BoneChat #105 - Bill Hunter (Canary Medical) - Predicting Risk, Less Readmissions & 14k Savings.

This is bone chat number 105. And this is Bill Hunter. He needs no introduction, but I'm going to quickly share how we found each other. Bill and I met four years ago. He read an article I wrote on smart implants that I wrote like 10 years earlier that said, why can't an implant tell the healthcare provider when there's an issue? Bill called me up and said, I'm doing your article at Canary Medical. We instantly hit it off. And he was on the podcast back in 2021. And now this is his second discussion with the Bone Chat members. Handing the mic over to Bill and make this Q and A. He's only got a few slides. Raise your hand anytime.

Yeah, please do.

All yours. All right.

Let'S see if I can get the full screen here. So I do want to call out Tiger. Tiger was the first person to find me here in the wilds of interventional cardiology and take an interest in what we were doing. You'll recognize this from an email he sent out maybe seven days ago, eight days ago, when he was interviewing John Palona. And a couple things I want to point out, because I made it the basis of my presentation, actually, that you have to be clinically significant and you have to have legitimate changes and things like length of stay or hospitalization or things that you can quantify and things that you can can prove. And it has to be more than just kind of buzzwords. And really importantly, even though it's in brackets here, the reimbursement and CPT codes, they have to be real or you really, really struggle. And so I'm going to talk about clinical data, and I'm going to talk about CPT codes, and that's really what I'm going to do today. Everything I'm about to show you has not been approved by the fda, although a lot of it has been submitted. And we have written 14 publications and 14 abstracts in the last three months, most of which are now accepted. And this is a summation of that data. So I know a lot of you know what we do. So I really only doing one slide as a quick review. But on the left is what the raw data looks like coming out of the device. So that's an X, Y and Z gyroscope and an X, Y and Z accelerometer. The simplest way to describe it is that we have learned to read those waves the way that a cardiologist would read a Holter monitor or an ecg in cardiology, we look at things like the P waves and The QT intervals, and they correlate with different parts of the cardiac cycle. We do the same thing. We look at these waves and they correlate with different parts of the gait cycle. Heel strike, swing phase, all that kind of thing. Step count, cadence, stride length, distance traveled, walking speed, range of motion. We learn to pull those out of the curves the same way that cardiology does. And so to go through four years and in two minutes, it takes us a really, really long time to figure out what a normal looks like, right? So until we know what normal range of motion is for a 55 year old female on day 18, we have no way of knowing whether, when we're looking at somebody else, whether they're ahead of schedule or behind schedule. So it literally takes us thousands and thousands of patients to do that. But now we can normalize every single parameter and we can tell normal recovery should look like in a patient that's progressing postoperatively. Tka, then the next step is equally obvious. Once you know what normal is, you start to look for the abnormal patterns. And can I match the abnormal pattern to the physiology that I'm looking for? So my electrophysiology isn't good enough to know what you know, why you get ST elevation. But I know if I see ST elevation, I'm worried the patient's having a heart attack. So that's what we start to look for. We look for these really, really rare events in the curves that match up to particular physiology. And when I went to medical school, they taught us that there were 30 some odd different gates. There's a cardiac gate and a COPD gait and a liver liver ataxia and all these kind of things. And spoiler alert, they actually all exist. And when you really understand how somebody is walking, you can do far more than just count their steps. This is the data asset. Right now it's about 15,000 patients. We have about 15,000 matching controls. We collect a million data points for every patient. We've monitored six millennia of recovery days. We've monitored billions of steps. And as a result, we now kind of have this really large objective database. And it's enough surgeons and enough hospitals now that I think it's fairly representative. And then what we did is we went out and purchased claims data. And so I don't have proms, for example, because that's not in the claims data. But I have the CPT codes, I have the ICD10s, I know whether they are readmitted, I know how money they spent, I know what their Medicines are, what their complications are. I know a whole bunch of things about them. And we also purchased 15,000 matched controls. So without an exaggeration, I think this is the biggest database that's ever been crunched in orthopedics for one particular abnormality. The first question we asked is, do monitored patients touch the healthcare system less? And this isn't as crazy as it sounds. We know that monitored diabetics, so you know, CGM patients have about a 40% reduction in ED visits. We know that monitored cardiac patients have about a 50% reduction in readmission. And so just the simple act of telling a patient how they're progressing and allowing them to make changes on the fly, if you will, by analyzing their own data, is, is in and of itself a therapeutic benefit. And so these are the actual numbers. About 15% of people will have at least one ED visit, typically at about day 40 to day 45. If they have a smart implant, that goes down to about 11%. And if the practice uses the smart implant, and by that I mean if somebody in the practice looks at the data once a week, that's it, just once a week, you can drop that down to about 8%. If you look at the readmission rates, it's basically 9, 6, 4 and a half. So you've got about a 50% reduction in readmission. These are all highly statistically significant. And importantly, we know that readmissions are typically 10,000 to $12,000 in costs. So just the act of not having people back and touching the healthcare system or using the ED like primary care is a pretty significant savings. So because it's claims data, we can dollarize that. And during the episode of care, you're looking at about a 4 to $5,000 reduction in the first 90 days. And then if you look at day 91 to day 365, you pick up another 6 to $8,000 in savings. So somewhere around 12,000 to $14,000 in per patient savings over the cost of year one. To put that in perspective, the Medicare billing for an entire TKA is $12,000. So you're looking at something on the order of the entire TKA market in terms of cost savings with respect to what we're seeing over the first year. You can also look at why the patients come back. And again, these are not maybe what you might think of as clinical diagnoses, but these are CPT codes. So this is stiffness, abnormal gait, effusion, manipulation under anesthesia, clots, pji. These are what actually get billed. And again, you see on almost every case a kind of dose response reduction. The data, the more you use it, the better the result you get from it. And it's pretty consistent across everything. So they're not touching the healthcare system as much and they're not having as many complications. And we can then dollarize that, of course, and we can tell you on a dollar savings what that means for each and every complication across the way. And that's how we end up with the dollar numbers that we have. So that's great. But how do I practice medicine using this? How do I use this in a meaningful way? And so here's what we did. Our data suggests that almost 100% of people show up for the two week visit. After that it's pretty dicey. About 50% of patients are lost to follow up. After that, they never see their orthopedic surgeon again. And so what we did was we said, let's look at everybody who was readmitted, let's look at everybody who had a fall or had a DVT or anything that ever happened. And then let's go backwards in their data pattern to the two week visit. And was there something in their two week visit that was different if you were a patient who had a complication versus a patient who recovered normally? So can I find silent ischemia before you have an mi? Right. Can I look for that ultra rare event in the data pattern that tells me that you're on the poor trajectory versus the right trajectory? And that's what we did. The idea being we wanted our clinicians on the very first day to be able to give an idea of how they were going to manage that patient going forward. So we looked at ED visits and readmissions. This is just in the process of being submitted to the fda. There is a pattern. The ED patients are ones that are not very active, but they also have poor range of motion. The hospital patients are actually generally inactive. When you dig deeper into this, you start to understand why people are coming back. The ED visits are a grab bag. These are patients who are having problem with their knee, but they're also patients who are having problems with medical issues. Atelectasis, pneumonia, cardiac decompensation, diabetes control. And so that's the grab bag that's coming back to the ed. On the readmission side, they're almost all medical patients. They're highly inactive. These are the patients who are being readmitted for pneumonia, cardiac decompensation and the like. They're not really orthopedic patients. And so when I tell orthopedic surgeons that, you know, almost 10% of their patients are readmitted, they say, no way. And I'm like, you're right, they're being readmitted under internal medicine. They're actually not coming back for orthopedic purposes. But you are seeing them in the edge. We can find them. They have a very characteristic pattern very early on. By day 14, they're already declaring themselves. As for the orthopedic patients, we're starting to have an understanding of what they look like. We collect range of motion, we collect cadence and stride length and things like that. Right now we have an algorithm that we're just in the process of submitting as well, that's about 97% accurate. So by day 14 we can tell with 97% accuracy which patients are gonna end up with less than 110 degrees range of motion, which patients are gonna end up with an abnormal gait pattern at 90 days. So already by day 14, these patients are starting to split off from the patients who are gonna go on to have normal kinematics at day 90. And we can build an algorithm for risk of limp as well. And that's starting to be quite encouraging. We've done one 800 patient series with John Dundon in New Jersey, and we looked at 800 patients. And by intervening early, once we have this kind of two week data, and then kind of increasing the PT and increasing the exercise of the patients who are falling behind, they've been able to reduce their MUA rate from around 3% to around 1%. So, you know, a threefold reduction in MUA. This is the ones that are really cool. We looked at the things that kill you. DVT and pji. The DVT data I really, really like because it makes medical sense to me. We've been taught from a very early age that you have to mobilize your patients postoperatively to reduce the risk of dvt. The question is, how much is enough? This is the cumulative directed step count over the first 14 days. And the answer is not very much, actually. You only need to take somewhere between 1600 and 2000 directed steps in the first 14 days to have a significant reduction in DVT rate. And if you look here, this is the Kaplan Meier curve. You can see that patients who are monitored have significantly lower DBT rates. But the interesting data is actually over here on the left. This is DVT by step count. And just like you would expect, all the DVTS are clustered here in the bottom, in the patients who are not moving around very, very much. And by the time you get to, let's say 1,000 steps a day, which would be 14,000 steps, you're kind of at the background rate of DVT. So this is a really good early warning of the patients who are just not active enough and who are two, three, maybe even fourfold increased risk of DVT just because of immobility. And it gives you a completely objective warning sign. This is being reviewed by FDA right now as kind of our first major flag. PJI is the only data I'm going to show you that is not statistically significant. We only have 71 infections right now to analyze, but it is pretty interesting. We're starting to see the pattern. And what we find is about five days prior to diagnosis, you see a really significant decline in patient activity. Their step counts drop dramatically. And then about two to three days before diagnosis, everything falls apart. So they're their step, their range of motion, their walking speed, their cadence, everything falls apart. And so what I think you're looking at is the evolution of a hot red knee. So what happens first is that the patient stops moving around because it hurts. And then they start to baby the knee. They cease to take it through normal range of motion. And our hope is that with enough math here, we might be able to give the surgeons 36, 48 hours notice because that might be just enough to salvage the implant to do a DARE instead of a total revision. Interestingly enough, because we just submitted this data yesterday, we're starting to see patterns even between patients who have a DARE versus patients who need a total revision. Very cool. In how the knee behaves relative to the degree of pathology that it's experiencing. So, last data I'm going to show you and then kind of a quick finish up. So these are what I call the PROMS killers. The patients in chronic pain and the mood disorder patients. These are the ones who historically report the worst outcomes from tja. You can pick them up quite easily. They're not very active, they don't do their pt, they don't move around very much. And to no one's surprise, that means they actually end up with physiologic abnormalities at 90 days. And so again, if you could get on these people earlier and help rehab them, maybe home care, something like that. You know, everybody's being measured by PROMS right now. And looking after the group that's most dissatisfied would probably be a good thing for a practice a couple Other quick little findings. There's a very specific gait associated with falls. They're short, choppy little strides. And when you see this, this is a patient that's got a very unstable gait. The average fall doesn't happen until day 90, but by day 14, you can already see them. They are actually primarily neurological patients looking at their charts. And last but not least, we have a pretty significant reduction in opioid use. And this is really, I think what you're seeing is that a patient who can look at their phone, see that what they're experiencing is normal, skips a dose. Like, if they're like, I don't really know, they look at their phone and everything looks normal. They're like, you know what, I don't, you know, I don't need to take my meds this time. And so we're seeing, you know, around a 20% reduction in opioid use, which is pretty important, as everybody knows. So what I do right now when I talk to people is say, all right, this is interesting and everything, but how do I use this on Monday morning? Like, what do I do with all this stuff if I'm a busy clinician? And I think what we really want is when you walk in on that 14 day visit, we want them to have like a single screen that they can look at. I spent my career in emergency medicine and if I couldn't understand it within two seconds, I wasn't really interested in it. So I wanted to have something really, really quick. So basically, if, you know, it's all green, so we're talking normal activity, normal range of motion. The top 70 percentile, if you will just leave them alone, whatever you're doing is fine. You know, I say this more to the payers, but, you know, your x rays and MRIs are all going to come back normal, by the way. So you may want to think about what you do in terms of, of, you know, additional investigative studies. On the other hand, if you're in the bottom 20, 30 percentile, right, take a look at why, right? If you're not moving around, if you have really poor walking speed and you're not very active, these are your medical patients. So take a look at the chart because these are your readmission risks, your ED risks, your DVT risks. You know, if they have COPD or if they've got diabetes, you know, they probably need to go see their internist because they're, they're just not recovering from the anesthetic very well. They're not recovering from the Surgery. Well, on the other hand, if you got poor range of motion, you know, your cadence is off, your stride length is poor, these are your orthopedic patients. So those are the ones you might want to increase the PT and exercise on, because if you can intervene at three weeks, it's way easier than intervening at three months in terms of rehabbing that joint. And so, you know, we're going to submit all this. We'll. We'll try and get, you know, approved flags so that the docs can read this the same way you'd read a blood test. But right now, I think this is the best way to use it. And then for Tiger's benefit, on the far right hand side here, I'll. I'll just finish this, and I'll go right to the question. Laura, we. These are the Medicare costs per patient by decile. A million patients that were looked at. This was Steve Lyons paper. And you can see that the top 20% spend a couple hundred bucks, and the bottom 20% cost a fortune. So if all you were to do was to identify those 20%, I think you'd not only practice good medicine, but you would save the system an awful lot of money. And so that's really how we kind of tie those two together. Go ahead, Laureline.

Congratulations. Looks like you've got a really great. A really great start. And it's. I think it's everything you imagined, maybe, and more. My question for you is the screens that you're showing us now. And I know Tiger said he downloaded the app, but he doesn't have a knee, so he's not a patient. So he can. What is the. What does the patient see versus what the physician sees? Are these physician screens or are they the same? Does the patient.

No, they're. They're different. So the physician sees the recovery curves right to the decile so that they know where they are. You know, for example, if you're in the 30th percentile, you're still okay, but we don't want people thinking that they need to go to the emergency room because they're in the 30th percentile. Also, incidentally enough, there's a Goldilocks effect to knee recovery. And by that, I mean you don't want to be in the top 10 percentile, either. You do not want to be the most active patient on Earth. You don't want to be inactive, but you actually don't want to. Want to be trying to run Boston on day seven. And so we don't want people Striving to be the 95th, you know, percentile. So we, we tell them if they're normal or not, and we give them their trends. But the doctors, yeah, but the doctors.

You know, you're at the top of the exact.

Yeah, you're doing great. You know, it's more like my peloton tries to encourage me to, you know, run a little bit further next day or whatever. Right. So, you know, interestingly enough, just from a regulatory perspective, the user interface is actually part of our regulatory files. So whenever we change the interface, you know, we have to interact with fda, but the doctors get everything. The patients, you know, what we're trying, what we're doing now is we're building it into a chatbot so that it'll be like, hey, Mrs. Jones, we won't say that you're not walking enough. We'll just say, hey, let's aim for an extra 100 steps today. You know, let's try and be a little more active or, you know, your range of motion. You know, let's work on that today. Let's make sure you do your exercises. Here's the things that we need you to do. So we're going to, we are actually doing now starting to tailor the rehab to what the data is showing the patient. And this is my last slide, tiger, just because I'm a science nerd. But we're starting to, you know, we have 160 Parkinson's patients and they walk different. And we have Alzheimer's patients and CHF patients, and we've diagnosed a couple of patients with stroke. And so all these gait abnormalities, which are very common in the 60, 70, 80 year old patients that are the bulk of our patients, we're starting to get an insight into using gait as an overall indicator of health, which is fun. That's the massive data dump. That's where we're at. We are working with a number of academic institutions because there are literally hundreds of studies left to be done. You know, we haven't looked at robotics versus non. We haven't looked at, you know, different surgical procedures versus others. We've got 160 bilaterals, we got 160 revisions. We haven't looked at any of them. So a lot of data to go through right now.

Yeah. Laurel Ann back.

Yeah, I'm full of questions. One question. When you showed your data at the kind of, at the very beginning, your bar chart data, where you showed the control and you showed the, the Persona IQs and the Persona IQs, was there anything different in how those patients were managed. So you said that they were, somebody was checking on them. But did they receive different information, one group to the other to know that they were in that? Did the clinician communicate with them to say, look, you guys are doing good, you know, stay the course, hang tight, don't do anything. Were they, were they treated differently or just. They were kind of, was there vase behind the scenes?

Yeah, no, no, there's no difference whatsoever. So everything you're looking at is, is retrospective data. So, you know, there's no, you know, there's no influence of pathway of care on that. It's, it's all, all retrospective. We do know when the, when the, the, the practice looks at the data. We do know when the patient looks at the data, we can tell when, when it's being accessed. And what I would say is that if you go through the compliance literature, and I have many times for previous products I've looked at, patient compliance is very, very consistent. Even in life saving drugs, it's very consistent. 25% of people do everything you ask them to do. 25% of people do absolutely nothing you ask them to do. They throw out the prescription on the way out of your office. And 50% of people are somewhere in between. So if I went into people's medicine cabinets, I'd find all kinds of half filled antibiotic prescriptions because they will have taken the medicine until they felt better and they didn't continue the entire course of antibiotics. That's exactly what we see. And so we see patients that never access the data, we see doctors that never access the data. We see patients who look at it every day, we see doctors who look at it every day. And what we have found is that there's a minimum effective dose of data. If you look at it about once a week, that's enough to, you know, to get more utility out of it. And if you don't look at the data at all, you do a little better, probably because your clinician's looking at the data. But you know, the therapeutic invention, intervention, if you will, is actually looking at the data.

Good. Jamie.

Yeah. Hey, thanks a lot, Bill. I've seen you present a few times and love what you do. We have a dynamic smart brace for scoliosis and we've done a couple times on it. So it's interesting for me to see the insights that you've been able to gain are massive. How many of those were you looking at originally? You knew about different gates and that. Were you actually trying to go that broad or Were you just saying, hey, we're going to try and nail two or three things, and then we happened to get the others? Yeah. You know what? We, we started out very. It's a very simple start to our story. When we. We knew that we could, we could collect the data. We. I think we always felt that when the data set got big enough and rich enough, we'd find things we didn't necessarily anticipate. But the way the product was designed was, you know, I'm not an orthopedic surgeon, but. But Fred Kushner was my founding partner, and he said, look, if you want to do remote care, if you want to be able to follow a patient remotely, here's what I need to know. I need to know if they're active or not. I need to have an idea of how they're functioning. So that's where step count and walking speed and all that kind of stuff came in in terms of are they on the couch or are they moving around? And then he said, what I really need to know is what their gait looks like. Right. Because when they, when they come in the office, the first thing I do is watch them walk. So, so tell me how they're walking. You know, are they walking normally, abnormally, are they limping? That kind of thing. And then, last but not least, I need to know what the range of motion is, because the last step of my physical exam is to put them on a table and go like this. Right. And so we designed the device to do that, and then, you know, it was our, our hope, and half of our company has always been data scientists, but it was our hope that as that data, you know, became more robust, that we'd be able to hunt for rarer and rarer events based on. On how people moved around. Right. Are you. You know, we've, we've been beta testing our. Our app and our pressure sensors on the scoliosis brace. I think you're right on, on the numbers there. You know, who's engaging in that. I think you also see if they're doing really well. They don't feel they have to check in if they're feeling all right. And, and that how often are you connecting to the device? And then is there, is that automatic or is that manual with the patient? So the device is on from 7 in the morning till 10 o' clock every single day, and the data transmission is entirely passive. So they have a base station in their bedroom. And every night at 2:00 in the morning, all the data is Relayed and then it's all processed and the next morning the doctor and the patient are looking at their data. So we don't do real time transmission. We'd be out of battery in, in no time. Yeah, that's what, yeah. And, and our power lasts the entire lifespan of the device. So we collect data for 20 years off of, off of the device. But you know, we're, it's, it's, it's not like we're an lvad. We're not life supporting and sustaining. You know, we're just kind of giving these kind of updates. But yeah, the patient doesn't have to do anything. The practice gets the data automatically. The patient gets the data automatically. So they're not connecting through their phone at all. It's all your external device. If they, they can connect through their laptop or their phone to access their data, but they do not have to do anything for the data to be uploaded. Yeah, that's probably a discussion beyond what we want to do on this. But the problem with phones is that you've got all kinds of updates across all kind of different platforms and all the rest. So we tried to make it as simple for the patient as possible. Our patients are not, you know, typically super tech savvy, so. Yep, absolutely. Great, thanks.

Thanks. Jamie, Good question. You're, you're, you're doing the same thing just later.

Yeah.

Here, if there's no questions, I've got a write in question from Patrick o' Clock Donnell, who's CEO of Nano Hive.

Nano Hive, yeah, yeah.

What has been the reaction or support or pushback from the following stakeholders. DPOs, payers, hospital CFO.

So, you know, let's, let's start with the payers. Initially, when we launched, we didn't really have a value proposition. It's like, you know, we're going to collect this data. We think someday it's going to be important. I can't tell you how. And you can imagine how well that went. So what we have been trying to do is to say, all right, look, how can we integrate into your cost structure? And I think one of the biggest learnings along the way was that we were asking the wrong people to pay for the product. So going to the ASC or going to the hospital and saying, hey, would you like to pay us a premium so that we can go save $10,000 for United Health is not the world's smartest business model. And so the other thing you need to realize too is that we actually don't make a better orthopedic implant. So the act of putting sensors into an implant does not make it a better implant. In fact, from a design perspective, you're trying just not to mess it up, right? You're trying to make sure you don't do anything wrong. And so, you know, it really shouldn't be the person buying the implant that's financially responsible for this. And we're really more like a glucose monitoring company. We're really more like cgm, where we're saying, look, you know, we don't make insulin, but we know that if we monitor your diabetics, you know, they don't. You know, they don't go the ed, they don't do these things. We're saving you a bunch of money. And if we do that, you know, can split some of the found money with us. So that has been how we've reached out to both hospitals now and to payers, which is to say, look, you're looking after your TK patients is actually more expensive than creating them, you know, than. Than operating on them. And a lot of that is costs that can be, you know, that can be avoided. And so if we're able to reduce your cost by 10, $12,000, would you be willing to, you know, to provide us compensation, some level of that, as compensation for doing that work and.

So.

Long winded answer. I know, but we only started having those discussions about four months ago. And that's why I was so taken by the quote that Tiger had. And what we do is we go in and we lead with those CPT codes, we lead with the numbers. And the best way that I've had it described to me by a payer is they said, look, Bill, we know how to quantify risk down to the decimal place, right? We know how to quantify financial risk. We have no idea how to quantify clinical risk. And so if you're building a product like this, I would strongly encourage you to, as quickly as possible, understand how you are able to quantify clinical risk. Because once you can quantify clinical risk, then, you know, the quants can take over from there and they can figure out how much it's worth. So it's. That's how we've pivoted, if you will, with respect to our business model. Bill, just to follow up on that, how much evidence did you need before you were comfortable going in for those discussions? I needed statistically significant dollar reductions. You know, so we've spoken to both payers and hospital systems, and, you know, the first thing they're asking us is well, let's do a pilot study and let's see if we can replicate those. And that's fine. I expected that. But that's kind of the next step from where we are now.

Doctor Maloney.

I want to say that I've enjoyed that talk more than I've enjoyed any other talk this year.

Thank you.

The, the sheer amount of information and data. You know, since I was in my 20s, my mantra has always been all information has value. And what you've done is amazing. You also use the word normal repeatedly. That's my praise when I'm talking about another human being. Normal. Great. I'm curious how you're. First of all, how do I get one of these as a wearable? I don't need a total knee, but I need one of these. Is, is that something that you're looking at for people that, you know, can, you know, use one as a kind of an exo measurement device? That's one question.

Yep.

Where is, where is AI Playing into your company now?

Yep.

Reminds me of a conversation I had in 2009 talking to Robin Young and of all places, Germany. We ran into each other and he had just started Pearl Diver, which was a giant.

Yeah, yeah, yeah, yeah, yeah, sure.

I'm sure you know that real well.

I do, I do, yeah.

Yeah, yeah. So I guess the, the most answerable question is, is it wearable in the future?

Yes. It's interesting. There's no animal model for the human knee. So when we were doing this, we actually had to make a wearable to go through the fda. And, and the way that we, we showed our efficacy, for lack of better term, is we went head to head versus a gate lab to show that we could replicate the gait as precisely as the gait lab could. So, yes, we have a wearable. It's funny, we have it as a research tool, and so I give it to a lot of our hospitals to use as a research tool. I never really thought of it as a commercial product. And mostly I am well aware that the whole wearable space is going to be owned by the big IT companies. We're never going to be competitive. So I didn't really want to try and compete with the big boys. Having said that, I, I think we're going to make them. I don't think it'll be a necessarily good business. But all of our clinicians want to know what the pre operative kinematics are like so that they can compare the post operative kinematics. So we are going to. We have a wearable version, but we're going to make it more widely available because I think having patient baselines will be very, very helpful and actually help us understand what we're doing more as well.

I know that you're, you're in, you're in Canada. I'm sure you're familiar with the Montreal company emoji. I was a consult, I was a consultant with them. So, yep, this is just a, a great, a great tool. And I, you know, as a consumer, I'd wear one like a, one week a year to see if I was going to get Parkinson's or something like that.

Yeah, Yep, yep. That's, that's, I think we're gonna, we're gonna make that available. That you, you bring up. Another interesting point is about halfway through this year I realized that we actually weren't monitoring knees. We're monitoring the way people walk. And so it doesn't really matter whether the sensor's in the ankle or it's in an ACL or it's in a hip. The same basic things apply. And even when we have a wearable that we put on spines for the same reason, to kind of look at that. And once you understand the patterns, if you will, again back to my ECG analogy. Once you understand the patterns, it doesn't matter where you detect them from. You're able to analyze them in fairly similar manners. And lastly, your question about AI. It's one of the most interesting things I've ever worked on is what we're doing in PJI right now. So not to nerd out too much, but three times a day we collect this ultra, ultra high resolution data. It's, it's 800 observations per second, so it's 40 times visual. And for the engineers out there, we're detecting one micron of movement. An X ray can only detect one millimeter of movement. And so we're looking at really, really tiny. We built it that way so we could tell if the device was osteointegrated or whether it was loosening. That's the purpose of it. But this high resolution data, it's a pig to analyze. It costs a lot of money, but it's giving us extra resolution on these PJI patients and it's giving us 40 times the data to go into to try and look for these really, really subtle differences. Super cool. And so, you know, a huge part of our effort is AI based right now.

I wish JR Toykin were on the call.

He would love.

Cool, good questions. Okay, Laura Lynn, then Kyle, then Andrea.

Okay, my Last question. Probably have you thought about potentially partnering? I know there are several companies in the space, remote patient monitoring companies. As opposed to doing this all on your own and just having, you know, the data, is there a potential for you guys to get maybe bigger, bigger access with, you know, with that kind of partnership and just something from a business model perspective, is there a monetation road?

So that's actually from a business perspective, very, very important question. So when I first met you and we first started talking about this, the only CPT codes that existed for, were for remote patient monitoring, right? You know, there's $50 for data collection and $50 for data analysis. You know, let's call it a thousand dollars a year. And so at least on the surface, if you're doing 500 knees a year, $1,000 per patient, you know, that's, that's worth doing. I, my attitude towards RPM has kind of changed a lot. Having been involved in it now in the last little while. I, I think the value of RPM for a company like ours is as a enticement for physicians to use our product. So if they have the ability to, you know, build a remote patients or you know, I can talk to a practice and say, look, you're, you're losing 50% of your patients to follow up. So if all you was a remote follow the patients who didn't come back, you'd still double your follow up income and blah, blah. So honestly, you know, it's more a fishing lure than it is an actual tool. I think it's a reason that docs might want to use the product. I don't think it's a particularly good business. And so, you know, we make a little bit of money on it, but trust me, it's a little bit of money where I think the value in this lies is twofold. One is that if you get this right, you know, you actually aren't doing diagnostics, you're doing prognostics. Right. You know, being able to tell what's going to happen is actually more medical value than telling people what's already happened. And so being able to run these diagnostic panels I think is, you know, where the reimbursement lies and then the ability to do patient management is where the risk sharing contracts lie. So if, if I can go to somebody, to a payer and say, look, you know, this is one of your biggest expenses, right? Remember tka, tja, the most common elective procedure in the United States, it's even more common than C section. So it's a really big cost. You know, the privates are paying somewhere between 20 and $30,000 in, in patient care postoperatively. You know, if I can take that to $15,000, you know, they're probably going to be interested in what we have to do, and that allows you to do risk sharing. So I, I believe RPM is fine. I'd be happy to work with a company to, to manage some of that. I don't. I don't look at it as a great thing. You probably see UnitedHealth has even stopped paying it out right now. They just announced that a couple weeks ago. But I think where you have to get paid in this type of business is, you know, in, in diagnostics slash prognostics and, and in patient care and patient management.

Just seems like it could go really well hand in hand, almost. Right, Because.

Oh, for sure, for sure.

You know, the algorithm to prevent a patient from having these, you know, complications, and if you can predict before they're going to have a PJI or before they're going to know that they have a pj, that's money. That is absolutely money. And so, again, I think you. I'm astounded, and I love to see how far you've come. It's amazing. It's incredible. Hats off to all.

Well, back to Tiger's last guest. Right. The problem is there isn't really a payment pathway for some of that stuff. Right. So what I really need this to be is like, instead of ordering an mri, you order a panel. Now, that would work really, really well. Right? So if I said, look, if you routinely are doing an MRI on six months and your hit rate is less than 1%, but we can identify which patients are falling behind. So Instead of paying $1,400, which is a reimbursement for an MRI, pay us $200 every time the doc runs a panel, that would be fantastic. And if our docs got paid $50 for reading that panel, that would be even more fantastic. And so that's the kind of way that we're starting to go through it and think of this more as a diagnostic prognostic tool and less as a, you know, remote patient monitoring. Although I think that's a huge part of the clinical story, I'm just not sure it's a big part of the financial story.

Thank you.

Great question, Kyle.

Yeah, thanks, Bill. So a couple questions for you. First is from. Not just from the data, but your overall experience in getting into your partnership and how this is integrated into the knee system, how has that experience so far colored in your potential next generation projects or where you go, where else you go in the human body. And I got a second question for you.

Sure, a lot. I'm going to give the preamble that this answer is not meant to be disparaging of Zimmer in any way, shape or form. It's an answer to your question of more or less of what it's like as a small company to work with a very large company. And I did that in a previous life when we were 70% of the revenues of Boston Scientific at one point in my life. So I've been through this drill a couple of times. And a big device company has very different motivations. Their goal is to sell more implants to more people at higher prices. That's really what they want to do. And the technology that we're developing certainly can help them sell more implants to more people, but it's probably not going to help them to do that at higher prices. And in fact, there's an argument to be made that under certain circumstances we may even cost you gross margin dollars by making the device more expensive. So that preamble has changed my view on what to do with next generation products. So I'll try to do this as succinctly as I possibly can. Let's say you've got a thousand dollar device and it costs a hundred dollars to make it and that's a 90% gross margin product. And you know, Tiger's on the board and everybody's happy with that. Then you go take one of these things, one of these sensor packages and you put it in and it costs $600. And so now that thousand dollars device has a cogs of $700 and your gross margin just went to 30%. And the CFO is sending me hate mail. Right, but, and I say don't worry about it, we'll make it up on volume because you'll take market share. And number two, we're going to get all this data revenue and it'll all iron out in the wash. And you can imagine how well that conversation goes. So for a big company with a existing product, that's difficult. But if I had a startup and I came forth with this new device and we had 0% market share and we only had 30% gross margins, but every percent market share we had was growth and it was growth with actual value. Our share price just goes up and up and up and up at infinitum and that's fantastic. And then if ultimately we make money at 90 plus percent gross margins on the data, then that's a fantastic story. Again, long winded way of saying that whenever possible, kind of going it alone or working with a similar size partner is probably the way to go. Going into an existing franchise is very, very difficult.

Right answer. Second question for you is I was talking to a startup about a year ago that was working on a sensor device for a suture anchor for road chair cuff repair. Idea being similar, like you can see the tension of the sutures over time. That can give an indication of how well the repair is holding up. What struck me as worrisome when I saw that technology was it may look like something is loosening and that may look alarming to a patient, but it may be perfectly normal within the course of things. It sounds like the risk of a false negative you manage with the user interface for your patients, is that right?

Yeah, you know, the, that's, that's a very good point. And that's probably the reason that you're not seeing me present data until we had 15,000 patients, because you've got to have a really large number and a really high degree of confidence of what normal is. Now, even then you're going to have outliers. I mean, human beings are human beings. If you look at the DVT slide I put out there, I've got somebody who basically, you know, was walking 10,000 steps a day who still had a DBT. So, you know, medicine's imperfect, but the reality is you have to have a really good grasp on what normal is. And the best thing you can do is just kind of give people an idea of what it is when, when you read a blood test, you know, you can have a normal, abnormal blood test, but that doesn't mean there's any pathology there. But you, you need to have an idea of what to look at. And I think that's what this data is meant to do. It's just meant to give you an idea of which people are outside of, you know, the two standards of deviations that we typically use in medicine.

Thank you.

Great question.

All right, Andrea. And then Patrick, we, we got to go fast.

Yeah.

So you answered some of the. So my first question was what, what other, you know, anatomical areas you're going to cover? And you mentioned some of them. But looking at your patent portfolio, there are definitely things that you, you haven't mentioned yet. So I'm interested in seeing how they, how you see that going. There's a lot of hemodialysis catheters and stents and heart fails and breast implants and sports equipment. It seems like it's a wide variety of applications.

Yeah, I mean, when we started the business, our belief was that, you know, over the next five to 10 years, every medical device is going to be able to provide some sort of feedback. And the nature of that feedback would really be dependent on the device. Orthopedics is all about movement. So you see a heavy concentration on accelerometers, gyroscopes, that kind of thing. Cardiovascular is all about pressure and flow. So you see things like ultrasound and pressure and flow meters. Our cardiac monitor is actually starting its pivotal trial right now with Scripps while I'm down here. So that is the next piece. And then oncology is all about imaging. So that's where you see the ultrasound and other stuff like that. And our goal is to basically populate these things in. These aren't going to be the cures for anything. But how wonderful to have some sort of objective data coming from the operative site. That's the vision of the business. And this is our first embodiment. But it is, but it is my hope that we will spread out. And again, we don't really want to be a device company. We don't really want to be a consumer electronics company. We really just want to connect those devices, you know, to the user interface is really what we'd like to do.

And so does do these monitors integrate with existing implants? Do you have to get the approval of the companies that make these implants or is it freestanding and the dog just decides, you know, is the one who, who makes the decision of using them or not?

Yeah, it, you know, the truth be told, the, the right way to do it is to work with the company that's, that's doing that. We did that as an exclusive the first time around for a whole bunch of reasons. Mostly because, you know, it was a very expensive to build that data infrastructure. I mean, we manage 15000 patients a night. We're actually a decent sized pacemaker company. But, but I also envision making almost like a cartridge where if a device company wants to have a smart version, they just put a void in their implant and the cartridge fits in and that gives them that opportunity. And then last but not least, we don't actually always have to be attached to the device. If we're just in the vicinity, we can probably do the monitoring that we need to do as a standalone monitoring, which is how we're doing it in cardiac, for example. So, you know, and that's better for us because then we don't have to run a partnership model at all. So it runs across the whole gamut, depending on what the clinical indications.

Thank you.

Great question.

Yeah. Pipeline. Patrick.

Phil, a lot of people may think of this as sort of like a high class problem that they'd be happy to have, but do you have, or have you had fundraising challenges because you have sort of a well endowed commercial partner at the same time, maybe you're able to bracket things off by products and that's the way you, you can sort of address it. But I could easily see an investor saying sort of like, why are you coming to me for money? Like just, just call them. You're talking to them every day. Can you speak to, to that? Maybe it's not a problem at all, but I could see that at least theoretically being a, you know, kind of an issue.

No, it's a huge problem. Yeah, very astute. It was a huge problem. We, we were quite successful raising early money because it was a pretty novel concept. And in a strange way, when we reduced it to practice, it became harder and harder to raise money. One, because the business model was pretty nebulous for one. And for two, a venture investor wasn't interested in investing in Zimmer Biomass. Right. And, yeah, and, and at this point in time, we were basically just a Zimmer biomet tracking stock. Right. We were, you know, as a mentor of mine used to say, you know, you can't be a two product company until you're a one product company. And, and we only, we only have one product. We only have one product. And, and so I had a very difficult time convincing the world that we weren't a knee company. Yeah. And not a lot of people want to invest in knee companies, quite frankly. So it was very tough actually. And it's only been as the data came out and people say, oh, okay. But also remember, we benefited from two things that we had nothing to do with. One is that in the 10 years I've been doing this, robotics has gone from a curiosity to mainstream. Right. So the ability for us to provide objective data back to the robot is actually pretty fascinating stuff. Right. And that'll be the next wave of stuff you see coming out of us. And then the second part is, you know, large language models didn't exist when we, when we started with this, but now you can put all this data into large language models and do crazy stuff with it. So now people say, okay, I can see how this might work in the long term. But the last four years have been very difficult to raise money for us. Okay, thanks.

Great question.

So my research, you can say yes or no. You've raised about 45 million total in three rounds, something like that. I don't know.

Yeah, yeah, a little bit, A little bit more. We've raised a lot more money non dilutively than then we've raised diluted.

So Patrick's question, Zimmer may just be the proof of concept for your second and third and fourth products. And that's kind of interesting.

I, I hope so. I mean the, in many ways and you know, I, I laugh because I, you know, I, I still am a newbie to, to orthopedics. It, if, if I had a cardiac problem, a product and I said I'll take aneurysm example. And I said I've got an implant that can go aneurysm and it can tell you if there's pressure flow in the aneurysm, which means I can basically tell you whether it's going to burst or not. That's it. It's one slide. I go the cardiologist, I'm like, here's the pressure flow sensor for the aneurysm. They're like, totally get it. Does it work? How much does it cost? We're done, right? When I try and tell the story about why you want to monitor a knee patient, it takes me 20 slides, right? Because we got to take the motion data, we got to do this, we got to do that. So in many ways I think we, we, we, we tried the hardest thing first, but that was because quite honestly it was the only thing we could make. It was that, you know, the big, the smallest thing we could make was this big when we started. I mean now we're making stuff like this. So you know, we can go into a lot of other things, but there are a lot easier clinical cases to describe than orthopedics.

And there's a lot, there's subs like specialties of medicine where like they are much more used to taking like getting like in, in different, you know, sort of cardiology or vascular type of fields. The markets will move like 10, 20 market share will move on the publication of a new paper and then it will move two years later on the publication of a new paper. I mean orthopedic market shares are measured in like basis points because they move so, so slowly. So, and that's a, that's a result of like, you know, sort of, there's a million reasons bad, bad data over time. Or very, you know, murky sort of data that's inconclusive. And, and, and so doctors are just very not.

They're just, just.

It takes so much to move orthopedic market share, especially for something like total knees.

It's true. When we launched drug coated sense, we took 70% of the world market in 70 days and we were the number four stent player when we started. So about as big a market move as you could possibly imagine. I'm biased, but I think this is a pretty important next step in orthopedics. I think we have something that's really highly differentiated, but it is still a very slow movement because surgeons are attached to their robots, they're attached to their implants, they're attached to their ways of doing things. And history has taught them that most new innovations don't result in dramatic changes in outcomes in orthopedics. So they don't shift very, very quickly. Interventional cardiologists will shift very, very quickly. Because in my career we've gone from open heart surgery to stents to putting in heart valves over a catheter. I'm like, are you kidding me? If you told me 20 years ago you'd rep heart valve over a catheter, I wouldn't have thought that was even remotely possible. So they're used to all kinds of innovation and they adopt it much more quickly. Yes.

Orthopedic businesses are addicted to the 510k process too. So like by definition, something new that comes out is only, can only be really so different from something that's already on the market. And even if the doctors don't express it in those terms, they can sense that and things move much, much slower.

And Tiger wrote about this recently and, and the data is all equivalency data. Nobody's coming out with data showing superiority. Nobody's going out doing big trials showing a difference between one thing or another.

Right? Yep.

Jamie, bring us home.

Yeah.

Are you having any consumer not. You're probably not doing direct to consumer, but do you have consumers sharing your story at all and getting any traction in ARP or other things like, yeah, I got this smart knee and any stories on that? Yeah, I need Arnold to come out and do a knee piece or something. I think this will be a great consumer story. If you look at Mako and how they really succeeded, they went out and they said, hey, be the first to have a robotic knee replacement. And, and the hospitals and the surgeons use that as a way to bring people into their practice. They marketed that as an innovation that people wanted to be a part of. I firmly believe that if you gave patients the option of having a smart implant versus a regular implant that a significant number of them would choose to do that. So I think there's a DTC story here, but you know, DTC and medical devices is always dicey and it's expensive and, and, and, and, and, and but, but yes, I think you could do very well with DTC on this product. Okay.

Yeah. I think the like to Mark's comment. I think the wearable, it would be your DTC play because it didn't matter what implant you got. Maybe patients or payers would pay more for that. I don't know.

Yeah, don't know. The, the major issue behind a wearable is unfortunately is that people don't wear it. But yeah, the patient compliance becomes a real problem and it's very difficult to get them to put it in the exact same place every day so that your data is comparable. It's good, but it's, you know, it, you just don't get the same level of fidelity, unfortunately. But point taken. I think it's worth doing. Yeah.

Mark.

Yeah. Malcolm Gladwell wrote that the only measurable prognosticator for a young person in athletics was speed. If I'm where, if, if, if I'm in the big time athletic world, what could a word.