Valencell was scheduled to present the results of a clinical study on its groundbreaking blood pressure monitoring technology at the American College of Cardiology conference in March, but unfortunately that conference was cancelled. So we decided to share that presentation and research here in a webinar format with an open Q&A session. Below you’ll find a recording of the webinar, the slides, and the transcript from the webinar.

You can find more information on Valencell’s blood pressure technology here: https://valencell.com/bloodpressure/

If you have questions or would like further information about this technology, please reach out at [email protected].

You can view the recording of the webinar here:

You can also access the slides from the webinar here:

 

Webinar transcript:

Calibration-free blood pressure monitoring using biometric earbuds

Ryan Kraudel: [00:00:01] Ok, thanks again, everyone, this is Ryan Kraudel at Valencell, and it looks like we’ve still got quite a few people joining, but in the interest of time, we’ve got a lot of information to cover, so I want to go ahead and get started before we dive right in a few housekeeping items. The if you have questions or any topics you want us to dive deeper on, please do submit those through the through the go to webinar control panel. We like to keep these as interactive as possible. So we will answer those questions as they come up throughout the session here. And then we’ve also got some time towards the end to to answer any questions we may not have been able to cover through throughout the course of the presentation. So that’s the first thing. The second thing is one of the first questions we always get is will the the slides in this material be made available after the webinar? And the answer to those questions are, yes, we will make both. We are recording this webinar and we will make the recording and the slides available to anyone and everyone who is registered for this webinar following the webinar. And we encourage you to share that with any of your colleagues or anyone else you think might be interested in in this topic who may not have been able to to make it here today. So with that, let’s go ahead and get started. We are very pleased to have with us here this morning Dr. Steven LeBoeuf. Dr. LeBoeuf is not only one of the co-founders and the president here at Valencell, but he is also the executive sponsor of Valencell’s blood pressure technology endeavors, which you will hear a lot more about today. So with that, I will go ahead and hand over to Dr. LeBoeuf and let him take over from here.

 

Dr. Steven LeBoeuf: [00:01:54] All right. Thank you much for that intro you all the all time. Much appreciated. I am very eager to talk with you all about our collaboration, free blood pressure monitoring technology, using biometric earbuds. And as Ryan discussed, we really were very excited when we got our abstract accepted for the World Congress of Cardiology. But, of course, I mean, covid-19 put a damper on that. It prevented us from presenting there and the event was canceled. So we decided to present this information to you virtually and try and replicate that environment here through our virtual connection. So I’m going to start by speaking about the outline, what I’m going to go through today. So the agenda includes the background of our blood pressure development technology, how we went about developing it, we also go through why the ear as opposed to other form factors. Then we’ll go through the clinical study, the overview and results. Will then go have a discussion about continuous BP monitoring with this technology. We will then talk about use cases, the variety of different use cases for this technology. We’ll talk about two very broad use cases there and we’re looking for partners to help us get these use cases into the marketplace. And how do we where we’re going from here? This presentation is very much about what we have, at least what we had as of the submission of the abstract back in twenty nineteen. We’ve gone beyond that already.

 

Dr. Steven LeBoeuf: [00:03:49] And where are we going from here with the technology? And then lastly, a lot of folks want to know how to evaluate the technology and we are very well equipped to do that for you and we’ll explain how you can have access to that technology. So first, let’s talk about the background. The it’s well known not just in the United States, but worldwide, that hypertension is a significant public health issue. And the CDC, for example, has estimated that in the United States, hypertension in twenty fourteen was the primary or contributing cause to death and over four hundred thousand folks. And that’s only risen over the years. And a third of American adults are estimated to have hypertension, but only half of them know that they have the disease. Now, the World Health Organization calls hypertension the silent killer because you don’t know you have it when you have it until it’s too late, which is the worst thing about it. But it’s estimated to cause over seven million deaths per year worldwide. So that’s roughly almost 13 percent of all deaths. What now? It’s known that combining prevention with regular monitoring can improve health and reduce the risk of heart attack or stroke. However, few households own a BP monitor, and even fewer will comply with regular monitoring due to the difficulties of properly using and donning a BP cuff. It is just using a BP cough or regular person. Using it the right way is is not the easiest thing to do.

 

Dr. Steven LeBoeuf: [00:05:25] I’ve seen many a time where doctors and even nurses have not even properly, don, to be careful when correctly doing measurements. And so it’s you can imagine it’s more of a challenge for those who are consumers in the end. And now you might see some references there in the bottom of the slide. We those will be available for you. And you could take a deep dive into the rationale behind these references. What I like to do now is this the motivation for why Vaillant’s decided to to investigate this area, to make a colorless technology for being able to without a calibration monitor blood pressure. What I like to do is give you an overview of our historical timeline on this. So we started working on this blood pressure technology all the way back in 2009. And it was very low level. It was it was an investigatory aspect here because we developed this entire monitoring technology which involved exculpatory analysis, optical analysis, inertial analysis, electrodes, all pretty much anything you can imagine you could shove into an ear. We evaluated this and one of the investigations was could we could we use p.g as a methodology for estimating blood pressure? And so we kicked that off and we completed a summary of that sometime around 2011. And we said, you know, this really is feasible. It’s something that can be done. We realized it would require machine learning in order to get there.

 

Dr. Steven LeBoeuf: [00:06:53] In twenty fourteen we were able to raise some funds to continue this research because it was really just a small effort in all the other things we had to do. And frankly, no one gave a damn about it at the time. You might remember in 2009, there weren’t any really any wearables around at the time and the wearable market didn’t start to maybe twenty eight at least as we know it today. But 2009, it was hardly anything in the marketplace. So people just having a hard time even understanding how do you integrate heart rate into wearable device, much less blood pressure. So we follow a foundational patent on this in twenty fourteen. And since then several patents have been granted based on that foundational filing. In twenty sixteen, we we got some some funding and we were able to collect data from a smaller group of folks. But it is a larger effort and we collected group data from about about one hundred, just under one hundred folks at that time. We started developing our model and we realized that we needed a whole lot more data and data is expensive. There are no databases that you can go to that comprise raw PBG data from a wearable device and combine that with blood pressure cuff data. There is a database that combines PPG data from a pulse oximeter with with data from an arterial line.

 

Dr. Steven LeBoeuf: [00:08:14] But that data is notoriously noisy and just not very usable for this application at hand. And so we had to develop that database. And in twenty eighteen we raise some additional funds and also collected a lot more data. And by twenty nineteen we collected over fifteen thousand data sets on many different parts of the body and we demonstrated that we could track blood pressure like a cough. And so now in twenty twenty we’re focused on launching with partners a journal wellness device and pursuing medical device partnerships. You see Valencell as many you’ll know, we don’t make blood pressure monitors. We don’t make wearable devices. We make technology that goes into them. And so we we license the companies. We provide solutions to companies. Since the modules either through ourselves or through distribution partners, we don’t have a never plan to make our own end product in products ourselves as far as in products that you would use Valencell. So we provide software solutions, we provide hardware solutions, but they go into that. So we rely on partners who want to put our technology into the products. So now that I’ve given you that that overview of how we got here at Valencell, so I’d like to discuss why we picked, you know, did I mention that we collected 15000 data sets about I want to say by the time we put this together, we had five thousand year data sets, three thousand of which were usable for this particular endeavor in the ear is is good because it’s one of the best locations, the best locations on the body for accurately measuring blood pressure.

 

Dr. Steven LeBoeuf: [00:09:48] And I’ll explain why we believe that is. The first is the ear has a very optimal vascular structure that the the ear comprises this high density of arterials which connect between the arteries and the capillaries, and they kind of blow up like balloons when they do this. And so you can imagine that amplifying that pressure signal that gets there. And it’s a high density in the ear of these with respect to other vascular structures, such as capillaries, finials and veins. And so you really get a nice amplification of that blood pressure way for. And also, there’s a fixed distance from the heart and if any, you all ever use blood pressure monitors before or develop them, you know, it’s really important to have that that that monitor on the arm near the brachial artery at the right distance with respect to the heart. And it’s a common problem, the mistake people make. And you can really screw up the measurement big time if you don’t keep that that distance stable on the ear, you don’t have that to worry about unless you really lean over too much. There’s really no difference there that you’ll see. I mean, in order to shrink that difference between the ear in the heart, you would have to basically destroy your spine, which would lead in death.

 

Dr. Steven LeBoeuf: [00:10:58] And we don’t recommend that. That’s definitely not the right way to do it. So it is just very stable in that that position there between the ear and the heart. And that that definitely helps. The next thing is blood perfusion. It’s well known that perfusion in the face and the ear is much higher than other parts of the body. I don’t know if you’ve ever accidentally cut your ear before. I’ve had many, you know, being from South Louisiana, many a zealous barber cut my hair and my ear. And that just feels like it’s never going to stop bleeding. In contrast, if you scratch yourself on the wrist, it clots up right away. There’s just a lot more blood perfusion at ear with than there is in other parts of the body. And that definitely helps. And lastly, another reason why the ear is so good is I don’t know if any of the part of the body where you can measure so much stuff at one spot. I mean, it’s the ear are the ear and the rear is probably not nearly as comfortable as the year, because in the ear you can you can measure heart rate. You can measure we’ve shown that you can measure blood pressure or body temperature, breathing rate, body sounds, exculpatory sounds. All these can be monitored at the ear. And you just can’t measure all these things at the wrist or other locations on the body.

 

Dr. Steven LeBoeuf: [00:12:08] So all these things together are a one of the reasons you can’t make me explain why the ear was so interesting to us. And indeed, when we looked at all the data that your data was better than any of the other data we collected, even though even though finger waves look pretty awesome, I got to say the ear has more pressing information in it. In these are the technical reasons why the IR is so awesome for this, but there’s also some market reasons why and as some of you may know, there have been several reports on this. IDC perhaps has been spearheading this is that the wearables market today is dominated by hearable. A lot of folks don’t know this. You think of wearables, you think of one of the devices. But the reality is already hearables top that of where wearables and the growth in here is higher than the rest of the wearables category. So it’s the biggest segment in the fastest individual growing segment, at least when you compare it to risk when devices. And the bottom line is when you’re hitting a moving target and what people are wearing these ear worn devices, these true wireless earbuds and also hearing aids, then you get to see how this really can have just a tremendous public health impact and a tremendous market impact to. Sort of like to do now is go through

 

Ryan Kraudel: [00:13:24] A few questions that have come in, if I can interrupt real quick on or so on the previous slide, you talk about the ear being a fixed distance from the heart. There’s a question here about why that’s important. Is it relevant to the movement or is it excuse me? Is it related to the movement component?

 

Dr. Steven LeBoeuf: [00:13:42] Well, that’s a great question. The reason why it’s important is because the if you if, for example, if you had a blood pressure cuff on your arm and you raised your arm up high, you would actually have less pressure going to that arm in in the thing is, what you’re trying to do with the brachial measurement is are you really trying to estimate the pressure of the heart? And when you raise your arm there, you really distance yourself from it. So if you raise your arm, you might get a 20 millimeter, a difference from what you would be if your arm were at heart level. So by having that fixed distance, you prevent any of the anomalies that could be associated with trying to model the heart’s pressure against the ear. So or because, you know, if if you could change that distance between the ear and the heart accidentally, somehow, then when you’re trying to develop these machine learning models, you have to factor for that. And that’s a hard thing to factor for.

 

Ryan Kraudel: [00:14:39] Cool. So few other questions. It looks like we will get to most of this in later slides, one of which is please comment on compliance to ISO 81060 Dash two, which I know you’ll cover later. So we’ll put that one to the side for a moment. Another question is, have you tried pectoral test areas to see how accurate your measurement is there?

 

Dr. Steven LeBoeuf: [00:15:04] So on the pectoral chest area, we haven’t done blood pressure measurements. We’re not sure we’ve done a few, but not enough to be meaningful. We’ve done a number of measurements there so that we can demonstrate the ability to measure motion tolerant heart rate and in breathing rate. And then also heart rate variability are interval, for example, but not blood pressure. By far, most of our data sets that we have, we have an equal number of data sets on the ear and on the finger and on the ear. We have data sets that comprise in the ear canal devices that comprise the conduit. And then also we have a number of datasets on the wrist as well, but not nearly as many as we have on the finger in the ear. But we have very few in other parts of the body.

 

Ryan Kraudel: [00:15:51] Ok, and then it looks like there’s another question about will the webinar be recorded for those of you who missed it up front? Yes, this is being recorded. And we’ll distribute the recording and a link to the slides out later after the webinar. So thanks. Why are

 

Dr. Steven LeBoeuf: [00:16:06] You going to transcribe this presentation from cajun English into American, British

 

Ryan Kraudel: [00:16:12] English? We will do our best for that. Yes.

 

Dr. Steven LeBoeuf: [00:16:16] All right, I know that some other questions and we will get to those in a bit with this presentation. So what I like to do is move over to the meat of this presentation, which focuses first with the abstract that we were to present at the the American College of College of Cardiology event in April. Actually, right now. So we have everything organized here in this one slide showcase and everything a snapshot. And this is this is somewhat of an eyesore for you to go through. So what we’ve done is we’ve broken this up into slides that are more visually appealing. So everything you see here, we’re going to talk about in a way where it won’t strain your eyes in. The key takeaway of this presentation is that we’ve shown that a biometric audio earbud can actually monitor blood pressure during free living conditions without the need for a cough in the next few slides I go through really are going to explain that. And since we’ve already gone through the background already, what I like to do is start with the methods. So our approach is comprised a machine learning algorithm applied to 15000 data sets from five thousand subjects. So we collected all this data and processed it and a lot of the data was unusable. And so we processed all this data, massive data collection, both in the United States and abroad, in Asia, Vietnam and the Philippines, to be exact. We then built our model around all of this training data we had developed.

 

Dr. Steven LeBoeuf: [00:17:46] And in folks, we had multiple measurements from multiple folks, so, for example, in that five thousand subjects, some people had been tested three or four or five or even in some cases six times. And we developed this machine learning model to connect the dots between the inputs and the blood pressure in the inputs to the model. And I’ll show you more about this in a bit. The inputs to this model included raw PPG data, inertial data from our PPG sensor, and then additionally, certain metadata, for example, especially the person’s age and the person’s weight and gender, were really important also to help getting to meet the ISO requirements. So then we did this. We built this data set up. But you can’t use that same data set to prove to yourself that your model works. There’s some things you can do as as data scientists know. For example, leave one out, which is a nonbiased way to do an assessment. But you never, ever really know how well your model really works unless you test it on a true a true unadulterated data set that is never, ever, ever, ever, ever, ever in any way been used to build a model. So we had to do this. So we adopted the ISO eighty one sixty dash to twenty eighteen protocol for for a metric just to gauge how well to still a metric would perform against a manual cuff. And in this case we, we hired some nurses to come to our facility, they collected data and recruited folks from all over to be participants in this new data set.

 

Dr. Steven LeBoeuf: [00:19:26] Now this data set is interesting. You see, when we developed the training here, the training model was done where people would come to a mall with a booth there and we had models across North Carolina or they would come to a facility in Vietnam or the Philippines and there would be a trained specialist who could measure that blood pressure through the exculpatory methodology, through the sphygmomanometer that was was there. And that was the ground truth. They also took a measurement simultaneously of and still metric on the opposing arm. And then they took a measurement on the ear. And that was done in the training in the ISO protocol. It’s more complex that people come in and there are multiple gustatory measurements done, multiple oscilometric measurements and a multiple ear measurements done. And it was all done according to that protocol. And there there it’s really strict. I mean, if somebody, for example, comes in and the blood pressure, as measured by three nurses, two of the nurses take individual measurements and they each have their own earpieces for the auscultatory analysis in their each measuring individually the blood pressure. And if they’re off by four millimeters of mercury as a as determined by the third nurse. So they write down on a sheet of paper what they measured. They sent it to the third nurse.

 

Dr. Steven LeBoeuf: [00:20:47] The third nurse reviews it and says, is this within four millimeters of mercury or not? And if it’s not, you can’t even use the dataset. So it’s a really strict protocol. What good it was. It was a really good way for us to judge how well we would do at passing and ISO test and then after this was done. The thing about it is you have these models, but for them to be extremely useful, you’d like them to be an embedded solution inside a device rather than something where you have to go to the cloud and hope there’s a connection that’ll do the processing for you. So we developed an embedded solution that took everything. We developed all the software in and put it into our perform tech biometric operating system. So this is biometric operating system. It it monitors your heart rate. It monitors your RR-intervals. It does all kinds of calculations. It calculates your blood pressure. All these things are done in real time and updates are given every second. And I want to talk a little bit about the embedded solution. I don’t want to get too technical in this discussion about it because that’s going to bore the hell out of everybody. But the the way that this embedded software works is you have your inputs, which is your inertial data, your PPG data in your static biometrics. And I mentioned, again, the static biometrics are things like age, weight and gender.

 

Dr. Steven LeBoeuf: [00:22:02] In the processing. The software has two components. There’s a and everything in here is autonomous. It is a qualifier. And as a BP estimator and I’ll explain with those due in a bit. And then from the BP estimator, it is the BP measurement that goes to the output. And there are many different ways to process the outputs that we provide at nodes for and software for the processing in the middle. The qualified the BP estimate. Let me talk a bit about that because it will help you with the rest of this presentation. The qualifier is something we developed that will look at that input, those inputs coming in, and it will say, do I believe that I’m really going to get an accurate BP measurement from this data? If so. It qualifies the data and then we know that when we estimate a BP, the statistics are going to be consistent with that of a curve. And we developed this. And it turns out that this is really important because in some cases you might get data where this person’s not going to have an accurate measurement and you don’t want to give them an inaccurate measurement. You want to give ideally only accurate measurements. Now, no sense is perfect, of course, but this qualifier helps substantially with this this this issue, which can be observed with PBG and really any technology, but also with PPG is particularly important because remember, now we’re not directly measuring blood pressure with this technology.

 

Dr. Steven LeBoeuf: [00:23:23] We’re measuring blood flow information in inertial information in the context of the static biometrics to estimate the blood pressure. So this this is the the how the embedded software is is is basically the nuts and bolts of it. So what I’d like to do now is explain the training data and the qualified test data set data. So the training data is on the left in the qualified test data set is on the right. The training data set is the data that was used to build the model. And you could see there were about twenty six hundred tests in almost seventeen hundred unique participants that were used to build this model. You can see the breakdown of men to women and smoggiest to not. Some folks were on medication. Some were not. Most were not. You could see the breakdown of the age in the range here. Also the breakdown of systolic and diastolic readings used to to build them up. All this stuff was done in the training data set, as I described earlier. At malls and facilities in Asia, in contrast, the qualified test data set was all collected using the ISO protocol. So this is the data where we had all the the three nurses, for example, collecting the data and putting it all together. And this is the data. This summarizes the data, only the data that was qualified. So, for example, if the qualified said, I don’t believe this is going to give a good result, that data was discarded.

 

Dr. Steven LeBoeuf: [00:24:54] This is only where the automatic qualifier said, I believe that this is going to give an accurate result. And so that’s what this data set is. And that ended up being about six hundred fifty four measures on one hundred forty seven participants. And the breakdown of the different ranges is here. Interestingly enough, we had to double check this the other day, that the qualified test data set was the exact swap of the agenda using the training data set, and that was just a coincidence. So we had thirty nine percent male to sixty one percent female in the qualified data set and it was swapped in the training data set. Kind of weird, but just how it turned out. So this qualified test data set, let me show you how we did compared to the old school district and also in context of the oscilometric auto-cuff because remember we measured ear, we measured finger, we measured wrist, we measured the auscultatory as the ground truth and also use the automated cuff as well. So this next plot shows the ear PPG and the auto-cuff data, systolic on the left, diastolic on the right, with the manual reference, the exculpatory reference being the ground truth. And when the results are so the on the systolic, if you look at the plots, the first thing that you’ll notice is that the trends for the PPG and auto-cuff are essentially identical. I mean, you really have to strain your eyes a bit, especially on the systolic side, to see any difference in the trend of how well the PPG did on the qualified data test data set versus the the auto-cuff.

 

Dr. Steven LeBoeuf: [00:26:34] Another thing, too, is if you look at the bottom of of each of these, you can see where I’ve summarised the statistics. Now, the ISO protocol, what it looks for is in order to pass, you need to have a mean error, no greater than five millimetres of mercury. And you need to have a standard deviation error, which is no greater than eight millimetres of mercury. And we were able to achieve that on this qualified test data set for both the systolic and the diastolic. You’ll note that on the systolic side, it looks like we did a little bit better than all the curve. That is not true. This is not within the margin of error. Seven point seven millimetres of mercury versus eight in this particular case is that small. Point three is just it’s in the noise. We can’t say we’re better than the cuff. Now, if you look at the right here in the diastolic side, that eight millimetres of mercury that we hit, that there definitely is the cuff is a little bit better in this case. It’s six point eight millimetres of mercury and that’s certainly a millimetre of mercury better. And that is significant. But still, we were able to hit the ISO requirements. So one of the things here in this data is that we wanted to say, OK, well, this is great.

 

Dr. Steven LeBoeuf: [00:27:54] Can can we use this as a way to gauge whether or not someone is hypertensive or not and how well would we do that compared to an auto-cuff? So we said, well, the simplest way to do this is just to use the A standard the American Health Association standard excuse me, American Heart Association standard. And one of the standards they give is they say you’re hypertensive if your systolic blood pressure is greater than or equal to one hundred thirty millimetres of mercury. And so we said, OK, well, let’s look at the auto-cuff and how well it did at this versus PPG. And they’re identical again here. It looks like wee wee. It looks like we’re a little bit better. That’s not significant. If you did this another time, it might be swapped. I mean, this is all the way to read this. Is there the same both of these, both at both the cuff and the people on this qualify test data set, they were identical at being able to predict whether or not someone would be classified as hypertensive by using a simple a very simple threshold of are you greater than equal to one hundred thirty millimetres of mercury in this test for the systolic blood pressure? So I want to talk about the qualified test data set more, because I mentioned that the qualifier rejects some data and it rejects a good bit of data.

 

Dr. Steven LeBoeuf: [00:29:05] And I want to explain more about how the qualifier works. So allow me to do that, please, so that the qualified and unqualified data sets that that you’ll see are based on how the qualifiers tune. So we tune to qualify. We found that we could not qualify to just decide. How confident did that software want to be in determining that your blood pressure was going to be extremely accurate with respect to a cuff? And when we were able to to tune it to where it can say, do I believe that this person is going to be in a statistical plus or minus eight? Do I believe that this this measurement is going to be a statistical plus or minus 10 or higher? And we did this we found was that we could we could tune the qualify to do this. And so all the data sets we collected, everything we collected, period in that ISO test, everything, 60 percent of the data sets were deemed qualified. So in other words, the data I showed you was 60 percent of all the data we collected, that qualifier went through all the data and said, here’s what I believe is going to be good flight data. It was right about that. It was certainly right about that. Now, it could be a little more liberal. And you say, OK, well, for a general wellness application and in and I should say that this is more than just, you know, well, as there are there’s another isostatic being investigated now where one of the ideas is maybe because of the fact that you don’t need a cough and that benefits there, they there may be willingness to reduce the precision that’s required.

 

Dr. Steven LeBoeuf: [00:30:35] So, in other words, have a more lax on the precisions where the standard deviation may be within plus or minus 10 millimeters of mercury. So if we take that liberty now, we assume that that’s that’s useful for public health. I believe it is. At least you don’t need to come to this. And 70 percent of the data sets will be within plus or minus 10 millimeters of mercury for a standard deviation of plus or minus 10. So with that with that knowledge, if you tune the qualifier to plus or minus 10, 85 percent of the data sets are acceptable in that ISO set and in that leaves 15 percent, which aren’t. So we went back and we look through all of the data that was rejected by the qualifier. So this 15 percent of data that was rejected as not being accurate enough and it certainly wasn’t it was not accurate enough. And we looked at longitudinal data we had collected, so we looked at the longitudinal data we collected. We have several data sets where people have come in for a few weeks, in some cases months or longer, and saw, OK, how how well we do it. We did an initial calibration of these folks, how well would we do? And we found that when we did that, these people, the stats for these people was plus or minus 10 millimeters.

 

Dr. Steven LeBoeuf: [00:31:49] So what that means is with what we have today, what I can say with 100 percent certainty today, what we have today. Eighty five percent of people who are tested, they’re going to see that the statistics are within five plus or minus 10 millimeters of mercury, 15 percent are not going to get that. But with a one time calibration, they will get that. And I feel really good about that, at least for four weeks of data on folks, because that’s where the majority of our longitude longitudinal data is. And so I feel really good about that today. If someone were to use this technology for a general wellness purpose, that is what you would get. Now, of course, there are R&D. We’re continuing to improve the R&D. We see improvements already. What I’m showing you is, is all the data. We have an update coming out in Q2. But the reality is that our goal is to, frankly, get to where 95 percent of subjects can see calibration for accuracy without a calibration at all. And you might say, well, why not 100 percent? It’s because, I mean, I’m a scientist and I’m real about this. I don’t know. I’m not aware of a single sensor in the world. I’ve looked at all of them. I’ve developed so many. That gets one hundred percent.

 

Dr. Steven LeBoeuf: [00:32:56] Everybody is always going to be some issues, even at even on roughly about five percent of the population. You can’t get a reliable reading either because the left bundle block other considerations. And so it gets ninety five is is is really good, believable goal and we believe we can get there, certainly the 90 percent and probably to ninety five percent as well with, with a calibration free solution about PPG. But today a product today can do these things. And because of the fact that a hardware is already set, the hardware is already set as it is, we can continue to provide firmware updates for improvement. Now, remember, what we were trying to do here was we were trying to do be as good as a cuff or as close to a cuff as possible for public health purposes. Our goal wasn’t to make a continuous blood pressure monitor. That was not the goal. However, it turns out we can do that. And I’ll show you some examples of this. So one of the series of tests that we did and I just have one example here, is we said, OK, let’s let’s do some. Reading studies on folks and monitor their blood pressure and see how are we doing, how are we doing in that in that study? And what we see is that we see exactly what you would expect to see. So in a healthy person, your your blood pressure and your heart rate in control, breathing should be one 80 out of phase.

 

Dr. Steven LeBoeuf: [00:34:15] And it’s because you’re that the theory is that your body is trying to maintain cardiac output. And as you inhale, your heart rate goes up, your blood pressure goes down, and as you exhale, your heart rate goes down and your blood pressure goes up. And this was important not only to see that we could truly track blood pressure with breathing, which we clearly can. And we’ve seen this time and time again. But but also, I needed to convince myself that we weren’t tracking heart rate. And let me explain, you see, and the PPG signal, the strongest signal by far when you’re in a in a wrestling environment, the strongest signal is heart rate. And so the data science mentality is part of my brain here says, well, wait a second. Now, what if heart rate is dominating these values? And that would not be good because that would mean you have a model that’s dependent on heart rate somehow. And so the fact that we see that the blood pressure is independent of heart rate was also refreshing to know that we really, truly are tracking blood pressure. And we’ve seen that time and time again. Another thing we’ve done is we’ve done some some exercise tests where we’ve done some some isometric induction of increased blood pressure. Here’s one example. This is typically what they look like. So this is a typical plot. It’s not our best. It’s not our worst. It’s which we typically see.

 

Dr. Steven LeBoeuf: [00:35:26] You would see those the the x axis. Here is time in seconds. The Y axis is blood pressure. And it’s a number of things to note here. The the red line is the our software’s estimation of the person’s blood pressure calibration free without a calibration. So here’s what we think your blood pressure is on the systolic side. Red line, the blue line here, that below it is the diastolic estimation from PPG. Again, not calibrated in any way. The pink dots are measured by the cuff and the blue dots are measured by the cuff, the pink dots being systolic in the blue dots being diastolic. And the square wave that you see just shows when the person does the exercise. So the person starts an exercise at the blood pressure goes up and how well do we track with that? And we see that. We certainly can. And we’ve done more work with this on this recently, a lot more work recently on this, because for different use cases, this can be really important. In speaking of speaking of use cases, what I like to do now is, is move on to discuss some of these use cases. So Valencell, one of the things like i mentioned, we rely on partners to get our technology into the marketplace. It’s what we’ve done is we spend some time thinking about how would you use this technology for a general wellness purpose versus a medical purpose. And so I’ll show you what we’ve conceived out here.

 

Dr. Steven LeBoeuf: [00:36:58] So in the case of general wellness, so it’s focused in our minds on monitoring blood pressure, using a biometric audio earbud, as, of course, as we discussed earlier, the use of of earbuds and smart earbuds and Bluetooth headsets and Bluetooth earbuds and in air pods and all kinds of things like this is proliferated. Now, this is a large group of consumers that are using these for exercise, listening to music, talking on the phone, a variety of things. And so if you start at the top left, you can see a situation where someone could put in the earbud and they get an update. They can get an update about how well they slept based off of their heart rate variability, for example, in that morning. And also get an update about what’s suggested for them to do that day. And also because we’re able to track your blood pressure and your stress as you as you drive or whatever you’re doing with the earbud we can even recommend. What’s the lowest stress route for you? What we learn we can we can then see and you appear to be stressed, we can recommend audibly a relaxation session. And there are several solutions available today in the consumer marketplace for doing this. There’s a company called Headspace, very popular for auto audio visual based stress reduction. And then also Calm is another one that many others and their people really do like these. So with this technology, you can be alerted to this if we sense that you’re more stressed.

 

Dr. Steven LeBoeuf: [00:38:20] And not only that, you get to see the results of the impact of that. So how well did that session help you reduce your stress and your blood pressure so that aetiology between stress and blood pressure can be better explained for the person there? And so and another thing here is perhaps someone needs to be relaxed and you could play a more relaxing track and music. And maybe you’ve learned this from how the music affects the blood pressure and the stress or for exercise. For example, you may recommend certain zones of heart rate exercise that are known to improve your blood pressure. And you can see the. The interaction between the diet and exercise choices you made and how well they’ve helped you reduce your blood pressure or keep it under control, and at the end of the day, you might look at your phone or look at some screen of whatever your choice and then see a summary of what you’ve done. And then you may want to help or relax yourself. And then you can be you can be you can institute a execute rather stress reduction or stress or relaxation therapy through the audio to reduce your blood pressure and get to see how well, by controlling your stress, you can reduce your blood pressure and then maybe get a better night’s sleep. So this is kind of a day in the life, different, different ways to look at how a wellness blood pressure mounting device might work.

 

Dr. Steven LeBoeuf: [00:39:37] I feel that it’s a great way to help people connect the exercise in lifestyle and diet choices they make with health improvements. All things that are generally well, well known to the public now on the medical side, it’s different. On the medical side, we see that we really the key here is to connect the person to a medical professional who can help them with managing the hypertension and in managing it with respect to diet and exercise, of course, but also potentially medication. And there’s a few ways to do this. We you know, the our technology can be integrated into hearing aids and it’s currently been integrated into hearing aid components. And so for somebody in the daily living, we can we can track various different vital parameters, such as in principle, heart rate, heart rate variability, blood oxygenation, blood pressure, even body temp is possible in the ear. And you can do this assessment and you can you can then without the person even knowing this has happened in a few times a day, but then provide an assessment that the doctor might be able to see maybe maybe just let them know how well they’re responding to whatever the therapy would be. Another use case, which is a little different than this free living use case, is, is an extremely important and valuable. We became aware of this issue from the VA a few years back where where there’s a need to triage people when they come into a hospital, for example, and and they may be sitting down waiting to get service.

 

Dr. Steven LeBoeuf: [00:41:07] And if you can place in your piece of them with some cover on topic or sanitary cover, which is plenty of your pieces today and places in their ear, and do an assessment of the heart rate, heart rate variability, blood oxygenation, blood pressure, body temp. With this device, you can do some great triage and get a lot information that could be useful in helping get them what they need sooner. And of course, now with with the heightened public awareness of diseases and how it might prevent you from being able to be mobile to either get to a doctor for telehealth purposes, this kind of device that can measure all of this in one spot and also audibly tell you how to do the tests could be really, really interesting for folks. So that’s how we see it on. That’s really how we see the use case on the medical side. And and this is just some of what we have done. And we’re looking for partners to help us to launch products in the general wellness side and the medical device. So. So where are we going from here? The first thing is expanding data collection, a lot of the reason why are the qualifier rejects so many measurements is because we believe and we’ve actually identified this, we know this is part of the problem is holes in the data collection.

 

Dr. Steven LeBoeuf: [00:42:19] So whenever you build a machine learning model, the model is connecting the dots between all these parameters in the blood pressure. And there are some holes there and in particular people who suffer from extreme hypertension. We have less of in our model. And so we are currently building that database. And I think, God, we collected all the data we did so far because it’s much harder now to collect clinical trial data because of the coronavirus issue we’re dealing with today. But of course, that will open back up again. We’ll be able to get more data, but we’ve got plenty of data already. And frankly, we have another path that we’re working on now for continual improvement with the massive data we already have. And as I mentioned, we’ve already seen improvements that we just simply not reporting right now on the on the blood pressure side. And then lastly, another thing is other body locations, because we don’t when we say it’s true that we have a lot of faith in the ear based of everything we’ve done and measured and also the science behind it. But we don’t mean to put down any form factors. So we have a lot of business in the wrist and arm in our business. You know, we provide solutions for the wrist and arm and also the chest, frankly. And the reality is that we love those form factors. We don’t think it’s impossible to make this blood pressure, modern technology work on.

 

Dr. Steven LeBoeuf: [00:43:36] These are the form factors. We don’t believe it’s impossible. We just think you need a whole lot more data in order to make that work. And it’s a difficult and challenging to get those data sets. Those data sets are not easy to come by. And and also the benchmarks you use to collect that data aren’t so hot. You know, it’s you really need somebody trained to get an accurate measurement from an escort that we monitor. And then a supplementary cuff isn’t always so hot either. And many all know if you take more than five measurements, a date on a millimeter cuff, then every six months or so, you’re supposed to take it back to to the manufacturer to get it get it retuned, to get a recalibrated to approach your source. So these are the benchmarks are not the easiest thing to use in building these models. So the last thing I’ll talk about here is, is how to get access to the technology we have. We’ve made this now available. We have a deep evaluation kit. It’s got three major components to it. It’s got some hardware, software and documentation. Now, the hardware has a wired earbud with embedded PPG in it. So all the calculations are done on this little chip that you can’t see on this board. And this board is designed to be large is just simply so you can test it. You can you can test your leads and see what the potentials are across the various different sides.

 

Dr. Steven LeBoeuf: [00:44:56] It’s a USB development board. It’s a there’s a everything is processed, all the metrics of process on this low power MCU that’s on the board. The interface is a PC interface. So you plug the dev kit into the PC, the PC software that does all the calculations and shows you all the diagnostics, for example. What’s the qualifier saying about the signal? What is the big wave look like? What what had there been too many motion artifacts, for example? And do you need a reset, a test? All that’s done there in the software and then the documentation that is provided is helps guide you on how to use Dev. Kit is videos as well and how to use the deficit and also how you would go about building a device with this technology. The sensor inside that earbud you see on the left is a commercially available sensor, which makes it great for a time, the market for the solutions. And we also have test procedures we recommend in a validation white paper on how we did our testing for you in this in this kit. And these kids are available. You can get them from just going to valences website. I Ryan, at the at the end of the presentation, maybe you could remind folks what’s the best email to use. It’s probably I’m guessing it’s benchmarked Vonzell dot com for for them to to get the information.

 

Dr. Steven LeBoeuf: [00:46:18] But you go to our website and on the website, get the information on how to get access to one of these evaluation kits. So inclosing here before open it up for Q&A. We’ve developed this calibration-free PPG based BP technology. It demonstrates cuff like performance without the cuff. We have followed the ISO eighty one sixty dash two protocol, the twenty eighteen version and the IR PPG blood pressure solution is demonstrated accuracy within the required limit for the qualified test data set. The current solution is we believe the current solution is ready for commercialization in a general wellness product today. In parallel, regulatory efforts are underway to enable medical grade solutions as well, and also continuous R&D to improve what we’re doing today. Now, although this isn’t designed for continuous monitoring, it wasn’t it turns out it can do this and it does it pretty well. And so you can also use this technology for continuous monitoring use cases. You know, originally what we envision was this. This technology would turn on every once in a while. I’ll take a measurement and turn it off. But it turns out you can get good, good measurements every second from the device. So, again, lastly, if you want an any kit, here appears the website right here [email protected], and we are extremely excited to be able to work with you to get this out. And now I would like to say that I’m extremely eager to open it up for Q&A.

 

Ryan Kraudel: [00:47:55] And we’ve got a lot of eager attendees as well, a bunch of questions have been coming in, great questions. Keep them coming. I’ve tried to we’ve had several overlapping questions, so I’ve tried to categorize some of the some of the main themes of the questions that are coming in, that one of the big ones, of course, is the regulatory question of what what would be required to go through a regulatory clearance process and what would be acceptable under general wellness criteria. So, LeBoeuf, can you talk a little bit more about our perspective on where to draw that line, what those what those different use cases and scenarios might look like?

 

Dr. Steven LeBoeuf: [00:48:42] You bet. So we’ve had consultation on this from our own consultants and elsewhere. And for let’s talk about the general wellness use case first. There is provision from the FDA for products that are used for general wellness purposes. It’s a general wellness exemption. It used to be I want to say it used to be called like a five to three B and you submitted that to the FDA. You no longer that provision. My understanding is no longer there to submit as a five to three B, rather, you don’t make any medical claims. You launch as a general wellness device and you only use it for general wellness purposes. And this means devices that are I meet with the FDA’s exact language, but in general, this means where it is generally known that exercise and diet and in lifestyle changes can improve your health. And so for this case, the way you would use the technology would be you could not launch this as a blood pressure monitor. In this case, you would launch it as a device that can estimate your blood pressure for general wellness purposes not intended for medical use. You could not launch that as a medical device. You would have to launch it as a general wellness solution on the regulatory side, where you want to make the medical claim that this is a blood pressure monitor and use it for whatever you want to use it for, for example, a blood pressure cuff equivalent or you want to use it maybe for hypertension indication or whatever you would like to use it for.

 

Dr. Steven LeBoeuf: [00:50:11] The mechanisms are five ten K or the mechanism could also be a de novo in some other mechanisms you could use as well now for software as a medical device. So if you want to launch, this is a software solution. The limitation that we have available is, you see, we don’t make the final product and whoever makes the final product is the one who needs to get approval. So the way we’ve been helping out here is we executed the ISO test ourselves. And the purpose of that was to have that data set to prove to ourselves into our partners that if they instituted or they rather executed and ISO test just like this, that they would be able to meet the ISO requirements. So we’ve been doing that ourselves. Also, we’ve been getting all certifications because the software we provide needs to be appropriately certified so that we can provide it to you. And so we’ve been through the process. We’ve got many of them already and we’ve got to think a few I’s to and T’s to cross to be fully certified. So we’re taking that upon ourselves. So providing all the data people need to get their own fighting capability to be substantially equivalent to a what we’re doing that and also and also the certifications we need to be able to be compliant with the the software that is is required to make this work.

 

Ryan Kraudel: [00:51:31] Great, several questions come have come in related to the process of taking a BP measurement. One question that keeps coming up is how long does it take to get one BP measurement with this technology?

 

Dr. Steven LeBoeuf: [00:51:49] Oh, a great question. You know, I didn’t mention that. So right now, we take 30 seconds. You don’t need 30 seconds. I’ll tell you, you really need six seconds. Bare minimum, you really need 60. The first measurement, because you need to get enough waves to make an estimate. The reason we pick 30 seconds is from the beginning, we envision we were going to go for a five 10 K on this with a curve and take about 30 seconds to go in. We had been instructed early on that the closer we can make this to how Acuff would measure, the better. And so, frankly, that’s what we picked the 30 seconds. You don’t it doesn’t need to be 30 seconds. Six seconds will do unless you’ve got bad PPG waves, in which case, of course, it will take longer. But six seconds of waves would do to get you an accurate measurement. But right now, the software takes 30 seconds. And then after that, 30 seconds, we updated every second.

 

Ryan Kraudel: [00:52:39] Got it. And then a few questions came in around the discussion of the the the ears static distance from the heart and questions more around it, does the signal impacted by whether the person is sitting up or lying down?

 

Dr. Steven LeBoeuf: [00:52:58] Yes, it is. And so what we’ve seen is that so so I’ve got to say, first of all, let me say this. We haven’t done extensive studies on position. And the blood pressure we have done is we’ve done some studies to see does the blood pressure change the right way depending on what people are doing. So. So in other words, if somebody is pushing their feet against a wall or some kind of weight to build pressure in the body, do we see the blood pressure going up? And it does. We’ve said, OK, well, if somebody lays down, does the blood pressure go down when they stand up, there’s a blood pressure go down. Does it go up? And we’re seeing that that is the case. The blood pressure is trending in the right in the right relationship. So we feel good about that, at least to date. But we I’ve got to say, we haven’t done it. If you wanted to test this for this and wanted to to market it for this purpose, you would certainly need to do more testing on that. And we’ve done our focus has been on making it meat equivalent to a cough, like a cough would be used, but without a cough.

 

Ryan Kraudel: [00:54:04] Got it, and a few questions on movement, does it does it require the individual to be still?

 

Dr. Steven LeBoeuf: [00:54:12] Well, it’s I’ll tell you this, the short answer is no. The longer answer is they need to be still enough. So which you can’t have is you can’t have certain in the ear. For example, we’ve so we’ve done test on this already. If the person is wearing this on the ear and they’re walking around and they’re not moving their mouths, we’re talking you’re going to get really deep values, really very sensible values. But when you talk or if you run, the noise is too strong. In Valencell, we are very skilled at removing motion noise and we do this for heart rate in our eye, probably the best in the world. That is, frankly, at least that’s what most people tell us that we work with. But for blood pressure, it’s still hard because the information is so caught up in that those PPG waves that you really need nice looking waves. And so on air, you can walk, you can move around as long as you’re not running impact noise is bad and also mouth motion noise is bad on if you use this technology on the thing that we have the technology as well, it’s just not as accurate as the ear on the thing that, you know, you can’t you can’t walk. You know, you have to be really still on the finger to make it work. And part of it is because, you know, you have modulations and pressure on your finger as you move and you just don’t have those in an earbud or or in a hearing aid.

 

Ryan Kraudel: [00:55:34] Got. Several questions have come in around the longitudinal data and does it continue to track over time? So I know you touched on that earlier, but can you expand on that a little bit?

 

Dr. Steven LeBoeuf: [00:55:51] The oh, yes, so with the longitudinal data, I see what you mean. So the longitudinal I’m I’m glad you asked this question because there’s an important point that I forgot to mention. So please allow me to mention this to quickly answer the question right now with confidence. Only have four weeks of data. We do have some folks that did go as much as a year away from measurement and they were fine, but it was only like seven folks. It’s just not enough statistical data. You know, I’ve got hundreds of folks who’ve been tested for at least four times over the course of four or five or six weeks. And we know that it holds up for that just fine. I can’t really say beyond that. But this is why this is another thing that’s important to point out. So one other thing. Violence is in the IQ test. Remember how to all this eight measurement. You taken this test? Well, up to eight measurements you can take. And we took all eight, which you can take all eight measurements back to back. You might take a whole hour for this person, but we’re going to take eight measurements back to back with the exculpatory telemetry cuff and the earbuds and all the other senses that are there. So we did this. So if you take that data in that session and you say that first data dataset, I’m going to make that the calibration. The statistics that you get are absurdly good, much better than any blood pressure cuff that I’ve ever seen. I mean, you get standard deviations around six millimeters of mercury. You get mean errors that are extremely low.

 

Dr. Steven LeBoeuf: [00:57:18] The problem is, I don’t see much public health use for that. And what I mean is an ISO test, an IQ test. Remember these people, this sitting down very calmly the whole time, not moving at all. And if another thing that the ISO requirement does is if your blood pressure changes by more than a rough, I want to say, 12 millimeters of mercury systolic over the course of those measurements, you’ve got to throw those data sets out so the person’s blood pressure ain’t changing hardly at all. And so now if I’m going to say, OK, well, I’m going to go ahead and judge myself by doing a calibration of this, I suggest with the blood pressure is hardly change. And you told me what the blood pressure is the first time and calibrate it and I do. Great. It’s not really fair. So what we did is we looked at we didn’t do that. We used longitudinal data. That’s where people came on separate days over the course of several weeks. And in that case is if we tell you if you told me the first blood pressure is has no bearing on what the next blood pressure readings are, for all practical purposes, there are two separate days and the person could have come back from exercise that could be hyped up one day, more in coffee one day versus the next, and you get to see that. And so that’s what was that the six millimeters of mercury from the ISO calibration was not nearly as interesting to me as a 10 millimeters of Mercury longitudinal.

 

Ryan Kraudel: [00:58:35] They’re speaking of calibration. There have been several questions that come in about what is the calibration protocol that we’ve used.

 

Dr. Steven LeBoeuf: [00:58:43] Oh, simple, in this particular case, everything I’m showing is extremely simple protocol, literally, for what I’ve talked about today, it’s only been an offset. And then we just keep applying that all set and that’s it, that we have much more sophisticated calibration protocols, we have not gone public with how we’re doing those yet and we haven’t made those available for folks yet. But this particular calibration was simply meant as a proof of concept to say if you took a measurement on day one and you said, here’s my all set, you kept that all set all the way for the rest of however long that calibration lasts, which we know it doesn’t change at all, or at least a month, then then you’re good to go.

 

Ryan Kraudel: [00:59:24] Gotcha. Also, quite a few questions on gender differences, given that hypertension tends to be more prevalent in males. Have we noticed any differences across genders or any any differences either in the results and or in the way forms across the different genders?

 

Dr. Steven LeBoeuf: [00:59:48] Well, you know, one thing I would recommend you all do here is we have a white paper online that you can you can download and has more information about this. And so what I’m telling you right now is from memory only, but I believe the white paper has the details people are looking for. But the reality is there is a slight trend where men have high blood pressure and women. And there’s also a slight trend where certain age ranges have high blood pressure than others we’ve seen. And and I’ll tell you from our data, it goes up for a while with age and then starts going down. And we’ve since checked this and seen that’s consistent with what’s known for the broader population of the planet. So so we do see some of those trends. And I think the white paper has information about some of those, either directly or indirectly in the documentation.

 

Ryan Kraudel: [01:00:38] Got it. Question around the skin tone and skin pigmentation, does that have any effect on the accuracy?

 

Dr. Steven LeBoeuf: [01:00:51] There’s none at all. There’s no we haven’t found any evidence whatsoever. We looked at this every which way from Sunday. Now, I’ll say this. The reality is that if you have darker skin, typically your PPG signals won’t have as much intensity as someone who has a lighter skin. It’s not always the case because there’s many factors that go into that. And frankly, perfusion is so much a bigger factor for what we do. That skin tone is not much of an issue, but it’s true. If everything else were the same and someone with darker skin would have a lower signal, but still well within the range, we can measure the blood pressure. So we haven’t seen an issue with that when where we have seen issues as people, where the perfusion is just really, really low or or maybe they were wearing in the earbud incorrectly. And the person that’s the person who is managing the study didn’t catch that or they hadn’t worn correctly. And then for whatever reason, it popped out of the year, that that’s the kind of thing we see. We really don’t see any skin tone issues here and bounce off of that.

 

Ryan Kraudel: [01:01:53] Got it. OK. Those are the major themes of the questions I know where we’re five minutes past the top of the hour. But LeBouef, if you’ve still got time, we’ll continue to go through these questions and I’ll get the more granular question. Do it.

 

Dr. Steven LeBoeuf: [01:02:09] This is how we do it.

 

Ryan Kraudel: [01:02:11] Ok, question next one around. Do you find more accuracy in the ear canal versus the concha placement?

 

Dr. Steven LeBoeuf: [01:02:22] Another awesome question, what I will say is that there’s a lot I can say about this, but a lot that is I would consider to be confidential. But what I can say is that we haven’t found any difference on the blood pressure waveforms in the inner ear versus the conscious. And so today, the technology we have that works in the culture will work the same way, all things being equal in the ear canal. The thing you have to be mindful of, of course, is you have more motion artifacts in the ear canal than you would have in the content. You would expect this if you ever stuck your fingers in your ears, for example, and talk. You’re going to see you’re going to feel a lot of a lot of motion there. And you just don’t have that motion in the content when you talk. I mean, you know, you can have our technology in the ear and talk and talk and talk for heart rate and not have an issue at all. And if you have our technical technology in the ear and talk and talk and talk, in some cases you’ll you’ll get an era because the tech that we just can’t accurately measure the blood pressure. So from the wave perspective, the information content is the same as the ear canal versus the coccia. As far as we can tell. From everything we looked at so far, we’ve looked at it pretty extensively. They they all look the same. As far as measuring blood pressure, that is not true for the thing. And risk those signals do not have the same content.

 

Ryan Kraudel: [01:03:44] Got it. The next question is, along with skin tone, quality, any impacts on age or differences and results from an age standpoint

 

Dr. Steven LeBoeuf: [01:03:56] Either are on the white paper. We show the trend in our eras. So we also show the trend in our residuals and compare it to the trends in the residuals of the Ocilla metric cuff. And what’s weird is what’s weird is we are where we get we are less accurate at the extremes. So the lower your blood pressure in the high blood pressure, generally speaking, the less accurate we are. And that’s true also for the telemetric in the study that we did, which is which is is odd. I wasn’t expecting it. And there definitely are some trends where certain ages are are less accurate than others. And I can’t remember exactly what they are because it’s not it’s not Lydia. It’s the certain age ranges where we’re more accurate than others. And frankly, we believe it has to do more with the data support used to build the model.

 

Ryan Kraudel: [01:04:53] And you’ve referenced the white paper several times and we’ve had a recommendation come through to to send out a link to the White Paper along with the recording of the webinar. And that’s a great idea. We will do that. So, Aronne, know where to get that straight away. Next question is on. Does hypertension medication impact the estimation, accuracy in any way?

 

Dr. Steven LeBoeuf: [01:05:18] What great question. And I’ll tell you, we have seen that it appears that some some medications have no effect at all. And this wasn’t just by our testing. We’ve had partners evaluate our kids already, other medications. It seems like they probably do. And we’re not 100 percent certain if that’s the case, because, again, getting this is a statistical game we’re playing here and getting enough data to be certain is is is one thing. But I’ll tell you that if you took out all the people so if you remember how I said we have this qualified data test data set with a qualified says, I like you, you, you and you measurement, but I don’t like you. Right. If you took every data set that came through in that ISO test and you took out just everybody was on medication, then the statistics for everybody, I believe, are better than 10 millimeters of mercury or thereabout. So without a calibration. So we know that that that, generally speaking, medication has some impact in the negative direction. But when you look at individual medications, some have no impact at all and other medications look like they might you. Part of that’s because, you know, the several medications for blood pressure and they work in different ways. And as you can imagine, a medication that might affect that transfer function between the the blood pressure and the blood flow could have an impact. And so in the future, one of the ways we can address this also is by filling in those gaps, frankly, with data collection. And that’s one of the things we really need to do, because we know that every time we’ve been doing this for years, every time we feel get that gap in data collection improves the model every time. So we need to do and find partners to help us collect more data on more different types of folks, but also as a fix. Additionally, you can do is had that be part model where the the medication someone is on is part of the model to improve the results.

 

Ryan Kraudel: [01:07:16] Gotcha, looks like several other themes are popping out here. One is a request to describe the sensor hardware and a little more detail and talk to the power consumption of the sensor hardware. So let’s start with that.

 

Dr. Steven LeBoeuf: [01:07:34] All right. So that since the hardware that you see there is from a BE 5.0, it’s a it’s an ear sensor that has the capability of monitoring your heart rate, RR-intervals and hence HRV, your activity parameters. So accelerometer, for example, cadence, it also can can detect your activity status. So we have software from that since a module that determines your activity status. For example, what activities are you doing, for example, there then additionally has the blood pressure estimation. It also has an output, a multi wavelength output that can be used for SpO2 estimation. We provide a little starter algorithm for you there. We don’t have a we have not done a medical evaluation of our SpO2. That’s not what we’re focused on. But we have put together a starter algorithm based off internal testing that you can build upon to improve the SpO2. And it’s also motion tolerant. Additionally, we provide information in that sense homology you can extract respiration rate and other respiration rate characteristics. It’s all provided from that. Since Emanuel is the best 5.0, it’s power consumption at full blast is somewhere, I want to say around four hundred micrograms of current and full blast. And so what which you typically do for for blood pressure is you’re not measuring continuously. You turn it on, you take a measurement, you turn it off. Yeah, gotcha. And if they want spec sheets on it, Ryan, that’s another thing they should ask. It would be on valencell.com.

 

Ryan Kraudel: [01:09:14] Yeah. Yep, agreed. Another few questions have come in around have we tested this technology specifically on populations with a fib or other arrhythmias?

 

Dr. Steven LeBoeuf: [01:09:28] We haven’t tested we have not tested this on populations with afib or other arrhythmias, anecdotally, what I’ll say is we have gone through. Good Lord, I’ve seen so many PPG waves in my life. And we do see folks where when you do a quackery analysis, they clearly have a of it. And in that particular case, we’ve looked to see can we determine whether or not they have blood pressure? In some cases we can. In other cases we can’t. We had one person come to our booth who is a well known in the field. He’s he’s a he’s been a speaker and presented in various areas and also has a journal. And there’s a lot of things in the media and wearables field in medical media. And he he was he was tested and we could not get a blood pressure off him to save our lives, at least with our current our current software. And so we know that there’s some cases where people have atrophied. We just can’t do it. We know this other cases where we can we have not seen or not looked at atrophied populations explicitly to see how to solve the issue for those persons. That’s another way that we could work with folks, folks who have access to those cords. Can we find a way to collect data and then develop a model to to solve issues with those particular populations?

 

Ryan Kraudel: [01:10:47] Gotcha, several. I’m trying to go back through the questions and see which ones we have here for the ISO data study normally requires eighty five subjects, three data sets, each for a total of 30 sorry, two hundred fifty five data sets in your validation study at one hundred forty seven subjects, 60 percent were qualified. How how to choose qualified data versus unqualified data without risking cherry picking.

 

Dr. Steven LeBoeuf: [01:11:23] Oh yes. I want to be explicit. This was not cherry picked so we had twenty thirty five data sets. The qualifier went through it. It picked who it thought would be accurate or not. We did not pick that at all. And that’s where that one forty seven that sixty percent of data sets came from. It came from those in and of that two fifty five I came exactly. I broke down. The white paper explains it. But the bottom line is that we didn’t hear anything. The qualifier did, its job was completely automated. We didn’t hear anything. It was a matter of fact, I should mention that the qualifier, if we want to cherry pick, we’d be hurting ourselves because the qualify isn’t perfect. So it rejects a lot of good data. So a lot of people’s data, a lot of people’s data was was very accurate, but they were rejected by the qualified because the qualified thought that the measurement was not going to be good. So it’s really good at saying if the measurements are going to be good, it’s not quite as good as some of the measurements going to be bad in. And so it’s if there was a baby with the bathwater, so to speak.

 

Ryan Kraudel: [01:12:29] Yep. And we we can’t have a webinar without question about blood glucose, so I’ll throw this one out there as your research continues. Is it possible to measure blood glucose? With this technology,

 

Dr. Steven LeBoeuf: [01:12:45] Though, I developed an algorithm from blood glucose and I sold it to Apple for about a billion dollars, so it’s done. No, I’m kidding. So I’ll tell you, the real story here with glucose is that Valencell has it has looked into this pretty thoroughly. And we’ve been looking into this for years. We as we stand here today, we don’t believe there’s a good way to truly noninvasively in a truly wearable way, a way that’s not so uncomfortable. You just really freaking prick yourself. We don’t believe there’s a good way. Noninvasively today without a calibration and without a painful God knows what waiting, however long process to to estimate your blood glucose that we think minimally invasive ways are great ways. They have their troubles, too. But we do believe in minimally invasive ways and invasive ways. We do believe that that’s viable. But we don’t see a truly non-invasive, truly wearable, low paying way to do this. However, we do believe it’s possible. We do believe this is very possible and we have data to support it, that you can predict whether or not someone may be about to have a spike or drop in their glucose. And we believe that this is possible based off the technologies that we have today. And we certainly are looking for partners who are interested in exploring that. Further in this can be useful because the reality is that you you you may want to predict ahead of time that someone’s about to have a spike or drop so they can take some kind of some kind of action, some kind of intervention, whether it’s the test themselves or to go to professional or to take a glucose dose if it’s going to go through low or take an insulin dose, if it’s going to go too high. But again, we have the internal data to support this, that this is possible. We don’t have the data. I would not be comfortable showing you that data today on a population of hundreds of people like for blood pressure.

 

Ryan Kraudel: [01:14:41] Got it and looking back through, we’ve had several questions or questions and comments coming in around sleep and wondering if we have done any testing of this technology to try to capture nocturnal dips and blood pressure or just blood pressure overnight during sleep.

 

Dr. Steven LeBoeuf: [01:15:00] Not intentionally. We have not done that intentionally. Accidentally. We had someone who did a longitudinal measurement wearing a data collected from us, and we saw that was the case. So it’s of one person. So we not put any credence into that at all. My best guess is that you would see that just fine. I mean, we see we are able to track continuously quite well, unless there’s a whole bunch of motion going on and we can sense that and stop the measurement. So during sleep, we would anticipate you would be able to see these dips. We would certainly anticipate that. And that’s one thing you could do with this kit. You can take this kit and you could you could go ahead and put it on folks and in various ways to see can you see that dip in sleep? And maybe that’s very useful for your use case.

 

Ryan Kraudel: [01:15:47] We’ve had several questions come in about our machine learning features that we’re using. I wanted to throw it out there because we’ve had several questions and comments come in, but not sure how much we can share about that. So I’ll leave that to you.

 

Dr. Steven LeBoeuf: [01:16:05] What’s the question? So I don’t

 

Ryan Kraudel: [01:16:07] Think all of the questions around our use of machine learning techniques and the features we’re using in our model.

 

Dr. Steven LeBoeuf: [01:16:14] Oh, yeah, so I’ll say that we we have filed a patent on everything, so we filed a patent application in twenty fourteen and we still use at least those methodologies today. We’ve additionally developed additional methodologies which we filed additional patent patents on. And our practice is that we, we, we explain how to do everything in the patents and we require you to read it. We’re not going to tell you anything extra so that the that that’s all there anything that we learn in the future that we feel is is is something that we can share. We will. But because of the sensitivity of this in the marketplace and the fact that we’ve had to enforce our patents before, we will leave it to just the patents.

 

Ryan Kraudel: [01:17:05] Gotcha. Well, we are way over time at this point, and I think we have covered the vast majority of the questions. My apologies if we didn’t get to your question specifically, but please do feel free to reach out to us at [email protected]. I think there may be one other slide with our contact information, specifically if you’ve got questions or want to reach out to us directly. A lot of a lot of things coming in from attendees for the webinar and the information shared. And so I’ll just say thank you to everyone for taking time out of your busy schedules to join us here today. We will be sending out the recording and the slides and also the link to the white paper as requested earlier. So please do share that with anyone you think might be interested and reach out to us if you have other questions or would like to discuss this further.

 

Dr. Steven LeBoeuf: [01:18:09] Yeah, and if you don’t mind, let me let me close by saying again, thinking, folks, we are eager to find partners to get this in the marketplace. We know we have something big. We know it’s very different from anything else is out there. We’ve got a lot of reports from folks of things similar to this. And when people test and they just don’t work as advertised and we’re finding from folks who evaluated this decade so far is that it’s working as advertised. So it makes me feel really good. And I want to say that a lot of the questions that were brought up today showcased the limits of what we can do it. And so, for example, the question about sleep monitoring, you might say, well, Jesus failed, so why don’t you just go ahead and do that test? Well, we’re 40 something people here at the sell and we’re focusing on a technology that’s groundbreaking. We can’t do those things ourselves to find partners who want to take it and move it that way. That’s the kind of thing that’ll help really get this out there in various ways. We want to help you be successful using our technology for use cases. You feel important and we realize there’s only so much we could do. And so that’s why we’re such a partnership driven company.

 

Ryan Kraudel: [01:19:08] Absolutely, yep, we look forward to working with our partners on it. So thanks again, everyone. Thank you. LeBoeuf, great job as usual. And everyone, have a great day. Stay healthy out there. That’s.

 

Dr. Steven LeBoeuf: [01:19:27] You all will go into. But.