They're terrible when they happen, but they're really really quite uncommon. And so most of the cost is medical management, just the things you're talking about. And that's the stuff that, you know, you want to be able to tease out from the data. Hey, Colleen's got it. Call risk is number 2. There you go.

Yeah.

Hi, Bill. Hey, thank you for the presentation. I think it was a really good one. Question I have is, how have you managed to overcome the practice adoption resistance that, you know, sometimes exists to any change in a treatment plan?

I haven't. You know, I— this is a gross generalization, but people who are dinosaurs like me, you know, it's not uncommon for them to be like, hey, I've been doing this for 30 years. My patients love me. My failure rates are really, really low. I don't do many manipulations. You know, I'm I'm perfectly happy with what I'm doing. On the other hand, you know, I have a daughter who's just in 4th year medical school right now, and she has never had a textbook. Everything she has was online. Her anatomy lectures, everything was on iPad. So when I talk to residents, or I talk to people who are new in practice, and I say, hey, we're gonna— they stop me. I don't even get through my first slide. They're like, oh, I totally get this. We're going to monitor the patients remotely. It's going to give me feedback. I totally understand. So I think it may be a case of almost having to grow up with the digital generation as we come into this. Now, there's obviously exceptions to every rule, but it's amazing that the generation that is used to having medical information 5 seconds on their phone totally understand this and others, you know, perhaps less so.

So the adoption rates, younger surgeons, probably.

Yeah, and this will be a surprise to no one as well, but a lot of younger surgeons are like, hey, I also, I'm, I'm starting up my new practice and fill in the blank, and I want to differentiate myself from the people I trained with. And so I want to offer robotics and smart surgery and kind of a digital solution for my patients, you know, 21st century approach to, to care. And that's one of the ways that they want to hang up their shingle and have something unique about the way that they're practicing medicine.

And one follow-up question, Bill, do you see a role for wearables in this continuum of care, either pre-, interop, or post-op?

I really do because at some point with enough data, you're going to be able to tease out who the best candidates are for the surgery in the first place. And that's really going to help with your overall, not just cost, but your overall patient satisfaction rates, right? And so, you know, I think, I think there's a real role there. We're a small company, so we're trying to stick to our knitting. And I don't want to be doing, you know, I know I've got a lot of small company folks here. I don't want to be trying to doing everything, but I think, I think there's a real niche there in terms of of understanding who is going to benefit most from the treatment.

Great.

Thank you.

Asher's up.

Hey, Bill.

Great—

sorry. Great talk. I apologize. I missed the first few minutes. So if this question was something you covered in the very beginning, I apologize. No worries. I think you guys recently got a breakthrough device designation for a spine— a smart spine implant. My background is spine. So I'd love to hear what you've got envision for a— you're the sort of the spine equivalent of what you have right now with Zimmer Biomet. And then long term, do you see a way for patients that have, you know, smart knees and smart spine implants for all these devices to kind of, you know, coexist and, you know, co-communicate together?

Yeah, absolutely. So I showed this before we started, but there's an interbody cage and we developed basically the same device that we have for the knee that will slide into the cage and convert that cage into a smart cage. It's actually easy, even easier to do with posterior instrumentation because all you have to do is actually be physically attached to the, to the rod and you can pick up all the data that you need. The way that I look at it is that it's kind of the anti-knee or it's the knee in reverse. How about that? So first off, you can collect all the things like you know, step count, cadence, stride length, distance traveled. So, you know, how, how well is the, the patient functioning? But in the knee, we're looking at range of motion in terms of— we want that range of motion to get progressively better as they recover. In spine, we want that range of motion to go away, right? So we want to go from 3 degrees range of motion down to zero. But there's a really interesting piece of physics, and that is If you think of this cage in the spine, every time the heel hits the ground, it sends a vibration through the entire skeleton. And so you can almost use that device like a tuning fork, right? So if it's loose, it's going to have one vibrating pattern. And then as you get osteointegration, you're going to dampen that until you, you eliminate that movement altogether. And so you can actually see it. You can see a loose implant. You, You can look at the vibration pattern, and then you can see that vibration pattern go away. Tiger and the engineers will recognize that as a Fourier transfer of the, of the data, of the vibration data, but you can actually look at it and tell when you've achieved anatomic fusion. So, I think, in my opinion, this is probably a much better use of the technology in terms of overall. It's, it's just., you know, a complicated area, but I think it's got a lot of potential.

And then as a follow-up question, do you guys envision licensing out your technology, like the underlying smart component, to companies with existing infrastructure in these, you know, other orthopedic sectors as a way to kind of expedite how quickly you can scale, or do you envision kind of coming out as all, you know, your own devices?

Yeah, that's a You know, that's 10 years of trial and error on our part. So we did an exclusive with Zimmer on the knee and no revisionist history. I'm happy we worked with them. I'm not griping, but, you know, you kind of have to when you're a little company, right? We building that data platform was triple-digit millions, right? And that's very hard to raise and you need big brother to help you get through that. And, you know, without Zimmer's funding, we— I wouldn't be able to show you these crazy things. But that, of course, is wonderful for Zimmer, and I hope it's a big successful product for them. But, you know, for us, you know, our opportunity set is Zimmer. In the future, if we're in a, you know, in a new market like spine, for example, and if money was no object, I would prefer to just go to the manufacturers and say, hey, look, if you'd like a smart device, just put a void in your implant and, you know, we'll create the cartridge that goes in. And if you don't want a smart device, fantastic. Um, and, and just kind of do it that way. And there's two reasons for that. One, of course, is that that gives us, you know, a bigger opportunity, particularly in some markets that are fragmented like trauma or others like that. But it, it does something else, which is that The people who ultimately pay the bills for us, which are really more the, the insurers, probably long-term for us than, than say the, the hospitals or, or the ASCs in the device, they would like to know if this product is better in, you know, obese men over the age of 65 and this product is better in, you know, slim women under the age of, of 40. That, that's— they really want to know which devices do well in, in which clinical settings and You know, in spine where there's an infinite number of diagnoses and an infinite number of approaches, you know, it would be useful to be able to look at that data across, you know, a wide spectrum of practices. So for us, we did an exclusive deal to start. It was the right thing for us to do. But in the future, we would rather be, you know, augment existing technology than, you know, be doing 100% on our own or being doing it with one partner exclusively.

Got it. What, what was your total raise to date, Bill?

Um, we as a company have raised over $100 million. Um, we've raised more than that non-dilutively, right? So, you know, building, you know, my company is, is 50% engineers, most who come from active implants. So they come from cardiology or they come from CGM or, or, uh, cochlear implants. And then the other half are all data scientists. So we You know, we have 30, 35 data scientists that crunch.

It's a data company. Yep. Much. Okay, Cope's up, and then, uh, Patrick.

Yeah, Bill and I have talked so many times offline. Every time I hear him talk, I'm like, I just have to sit back and I'm amazed. Um, so Bill, this may be a dumb question, and I also live in Pennsylvania Dutch territory where we have people who still hook their horses to Yep.

Right. Yep.

So I get a little bit of that around here. Yeah. Does anybody ever say, ah, I'm not so sure I want something in this, in my leg telling my doctor where I'm at?

Yep. Right.

Yep.

For sure.

There's, so, you know, speaking of Amish country, about 3% of our patients don't have Wi-Fi.

And that, right. Right.

You know, that's a non-starter for us. If, if, if the house isn't connected, we're not gonna be connected. I would say for about 10% of the population, this is too big brother. I explained to them and I'm like, look, we don't have enough power, you know, to put a tracker in you. You know, we can't power GPS. So I know you took 2,000 steps, but I got no idea where it was.

You're already carrying this around.

Yeah, and this is the greatest tracking device in history.

Exactly.

So, but yeah, for some people they're like, you know, get that thing away from me. But for most people, the idea that they can pick up their phone on a Wednesday and see if they're recovering ahead or behind schedule is incredibly powerful.

Cool. Patrick.

Well, you, you mentioned having a cardiovascular training and background. I'm curious what has— what you've learned, positive or negative, or what surprised you, good or bad, about working in orthopedics now, at least, you know, comparing those two therapeutic areas.

Well, at the risk of ruining my career, I'll answer that brutally honestly. So what I have found, the big difference is in interventional cardiology, she with the best toy wins. If you make the best technology, you take share. I was fortunate when I was a kid, I worked on what became the TAXUS stent, which was kind of the first big drug-eluting stent. And we took 70% market share in 70 days. So we took the failure rate of stenting from, call it 25 to 30%, down to 1 or 2%. And they switched overnight. In orthopedics, it's very much different market dynamics. I think even if I was able to show you right now that we could, you know, I think by week 3 right now, I can already tell you whether a patient's going to do well or not. I think we can already do that. You know, fast forward a year from now and we can show you every complication. That's not going to translate into massive switch in market, right? There's so many interesting— I'll call them interesting— dynamics in orthopedics where there's loyalty to the particular manufacturers. You know, if you've spent a bunch of money on one company's robot, you're not going to be in a big hurry to switch to another company's implant. So it is a very different dynamic. And in fact, Efficacy alone is, is not a good enough part of the story. You, you have to do so much more than that. And I don't, I didn't fully appreciate that. I'm, I'm a geek scientist by training and I thought, all right, if, if we do something really, really cool, you know, the, the old Field of Dreams mentality, if you build it, they will come. That happens in cardiology. It, it just, it's not the same in orthopedics. In orthopedics, it's a, a lot more nuanced is the polite way to say it.

Yeah.

Yeah. Do you think some of that is because they're— certainly in hip and knee replacements, they're measuring outcomes and over such a longer interval. I mean, cardiology studies are, you know, they're like 30-day studies or 60-day studies or 1-year studies. They're not decade-long types of things.

I think that's a really good point. I mean, I've heard it said in orthopedics, don't be first and don't be last. And so I think that's one of the— the mentalities. The other thing too is, you know, the stuff you do in cardiology is life-supporting or life-sustaining. So if, you know, if you've got a better outcome, that matters. In orthopedics, it's not the same. And so you're looking at longer endpoints, you're looking at how somebody's doing at a year or 2 years. And we've seen in orthopedics before, the complications may not show up for 4 or 5 or 6 or 7 years and beyond. So I, in fairness to my orthopedic colleagues, I think it's a different animal. And as a result, the market behaves quite differently. And so you can't just sell with technology. You've got to sell with more than that.

You're selling cement, not milk.

Yeah, pretty much.

Yeah.

I know I invested a lot in surgical navigation early on because I had gone to Europe and saw the outcomes were better.

Yeah.

And I came back to the United States And I learned that orthopedic surgeons would rather do 6 surgeries in 1 day without it than 4 with it.

Yep.

Because, because the outcomes were 20 years out. It's 20 years old.

Yep.

Now, you know, having said that, you know, I, I've got a few practices that believe that they've cut their complication rates by 200, 300%. And probably more importantly, their patient satisfaction rates are high because people like to know what's going on. You know, you know, waiting until your 3-month follow-up with no idea whether you're good, bad, or indifferent is stressful. And so being able to have some sort of a metric is useful. So the flip side of orthopedics is that if a practice converts over, it tends to be stickier technology because the reverse is true in interventional cardiology. Somebody else, as soon as they have a better toy than yours, then you know, you're in big trouble too. So it's, it's got a shorter uptake, but it's got a longer tail.

Psychological aspect of patient recovery and patient success is totally underplayed, probably because it's much tougher to sort of put in an academic paper.

Yeah.

But like when people feel that they're doing good and feel that their doctor has not only their best interests at heart, but has their sort of, you know, their doctor's using the sort of latest and greatest technology, they're motivated, they sort of start on day zero after surgery, like from a better frame of mind, I think that matters a lot more than ever than it like gets discussed certainly on the podium.

Well, and, and it's, it's been proven in any disease state you ever wanna see.

There, there would be placebo, different trials.

There wouldn't be a placebo effect if, if, if that was not true.

Yeah, true.

So, and you know, I think, so I, I, I practiced emergency medicine before I lost my way and ended up in this crazy business and And I would say that 75% of the people I saw were there for reassurance, right? It's, you know, it's the pizza. You're not having a heart attack, right? But everybody wonders what's going on. Is this normal? Is that not normal? And so in my opinion, just the reassurance of looking at, say, oh, I'm at the 70th percentile. Fantastic. Maybe keeps me out of the emergency room. Maybe that's an MRI that, you know, I don't go to my doctor saying, oh, I'm kind of worried. And I end up getting an X-ray and MRI, a bunch of other things. Or I go to the ER and some guy like me puts a needle in the joint, right? And so I think—

Interesting to think about.

Yeah.

So, I think the value of just that reassurance is incredible for the patient, but I think it's actually really valuable for the system too. And if you do it right, you know, I think there's a lot of dollars left on the table in terms of treating 80% of people who, are probably just fine.

If you can turn ER visits into office visits and office visits into phone calls and phone calls into nothing, it's tough to quantify, but there's definitely financial savings there.

And this is going to shock you, but you know who really gets that is the payers. When you go in and you tell them that, and they'll look at that and they'll be like, oh, I totally get it. This is just like CGM, right? Like, my goal of CGM is to predict what's going to happen with my patients so I keep them out of the emergency room, right? I have, you know, when I used to finger stick like you did when I was a kid, you know, all you had was this single data point. But now that I've got, you know, a chain of data that's valuable, now I can use an algorithm to predict what the next 24 hours is going to look like. And then, you know, kick in with my closed-loop pancreas, and now I can administer care before I get into trouble. They totally understand how that works, right? And so I think that's the way things are going. So if anybody's interested in this whole monitoring, it doesn't even matter what field you're in. I think the business model to follow is CGM. I think that's really where it goes.

Yeah.

Great stuff. Vinay's got a comment or a question.

Yep. Thank you. Thank you, Bill. Bill, have you seen any difference or impact on the readmission rates in the patients who received PersonaIQ?

Not, not yet. And that's just because that's a bit of an EMR challenge where we have to go in and hunt people down. So we're, you know, we are in the process of doing that. There's, what's really interesting, kind of one quick segue and then back to your original question. If you look at our original data, you know how we've been saying 80% of patients are, you know, happy with their knee and 20% aren't. If you look at just the global population, 80% of patients have good kinematics and 20% of patients don't. Now, I don't think that those two circles are a perfect overlap, but I bet you they have a pretty decent, uh, shared amount. What we need to do is then match those up to the patient-reported outcomes and see how that works, you know, how predictive are kinematics versus, you know, reported outcomes. And then of course the next step is to look at all the complications, readmissions, etc., and see if we can start to tease out which patients are at risk for all of that. But, you know, I always go back to, without the data, you're never going to get there. And so this was a little bit of a, you know, a search the new world, you know, Christopher Columbus experiment where you've just got to go get the data and try and figure out, you know, whether the Earth's round or not after. And that's what we've been doing. And I think we now have probably follow-up on 5,000, 6,000 people, you know, well out. And so now we can go back and start to do that kind of work.

I have a follow-up question regarding the stroke that you identified. Can you tell a little bit more what kind of gait patterns help you identify and can it be done consistently?

Sure. So what we encourage the practices to do is if if one of our alarms goes off and a patient's like got a really abnormal gait pattern, we don't just assume the device is right. So what we'll typically do though, not we, the practice will send an email or whatever the patient say, hey, will you take a cell phone video of, you know, the person walking? And then they'll look at that cell phone video and compare it. And so that's how it gets picked up, right? So the clinician looks at that and says, Oh, that's a hemiplegia, right? And that's how we do it. So we use this as an early detection system right now. We don't do the diagnosis. My hope is that in the future we'll be able to do the diagnosis when we have enough data. But ours is just to identify something that's way off and then, you know, kind of channel it into the medical system.

Do you think something similar can be done for patients, for cardiovascular patients, especially those with stroke?

Without question. I read a paper the other day that iWatch, if you use it correctly, you can diagnose subclinical heart failure with iWatch, right? Because you have all this patient activity data, right? So you have their step count, cadence, and stride length, and all that type of thing. And with enough continuous data, you can see those changes in the patient's overall function. In fact, in the cardiac monitor that we're building, we put an IMU in it. Because we want to see what their activity levels are, and we want to even see if they're sleeping upright or flat, because these are the types of things that are going to tell us, you know, whether or not they, they might clinically be in failure. So I, I actually think heart rate is going to be— or excuse me, I think, uh, failure is going to be within the next year. I think we'll have enough data to do that, and then I'll at least feel like I didn't waste my life outside of interventional cardiology.

Gonna use that medical degree.

Okay, Elizabeth.

Yeah. Hi. I was just going to reiterate, and I, I was driving. I'm not sure who said it. Um, so many of you know I'm a psychotherapist and I've thought a lot about, uh, stress and, and illness. And I think you're— there's a lot there that you can dig into and research and whatnot as far as people needing comfort. They need information, they need data. I mean, that's gonna And if it's coming from a device that their doctor is overlooking, or it's really, it's like another physician extender.

It is. And what they really like is that they know their doctor's seeing their data. Right. And it works in both ends, by the way. So the people who are doing really well want to show off. They want to show their doctor they're recovering well and doing well. And so for the 25%, 30% who are real overachievers, believe it or not, there's a real therapeutic benefit to that. And it also means you're not doing X-rays and MRIs on people who definitely don't need them.

Right.

On the other hand, you know, if they're not doing well, you know, it's nice to know that, you know, there's a failsafe in there and somebody might pick it up. Because as you know, a lot of older patients in particular are really reluctant to go see their doctor. And they will often minimize you know, the extent of their symptoms. And so having an objective view of that, I think, is potentially quite valuable.

Yeah, congratulations.

Thank you.

Well, you could almost gamify that like a Peloton, and you could have patients competing against each other. I mean, man, this is—

well, you know what's interesting about that, Tiger, is that so when we first spoke years ago, if you had asked me, you know, how we were going to do the user interface, I would have been, all right, we're going to design this and we're going to do that. And we do, we have a Canary interface. But in actuality, I don't think that's going to be us. I think what's going to happen is we're going to plug into everybody else's user interface. Because if you're at HSS, you don't want your patients looking at the Canary app or Zimmer app or anybody else's app for that matter. What you want them to do is look at the HSS app where the patient can see their blood test results and their X-ray results and their recovery curves. And when their next appointment is with Dr. Kushner and, and, you know, that type of thing, they want to have it there. And so we're going to run more of an open architecture where we plug in. And in fact, a lot of our big docs, our big ASCs, are developing their own patient engagement tools. And, you know, I think what we're going to do is just be a data provider as opposed to, you know, the actual user interface, because everybody you know, has their own way of, of managing their patients.

Yeah.

Can I cope?

Bill, is anybody talking to you about using this as a flat-out quality measure?

Yeah, for sure.

Like, just, I mean, like, for insurance or an ambulatory surgery center that, or, I mean, or the orthopedic groups to say we do it better and saying this is a way we can prove it.

Without question. So there, there's, There's two things I can tell you for sure. One is that different docs have different recovery curves. Um, so, and, and by the way, that's not to imply that they're bad. They, they may, they may end up, you know, with identical results, but we're starting to compare those because like, hey, your patients do well at the beginning, really, really well. So what are you doing that I'm not doing that's getting those in? And by the way, your patients seem to do really well on week 3 to week 8. So what are you doing that I'm not doing that, that that's helping there. And what I can tell you is that a lot of TKA happens between week 3 and week 9, um, which is precisely when most patients don't see their doctors. Um, and most of that recovery, that soft tissue recovery, if you hit that window and you, you get it right, um, you end up doing fine. And if you miss it, you know, you, you've maybe missed an opportunity to, to intervene. And so Yes, the people who are doing 600, 700 knees a year and are getting really good results, this is an objective way for them to, you know, point to that. And, you know, for a surgeon who just came into practice, if you do all your cases robotically and you have digital follow-up on all your patients, you're actually going to have a career record. You're going to have your own baseball card of how you did over the course of your career. And actually see the trends and the differences in the things that you do as you make changes to your technique. You're going to actually be able to see those or at least compare them to what you were doing before. So I think it's really important, actually.

Surgeons, just like everybody else, want to get better. Absolutely. Another way to measure your own self-improvement.

Yep. Charlie, Charlie DeCook says that all the time. He says, this has made me a better surgeon because I look at it and if my results get better, then I keep that. If I do something and my results get worse, you know, I question whether or not that's, that's the right thing to do.

It's like a TrackMan in golf where you're watching the ball flight and the RPMs and curve.

Yep.

And, and it's telling you immediately response. It's response to any changes. Yeah, it's fascinating. Uh, this has been great.

Thank you. Thank you.

Yeah, I want to be respectful of time. We're at the top of the hour, but this, this is terrific. We may have to have you back for part 2, Bill. I think if we sleep on this tonight, we'll have even more questions.

Well, thank you for being kind of the first to discover us. You know, I, I think we're probably a, you know, might be another 2 years before I can come and show you the more important stuff, but it is starting to come together.

Yeah, awesome. Great strong work in the last 3 or 4 years. So keep up, keep up the good work. Thanks for your time on BoneChat. And, uh, yeah, we'll talk again soon.

Awesome.

Thanks, guys.

It's a wrap.

Very fascinating.