The Peter Attia Drive - #54 – Kevin Sayer, CEO of Dexcom： Continuous glucose monitors – impact of food, sleep, and stress on glucose, the unmatched power of CGM to drive behavioral change, and the exciting future of CGM

00:00:00.000 Hey everyone, welcome to the Peter Atiyah drive. I'm your host, Peter Atiyah. The drive

00:00:10.880 is a result of my hunger for optimizing performance, health, longevity, critical thinking, along

00:00:15.940 with a few other obsessions along the way. I've spent the last several years working

00:00:19.660 with some of the most successful top performing individuals in the world. And this podcast

00:00:23.620 is my attempt to synthesize what I've learned along the way to help you live a higher quality,

00:00:28.360 more fulfilling life. If you enjoy this podcast, you can find more information on today's episode

00:00:33.000 and other topics at peteratiyahmd.com. Hey everybody, welcome to this week's episode

00:00:43.360 of the drive. I'd like to take a couple of minutes to talk about why we don't run ads on this podcast

00:00:48.580 and why instead we've chosen to rely entirely on listener support. If you're listening to this,

00:00:53.820 you probably already know, but the two things I care most about professionally are how to live

00:00:59.060 longer and how to live better. I have a complete fascination and obsession with this topic. I

00:01:04.760 practice it professionally and I've seen firsthand how access to information is basically all people

00:01:09.920 need to make better decisions and improve the quality of their lives. Curating and sharing this

00:01:15.100 knowledge is not easy. And even before starting the podcast, that became clear to me. The sheer volume

00:01:20.380 of material published in this space is overwhelming. I'm fortunate to have a great team that helps me

00:01:25.800 continue learning and sharing this information with you. To take one example, our show notes are in a

00:01:31.600 league of their own. In fact, we now have a full-time person that is dedicated to producing those

00:01:36.260 and the feedback has mirrored this. So all of this raises a natural question. How will we continue

00:01:42.440 to fund the work necessary to support this? As you probably know, the tried and true way to do this

00:01:47.980 is to sell ads. But after a lot of contemplation, that model just doesn't feel right to me for a few

00:01:54.200 reasons. Now, the first and most important of these is trust. I'm not sure how you could trust me if I'm

00:02:00.400 telling you about something when you know I'm being paid by the company that makes it to tell you about

00:02:05.080 it. Another reason selling ads doesn't feel right to me is because I just know myself. I have a really

00:02:11.280 hard time advocating for something that I'm not absolutely nuts for. So if I don't feel that way

00:02:16.460 about something, I don't know how I can talk about it enthusiastically. So instead of selling ads,

00:02:21.580 I've chosen to do what a handful of others have proved can work over time. And that is to create

00:02:27.420 a subscriber support model for my audience. This keeps my relationship with you both simple and

00:02:33.480 honest. If you value what I'm doing, you can become a member and support us at whatever level

00:02:39.460 works for you. In exchange, you'll get the benefits above and beyond what's available for free.

00:02:44.300 It's that simple. It's my goal to ensure that no matter what level you choose to support us at,

00:02:50.040 you will get back more than you give. So for example, members will receive full access to the

00:02:57.520 exclusive show notes, including other things that we plan to build upon, such as the downloadable

00:03:04.240 transcripts for each episode. These are useful beyond just the podcast, especially given the technical

00:03:09.180 nature of many of our shows. Members also get exclusive access to listen to and participate

00:03:15.940 in the regular ask me anything episodes. That means asking questions directly into the AMA portal

00:03:22.660 and also getting to hear these podcasts when they come out. Lastly, and this is something I'm really

00:03:28.000 excited about. I want my supporters to get the best deals possible on the products that I love.

00:03:32.800 And as I said, we're not taking ad dollars from anyone, but instead, what I'd like to do is work

00:03:37.360 with companies who make the products that I already love and would already talk about for free and have

00:03:43.220 them pass savings on to you. Again, the podcast will remain free to all, but my hope is that many of

00:03:51.000 you will find enough value in one, the podcast itself, and two, the additional content exclusive

00:03:57.900 for members to support us at a level that makes sense for you. I want to thank you for taking a moment

00:04:02.960 to listen to this. If you learn from and find value in the content I produce, please consider

00:04:08.480 supporting us directly by signing up for a monthly subscription. My guest this week is Kevin Sayre.

00:04:14.140 Kevin is the executive chairman of the board of directors and the president and CEO of Dexcom.

00:04:18.900 During his tenure with Dexcom, he's been instrumental in leading the development of new technology,

00:04:23.260 including the G5, G4, and of course now the G6. If you've followed me for some time,

00:04:29.440 you probably know how much I love my Dexcom glucose monitor. I consider it to be one of the

00:04:34.520 two most important devices that I wear, and I wear it pretty much all the time. I would say

00:04:39.920 300 to 330 days a year I am wearing my CGM. It's offered me great insight, not just into the obvious

00:04:47.920 things like how does eating food X impact my glucose, although I still continue to learn a lot

00:04:55.440 through that lens, but also the impact of exercise, sleep, and stress on mitigating those things or

00:05:02.160 amplifying those things. For example, one of the things I've noticed much more recently is the role

00:05:09.260 that cortisol in the evening plays on my glucose. In this interview with Kevin, we talk about a bunch

00:05:15.280 of things. We talk about how we met and his path to becoming the CEO of Dexcom. We talk about how the

00:05:19.640 CGM technology works and what makes the G6 what I conceive to be is the best model yet. We talk about

00:05:26.240 the challenges of working in the US healthcare system after getting CGM insured. We talk about

00:05:30.660 developing an even smaller, more user-friendly, less costly product while maintaining performance

00:05:35.700 and accuracy. Kevin does a really good job of explaining his framework on what are the

00:05:41.880 constitutive elements that would drive a CGM device. You'll see where this performance and

00:05:47.100 accuracy become important. We talk about what's the next thing on horizon for patients with type 1

00:05:51.160 diabetes, the accuracy of the G6, software improvements for people with type 2 diabetes,

00:05:56.460 and the remarkable benefits of real-time feedback that drive behavioral change for anybody using this.

00:06:01.660 Talk a little bit about Dexcom versus the competitors. There's basically three large companies

00:06:06.220 that do this, Dexcom being one, Abbott being the other through the acquisition of Libre, and Medtronic.

00:06:12.000 And we talk about less invasive options that people have asked about a lot. Talk about the cost,

00:06:16.920 why you need a prescription, when there might be an OTC, that's over-the-counter option.

00:06:21.960 We talk about smartwatch integration, Bluetooth, exciting collaborations and partnerships,

00:06:26.660 and future places for CGM, such as in the hospital. We talk about lessons learned through failures

00:06:30.880 and successes. Before we get to this, I want to read a disclaimer that Kevin's team has asked me to

00:06:35.340 read, and that's of course because he's the CEO of a publicly traded company. This presentation

00:06:38.720 includes forward-looking statements. All forward-looking statements included in this presentation

00:06:43.040 are made as of September 21st, 2018, that was the date of the interview, based on information

00:06:48.060 currently available to Dexcom and are subject to various risks and uncertainties, and the actual

00:06:53.180 results could differ. Dexcom undertakes no obligation to update such statements. With that said,

00:06:58.340 and without further delay, please enjoy my conversation with Kevin Sayre.

00:07:01.600 Kevin, how are you? I'm great. Thanks so much for making time. I don't often come into somebody's

00:07:10.200 office, take them away from their work, and ask them to sit down and talk. You know, our discussions

00:07:15.160 have been so interesting over time. I was thrilled when you asked me to do this. I can't imagine what

00:07:20.240 journey will go on as we talk for the next while, but I'm really looking forward to it. So speaking of

00:07:25.280 which, I think I've told the story a couple of times to some friends, how we met, but it's sort of

00:07:29.460 funny to me because anyone who knows me knows I'm, when I'm on an airplane, I'm there to mostly just

00:07:34.660 work, and I don't really talk to people. Occasionally, if somebody talks to me, I'll

00:07:39.020 certainly be polite, but I'm really focused. And so, of course, it was about three years ago,

00:07:43.780 you and I met on an airplane. The funny part to me is how it happened, which is I'm taking a little

00:07:49.880 break from work to look at one of my 12 favorite watch sites, and you look over. Do you remember what

00:07:56.260 you said? Oh, I do. I took my headphones off. I said, are you sure you want one of those?

00:08:01.720 You know, and I'm the same as you, Peter. I get on a plane, and my wife claims I am the most

00:08:05.820 unfriendly individual in the world because I sit down, I wait for that moment when I can open up

00:08:11.360 my computer, I'm hooked to Wi-Fi, I start working, and I'm thrilled. Or I just take a break from life

00:08:16.820 and watch something on the screen. So we hadn't said a word, and all of a sudden, you open up to watches,

00:08:21.860 and I go, I've got to talk with this guy. We've got to look at these things and make sure he picks

00:08:26.660 the right one. It was a very fun meeting, and then neither of us worked the rest of the flight.

00:08:31.800 Yeah, but the best part is we talked about watches for an hour, and then we're about an hour outside

00:08:37.480 of New York. And at some point, we literally hadn't talked about anything but watches, but I just sort

00:08:43.300 of said, oh, are you heading home or heading to work? And you said, oh, you're going to work. And then

00:08:47.020 it was like, what do you do? And then this is the part I don't think I told you. I had only heard

00:08:51.800 about Dexcom two months earlier from a guy named Jake Kushner. I don't know if you know Jake, but

00:08:56.480 Jake is a pediatric endocrinologist. Yeah. And so I was having dinner in Houston with Jake a couple

00:09:00.960 months earlier, and he was showing me his data on his kids with type 1 diabetes that are walking

00:09:06.420 around with hemoglobin A1Cs of 5.6. And he's saying to me, it's a game changer. We're done with

00:09:12.880 telling kids with T1D that 7.5 is normal. The new normal is 5.6, and we're having no hypo

00:09:18.700 events. And he goes, but we can't do it without this thing called CGM. And I was like, tell me

00:09:25.200 more. And he said, so we use this thing. I think it was the G5 that they were using at the time. And

00:09:29.280 I remember sort of thinking about it and going, God, that sounds so interesting. That's on my list

00:09:34.100 of things I need to learn more about. And then I ask you what you do, and this is what you – they

00:09:39.540 couldn't have made it up. No. And this is all we do here at Dexcom. And we do see remarkable

00:09:44.960 results in type 1 diabetes with these kids and with everybody who uses the technology.

00:09:50.360 Well, I've been wearing one now for almost three years. I think after that flight, you said, well,

00:09:54.040 you got to come in one day and we'll shoot the breeze. And within like a week, I was wearing one,

00:09:58.420 and I've never stopped. Well, and I also use you as a critique because if we give Peter technology,

00:10:03.780 you're a very, very good critic and very objective versus our diabetes patients who,

00:10:10.040 while they love the technology, it's life-saving for them. For you, you kind of look at it from

00:10:14.460 a different perspective. And I use observations like that all the time when I talk to engineers.

00:10:20.040 And as we design future products, I love different perspectives on things. And

00:10:23.960 one of the things in our culture that is very unique, we are very curious as an executive team

00:10:30.580 by nature. And curious is the word that I would use. We love to learn, explore. We all wear these

00:10:36.280 things. We all wear the new ones. In fact, I'm supposed to get a prototype of a new one next week

00:10:41.340 that's going to communicate with some other devices. It's really fun. And it does allow for

00:10:47.400 learning. It keeps your mind going very quickly. We put Peter on our G6 technology right as it was

00:10:53.220 approved. Yeah. Well, actually, I mean, I was on that G5X about a year ago, which was almost a G6 minus

00:10:59.580 the calibration issue. That's right. So now let's go back a little bit. What were you doing

00:11:03.840 before Dexcom? So if you go back in the history of my career, I started in diabetes in 1994. I met

00:11:11.820 an individual named Al Mann, who's one of the most remarkable entrepreneurs ever. And he had designed

00:11:18.040 the insulin pump that is now Medtronic Diabetes. And we had a glucose sensor there. That's where I saw

00:11:24.300 the first one. And I worked in diabetes there as the finance person and very much a business person

00:11:30.380 till we sold the company to Medtronic. I stayed on with Medtronic for a while. And we launched the

00:11:35.380 first continuous glucose monitor at MiniMed and Medtronic back in those days. And then I decided

00:11:40.800 I didn't love big company. So I left and I found a couple of other jobs that were interesting, but I

00:11:48.100 didn't love them. My job right before this company was culturally fascinating. It was a Singapore

00:11:52.640 headquartered Singapore public company, and I was head of the U.S. operation and the chief financial

00:11:59.400 officer of the entire company. But we had operations in China, Japan, and all over Europe. At the same

00:12:05.460 time, I joined the Dexcom board. My dear friend and associate from MiniMed, Terry Gregg, was the CEO

00:12:11.800 here. And as we wrapped our other company down, as it became obvious we weren't going to be a U.S.

00:12:17.760 presence, the opportunity came to move to San Diego and go to work at Dexcom. So I went home after our

00:12:25.740 board meeting in December of 2010 and told my wife, I think we're leaving our dream house here in North

00:12:31.780 L.A. We're going to go have another adventure. And we moved down here in June of 2011, and it's been

00:12:38.700 just a tremendous ride ever since. Now, between the time you spent at Medtronic, where you were a

00:12:44.240 finance guy, but obviously, as the listener is going to learn soon, you have pretty good technical chops

00:12:49.260 on this stuff. But you were working on CGM then, and you probably just by osmosis absorbed a lot of

00:12:54.560 it. But between what sounds like the mid to late 90s, and then your return here via Dexcom, were you

00:13:00.860 still in a CGM space? Nope. My position right after Medtronic was with a specialty laboratory

00:13:07.040 that ran esoteric lab tests and did that for a little while. And my company right before here,

00:13:14.080 the Singapore-based company, was a drug-eluting stent business. And we had a proprietary drug

00:13:20.160 formulation and a proprietary polymer that was bioabsorbable. So when the drug-eluting stent

00:13:26.800 was expanded into the vessel, both the drug and the polymer would absorb for better safety rates.

00:13:32.260 But getting a stent approved in the United States was too big of a deal for a company as small as

00:13:37.960 ours. Well, especially when Abbott, Medtronic, J&J, and Boston Scientific, the four of them own the

00:13:43.880 entire market. The law of four. Medical device companies, you see technologies, you start with

00:13:49.380 a couple, they swell up to 10 or 11, and then you go back down to three or four.

00:13:54.360 So this is amazing. I didn't even know this about you. So once we're done with this podcast,

00:13:57.600 when we go grab coffee, we'll talk about my previous life where I had to know everything

00:14:01.960 about drug-eluting stents. I spent two months interviewing 150 interventional cardiologists

00:14:08.120 around the country when I was at McKinsey, trying to learn everything about this for our client,

00:14:12.960 which was obviously, I can't say who, but it was one of those four. And it was at the reintroduction

00:14:17.740 because remember there was a period of time when they had to go back to bare metal and then they

00:14:22.080 had to reintroduce. So fascinating topic. So basically you came back to this company and

00:14:27.860 what I assume you were bringing that was of most value was you understood medical devices,

00:14:32.380 you understand durable medical goods, you understood the law of healthcare. It wasn't that

00:14:38.300 you were the world's expert on CGM.

00:14:40.480 No, I'd been on the board and I really will never be the world's CGM technical expert. These guys here

00:14:45.920 are spectacular, but what I did bring, or at least I believe is very much that curious factor I talked

00:14:53.740 about and a very open and aggressive mind to grow our business. This company has gone through tremendous

00:14:59.000 change. I came in for a couple of reasons. A, Terry is exactly 10 years older than me and I was supposed

00:15:05.240 to succeed him at Medtronic and that one didn't work out the way we wanted to. And when I joined the

00:15:10.760 board, that was something we'd talked about early on. And I know, knew all these guys, I knew the

00:15:16.600 technology very well. I knew how bright they were. And I just thought it would be a tremendous

00:15:20.680 opportunity. Let's take another step back just for folks to give them a basic, because there's so many

00:15:26.580 questions I get asked about this. In fact, I, I went out to Twitter a week ago and said, you know,

00:15:30.920 I was going to be sitting down with you and hey folks, if you have questions, shoot them over. And I was

00:15:34.740 really amazed at how many people ask questions. And there are a lot of questions that I could answer

00:15:39.480 truthfully, but there are some that I can't. So I'm going to try to assimilate the question.

00:15:43.120 So I think the first question that I get asked a lot is, how does this actually work? In other

00:15:47.700 words, why is it that even if you don't look at CGM, if you just talk about point of care glucose,

00:15:52.300 why is it that with one prick of blood, you can measure something like glucose, but you can't

00:15:57.700 measure a number of other things? What is it about measuring glucose that enables that?

00:16:01.780 Well, our technology is designed to measure glucose only. And the way our sensor works

00:16:06.300 is there is a small wire that's thinner literally than a human hair or as thin as a human hair

00:16:12.020 that's inserted subcutaneously into your tissue. That wire is, is very special metal alloys and it's

00:16:19.120 coated with a number of membranes. That wire and those membranes then generate an electrochemical

00:16:26.780 signal that goes up into a transmitter that sits on top of the sensor. You, you, you've worn it,

00:16:32.400 you know what it looks like. And inside that transmitter is an algorithm that converts that

00:16:36.980 electrochemical signal to a glucose value. So that is the only thing that we have done with

00:16:43.100 this technology. There probably are other analytes we could take and develop enzymatic layers that

00:16:50.540 would sense that type of chemistry. In fact, you and I have talked about some of the things that

00:16:54.500 might be meaningful and useful over time, but we've grown so quickly. We just, we haven't addressed

00:16:59.720 those. But I think the key is it's, it's an enzymatic reaction. And I think that's the point

00:17:03.940 I try to make to folks is if there's an antibody that can be linked to it, if there's an enzymatic

00:17:08.580 reaction that can be done without multiple washes, that's the kind of stuff you can measure in point

00:17:13.300 of care. Whereas many things that are complex, you can't, you know, you can't measure, you know,

00:17:17.960 the number of lipoproteins or something like that because that requires NMR. It's a very separate

00:17:22.040 test. You can't even measure insulin. You're either going to do it through a radioimmune assay or

00:17:25.740 you're going to do it through multiple washes, something called an ELISA-based assay where you

00:17:29.940 have to rinse and repeat and stuff like that. So out of the gate, there are some things that can

00:17:34.700 be measured in a drop of blood, which again is different from what you guys are doing. And,

00:17:38.640 you know, glucose is one of them, electrolytes, lactate, beta-hydroxybutyrate, which you and I

00:17:42.680 have talked about a lot. So, I mean, I guess let's go back and so how deep is that, that little

00:17:48.560 filament that sits inside the patient? It goes in at a, at an angle. Now it sits in probably a half

00:17:54.220 inch under the skin. The insertion technology of the sensor, it goes in with a needle and the needle 0.71

00:17:59.880 comes out. There is so many engineering hours into the insertion of that sensor. And one of

00:18:06.660 the first things you notice about the Gen 6 product, it's an automated insertion. The needle is in your

00:18:11.780 skin, as the engineers tell me, less time than it takes a hummingbird to flap its wings a couple of

00:18:18.040 times. So it's in and out very quickly. So a patient doesn't feel it. And then the sensor rests in

00:18:23.680 the subcutaneous tissue. And you really don't feel that sensor at all.

00:18:28.060 I think I've commented to you probably the very first time I tried this, which was on the prototype

00:18:32.340 that with the G5, of course, you are the one as the patient who is responsible for the velocity

00:18:38.260 with which that needle goes in. And if you're a human being, which we all are, I think we're just

00:18:42.780 hardwired to do it a little slower than we'd like because it's so, you know, it's a little

00:18:46.900 uncomfortable. Look, it wasn't the most uncomfortable thing in the world, but it's...

00:18:49.980 If you're doing it, it's scary.

00:18:51.480 Yeah, yeah.

00:18:51.900 It is not an appealing thing. You look at that, our old device, which was very functional,

00:18:56.600 but you look at it and it truly looks like, I don't know if I want it. And we've had patients

00:19:02.500 send emails saying, I've stared at this thing for six weeks and I don't dare put it in my body.

00:19:08.040 We're not having that reaction with the new one.

00:19:10.260 Yeah. And I had that reaction with a number of my patients because as you know, a lot of the

00:19:14.860 patients in my practice, though they don't have type one or type two diabetes, like me,

00:19:19.420 they understand the value of even people with, you know, quote unquote, normal glucose levels

00:19:23.960 being able to track this. But occasionally folks would freak out and they'd say, you know,

00:19:28.420 we'd go through the, all the process of getting them a three month supply and they, you know,

00:19:33.180 they've already paid for it because of course their insurance isn't going to cover it. They don't

00:19:35.920 have diabetes, but then that needle just sort of scares them away. Of course, the other thing that

00:19:41.240 we speculate, or maybe it's not a speculation, maybe it's just a given is the lack of trauma

00:19:46.420 that's involved with the G6 insertion seems to make for a much more accurate device, which the G5

00:19:53.220 itself was remarkably accurate, but I don't know if it's anecdotal, but in my hands, at least the way

00:19:58.360 I track it, the G6 seems to be so accurate. It's almost hard to believe. There's so many elements

00:20:05.080 that go into the performance of a sensor. I believe the insertion process does make it more accurate,

00:20:10.000 but that G6 sensor has new membrane technology versus the old one. There's also a new algorithm.

00:20:17.800 In fact, one of the things that has made our company successful during this G6 project in

00:20:23.420 the middle of it, our algorithm engineers came and they said, we have a better way to do this,

00:20:28.560 but it's going to delay the project if we do it. We looked around the room and said,

00:20:33.260 but if it's going to be that much better, let's delay it and put it in. And so we did. And we made

00:20:39.080 that decision. That algorithm has helped tremendously with the accuracy. So all those

00:20:44.100 things together, the easier insertion, the new membrane technology, the new algorithm. I think

00:20:50.640 even the fact that it's flatter on the body and doesn't pull as much as is helpful as well.

00:20:55.180 I think on three occasions over two plus years, my kids kicked out my G5, you know, just, and

00:21:04.800 including once on the very first day I put it in, which really bummed me out when I picked up my son

00:21:09.660 and he obviously just not playing around. He horsing and he kicked me and the thing came out.

00:21:14.040 Whereas the G6, yeah, it's a lower form factor. And obviously this is only going to continue as the

00:21:20.640 evolution of these things goes. So you talked about a couple of other things that could be measured

00:21:25.440 besides glucose. Are there any that you think are commercially viable? There's lots that could be

00:21:31.860 interesting, but do you think there's anything else that's commercially viable?

00:21:34.620 We don't know yet, but we have so many markets to take our technology to now that we need to

00:21:42.440 capitalize on this opportunity and take care of the people that we serve and open the other markets

00:21:47.660 where glucose can just be a wonderful tool that we really haven't done that much of yet.

00:21:53.420 So approximately how many patients in the United States have type one diabetes or what

00:21:58.080 percentage of the population? How do you think about it?

00:21:59.540 About a million. Anything you read in the literature says anywhere from 1.3 to 2 million

00:22:05.760 in the United States. A common number is 1.5. And there are another 1.5 million people

00:22:12.000 with type 2 diabetes who intensively use insulin. So they are really in need of a continuous glucose

00:22:18.840 sensor to manage their health.

00:22:20.680 Now going back to those folks with type 1 diabetes, what percentage of them use one of the commercially

00:22:27.540 available continuous glucose monitors?

00:22:30.000 Most would estimate somewhere between 20% and 30% right now. And again, these percentages have

00:22:35.200 grown rapidly over the past couple of years as we've gotten better and as other products have

00:22:39.280 come to the market.

00:22:40.580 So you and your competitors, and I guess your main competitor is Abbott?

00:22:44.580 Abbott and Medtronic.

00:22:45.880 Yes.

00:22:46.060 The three of you collectively aren't even remotely close to saturating the most obvious market

00:22:51.540 for this product, which is type 1 diabetes.

00:22:53.460 That's right.

00:22:54.520 Again, hard to speak speculatively about what others think, but I assume everybody's also

00:22:59.280 looking at that second group you just mentioned, which is another market, which is the insulin

00:23:03.480 using patients with type 2 diabetes.

00:23:05.720 Yeah, in fact, when Medicare and CMS gave us approval in January of 2017, they approved

00:23:12.840 the use of CGM for intensive insulin using type 2 patients. And that was a step that we

00:23:18.560 really needed. One of the barriers, and it's not really a barrier, but it's very interesting

00:23:24.540 as you grow a business. And one of the fascinating things about our journey here, it makes such

00:23:30.300 common sense. Everybody thinks, well, everybody should just have this. And the insurance companies

00:23:35.660 were very well insured and very well covered, but it's not always easy to get. I was in

00:23:41.320 a state not too long ago, for example, where a patient had to document a certain number of

00:23:48.000 events where their blood sugars went below 50 before they can get a Dexcom, even if they

00:23:53.720 have type 1 diabetes.

00:23:54.840 Which is very dangerous for a type 1 diabetic.

00:23:57.340 Very dangerous for anybody. Yes. And yet, because of fear for spending money on devices and

00:24:04.580 treatments that aren't useful, payers are hesitant sometimes. Not all of them. We get

00:24:10.300 very, very good coverage, but it is cyclical. You know, a payer will cover us very well for

00:24:16.240 a two-year period and they'll look and, man, we've added so many people to this technology.

00:24:20.180 What are we doing? Let's slow it down. So they'll slow it down through criteria.

00:24:24.240 We have a full-on sales team that does nothing but call on those who reimburse for the product and

00:24:31.920 we try and make it as easy as we can. There are trade-offs for price and access and we try and

00:24:38.200 negotiate those things on a regular basis. I believe over time, at least 80% of the people

00:24:44.120 with type 1, and I've said this in public forums, 80% of the type 1 intensive population should be

00:24:49.660 using continuous glucose monitoring. Has there been an economic analysis done

00:24:54.580 ideally independently, but it might be done obviously by one of the companies or more,

00:24:59.880 that does the equivalent of what we would think of as an NNT for a drug trial? So for the listener,

00:25:04.720 an NNT is a number needed to treat. This is an analysis that's not done economically. This is

00:25:10.060 more of a life-saving. So how many patients do you need to treat with a certain drug to save a life?

00:25:14.300 But you can also start to do these things economically. And is there a question of how

00:25:19.240 many patients need to be on a CGM before you either save a life through a hypoglycemic event

00:25:24.740 or maybe a different analysis would be the cost of the CGM versus the cost of the complications

00:25:31.500 from mismanagement? So has any of that been done? We have done some economic models and studies in

00:25:37.380 that area. What we find is the major cost that CGM prevents is exactly what you talked about in the

00:25:43.280 acute setting. They prevent hospitalizations. So you end up with a specific set of patients who are

00:25:49.740 what are often called hypoglycemia unaware, who aren't aware when their blood sugars go low. The

00:25:54.940 CGM gives them alerts and alarms. And our system also alerts and alarms others who give care to those

00:26:01.000 patients to whereby they can be taken care of. That share feature is really one of the most

00:26:06.960 remarkable things we've done here. You could be in Singapore and your wife could be here if

00:26:13.160 you had type 1 diabetes and your blood sugars went low. Your wife would know. And she could call the

00:26:18.080 hotel and wake you up. And we hear stories like that all the time. Does Medtronic's product or does

00:26:22.820 Abbott's product do that? I believe Medtronic's new product does. I believe they do have some sharing

00:26:27.820 capabilities. But I know we pioneered it. We're the ones who got it through the FDA. And it took a lot

00:26:33.880 of effort. So that sharing capability is really good for hypoglycemia unaware. That category, one

00:26:40.820 hospitalization, I've seen numbers ranging from $10,000 to $30,000 for that one hospitalization. But

00:26:47.540 not everybody's hospitalized. We certainly save on the long-term complications of diabetes. And that has

00:26:53.220 been measured historically. And you and I have talked about this with A1C, which neither of us believe is

00:26:58.960 the perfect measure. But in our studies, we see at least a half a point drop over the course of a 12-month

00:27:05.780 period compared to multiple finger sticks. A full point drop for the patients, if you just look at

00:27:10.800 them independently, on continuous glucose monitoring use for a six-month period that is sustained. That

00:27:16.980 drop in A1C leads to reduced costs in the healthcare system. And there are very, you know, well-known

00:27:23.320 economic models to build that out. But here's where the rubber meets the road. A 25-year-old

00:27:28.800 probably isn't generating all those complications yet. And a 25-year-old is with an insurance plan for

00:27:33.920 two years before they move to their next job. And they're on another insurance plan for two years.

00:27:39.940 So we have to do a better job as a company documenting these economic benefits. We run more studies

00:27:45.420 to see how, you know, how much quality of lives improve, to see what we can prevent. We are currently

00:27:53.220 building outcome-type based models to whereby we can, look, we can go into an insurance company or a

00:27:59.820 payer and say, look, here's what we think we can save you if we can put everybody on CGM. But we're

00:28:05.340 creating a new industry here. It's a completely new business. So every day is a challenge like that.

00:28:11.940 And every day there's more thoughts and more things to address along those lines. The Medicare one is the

00:28:18.140 most interesting. I spoke at a conference on a January Wednesday where I told a group of our

00:28:24.760 investors we would not have Medicare approval for 18 months. And we got it Thursday.

00:28:31.280 That makes you look not terribly well informed.

00:28:34.260 That is correct. And the way it was approved, our device talks to a phone or a dedicated Dexcom

00:28:40.620 receiver. The way it was approved for the Medicare patients is they couldn't use the phone app.

00:28:45.620 Yet our Medicare patients are the ones most likely, other than children, to need to share the data with

00:28:54.300 somebody else. It took us, I want to say, 18 months. We finally got the phone use approved for

00:29:01.220 the Medicare patients. When you're creating an industry like that, there are all sorts of hurdles

00:29:06.160 that you run into that you don't contemplate. You just kind of have to react.

00:29:11.340 I mean, I think that's one of the challenges that folks face in this industry in general,

00:29:15.060 which is the policy piece of it is hard because you don't exactly know from a long-term planning

00:29:21.580 standpoint exactly what changes are going to come about. The challenge that you raised,

00:29:25.980 I think is, I'm glad you brought it up because I think it's one of the most misunderstood issues

00:29:31.820 in the US healthcare system. I grew up in Canada and I'm actually not incredibly that impressed with

00:29:37.120 the Canadian healthcare system. I know that's going to make me a lot of enemies out there,

00:29:40.000 but I don't think it's like a panacea. I don't think necessarily Canada is doing it right either.

00:29:43.400 But there's one thing that I think a single-payer system, at least one thing a single-payer system

00:29:48.560 brings over a system that we have in the United States, which is the incentive is aligned between

00:29:54.580 the payer, the patient, the provider over a long-term. It's easier to align incentives because

00:30:01.600 one way or the other, that payer is going to pay. It's pay now or pay later.

00:30:07.060 Yeah, that's right.

00:30:07.620 Which by the way, we were talking about children before we went on the mic and you had great piece

00:30:13.080 of advice, which I'm going to ask of you to share with others. But my brother's advice, which is,

00:30:17.460 my brother has five kids. His piece of advice is pay now or pay later. Pay now or pay later.

00:30:23.080 You can either discipline your kid when they're five or you're going to discipline them when they're

00:30:26.240 older. But this thing with healthcare is if you have that 25-year-old and the difference between

00:30:32.820 them having a hemoglobin A1C of 8% versus 6%, in the next year, it's nothing. It doesn't matter.

00:30:39.280 But in 20 years, it matters beyond words.

00:30:42.220 That's right.

00:30:42.400 And if you as the payer are on the hook for that one way or the other,

00:30:45.780 the economic model becomes so much more straightforward because you don't have the

00:30:50.380 portability of healthcare.

00:30:51.580 That's right.

00:30:52.780 Again, I'm not close enough to this to know if a single-payer system is even something that's

00:30:56.780 viable in the US. It strikes me as politically very unviable, but who knows? That said,

00:31:01.900 it's not what you're dealing with now. So you have a harder challenge to make the case

00:31:05.720 economically, I suppose.

00:31:07.240 No, we do. But again, we're covered by 99, literally 99% of all reimbursement authorities

00:31:13.960 where our coverage is a little spotty still is in the Medicaid programs. We've got several states,

00:31:19.040 but it's still, there's a lot of documentation required. We're covered where it gets difficult

00:31:24.840 is the documentation the patients have to provide or the physicians have to provide

00:31:28.760 and consistency of co-pays and things of that nature. We're also classified as durable medical

00:31:33.840 equipment. So our patients have a co-pay and a deductible, which leads to a very seasonal

00:31:38.920 business for us because in the fourth quarter, most of our patients are maxed out on their

00:31:44.360 deductible and co-pays and they want to buy everything they possibly can.

00:31:47.760 It's like Christmas.

00:31:48.460 It's like Christmas. So from October to the end of December here, it is all hands on deck.

00:31:52.980 That's interesting. What percentage of your patients, I'm guessing it's very small,

00:31:57.840 but do what me and my patients do, which is it's a cash pay business. There's no dealing

00:32:02.520 with insurance. It's directly from, you know, physician writes the prescription and the patient

00:32:05.920 comes straight to Dexcom.

00:32:07.620 Not very many. Again, like you do, we have cash pay patients and everybody in your system uses

00:32:13.100 prescription. It is important for the regulatory folks that I point that out. Not as many as we used

00:32:17.640 to. We used to have a lot of Medicare patients who paid cash.

00:32:19.920 I see.

00:32:20.840 But once Medicare covered, we actually had to refuse the Medicare cash patients

00:32:25.080 because if we took cash Medicare while Medicare covered a device, we were in trouble. And it took

00:32:30.700 them seven or eight months to figure out how we could ship and process this new product code. So we

00:32:37.060 had upset patients for, I don't know, somewhere between six and eight months before we could start

00:32:42.100 serving them properly.

00:32:44.840 The people who don't live in the United States that are listening to this are scratching their heads

00:32:48.600 thinking, what is going on over there? It is a, we have a really complicated healthcare system.

00:32:54.580 Yeah, we do.

00:32:55.420 But look, there are some things that I think it does well. And I think, you know, one of the things

00:32:58.520 it does well is it seems to set up an economic engine that allows for an innovation that I think

00:33:03.000 is hard to do outside of the United States.

00:33:05.020 Oh, I'd agree with that.

00:33:06.360 And I don't think it's a small accident that the disproportionate share of the innovations come

00:33:11.600 out of the U.S. So, you know, kind of going back to some of the technical stuff, you have a

00:33:17.620 partnership, you have a collaboration with Verily.

00:33:19.560 We do.

00:33:20.060 So tell people what is Verily, how does it have, what does that have to do with Google

00:33:22.940 and how did that collaboration come about and how does it work?

00:33:26.120 Well, the way it came about was interesting. Back in 2015, there was a healthcare group within

00:33:32.480 the walls of Google before Alphabet came around and they analyzed technologies where they felt

00:33:38.460 they could develop technology that would make an impact on future products. And they also analyzed

00:33:43.940 disease states where they felt they could use Google's or Google Healthcare's analytic

00:33:49.300 capabilities to develop better care models. Google has made a couple of attempts at healthcare and

00:33:55.620 they came down and met with us and said, look, we've looked at your product. We've looked at what

00:34:01.080 you do. We have a lot of electronics expertise. We have expertise in data analytics and a bunch of

00:34:06.260 other areas. We would like to partner with you. They also told us they were going to talk to other

00:34:10.460 glucose sensor companies as well. We looked at that and said, you know what, this Google is a big

00:34:17.500 company and they do a lot of things. And we looked at their concepts and their thoughts and they looked

00:34:25.300 at the problems different than we do. They look more at miniaturization and size and convenience

00:34:30.580 and the things that consumers look at. Because we as a company, again, serving our population of

00:34:36.920 people with type 1 diabetes, and anybody would tell you this, our core principle has been performance

00:34:42.200 and accuracy. Performance and accuracy, we always want to build the best. We were the first to be

00:34:47.420 connected. But as far as really getting the thing tiny or something that somebody without diabetes could

00:34:53.600 wear, that had not been a focus here that they could wear for fitness. So when they came down and

00:34:58.960 presented this to us, we decided that as an electronics and design partner that we would

00:35:04.780 like to work with them. So after several months of negotiations, we, not several months, a couple

00:35:10.460 of months, we signed up a deal and we're co-developing future products for Dexcom. Those products will be

00:35:16.140 Dexcom sensors. Those membranes and those algorithms that we've developed to make our sensor accurate

00:35:21.220 and perform very well will come from our core technology that we've always developed.

00:35:25.980 The electronic side and some of the communication and the Bluetooth and those other things have been

00:35:31.840 driven by the Verily engineers. That's where they have great areas of expertise. The overall package

00:35:37.700 and mechanical part of the system, we've done together. We've talked about this, Peter. We focus

00:35:42.740 on performance, patient convenience and usability, and then taking cost out of it. We kept all three of

00:35:48.960 those in mind as we did that.

00:35:50.760 Yeah, it seems to me, and I like your framework, right? So let me see if I remember what you said. It's

00:35:54.260 going to be performance, which is how accurate is this thing? End of the day, that's the single

00:35:58.480 most important thing. If it doesn't perform, who cares what else it does? Then the patient

00:36:02.680 experience. How big is it? How easy is it to use? How often do you have to change it? Do you have

00:36:06.580 to separate the sensor from the transmitter? You know, all those things.

00:36:08.600 What does the software look like?

00:36:09.860 Yeah, yeah. And then cost.

00:36:11.800 And then cost.

00:36:12.720 It seems to me, having been fiddling around with this thing for a while, that the step between G5 and G6,

00:36:20.240 you're now on the asymptotic part of the performance curve. There's not a lot of performance

00:36:24.480 gain at this point, is there?

00:36:26.500 There is some performance to gain. Boy, we spend a lot of time studying this and talking about this.

00:36:32.620 As far as just bringing the overall accuracy percentage down, there's not a whole lot to gain.

00:36:37.580 Where there is to gain on the performance side is bringing that curve of performance. You have a

00:36:44.440 bell curve of performance for every sensor, pushing that bell curve in to whereby every experience is the

00:36:49.720 same. Most of them are the same now, and we believe it's very consistent. But that consistency and

00:36:55.160 reliability enables us to do two things. Number one, it enables us to do very sophisticated things

00:37:00.860 with people who want to control their insulin dosage with our sensor, whether through an

00:37:05.900 artificial pancreas system or on their own using our data. But the other thing it enables us to do,

00:37:11.220 if that performs the same all the time, it will enable us to go out to broader markets.

00:37:14.740 Because people won't question or challenge the data. They will see it and go, okay, this is right.

00:37:20.400 This one works. We call it internally eliminating the outliers. And I know every company does this

00:37:26.340 with what they build. And so performance-wise, that's where we're headed there. And those are

00:37:31.540 the things that we look at in our meetings all the time.

00:37:34.180 You know, I don't even know if I told you the story. It's kind of funny. About a year and a half

00:37:37.900 ago, a mutual friend. So it's a friend that I introduced to you, but I won't call him out.

00:37:41.540 I'll keep his identity private. He introduced me to Andy Conrad at Verily. But it was through

00:37:46.440 some completely unrelated, there was a different reason for us to meet. So I'm up there. And I didn't,

00:37:50.840 I, at the time, I just, I either didn't know that you guys had worked with Verily or I just

00:37:54.560 hadn't put two and two together. So I'm sitting in his office and I see these mock-ups of things

00:37:59.380 that look like next-gen CGMs. And I said, Andy, what are those? And he goes, oh yeah,

00:38:04.640 we work with this company called Dexcom. And of course I lifted up my shirt and showed him my,

00:38:08.700 it was at the time it was a G5. And oh my God. I mean, it was, I felt like Dorothy. This was like

00:38:14.500 Oz. I couldn't believe what the future could hold. So going back then from a technical standpoint,

00:38:21.060 because you alluded to what the future could hold, right now you are measuring glucose in

00:38:26.900 subcutaneous fluid. Correct? Yes, that's correct. Not in blood.

00:38:30.220 Not in blood. People need to understand that. So even a relatively lean individual has enough

00:38:34.480 subcutaneous tissue that the risk of using this thing, it's like, is it an issue in kids? Because

00:38:39.840 I've seen some really tiny little kids with type one diabetes. Sometimes it is. And parents learn how

00:38:44.920 to make sure they insert it and that the needle doesn't go into the muscle tissue. There will be an

00:38:50.820 occasional instance where a sensor may go too far in the tissue. And, uh, but by and large,

00:38:56.620 no, it's very successful. There's a meeting in Florida every year, a group called children with

00:39:01.560 diabetes. And look at all the kids who wear this. It's not an issue for children. They do extremely

00:39:07.320 well with it. You know, and I think I'm, I don't know if I ever told you the story, but I was at a

00:39:10.820 swim meet. It was my daughter's swim meet. I had just finished my swim workout at the same pool. So I'm in

00:39:15.360 my bathing suit, finishing my workout as the kids are about to start the meet. And this kid comes

00:39:21.480 over to me and his eyes light up like saucers. Cause he's, he sees that I have a sensor on and

00:39:26.100 he's got a sensor on and he is tiny. He is the littlest kid I've ever seen. And he was like,

00:39:32.400 do you have type one diabetes? And I said, no, why? And he, and he looked, I didn't even realize

00:39:37.840 that what he was looking at. And I said, uh, he said, cause you're wearing the same thing I'm

00:39:40.800 wearing. And I said, Oh, Hey little buddy, let me tell you, I volunteer to wear this thing.

00:39:44.900 That's how cool it is to wear. Oh, that's great. Yeah. He loved it. He loved that there was some

00:39:49.000 adult who was wearing this, who didn't quote unquote have to wear it. But I was like, yeah,

00:39:52.580 we're the coolest guys here. I assure you. One of the most humbling things about this job

00:39:57.860 is when I go out to a diabetes function and meet the children or the adult patients, whoever, who

00:40:05.740 has had this device save their life or change their life the way that it has. They want their

00:40:11.760 picture taken with me. I said, how about we just take care? No, we want you. Nobody wants, 0.60

00:40:16.840 needs their picture with me. No, they want to take a picture. They are so gracious. They are so,

00:40:24.240 I know that they're so indebted to our company and it's really why we come to work every day.

00:40:28.380 What keeps us going? We had my favorite meeting ever. An 11 year old young man comes in,

00:40:34.200 we sit down and people come to San Diego on vacation and they come to Dexcom.

00:40:40.440 So we got Legoland, SeaWorld, Dexcom. And parents will bring their kids in and if

00:40:46.380 whichever the executive team is here, we'll take them on a tour of the factory and then we'll sit

00:40:49.960 in a conference room with them because that is the best market research in the world to sit and

00:40:54.160 talk to somebody who uses this each and every day. This 11 year old young man said, I have some

00:40:59.520 questions. And I said, okay. Tougher than any analyst. He pulls like a legal pad out of his

00:41:05.280 notebook, out of his backpack and he has 40 questions. He wanted to know the chemical composition

00:41:11.160 of the membrane 11. And we just looked at him and go, no, we're not going to tell you all those

00:41:17.640 things. But he had thoughtfully written out all these questions and then had several suggestions for

00:41:23.980 us. It's just amazing. And your story, it happens all the time. Yeah. It's, it's really special. And

00:41:30.340 I, you know, I do have a friend with type one diabetes and she did say something to me that I've

00:41:34.520 thought a lot about since she said, you know, Peter, you have the luxury of, if you don't want

00:41:39.880 to wear that sensor for a couple of weeks, you don't have to wear that sensor. She goes, I don't

00:41:43.320 have that luxury. Like, and again, she wasn't saying that to be critical of how much I loved using it.

00:41:47.980 She was just saying like, understand that my relationship with this CGM is different from yours. This is my

00:41:53.740 lifeline. For you, this is just a gain in your health, but it's not necessary for you to,

00:41:59.800 for you to function. And I, I was sort of humbled by that a little bit. And I realized, you know,

00:42:03.980 I can understand why someone who's got type one diabetes has sort of a, a different relationship

00:42:09.300 with this on both extremes. On both extremes. It, it is love a lot of time. It's sometimes hate.

00:42:14.780 And the hate I think is the, I have to do this. I have, I have to have this thing on me.

00:42:19.040 I had two close friends in residency who had type one diabetes and this was back in the wild west. I

00:42:25.360 mean, it was, I couldn't believe how archaic it was and how hard it was for them to manage

00:42:30.200 their diabetes. Cause you know, in a surgical residency, it's very difficult to predict when

00:42:36.940 you're going to eat. Oh yeah. And therefore it's very difficult to predict when you're going to use

00:42:41.680 insulin and which insulin you're going to use. And I remember one of my friends who's now a cardiac

00:42:46.040 surgeon in Minnesota, there would be times when he always had to have a thing of orange juice

00:42:51.460 and candies in the OR next to his pager. And you know, he would hopefully get to the point where,

00:42:57.740 you know, he could say to a nurse while he's operating, Hey, I need some orange juice quickly.

00:43:02.040 And she would bring over the orange juice and tuck the straw behind the mask and let him

00:43:05.980 sip some orange juice or stick a candy in his mouth or something like that. If he was, you know,

00:43:10.280 this was an operation that was supposed to be done in three hours and now it's five hours,

00:43:13.260 that kind of thing. We take for granted in many ways, like what this disease meant before banting.

00:43:18.380 Oh, I don't know how people did it. Well, they did it. They managed and people have lived long,

00:43:25.300 healthy lives, but this tool is so much better than what we had before. And the things that can

00:43:31.140 come in the future, we made an acquisition a couple of weeks ago. The next thing on the horizon for type

00:43:37.120 one diabetes, let me take a step back is to develop algorithms to go with the sensor. Because right now

00:43:42.340 we tell you your glucose value, we tell you your trend, we tell you how fast you're going up or

00:43:46.900 going down. We will give you alerts and alarms based on what's going on, but we don't tell you

00:43:52.780 what to do. And algorithms have been developed for automated insulin delivery that will automatically

00:44:00.660 do things for you. One product that we're partnered with is Tandem's new insulin pump and

00:44:05.780 it shuts off when your glucose goes too low and then turns back on based on sensor signals.

00:44:11.180 And the early read on that product from consumers has been wonderful from patients. But these

00:44:16.500 algorithms are getting more and more sophisticated.

00:44:18.740 Explain for the listener where that product sits. How does it actually work?

00:44:22.540 So there's an insulin pump that is on your body. It's about the size of a good old fashioned

00:44:28.120 pager that you used to carry around in your hand. And there's an insulin cartridge inside.

00:44:32.900 That pump contains software that regulates the delivery of insulin through that insulin cartridge

00:44:39.620 to a small infusion set that is usually placed on your abdomen, a little Teflon cannula that goes

00:44:45.600 under your skin. And the sensor resides another place on your body. And the sensor signal is read by that

00:44:51.740 pump, which then takes all the information about insulin that it knows how much has been delivered,

00:44:58.260 your glucose value, and will determine, at least for this system, if your blood sugars go too low,

00:45:04.860 it turns off insulin delivery for some period of time and then turns it back on when it sees the

00:45:10.680 trend that you're coming back up. Algorithms over time will do more than that. They can regulate your

00:45:16.760 insulin delivery all throughout the day, whether you're high or you're low, and detect things like

00:45:22.620 meals. I mean, this is very futuristic, but there are algorithms that will detect, okay, it's very apparent

00:45:28.580 that Kevin just ate. Let's put some insulin in here and see how it goes. There are algorithms, again, to

00:45:34.440 take care of you more during the nighttime. Because during the nighttime while you sleep, you're not

00:45:39.920 eating and you're not exercising. So there's not as many variables involved. So we acquired one of these

00:45:45.980 algorithms two or three weeks ago just because it's an asset we felt we needed to have and develop.

00:45:52.800 Not everybody uses an insulin pump. Not everyone will have access to that technology. It's not

00:45:57.640 inexpensive either. We think a lot of people will, but we also think we can take these algorithms and

00:46:03.220 apply them to people who use multiple daily injections as well to develop decision support.

00:46:08.600 You know, I always tell the engineers, I want the Staples button. I want the easy button. I want to

00:46:14.740 take a shot and I want to hit something and say, how much insulin do I take? Or I want at 10 o'clock,

00:46:19.640 I want to hit a button that says I'm about to go to bed. What do you predict is going to happen?

00:46:23.960 Those types of things are an engineer's dream, which is, again, why it's so fun to work here.

00:46:30.240 We have chemical engineers, software engineers, electrical engineers, mechanical engineers,

00:46:34.360 packaging, any engineer you could think of. Well, the algorithm guys, and along with the wonderful

00:46:40.400 asset we acquired, are going to look at developing those type of decision support tools to help make

00:46:45.340 this easier for patients to make good decisions. Now, the question becomes, how much benefit does that

00:46:50.640 add to their care? If they only go from 8A1c to 7.8, it's probably not worth it and nobody's going

00:46:57.400 to pay for that. But we think we can build a very strong outcome case over time.

00:47:02.180 But even if you went from 8 to 8, but reduced bad outcomes, like hypos. 1.00

00:47:07.420 Yeah, and reduced hypos. Yeah. That oftentimes happens. Your A1c may stay the same, but your time

00:47:12.840 hypoglycemic may go away.

00:47:15.800 That's the thing I was impressed with Jake's data way back, even before you and I met, which was

00:47:20.680 they dropped hemoglobin A1c, I want to say like 2% on average, but they reduced the hypos too.

00:47:27.140 Oh, yeah.

00:47:27.840 And that's the part that, like I remember in medical school, the standard teaching was,

00:47:32.080 you don't care about, the hypers are fine. Yes, you can go to hyper, but the thing that you have

00:47:36.640 to ward against is the hypo and you'll accept a higher A1c to do it. And I think the tide is changing.

00:47:41.620 I think people are understanding that because of how long a person is going to live with type

00:47:46.260 1 diabetes, the microvascular complications are real. And as you probably know, there are data

00:47:51.720 that actually suggest, and I saw this through Jake's work, that better glycemic control in kids

00:47:58.540 when they're young, these are not randomized. So there's an obvious confounder here, but it can

00:48:03.460 predict better intellectual performance in school and stuff like that. Now, again, the obvious

00:48:08.300 confounder there is maybe the kids that have better glycemic control have better glycemic control

00:48:12.980 because of some other variable that's impacting their performance later on, socioeconomic status

00:48:18.860 or the education of parents or the attention of parents. So there's lots of other things there.

00:48:23.020 I don't follow this literature closely, so it's possible that there's actually been a randomization

00:48:26.460 that would allow for determining that. But the other thing, we talk a lot about average blood glucose

00:48:31.960 and instantaneous blood glucose, but another variable that I find very informative is the

00:48:36.580 standard deviation. So every day, I look at my report. I always do the seven-day report. I want

00:48:42.300 to see the seven-day report every day, and that's showing me my average blood glucose for the last

00:48:46.340 seven days and my standard deviation. And then every month or so, I want to see my 90-day trailing report.

00:48:52.400 And you brought up hemoglobin A1c earlier. I'm one of those people for whom hemoglobin A1c is

00:48:57.640 categorically useless because I have beta thalassemia minor. So my red blood cells are very small. They stick

00:49:03.340 around for very long periods of time. My hemoglobin A1c on a blood test generally varies

00:49:07.980 between 5.6 and 6.0. Basically, I'm pre-diabetic on that test. But using CGM for the last few years

00:49:18.420 and calibrating it, because even the G6, I still calibrate every day just to-

00:49:22.960 We left that option in.

00:49:24.220 By the way, that is a very elegant option. That's basically when I took my Libre and threw it out

00:49:28.880 the door. Because I used to use the two in parallel, and I just gave up on the Libre because

00:49:33.860 you couldn't force the calibration. And it was so inaccurate. And you couldn't get real time. I mean,

00:49:38.540 there was just too many problems with it to even discuss here. But using these forced calibrations,

00:49:44.700 I realized this thing's plus or minus 3%. There are many days, Kevin, when it is the exact same number.

00:49:53.100 And I think to myself, this can't be happening. The meter says I'm 88, and I poke my finger,

00:49:58.460 and it says I'm 88. And I'm like, how is that even possible?

00:50:00.740 Yeah, we call that the unicorn.

00:50:02.680 Yeah. But I get a lot of unicorns. Today, before I came here, I did a calibration. And

00:50:08.280 I don't know, the device said 89, and I was 88 or something. I mean, it was staggering.

00:50:12.620 It's an interesting dynamic because you are very focused on accuracy. I would say the biggest

00:50:17.580 benefit of G6 to our users has been the accuracy without the calibrations. We had somebody in

00:50:23.580 yesterday who's been on our product for a very long time, one of our Dexcom warriors.

00:50:28.460 And I asked him how it was going, and he held his hands out to me, and he goes,

00:50:32.280 my fingers have not felt like this for 15 years. And that's how long I've had diabetes.

00:50:37.640 Well, that's another interesting point you bring up, which is when I do my fasts,

00:50:41.960 I'm calibrating twice a day. I'm checking ketones two or three times a day. Just 14 days of that kind

00:50:49.380 of checking, my fingers are all blue.

00:50:52.240 Oh, yeah.

00:50:52.720 And they hurt like crazy. And that's another realization for me, which is think of all those

00:50:57.600 kids, that that is their life.

00:51:00.460 That's right.

00:51:01.560 It sounds like, how big a deal is that? It's really painful. In fact, you start to lose sensation.

00:51:08.040 One day, I was upset about some clinical results I saw, because we were calibrating the sensor

00:51:16.140 with a meter, and it became apparent that the meter was off. So I got my car, I drove to the

00:51:21.920 drugstore, I bought three meters. And I stuck my finger 60 times in a day, 20 times with each of

00:51:28.400 the three meters. And I made my own little spreadsheet, recorded all the meters, recorded the CGM reading to

00:51:33.620 see what had happened. Because in the old days, when you had to calibrate this device twice a day,

00:51:39.500 a bad finger stick would set it off for a while. Now, we have since revised the algorithm to identify

00:51:45.480 a finger stick that we just think is inaccurate, and we'll wait a while and ask for another one if

00:51:50.460 somebody enters them. But it was very discouraging, because we'd submitted a clinical trial to the FDA

00:51:55.260 based on a meter from a large company that we were told was fabulous, and it turned out to be

00:51:59.260 high all the time. And that drives our accuracy level off.

00:52:04.620 Do you have a point of view about which of the commercially available drugstore meters do you

00:52:10.080 find the most accurate?

00:52:11.200 I wouldn't offer that. I have ideas. I think most of them are okay. I would tell you my experience

00:52:16.880 with meters is also about the consistency. So oftentimes, when I'm wearing our sensor and go

00:52:23.740 to check the accuracy, now, I don't enter the calibrations like you. I like to see what the

00:52:27.820 patient is experiencing. But I'll stick my finger twice in a row and get like, some meters, I'll get

00:52:32.820 a big difference. Others, I don't get very much.

00:52:34.820 And do you do two sticks or same blood?

00:52:38.400 I vary it.

00:52:39.180 Okay. So here's a little tip that I've learned, and maybe you've already heard this from patients,

00:52:42.760 but by washing your hands before you do it, then using the alcohol, you can get some false

00:52:49.300 elevations if you have any food on your hands. So if you have just a little, you know, again,

00:52:53.240 like a crumb on your hand can actually raise the glucose. So, and I remember my friend in

00:52:59.540 residency, I don't, again, I'm glad I'm not identifying because I don't want to call him

00:53:02.800 out. I mean, he wouldn't even wash his hands before he'd be poking his finger because he

00:53:06.520 was just like, you know.

00:53:07.920 On the go all the time.

00:53:08.460 Yeah, on the go. It's like, I don't have alcohol pads. I'm just going to poke my finger

00:53:11.540 and, you know, but you know, those little things do, I think, add up to a little bit of benefit.

00:53:15.440 For what it's worth, I guess I can say what I like the most. For glucose, I like the

00:53:18.840 One Touch Ultra. I've switched to that. I used to use an Abbott meter or a different meter by,

00:53:25.440 I think it was by Abbott. And I don't know, I found the One Touch Ultra to be pretty darn good.

00:53:29.460 And I've done the same sort of experience, the thing you've done where I've gone multiple sticks,

00:53:33.160 multiple fingers, blah, blah, blah, blah, blah. Going back to finish the A1C thing,

00:53:37.480 my 90-day trailing average glucose under these very tight conditions shows an imputed A1C of 4.5 to 5.1,

00:53:47.620 depending on how tight my nutrition is. So in other words, my hemoglobin A1C is a full percent

00:53:55.260 higher. And that's actually posed real difficulties for me when applying for life insurance because

00:54:00.240 they say, you're a pre-diabetic. And I say, I'm not a pre-diabetic. And they say, yeah,

00:54:05.180 but your hemoglobin A1C is 5.8 or whatever it is. And I've had many interesting discussions with

00:54:11.320 the actuaries as I explain the limitations of hemoglobin A1C. But if I could be health czar for

00:54:18.820 a day and wave a magic wand, I would love to have something that meets all of your criteria,

00:54:25.040 right? So it does the performance, it's got the patient, it costs very little. But I really think

00:54:29.460 that if everybody had a CGM on every minute of every day in some completely fantasy-based way where

00:54:35.220 it doesn't hurt, it doesn't cost anything, but they had that data. It just drives such a behavioral

00:54:41.000 change. And I think that that's why the use of CGM, even without a deliberate intervention,

00:54:47.820 superimposed, improves outcomes. Because it is really hard to not pay attention to how activities

00:54:56.800 in your life, mostly food, but not just food, stress. It's amazing what stress can do to your

00:55:01.920 glucose level. I think one of the first insights I remember coming and talking to your team about

00:55:06.020 a couple of years ago was how high nighttime cortisol levels were raising my glucose levels

00:55:12.300 at night. I'd go to bed with a glucose of 90 three hours after eating dinner, and I'd wake up with a

00:55:20.080 glucose of 105. And once I started measuring nighttime cortisol levels by collecting urine overnight,

00:55:27.840 I could sort of correlate this amount of cortisol produced at night to how much that glucose level

00:55:33.660 would rise in the morning. And today I'll tell patients that fasting glucose is not nearly as

00:55:40.240 helpful as people think it is inside of a reasonably physiologic range. So someone who has a fasting

00:55:44.320 glucose of 130, that's a different situation. There's something going on there. But I have patients

00:55:49.320 that get very upset if their fasting glucose is 100 instead of being 90. And I say, it's very difficult

00:55:55.500 to understand what's going on there. That can very easily be explained by cortisol or hepatic

00:56:00.700 glucose output for some other reason that's not a function of insulin resistance. So that's why I

00:56:06.100 just think that these tools are so valuable. And like even yesterday, I was on a late flight coming

00:56:10.260 back from, I was in Chicago, and I just hadn't eaten much. And my flight was delayed. That's the other

00:56:15.880 thing. So I went into like the little store near the gate, and they don't sell anything but crap. 0.99

00:56:20.180 I mean, that's basically. Absolutely. Yeah, yeah. And I've already, like there are certain 0.99

00:56:25.480 snacks that I love, and they're kind of my comfort food when my flights are delayed. Like I feel like

00:56:30.040 I deserve a treat because I'm stuck in the airport. And I bought one. I bought my treat. My treat is

00:56:36.680 trail mix. I love me some trail mix.

00:56:39.020 Oh, trail mix. Trail mix will spike you as fast as anything you can eat.

00:56:42.120 Yep. So I get on the plane, and I buy the big trail mix because I don't do anything in moderation.

00:56:46.320 So it's got, you know, it's full of M&Ms and raisins and peanuts and little whatever.

00:56:51.680 And I get on the plane, and I'm looking at the bag, and I'm looking at my glucose,

00:56:57.120 and my glucose level's great. And I just decide I'm not eating it. So I didn't eat it.

00:57:02.280 Oh my goodness.

00:57:03.280 But I promise you, if I didn't have that CGM, I would have eaten it. But I just didn't feel like

00:57:07.980 looking at that glucose of, because I know I've done it in the past. I eat that whole bag of trail mix,

00:57:12.240 my blood glucose will be 130, easily.

00:57:14.840 Oh, then that shows what good shape you're in, because that'll get me up to 150 or 160.

00:57:20.000 It'll get me higher. The worst glucose experience I've ever had was in an airport.

00:57:26.720 And I had been with investors all day long, and I had not, my glucose hadn't spiked at all,

00:57:32.980 because when I sit in those meetings all day, I don't eat very much. And I'm in the airport,

00:57:36.940 and I'm in Boston. I say, you know what? I'm going to try something here. The yogurt covered

00:57:43.660 raisins in the little thing where you turn the wheel like the bubble gum machine, got myself a

00:57:50.200 cup of yogurt covered raisins. I think there's some chocolate raisins in it too, and ate the whole

00:57:55.660 little cup of them. And I'm telling you, that was my only ever above 200 glucose spike in my life,

00:58:02.600 and it only took 30 minutes. It just, right up as fast as possible. And it came down very quickly

00:58:09.060 as well.

00:58:10.180 Yeah, and then it overshoots. It probably came down to 60.

00:58:12.320 And then it overshot, and I was, yeah, overshot down in the low 70s, because I'd put so much

00:58:17.900 glucose into my body, that's how my pancreas reacted. But all of our bodies behave differently.

00:58:23.800 We've talked a lot about type 1 diabetes today, and one of our visions going forward is to get into

00:58:29.280 other conditions. And you're certainly a pioneer in this, and really one of our first. We've started

00:58:35.960 doing work in type 2 diabetes. So again, let's paint a picture. We've talked about kids with type 1,

00:58:41.860 adults with type 1 who have to deal with insulin their entire lives or they die.

00:58:45.960 And the device is a love-hate relationship because it always reminds them.

00:58:50.580 Now let's look at somebody with type 2 diabetes. They're my age, 60 years old, may have had type 2

00:58:56.740 diabetes for 5 years. All they do is they go to the doctor, they get some pills, they get a meter,

00:59:01.280 they go home, and nobody tells them what's going on. And they say, you need to eat less and exercise

00:59:05.120 more and take your pills. And after 5, 10 years of this, all of a sudden your A1C's at 11, and what do

00:59:13.200 you do? So we start adding more drugs. And you can turn on the television anytime you want to.

00:59:20.800 And you can see more type 2 diabetes drug advertisements probably than any other

00:59:24.540 pharmaceutical product. Because these drugs are expensive and there is a big population here to

00:59:30.560 be addressed. So we start prescribing other compounds, hoping the A1C will come down. And it might come

00:59:36.340 down a little bit. It might not. And if it doesn't come down enough, we prescribe another compound

00:59:41.680 and another compound and another compound. And some point in time, that patient gets the dreaded

00:59:47.520 information they've got to go on insulin. But type 2 diabetes, I'm not as near as scientific as you,

00:59:53.300 but it's almost not diabetes. It's almost a completely different disorder with all the

00:59:58.120 insulin resistance and all the other things that go on in your body.

01:00:02.240 Yeah. The biggest disservice that I think has been done to this field, this is hyperbolic. I mean,

01:00:06.940 there have been many disservices. It's the fact that we refer to them both as diabetes. Now,

01:00:11.500 I understand why historically that's the case, but you're absolutely right. Type 1 diabetes and type 2

01:00:15.680 diabetes are so different. And I try to be disciplined about always referring to them that way and not just

01:00:21.440 saying diabetes because it really does a disservice to them. What you point out is very important,

01:00:27.400 Kevin, which is that the real problem in type 2 diabetes is that we have fixated so much on

01:00:34.140 just the glucose and not worrying about how we get there. So when you use a class of drugs that

01:00:40.320 increases endogenous insulin production and or when you use more insulin, what you reduce is the

01:00:46.480 glucose and what you reduce are the microvascular complications. You do not save lives. They're

01:00:51.400 still getting cancer. They're still getting heart disease. They're still getting Alzheimer's disease

01:00:54.960 at a much higher rate. The challenge in type 2 diabetes is the combination of the hyperglycemia

01:01:01.740 with the hyperinsulinemia. So the therapies that save lives in type 2 diabetes would be things like

01:01:08.080 metformin, where you're actually fixing the glucose side of the equation and not just adding more

01:01:13.300 insulin to the system. The SGLT2 inhibitors, these are the things that are getting the glucose out of

01:01:18.040 the system, which actually brings me to a question based on something you said earlier.

01:01:23.000 Is there any way it is algorithmically possible to estimate, let's say, the average insulin that a

01:01:32.120 non-diabetic would make over the course of the day based on glucose if you had some data points? So if I

01:01:37.100 came to Kevin and said, look, let me go and get under test conditions, certain levels of glucose and

01:01:43.580 insulin that are associated, we develop a, you know, a kinetics model for it. And then by tracking my

01:01:51.080 continuous glucose data over time, could we ever try to estimate what my insulin levels are in between?

01:01:56.540 We have some studies that are very preliminary and I really can't talk too much about it, but we do

01:02:02.620 believe that it's possible to measure how quickly insulin reacts and how much by the slope of the

01:02:10.200 glucose curves, but not at our five-minute measurement intervals. So we're looking at other ways to record

01:02:16.420 data, to develop algorithms to do that, but it's very preliminary right now. I think over time, and

01:02:24.400 this gets back into another thing that Dexcom will do over time, we have an open architecture platform as

01:02:29.860 well. So we've developed APIs to whereby we will provide data to other software platforms if they

01:02:36.640 can analyze glucose values better than we can. One drop, there are a number of them. They haven't

01:02:42.540 produced a lot for us commercially, but we've made the decision that this is the type of company we

01:02:47.300 want to be because we just can't do everything. I think over time, that might be something that

01:02:53.480 we'll look at, but it also might be something somebody might be able to look at better than us.

01:02:57.660 Your raisins one, by the way, is so funny because I've had a number of those experiences. My similar

01:03:03.200 to that one where I was shocked was grapes.

01:03:06.900 Oh, grapes are awful for glucose values. 0.93

01:03:09.440 No, I had not eaten anything in about 18 hours.

01:03:12.580 And you had grapes with the glucose sensor on.

01:03:14.780 I had grapes, and I was so pissed off. I was like, I thought the thing was broken. I was like ready to

01:03:20.800 rip it off my body. I was like, there is no way my glucose is this high after a... I haven't had a grape

01:03:25.760 since. That was about six months ago. And I was on an airplane about a month ago, and they had like

01:03:32.520 a snack tray, and it was like some cheese, a strawberry, and some grapes. And it had like

01:03:37.740 five grapes. And I was actually really hungry. I ate everything, but I didn't eat the five grapes.

01:03:42.160 Now, I'm sure the five grapes wouldn't have done anything, but I'm boycotting grapes. I am done

01:03:46.840 with that what I call quasi-fruit.

01:03:49.180 Well, I have not eaten many yogurt-covered rice since that episode in Boston either. So there you go.

01:03:55.080 But see, that's my point, right? I think

01:03:56.860 CGM has the potential to change the way people eat more than any other technology I have

01:04:02.460 ever laid eyes on. I'm heavily biased. I mean, I've stated my bias many times.

01:04:06.860 I'm a very big proponent of CGM.

01:04:09.040 But I also believe in real-time feedback.

01:04:12.220 I think it's very difficult to curb a person's behavior

01:04:15.180 when the feedback loop is so long. And the weight on the scale,

01:04:19.240 the blood test that you get at your doctor every six months, those things are valuable.

01:04:24.820 They help. But nothing trumps the 30 minutes after I eat the chocolate-covered raisins.

01:04:30.880 A, I see what it looks like. And now I can correlate that feeling of,

01:04:34.960 I feel like crap an hour later because my glucose has gone from 200 to 60. 0.93

01:04:40.520 That's a crappy feeling. That's not physiologically reasonable. And now people see why. 0.97

01:04:47.320 So over time, where this gets very interesting is that core technology platform I talked about

01:04:52.720 where we make products that perform well, that patients will use, and that cost less,

01:04:59.080 never goes away. But there's a software element to what we do that we can change dramatically

01:05:04.520 given the experience we want a patient to have or that a patient needs.

01:05:08.480 And again, I'll give you the perfect example. Our type 1 patients, the alerts and the alarms

01:05:13.500 at nighttime while they sleep are absolutely critical. But somebody with type 2 diabetes

01:05:19.200 who's taking these compounds, that may not be as critical for them. But what would be critical

01:05:25.300 is an analysis at the end of the day where we compliment somebody on,

01:05:30.720 gee, you kept your glucose within range extremely well. Our insights, as we look back,

01:05:36.100 what did you eat for dinner today? Or integrating the CGM data into a food database to whereby

01:05:42.340 you take pictures of what you eat over time. And when you go back, you look at that food and say,

01:05:47.920 wow, I ate those scrambled eggs this morning and my glucose values didn't go up. Maybe scrambled

01:05:53.160 eggs for breakfast is a good thing for me. And then you take a picture of pancakes.

01:05:56.380 And the next time you sit down to eat, if you take a picture of those pancakes, you get reminded,

01:06:00.900 you know, you had pancakes a month ago. And that may not have been the outcome that you're looking

01:06:06.100 for. And I think this is the kind of stuff Neutrino is working on. And I agree that that stuff is so

01:06:10.800 powerful. I feel like, you know, people assume I know a lot about food because of, you know,

01:06:15.320 sort of the work I've done and the things I've talked about and written about. And, but the reality of

01:06:19.100 it is, I'm still humbled by how much I don't know, because it's not just what you eat, but it's the

01:06:24.800 physiologic state you are in when you eat it. It's very different what happens if you eat a bowl of

01:06:29.800 pasta after you've exercised, you know, 30 minutes after a hard workout versus, you know, eating it

01:06:36.280 after having not exercised. It's a very different glycemic response. Stress, we've talked about this

01:06:41.340 and it can't be understated. How much higher my glucose levels are when my sleep is crappy.

01:06:45.900 Oh, or the correlation between high glucose levels and crappy sleep. Because if you go to

01:06:53.060 bed with very high glucose levels after eating a very high carb dinner, I've seen a direct

01:06:58.340 correlation between my sleep and my glucose values. Yeah. Personally. Anyway, I can't make this case

01:07:02.980 enough. Now we talked about, there are a couple other companies out there that are doing this. Is

01:07:06.380 there any publicly available information that compares the performance and fidelity of the Abbott

01:07:11.780 product or the Medtronic product to your current product? We all have accuracy tables in our user

01:07:17.840 guides, but they're biased towards the way that companies want to project themselves. We have

01:07:23.600 studies that we've published comparing ours to the other products. We maintain our performance all the

01:07:28.860 time. Yeah. It's not subtle in my hands. I mean, I've been pretty vocal about it. It's not a subtle

01:07:33.260 difference. And it isn't. We perform extremely well across all the range of glucose in the low and in

01:07:37.920 the high range. I won't speak as much to the other guys. Our patients know. Our patients know what a

01:07:43.900 Dexcom is versus the others. And we can live with that one. Do you think that you will be able to

01:07:49.000 maintain performance as the sensor gets smaller and smaller? So in other words, is there a day when

01:07:54.080 there's going to be a one millimeter sensor? So something that really doesn't even get into the

01:07:57.860 interstitium, but is still sort of in the dermis itself that can perform this well? Or is there

01:08:03.640 just a technical limit to being able to do this without accessing interstitial fluid?

01:08:08.200 For us today, we've only gone down to the interstitial fluid. We have looked at technologies

01:08:12.120 that would go down more shallow. We have within our R&D group, an advanced technology group who

01:08:18.140 looks at nothing but things like that. Our second Verily product, one of our goals is to have a

01:08:23.260 shorter sensor and to make it smaller. But there are complications of a shorter sensor.

01:08:28.440 What do the membrane layers look like? So you can get an extended wear on a 14-day sensor.

01:08:32.540 How's the electrochemical reaction on a longer life sensor if it's shorter versus the one that

01:08:39.260 we have now? So we go through a lot of trial and error as we look at these things to determine what

01:08:44.580 is the best combination of all of these features. I think we'll have a shorter one. Right now,

01:08:50.860 we're very comfortable with the interstitial fluid.

01:08:52.980 If the pain of insertion is effectively eliminated, which it is, once it's in there, I don't actually

01:08:58.820 know the difference between it being in me versus on the skin. And I've played with other sensors out

01:09:04.440 there that there are some that are as short as 400 microns.

01:09:07.460 There's some very short ones. And we evaluate those technologies. We look at them.

01:09:12.020 One of the beautiful things of being a public company and having resources is if we saw one

01:09:15.940 that we thought was remarkable, we'd talk to them and see if there's something we can do to further it

01:09:20.520 along. But right now, we're comfortable with what we're doing and what we have.

01:09:23.940 Now, another question I get asked a lot is why can't this be done optically? Why can't this be

01:09:29.700 done the way a pulse oximeter where you put it on your finger? So for someone listening to this who

01:09:34.860 doesn't know what that is, but if you've ever had surgery or anything like that, when you're

01:09:37.740 under anesthesia, there's a little device they usually put on your finger or on your earlobe

01:09:41.380 and using light, it's actually measuring or estimating the saturation of oxygen. And can't we

01:09:48.380 just do that to measure glucose? Are there limitations to this that are just too great?

01:09:52.680 It has been tried on numerous occasions to do that. Several technologies have tried non-invasive

01:09:58.840 sensors. In fact, the first glucose sensor that I saw somebody wear, other than one we developed at

01:10:04.620 MiniMed and Medtronic, was one called the GlucoWatch, which actually was a watch that sat on your wrist

01:10:10.020 and would take fluid out of your skin through reverse anephoresis and measure it that way.

01:10:17.160 Wow.

01:10:17.640 It would leave a red mark on the outside of your arm and it was hard. We've seen a lot of

01:10:23.240 non-invasive glucose sensors and people call our sensor invasive because there's a needle and a

01:10:28.940 needle that goes in your body and comes out and a sensor that remains. So oftentimes people say,

01:10:33.880 we're going to develop something that is non-invasive. If it's two inches by three inches and sits on

01:10:39.540 the outside of your body and you can say, yeah, it was non-invasive. Well, how do you find invasive?

01:10:44.740 To me, invasive is does it invade your lifestyle and does it produce the result that you're looking

01:10:50.680 for? A lot of people have worked on that type of thing. They've worked on just measuring on the skin.

01:10:57.420 There's one company that has one that measures in your tears that you put inside your eye that I've

01:11:02.460 seen.

01:11:03.580 Are these accurate?

01:11:04.520 Not accurate enough and none of them have been to market yet, but only time will tell. People will

01:11:09.780 continue to look, but they're not solving the fundamental problem. I'll go back to those three

01:11:14.940 things we build our product around and I'll add a fourth. If it's accurate, if you solve the accuracy

01:11:20.600 problem, if you make it more convenient and solve the convenient problem, if you lower the cost,

01:11:26.500 and finally, if you can produce a healthcare outcome, you have something that's meaningful.

01:11:30.940 Well, not having the little wire in your skin while one can try and paint a horror story around that,

01:11:37.740 it's not really that invasive to your body. So anything that comes has to solve one of those

01:11:44.420 problems better than the technologies that are on the market. And if somebody can do that with an

01:11:50.680 infrared sensor, optical sensor, then we've had a wonderful technological step and we'll all figure

01:11:58.460 out where we move from there. We just haven't seen anything yet that's accurate enough, performs well

01:12:03.520 enough. Yeah. I suspect it's going to depend heavily on how much this stuff matters and meaning how much

01:12:10.000 accuracy matters. Obviously for me and for my patients, that's the highest priority. And what

01:12:15.460 you give up right now is cost. I think of the four metrics you just identified, the place where this is

01:12:21.180 by far the biggest barrier is the economic barrier. Is the economics. I would agree. So on that thread,

01:12:27.860 anyone listening to this who says, gosh, I got to have a CGM, they have to go get a prescription

01:12:31.100 from their doctor and very likely they're going to pay out of pocket for it. Unless they have

01:12:35.780 diabetes. If they have diabetes, they... Including type 2? Depends. A few payers cover type 2,

01:12:41.740 not many, but a few. Do you have the liberty of saying which payers? Do you know which payers cover

01:12:45.460 type 2? Not off the top of my head. Because it's going to depend by region as well. It's very regional.

01:12:50.720 And most of the type 2 patients are patients who use intensive insulin, such as the CMS Medicare

01:12:55.480 patients. And this gets into our payer negotiations. Sometimes we sit and we discuss,

01:13:02.240 we want to make this easier for everyone to get. We want to cover type 2s. We want to cover type 1s.

01:13:07.060 We want all these things. And they go back to their economic models and say, well, we can't do all

01:13:12.040 those things, but we can give you this and we can give you this for the type 1s. And it is very much

01:13:17.360 a business negotiation. And then based on your insurance plan, typically our patients pay 20% of the

01:13:23.360 cost. Standard copay for durable medical equipment. And depending upon the timing of the year, if you

01:13:29.040 buy it the first of the year and their copayments and their deductibles have not been met, it's more

01:13:33.200 expensive for them. And is the cash cost fixed across the country or is that regional?

01:13:37.080 We have cash prices here that are pretty much national.

01:13:40.440 You can say what the cash price is?

01:13:41.280 I don't even... I'll be honest with you. I don't even know what they are.

01:13:43.500 I should know this. My office does all the buying. The sensors come in boxes of threes, right?

01:13:48.580 Three sensors.

01:13:49.040 And we typically buy them nine at a time. So that's about a three-month supply because it's 10 days per

01:13:53.720 sensor. That's 90 days. You know what? I used to do the math on what are you paying per day. I'd

01:13:58.400 amortize the whole thing over a per-day cost. Does nine bucks a day sound about right?

01:14:04.280 That's about right as to what somebody pays on the outside.

01:14:06.820 Yeah. I think that's what we pay. Nine or 10 bucks a day.

01:14:09.940 Between nine and $11 a day. We say on our earnings calls, for example, our average price per sensor is X.

01:14:15.880 We used to be only seven-day sensors. So our average price is always between $70 and $75

01:14:20.980 that we have recognized in our revenue. Our 10-day sensor, we have different models as we've gotten

01:14:27.040 that reimbursed. That cost per day has to come down.

01:14:30.900 What do you think, or have you guys done an internal analysis that says, look, at this point,

01:14:36.600 this is a product where if we have 1% of the population has type 2 diabetes, you've obviously

01:14:42.140 built a very successful company where your customer base is 1% of the US population or

01:14:46.380 less. You've got a type 2 diabetes population that I think would benefit from this as much

01:14:51.040 as the type 1 population would, even though, as you've pointed out eloquently, different

01:14:55.380 diseases basically. So that takes you to 10% of the population. Then you've got the pre-diabetic

01:15:00.960 population.

01:15:01.800 Another 10.

01:15:02.660 Exactly. But then you've got like sort of half the people are kind of, I hate the term wellness,

01:15:07.760 but these are people that we're trying to optimize health for and they don't have a disease,

01:15:12.000 but it's not just that we're trying to prevent them from getting diabetes, but we're trying

01:15:15.360 to figure out if you can actually, you know, make them live longer through better glycemic

01:15:19.380 control. What do you think, or what's your intuition about what does that need to cost

01:15:23.640 per day to make that group say, look, out of pocket, I'm going to pay for this. Because

01:15:27.620 there's no way an insurance company is going to pay for, you know, healthy people wearing

01:15:31.060 these things.

01:15:32.180 We drive those discussions all the time. It's interesting. I had a discussion with a payer once

01:15:37.560 and I said, look, we want this to go to everybody that uses insulin intensively. Tell me what

01:15:44.340 the price is. Just tell me what to make it. Give me a number. And the response I got was

01:15:51.360 how low would you go? So I don't know what a person would pay out of pocket because all

01:15:56.620 we do is speculate on these kinds of things. Is it $25 for a two-week sensor? Is it $50 for

01:16:01.540 a two-week sensor? A lot would depend upon the information and the things that you learn

01:16:06.000 from it. There's a whole regulatory process we'll have to go through to get a health and

01:16:10.700 wellness product. For example, our device is a prescription device now. We've been very

01:16:15.140 clear in our discussions. Your patients who use this in its non-diabetic state, they have

01:16:20.580 a prescription from you and you pay cash for that.

01:16:22.600 And we write them a new prescription every three months.

01:16:24.380 And they get it. To make this a non-prescription product, there's a lot of steps we'd have to

01:16:29.780 go through at the FDA. And one of the things that we look at with respect to our human factors

01:16:36.440 and our software design, if it weren't on prescription, what would that software experience

01:16:41.220 have to look like?

01:16:42.440 So let's explain for the listener why that's the case. So right now, everything we've talked

01:16:46.620 about, as you said, is still a medical device.

01:16:49.100 It is a class two medical device.

01:16:50.780 A physician writes a prescription. And I'm only just talking about the cash patient. So this

01:16:54.480 is not about insurance anymore. Physician writes prescription, patient gets it filled,

01:16:58.180 and their all-in cost is about nine bucks a day to wear this thing. And they are seeing

01:17:03.200 the exact same data using the exact same device with the exact same algorithms that a patient

01:17:09.060 with type 1 diabetes is using.

01:17:10.220 And the exact same, yeah.

01:17:10.920 Everything is the same. I am using everything that someone with type 1 diabetes is using.

01:17:15.220 So explain, Kevin, why if tomorrow a little elf jumped out of your desk and said, voila, we 0.99

01:17:22.260 can make this thing for a dollar and just say that the dollar was the price where like at a

01:17:27.640 dollar a day, you could penetrate this new market, that the FDA doesn't say, great, go ahead,

01:17:33.520 go start selling it. Like there's a fundamental issue. What is that issue?

01:17:36.640 We have to go through clinical trials, particularly if we're labeled for use for people with diabetes,

01:17:41.460 to demonstrate the accuracy and consistency of the product across all glucose ranges.

01:17:46.500 The FDA, in all candor, has been very progressive with Dexcom. I mean, we're the first company ever to

01:17:51.460 go directly to the phone. We're the first company to share data. We're the first company ever to have a

01:17:56.700 sensor perform at the accuracy levels that we have. Most every first with the FDA in our industry

01:18:02.200 has come through us partnering with them. Most recently, they created a new category for continuous

01:18:08.800 glucose monitoring, and this is in the intensive diabetes world, called ICGM versus CGM, where they set

01:18:14.520 accuracy standards, where their products have to demonstrate on a statistical basis they perform.

01:18:20.800 And you have to have a certain number of values in a study in the low range, in the high range,

01:18:26.360 in a certain number of points to statistically prove that that product is capable of doing the

01:18:31.520 things that we want it to do. In our intensive world, that's dosing insulin, either through a manual

01:18:37.400 calculation or thinking by a patient. But also with this ICGM category, which stands for interconnected,

01:18:44.140 we can connect to other devices. And they've made an easier path for other devices to connect to our

01:18:50.020 product. So for example, if you have an insulin pump and an algorithm, and would like to run that

01:18:55.520 algorithm of glucose sensor, your regulatory path with Dexcom is just to show you get the signal.

01:19:01.040 You don't have to do a study to show that Dexcom's accurate. We're done. It's interconnected.

01:19:06.480 And Tandem's recent approval is along those lines. But if you're displaying a glucose value,

01:19:11.720 and that is a real-time glucose value, and that is where we've been in our discussions with the FDA,

01:19:16.300 hey, that's a class two medical device. And as of today, that requires a prescription.

01:19:22.560 At some point in time, again, as we look to the future, and as you and I have talked,

01:19:27.460 I'd like a device we could sell over the counter, or we could sell through nutrition programs

01:19:31.900 that a patient couldn't use to dose insulin. But what would be the real-time feedback you'd give

01:19:37.080 somebody?

01:19:37.580 So explain why that's necessary. That's what you just said is the important point,

01:19:42.060 which is if you want to sell this over the counter, the FDA is going to say it cannot be

01:19:48.680 used by someone to dose insulin.

01:19:50.680 For today. For today. And maybe that will change over time. But for now, those are the rules. And

01:19:56.220 those are the rules that we play by.

01:19:57.720 And that's very important for, I think, people to understand, because that's one of the questions

01:20:00.800 I get asked the most, which is, why can't this thing just be over the counter now? I'm willing to

01:20:06.640 pay the full price. Why do I have to? Because not every patient has a physician who understands

01:20:11.920 the benefit of this. And I know, just from what I see on social media, a lot of people

01:20:15.800 go to their doctor and say, hey, you know, Dr. So-and-so, please write me the prescription.

01:20:20.680 And he or she says, what the hell for? Get out of here.

01:20:24.400 Yeah. What are you talking about?

01:20:25.440 Yeah. And so now that patient who's actually says, look, I'm willing to make an enormous

01:20:29.280 investment in this because it's my health and I believe that this is going to matter.

01:20:32.540 Well, they can't do it. So what I'm hearing you say is, there's one regulatory path that

01:20:39.940 says, maybe down the road, the FDA says, you know what? This will no longer be something

01:20:44.960 that requires a prescription. If the patient's willing to pay for it, the patient can have

01:20:48.840 it.

01:20:49.560 I can't speak for them, but they have been so progressive with us so far. What we would

01:20:54.940 literally have to present, again, is a case as to how somebody would never harm themselves.

01:20:59.180 And remember, there's a needle involved in inserting our product as well. There may be

01:21:03.540 standards and rules around needles, things like that. So I can't speculate how we get

01:21:08.420 there, but I think over time, that's something we would like to take on. But again, it would

01:21:14.040 not be a device used for diabetes treatment. It'd be a device used for other things. We'll

01:21:18.780 have to figure out what that looks like. We don't know yet.

01:21:21.340 And what you were alluding to a moment ago was the other option for making this OTC is not to

01:21:28.880 take the same device and just deal with the FDA changing a rule, but it's to create a device that

01:21:33.940 would fit into the current sort of regulatory environment, which says the device would basically

01:21:39.820 have to be dumbed down. You'd basically have to strip functionality out of the device, right?

01:21:44.460 Strip functionality and the user experience. Yeah. Not necessarily accuracy, not performance,

01:21:51.160 but strip down the information that's presented to the individual.

01:21:54.220 So the ways that that can be done is one, you could make it such that they don't get real-time

01:21:58.840 information. Because if you don't have real-time information, if you see the perfectly accurate

01:22:02.480 identical data a day later, it doesn't help you dose insulin. Another way is, as you said,

01:22:07.620 you just show ranges.

01:22:09.340 Colors. The circles on your Apple Watch.

01:22:11.580 I don't wear one of those things, but...

01:22:13.540 I know you don't. You and I both like to wear our regular watches. I do wear the Apple Watch

01:22:18.940 frequently just to see what the user experience is.

01:22:21.760 Okay. So what do the circles do on that?

01:22:23.760 It's for training. You track how many times you stand up, how many steps you take a day.

01:22:27.600 I see. Okay.

01:22:28.740 They have some very simple measures. It's right there. It's pretty easy to see and very easy

01:22:33.540 to use.

01:22:34.140 This is a totally unrelated question. Does the Apple Watch have a little app or plugin that allows

01:22:38.580 you to put the face of really cool watches on it so that you... Like if you really want

01:22:42.980 to have a paddock, you could have like a little paddock Apple Watch. Because I mean, that would

01:22:47.300 be easy to do. You just need a picture of the paddock and you could have it like looking like

01:22:50.840 a... Pick your favorite GMT or whatever.

01:22:54.400 See, Peter, we start talking and it gets back to watches and other businesses we'd like to

01:22:58.260 go start and do. I don't know that that exists, but that would be fun.

01:23:01.680 Who... Why haven't... Apple, if you're listening to this...

01:23:03.820 Maybe somebody has.

01:23:04.600 Yeah, yeah. Someone's got to have developed a plugin that allows you to have any watch

01:23:07.700 you want as your Apple Watch.

01:23:08.900 As your Apple Watch.

01:23:09.520 Apple Watches cost, what, a few hundred bucks?

01:23:11.500 Yeah.

01:23:11.700 But you could put the face of a million dollar watch on it if you want.

01:23:14.720 Well, and what we have found within our company and with our patient base, there's a very high

01:23:19.380 number of Apple Watch users because our device goes from the phone and then the phone,

01:23:24.280 there's a watch app.

01:23:25.220 Oh, so I didn't even... Because I don't use the thing, I've never noticed that.

01:23:27.440 Yeah, there is a watch app and then we are developing a direct-to-watch transmitter communication

01:23:32.580 protocol. What that would do for a patient, again, the new Apple Watches have cellular

01:23:37.300 capability as well, as will some of the Android Watches and some of the new Fitbit products.

01:23:42.800 But with that direct cell capability, that watch then can become a patient's receiver and the

01:23:48.740 data could still be shared. You wouldn't have to have it go to your phone. And again, we're

01:23:53.840 thinking about those criteria I talked to you about earlier. As far as convenience, it's a lot

01:23:58.420 easier for some of our people to pull their sleeve up and look at their glucose than to get their

01:24:03.240 phone out during an important meeting or to even take their phone to bed. They could wear their watch

01:24:09.200 to bed.

01:24:09.680 Well, that's another new thing I like about the G6 a lot. We didn't talk about this. With the G5,

01:24:14.920 I had to keep my phone in my room to see all my nighttime data. The G6 seems to store the data

01:24:22.200 so that if I keep my phone in the office, it backloads my file.

01:24:26.880 There is some backloading if you do lose connectivity. The transmission range of our product,

01:24:33.720 I believe it's labeled for 20 feet. And if we've learned one thing about going to the phone,

01:24:39.040 Bluetooth is not an exact science. And I think we've all experienced that where some headphones

01:24:44.100 connect really well. Other headphones do not. Sometimes you get your car, it works.

01:24:49.060 Sometimes it doesn't. It is very hard to explain.

01:24:51.740 But is that why you went with Bluetooth as opposed to near-field communication?

01:24:55.020 We went with Bluetooth because we didn't want our patients to have to put the phone close to the

01:24:58.620 device. We wanted the continuous feed. I think there is a place for near-field. And it's quite

01:25:04.000 possible we'd have both chips in a future one. But for now, we're very BLE oriented. And it works

01:25:10.860 pretty well. Are there any, and again, I want to be sensitive to what you can and can't talk about

01:25:14.860 publicly. Are there any other partnerships or collaborations that you're able to speak about

01:25:18.680 that are maybe interesting to consumers, especially in this sort of non-diabetes, non-type 1, non-type

01:25:25.080 2 market for the people like me who are thinking about this for quote-unquote wellness?

01:25:30.500 Well, nothing real concrete. I will tell you that Apple has been a wonderful partner for us in

01:25:35.580 getting the device to the iPhone and getting our apps working on Apple Watch. In fact, at the

01:25:41.000 developer conference one year, we got a VIP invitation. Our head of R&D couldn't go, so I

01:25:46.260 went. And Peter and I are dressed differently. I'm a total business person, and Peter's Peter. And so I

01:25:52.580 show up at the Apple developer conference in my sport coat, in my shoes. And of course, the guys next

01:25:57.560 to me are in shorts, and they're way younger than me. And they're looking at me like, who's this?

01:26:02.100 Who's this? Nobody dresses like this here. And I'm sitting there feeling pretty uncomfortable. And

01:26:07.960 they came out and talked about Apple Watch. And lo and behold, Apple's got Dexcom Watch on the screen

01:26:13.720 up there. And I elbowed the kid next to me. I said, do you see that? He said, yeah. I said,

01:26:20.260 I run that company. I thought I was pretty special for the day. Our partnership with Verily includes

01:26:28.520 work with a company that they have called OnDuo that is developing treatment for type 2 diabetes

01:26:34.200 and developing apps that combine everything they've learned about diabetes with a lot of

01:26:39.720 Google technology. And they're in pilot phases with various payers on the type 2 side. We're

01:26:45.440 also doing a lot of work with UnitedHealthcare on type 2 diabetes. And I think in the future

01:26:51.320 with them, we're also going to develop some pre-diabetes models as we take this Gen 6 technology

01:26:56.880 out. You know, we've talked about going to the consumer. Peter, you wouldn't have taken

01:27:02.280 G5 to a consumer anywhere. It's too hard.

01:27:05.300 Yeah, I had a hard time getting patients to do the G5.

01:27:09.600 G6 is, it's not all the way there, but it's close. There are things we can do to make it

01:27:15.580 a little bit easier, like pairing the device with the transmitter and some of the other things

01:27:19.820 that have been inherent within our system. But it's close. As we get this thing closer to

01:27:26.160 consumers, I think you'll see us try more things, more and different things. A couple

01:27:31.360 of other applications where we'd love to take our continuous glucose monitoring technology

01:27:36.160 where it's not. I've had numerous physicians ping me on LinkedIn saying, why aren't you in

01:27:42.420 the hospital? Why aren't you in the hospital?

01:27:45.700 That's a good question. Why aren't you in the ICU of all places, right?

01:27:48.640 We tried many years ago. We had a joint venture with Edwards Life Sciences. And then the FDA was

01:27:54.300 extremely strict on accuracy and performance. And by the time we built what we built, it just,

01:28:00.860 it didn't work out. Our technology has come so far and we've learned so much through projects like

01:28:06.460 that in the past that we will make a run with G6 in that type of market to help patients. I mean,

01:28:13.080 think about it. You connect the Bluetooth to the nurse's station. I was in one medical center where

01:28:17.640 patients in the ICU are tested every 30 minutes. 48 finger sticks in a 24-hour period.

01:28:23.380 Not uncommon. Intensive insulin therapy, of course, has had its ups and downs resurgence in the

01:28:28.860 literature, in the critical care literature. But basically, there was a paper that came out in 2001

01:28:33.580 or 2002 in the New England Journal of Medicine that really changed the face of how glucose and insulin

01:28:39.760 levels were managed, or glucose levels were managed using intensive insulin therapy in the ICU.

01:28:43.540 That has been questioned. Obviously, there have been subsequent trials that have suggested that

01:28:47.640 maybe the benefits on the glucose side come at the cost with an insulin side, et cetera. But

01:28:52.340 you're absolutely right. Now, that reminds me of something, though, which is

01:28:56.160 acetaminophen, Tylenol, has an effect on the performance, at least of the G5.

01:29:01.380 Not G6.

01:29:02.500 Oh, it doesn't have the effect anymore.

01:29:04.000 It does not affect G6. We ran a very, very, very large study and put a lot of acetaminophen

01:29:11.320 into patients to whereby we have no contraindication for acetaminophen anymore.

01:29:15.200 So that's interesting. But the G5 would over or underestimate?

01:29:18.940 It would go high.

01:29:19.860 It would overestimate.

01:29:20.520 It would go up a little bit.

01:29:21.100 Yeah. So one of the challenges, I suspect, in the ICU patient is all of the drugs that

01:29:26.080 they're on, also they third space like crazy. So their interstitium looks totally different

01:29:31.440 from a normal person's interstitium. It's almost like there's a whole new calibration, right?

01:29:35.620 We're going to have to go run some studies. But that being said, I'll go back to what

01:29:41.600 you said your friend told you, you can pay me now or pay me later. If we can get this

01:29:46.180 in the right configuration in the hospital on most every patient, at least every patient

01:29:51.020 over 50 that comes in, forget whether they don't have diabetes. That hyperglycemia effect

01:29:57.360 on healing, we know, is a very difficult thing. If we can get it on everybody and make this

01:30:02.580 affordable and fit within the normal workflow of the hospital, that's a tremendous market

01:30:08.520 for us. Another place we haven't been or been labeled, and you're going to be a little

01:30:12.340 bit, probably chastise me after this. We've never been labeled for gestational diabetes.

01:30:18.040 We had this discussion once. I didn't realize you weren't actually labeled, but I have a patient

01:30:22.500 whose fiance is pregnant and she had a device in and it turned out she had an old device.

01:30:29.460 So she had the device and then never used it for, with one pregnancy, then used it with another

01:30:34.440 and it wasn't working. And we did some troubleshooting and to make a long story short,

01:30:38.420 she said, well, I called Dexcomer, but they said it's not approved for gestational use.

01:30:43.200 And I remember saying, what would be the difference?

01:30:46.820 And I'm very proud of my people for saying the right thing. It's just, we haven't run a study in

01:30:52.120 gestational diabetes. We'll have to run a study there to get that label. One of my daughters-in-law 0.95

01:30:57.020 contracted gestational diabetes while pregnant with twins. So her OBGYN wrote a script and we

01:31:03.800 procured a Dexcom for her. And he just looked at it and said, this is remarkable. We should all be on

01:31:10.480 this. And this gets back to A1C and other measures. As we look at evolving this Gen 6 technology,

01:31:16.940 one of the things that we look at as this product is a diagnostic as well. Right now you drink that

01:31:22.740 real sugary drink, go way high, you go way low. It's kind of awful. What about a blinded sensor

01:31:28.780 on somebody for a week? Yeah. Infinitely better. And then develop the algorithms to see what we can

01:31:34.700 predict with respect to gestational diabetes. Those are the type of studies that we want to run in the

01:31:40.420 future and the way we'd like to go about that market. And there's others as well. Again, I'll go

01:31:45.660 back to your health and wellness one. Combining this with some of the nutrition apps out there,

01:31:51.280 the Weight Watchers app, combining if that had glucose data and you ate your points per day and

01:31:57.720 you ate your 32 points today, but your glucose values were still high, maybe we take you down to 26

01:32:02.700 and vice versa. You can see how that can integrate with it wonderfully and integrate without showing

01:32:10.880 a glucose value that you can dose insulin off of. I just think in the end, it's a much better metric

01:32:16.040 for health. And I think health matters more than weight. Weight is correlated with health,

01:32:23.900 but the correlation is not exceptional. It's reasonable, but it's so far from great. And

01:32:33.020 my prediction would be when those trials are done, your standard deviation of blood glucose over a three

01:32:40.440 month period is going to tell you infinitely more than whatever the fluctuations were in your weight.

01:32:46.960 And it's frankly going to help you identify probably the type of weight that matters more.

01:32:51.900 Obviously there's been a big discussion about the difference between subcutaneous fat and visceral

01:32:55.620 fat. You could have two people that weigh exactly the same and you could even have two people that have

01:32:59.980 a similar mass of fat, but where that fat resides probably tells us much more about their metabolic

01:33:04.740 health. So I don't know. Like I said, if I could be czar of whatever for a day, I just don't see why I

01:33:11.480 wouldn't want someone to have this information continuously in real time, every minute of every

01:33:16.180 day under any physiologic circumstance, because what you learn and not just someone who's got type

01:33:22.940 two diabetes or type one diabetes or gestational diabetes, it's going to alter your baby. It comes

01:33:27.100 back to the airport stories, you know, when we're in these environments that are at our worst. And you learn

01:33:33.080 other really interesting things. I mean, these things, you know, you get a sense of how much even exercise

01:33:38.420 can transiently raise your blood glucose, but what the benefits are after the fact, which is your ability

01:33:44.120 to dispose of glucose goes up significantly.

01:33:46.840 Yeah. I get about a 30 point spike in a workout, but for the rest of the day, my average glucose is about 10

01:33:52.880 points lower. Yeah. It's so amazing to think that such a simple biomarker is glucose, something we

01:33:59.660 take for granted on every chemistry test we do on patients. You know, no one goes to their doctor and

01:34:05.640 doesn't get their glucose measured, but what the utility of that is when you can measure it every

01:34:10.440 five minutes. Kevin, this has been super interesting. I want to be respectful of your time. I know we don't

01:34:16.120 have all day here and you have to get back to work. One of the stories you just told that I love so much

01:34:20.980 that I didn't know was the going to the pharmacy, getting three of these things and poking yourself

01:34:25.540 60 times. To me, that's like, you know, Dexcom is a pretty big public company. Your market cap's

01:34:30.740 like what? $12 billion? $12 billion. Yeah. And, but yet you're still kind of a startup company

01:34:35.520 because that's a startup CEO move. It really is. And I've had to learn to be less startup

01:34:41.940 as we've grown. We've all grown with the company, but that culture. And again, one of the things that I

01:34:47.440 manage as CEO here is we've had to bring in skills and supplement our management team

01:34:53.080 from that startup mentality to the midsize company mentality to whereby we have more structure and

01:34:59.500 systems of a large company. So we're trying to mix all of those things together. We just had our

01:35:05.260 employee survey and it's interesting. I read every comment. It's almost 200 pages of comments. I read

01:35:11.720 everything all the employees write, the good ones, the bad ones, and the funny ones. I will not quote

01:35:16.900 the funny ones, but there, there, there's some pretty hilarious things people say. The most

01:35:22.620 contradicting comments are, we all need to go back to being a startup versus we need more structure to

01:35:30.120 be bigger. And both are true. Both are true. The, the nimbleness of a startup in an industry where

01:35:37.060 you're creating everything is absolutely essential. You've got to remain nimble. You've got to be willing

01:35:42.840 to fail. You've got to be willing to start over. You've got to be willing to do what's right.

01:35:48.320 Conversely, that big company structure where you have systems and, and balances and checks to make

01:35:54.820 sure you're doing the right thing. And you start watching dollars more is a different culture. So

01:36:00.480 we're mixing those two together. We're kind of like a mixed family here. We will have added more

01:36:05.300 than 1500 employees in the last two years. Wow. I didn't realize you were that big.

01:36:10.020 I was employee three, about 300 in 2011. I think we have 2,500 full-time employees and several hundred

01:36:18.180 temps. We do a lot of our manufacturing work temp to hire and bring them in and grow and shrink with

01:36:24.980 volume. There was a temporary workforce, but no, we have that many people. Can I add one more question

01:36:29.140 for your internal survey that I think will produce awesome, interesting results? Every month have a,

01:36:35.280 or every quarter or whatever frequency makes sense, have a tell us your best surprise glucose spike

01:36:40.780 story. Because I mean, honestly, this, the grapes, the stupid chocolate covered M&Ms, whatever. I mean, 0.99

01:36:47.580 I think, I mean, there's going to be the obvious ones. Like I ate a bag of Swedish fish. Look what

01:36:51.860 happened to my blood sugar. But there's these surprise ones either in the velocity of glucose escape

01:36:56.980 or the peak. Because I'm guessing most of your employees are wearing these things, right?

01:37:00.800 No, no, no, no, no. There's too many of us. There's too many of us now. But you know,

01:37:08.340 if we run an internal study under IRB protocol and everything, we will recruit volunteers within

01:37:14.000 the company. We have several people with type 1 diabetes that work here because they can lend us

01:37:19.060 perspective for our patients that we don't have on our own. So that's important to us as well.

01:37:24.440 A lot of it wear them. And a lot of the things that we learn are through those in-house studies

01:37:29.400 and through that in-house wear process. You said something a second ago about

01:37:33.420 failing quickly. You have five kids, right? I have five sons.

01:37:37.740 And just before we were getting ready to start, we were talking about our kids. And you asked me if

01:37:42.240 I wanted some advice on being a parent, to which I couldn't say no. Of course I did. Do you remember

01:37:46.840 the advice you gave me? I do. And I was recounting a story of how things have changed. There's a big gap.

01:37:53.260 My oldest is 15 years older than my youngest. And there's an eight-year gap between number four

01:37:58.140 and number five. But what I told you is we can't be afraid to let our kids fail.

01:38:05.120 I know I'm going to sound horribly mean, but everybody getting a ribbon and everybody getting

01:38:09.240 a trophy. I once told a coach to cut my son so he would learn something. If he wasn't going to play

01:38:16.300 him, go ahead and do that. We learn in life as much from our failures as we do our successes.

01:38:21.720 Sometimes we learn more. And sometimes they motivate us to be great and motivate us to do

01:38:28.020 great things. But we oftentimes as parents shelter our kids so much from that, what appears to be

01:38:34.540 traumatic. And I'd given that advice to somebody who didn't have any children. He came in my office.

01:38:39.500 He goes, give me one bit of advice. And I said, all right, I'll give it to you. Because we just try and

01:38:44.720 manage and regulate everything so much all the time. And look, I will tell you with number five,

01:38:50.020 I did a lot more of that than I did with number one. Number one would tell you that I was very

01:38:55.160 tough on him, but I let him carve his own way. And it's a good thing to let him learn.

01:38:59.460 Last question, Kevin. Is there a failure that you point to in your professional life leading up to

01:39:05.280 where you are today that has taught you more than the others about what you're doing now?

01:39:10.820 Boy, that is a great question. I would tell you one in particular at a corporate level. I'm going

01:39:19.860 to go big picture. I was at one company where I knew exactly what to do. And it was really the

01:39:30.200 only course of action that would have made that company successful. And I presented to my board and

01:39:34.460 they said no. And I then rolled up my sleeves and worked hard for several months and several months

01:39:41.940 and several months. And I still got the answer, no. And I finally left after beating my head up

01:39:49.120 against the wall. And that taught me a lot of lessons. Number one, to be better prepared in how

01:39:55.040 I give my message. And I attribute prayers like that to me. I don't blame the others. To be better

01:40:00.980 prepared on how I deliver my message to be more thoughtful. But if the situation isn't going to

01:40:06.320 work out, and what I ended up doing, and my wife would tell you this, my commute was five miles at

01:40:11.240 the time. It's the most miserable I've ever been in my entire life. She would look at me and say,

01:40:16.660 what on earth are you doing? I'm going to fix this. And we're going to get this. And I couldn't win.

01:40:22.780 So I left. And that was the one time where it felt like I didn't get done what we could have got

01:40:28.340 completed. But I didn't see all the signs on the wall. So I've looked at ways I can be better from

01:40:34.900 that. When I see something strategic like that, I do more homework. We have the people around us do

01:40:40.900 more homework. We're more thoughtful. So that's one of them. We've all had numerous little failures.

01:40:46.900 Sometimes, again, our failures are our biggest blessings. I would say the failure that's been my

01:40:51.820 biggest blessing are all the jobs I didn't take. Because you get opportunities and you get mesmerized

01:40:56.940 with an opportunity. I'm going to go do this. And countless times in my life, those didn't work

01:41:01.640 out. And then lo and behold, an opportunity comes along and it's, wow, I'm so glad that didn't work

01:41:07.800 out. And then you learn so much from that. Now, and I would say the other thing that's important as

01:41:13.700 far as success, you're only as good as those people that are around you. And it becomes clearer and

01:41:19.780 clearer to me the longer I work. This team we have here at Dexcom, they are committed. They work hard.

01:41:25.300 We are all engaged in our business each and every day. I never go to a meeting with another company

01:41:31.460 where their people are as engaged in what they do as ours.

01:41:36.160 Well, that's an interesting point to end on because I do think that Dexcom, obviously I'm

01:41:39.780 very biased. I've been very, very vocal about my bias to Dexcom over the other two, Abbott and

01:41:45.200 Medtronic. But I think you guys have an advantage that on the one hand is a disadvantage. Let's be

01:41:50.360 honest. Diversification, right?

01:41:51.880 Those companies have a million products to fall back on. CGM is but one of a hundred things they

01:41:57.280 do. But that disadvantage is actually your biggest advantage, which is everybody shows up to, and I

01:42:03.260 feel like I know half the staff here. I'm here so often shooting the breeze with you. There's

01:42:07.840 something to be said for, what's the expression? You burn your ships, right?

01:42:10.960 That's right.

01:42:11.280 Yeah, yeah, yeah. You guys have burned your ships. You're all in on CGM.

01:42:14.280 We are.

01:42:15.180 Well, Kevin, I want to thank you very much for your time, for your insights, and for your

01:42:18.760 friendship. It's been amazing to think that we met on that airplane almost three, four years ago.

01:42:24.800 And to this day, we manage to do a good job of not talking watches, this whole thing. Once we turn

01:42:29.240 the recording off, I want to talk to you about the one I'm wearing right now, which I think you're

01:42:32.400 going to love.

01:42:33.140 It is beautiful.

01:42:33.860 Yeah, thank you. I wore it today because I knew I was going to see you.

01:42:36.780 Okay, great. 1.00

01:42:37.280 This is what watch idiots do. We do show and tell. 1.00

01:42:40.520 Does your wife make fun of you as well? 1.00

01:42:42.480 Oh, that's ridiculous. Now, I've learned a good trick, which is you start giving your 0.99

01:42:47.460 wife nice watches, and she now understands and appreciates why a Daytona is a nice watch 1.00

01:42:53.460 and what Paul Newman meant and blah, blah, blah, blah, blah, and all those other things.

01:42:57.560 So now she's a little bit more tolerant of my nonsense.

01:43:01.480 All right.

01:43:02.640 All right, Kevin. Have a great afternoon.

01:43:04.360 Thank you, Peter.

01:43:04.780 Thanks so much for your time.

01:43:05.520 We'll see you.

01:43:08.180 You can find all of this information and more at peteratiamd.com forward slash podcast.

01:43:13.460 There you'll find the show notes, readings, and links related to this episode.

01:43:17.580 You can also find my blog at peteratiamd.com.

01:43:21.040 Maybe the simplest thing to do is to sign up for my subjectively non-lame once a week email

01:43:25.540 where I'll update you on what I've been up to, the most interesting papers I've read,

01:43:29.320 and all things related to longevity, science, performance, sleep, etc.

01:43:33.060 On social, you can find me on Twitter, Instagram, and Facebook, all with the ID peteratiamd.com.

01:43:39.080 But usually Twitter is the best way to reach me to share your questions and comments.

01:43:43.100 Now for the obligatory disclaimer.

01:43:44.680 This podcast is for general informational purposes only and does not constitute the practice of medicine,

01:43:49.860 nursing, or other professional healthcare services, including the giving of medical advice.

01:43:54.980 And note, no doctor-patient relationship is formed.

01:43:57.920 The use of this information and the materials linked to the podcast is at the user's own risk.

01:44:02.720 The content of this podcast is not intended to be a substitute for professional medical advice,

01:44:07.500 diagnoses, or treatment.

01:44:09.200 Users should not disregard or delay in obtaining medical advice for any medical condition they have

01:44:13.780 and should seek the assistance of their healthcare professionals for any such conditions.

01:44:18.860 Lastly, and perhaps most importantly, I take conflicts of interest very seriously.

01:44:23.080 For all of my disclosures, the companies I invest in and or advise,

01:44:26.920 please visit peteratiamd.com forward slash about.

01:44:32.720 Thank you.

The Peter Attia Drive - May 20, 2019

#54 – Kevin Sayer, CEO of Dexcom： Continuous glucose monitors – impact of food, sleep, and stress on glucose, the unmatched power of CGM to drive behavioral change, and the exciting future of CGM

Episode Stats

Summary

Transcript