The hidden health benefits of wearables technology with Michael Snyder

[00:00:02] Right now, while you're listening to this, your body is keeping a record. Your heart rate, your temperature, your breath slowing as you sleep, your blood sugar climbing after lunch and falling again before you notice a thing. Somewhere in you, a virus landed this morning. Your immune system already knows, even if you won't for another three days.

[00:00:28] For most of human history, none of that was written down. The body kept its books in private. You found out something was wrong when it started to hurt, but then whatever it was had a head start. Weeks, sometimes years. We've built a system around the delay, annual checkups, blood draws, measurements, but it's still just a snapshot.

[00:00:51] And a snapshot can't tell you which direction you're moving, only where you stood on one Tuesday in a paper gown. But the record was always there. Every heartbeat, every spike, every night of bad sleep. We just couldn't read it before. Now we can.

[00:01:12] This is TED Tech, a podcast from TED. I'm your host, Sherrell Dorsey. Today, we're welcoming a special guest on the show, Michael Snyder, genome researcher and former chair of genetics at Stanford. Michael gave a talk at this year's TED conference in Vancouver about why he's obsessed with wearables.

[00:01:33] He opened with the diagnosis. Our health care system is broken. Once a year, you go to an office that hasn't fundamentally changed in 40 years. They take a lot of blood. They make a handful of measurements. Then they treat you based on an average of everybody else. In his talk, he called this practicing sick care, not health care. Michael's vision for the future of health care is a personalized one.

[00:02:01] Instead of relying solely on averages and annual checkups, we have a full view of our own bodily functions at a moment's notice. A dashboard. Your body monitored the way your car is monitored. This future is powered by wearables, tracking everything from sleep to stress. And for Michael, that future is already here. Michael wears eight devices a day.

[00:02:28] Four smartwatches, two rings. Even his hearing aids are sensors. They track how much he's talking to other people because isolation is a health risk too. A smartwatch caught his Lyme disease before he had a single symptom. His lab built a system that flags COVID a median of three days early, 80% of the time. And the number one thing that sets off those alerts? Not a virus.

[00:02:59] Workplace stress. He put glucose monitors on people who thought they were fine and found spikes as bad as diabetics who had no idea. He found that rice hits most people harder than ice cream. And with some help from machine learning, he took this data a step further. Finding that the shape of your glucose curve can tell you which subtype of diabetes you have.

[00:03:23] Michael's pitch is that using wearables, we can catch things early and effectively. And it's the cheaper part of medicine. A $50 watch. Something many in the world could wear. And if you combine them with other procedures like genome sequencing and blood biomarkers, with tech like personalized AI and machine learning, we have the tools to make healthcare more transparent than ever. And yes, Michael's done that research too.

[00:03:53] Fresh off the stage in Vancouver, I sat down with Michael to learn more about the possibilities of wearables and how this data could change how we interact with our healthcare system. But first, we're going to take a quick break. Welcome back to TED Tech. I'm here with Michael Snyder. Michael, you describe yourself as medicine's most measured man.

[00:04:18] So I guess tell me at the point of which you decided you could not live without this measurement and monitoring as part of like your life. Yeah, well, I do have two petabytes of data on me, which is a lot, by the way. Most people would have probably less than a millionth of that. Anyway, it started when I moved to Stanford with the idea of bringing big data into better managing health.

[00:04:45] And then I just guess I got more and more sucked into it. I can't pick one particular time. I believe at some level, this is here to stay, period, and should be for all of us. We should all have smart watches and be doing continuous glucose monitoring periodically to track all that. So I think that stuff is indispensable. And I will emphasize, I think it's really important for people to get themselves measured while they're healthy so they know what their baseline is. You have a lot of wearables on right now.

[00:05:14] You have watches, you have rings on, and so you are a living, breathing test subject for yourself. And you were also one of the first people to have your own genome deeply profiled and monitored. What did you discover about yourself or what most surprised you? What did you learn? Yeah, so when we sequenced my genome back in 2010, I guess, we did see that I was at high risk for type 2 diabetes.

[00:05:39] Now, I grew up in rural America, so people died of old age. That's how they did. Yeah. You didn't really know what they died of. And I think in hindsight, there might have been some evidence of diabetes in there, but it really wasn't evident to me. So that was a surprise. I actually had a mutation in a telomerase gene, which could have put me at risk for anemia and things. And we did follow-up experiments. I didn't have anemia, so I wasn't too worried about it. My mom had the same mutation. We sequenced her, too. Yeah, yeah, yeah.

[00:06:07] So the bottom line is we saw those two things that were there. And the diabetes I thought was interesting, but it put me on the alert to look for that as we did deep data profiling on me. And I want you to also walk us through the layers. So you're combining genomics with blood biomarkers and then these wearable signals, which, I mean, I'm a watch wearer, and I like to count my steps and maybe figure out my heartbeat.

[00:06:31] But how did these three things talk to each other to build a picture of your health without you going crazy about maybe I have this disease, maybe I don't. I'm not a doctor. I don't have a medical license, but now I'm trying to diagnose myself. Yeah. Well, on one hand, maybe you could call me crazy because I do go so deeply on me and some of the other people we study. But it did kind of evolve, meaning we sequenced the genome first, and because of the nature of the work, we did these deep, we called them omics profile,

[00:07:01] looking at all my proteins, proteomics, my metabolomics, meaning all my metabolites and so on. We would profile several people to see just how similar or different we were. Then we discovered people are pretty different, more different. There are a lot of just basic questions we didn't know. And then the werewolves were just coming out when we started, maybe a year or two later. And once they came out, we sort of recognized, well, these are pretty powerful health monitors. They're being used as fitness trackers.

[00:07:29] We said we should put them on these folks we're studying and see what's going on. So we're collecting all this data types because we're not sure what data are going to be the most valuable, the most useful. And at the end of the day, genome turns out to uncover a lot of things. Metabolomics uncovers a lot of things. So all your metabolites and the wearables uncover quite a few things. So it really, in the end, you do want the whole picture because it gives you a more complete picture of your health.

[00:07:55] If you just do one thing like the wearables or the omics and say you don't do the genome, you will miss some information. So back then, though, it was clunky. And still to some extent is to analyze all these different data types and interpret them and try and figure out how to combine them. And I think most people, they look at individual data types and then they try to make judgments based on those. Don't really do proper integration, in my opinion.

[00:08:21] But now with AI, of course, it's pretty easy to do this and build dashboards. And our lab's a little bit unusual in that we do basic research and we do proof of principle experiments. But then we spin off companies. I mean, I think that's what academic labs are good at. They're good at, you know, discovery proof of principle. They're really not good at scaling. So these companies can go off and then, you know, take some of these things and scale it.

[00:08:45] Like we have a continuous glucose monitoring company that actually will, you know, order glucose monitors and they have a whole 54 million food database so they can make, you know, see what spikes you, see what doesn't spike you, make all these predictions. And with AI about what you should do, whole behavioral modification. Fascinating because as you're talking, I'm thinking about the broader implications and also what the end goal is. Are we looking to improve medicine through more biomarkers?

[00:09:15] Are we looking to ensure that people have this sense of agency over their own health, right? And can either share more with their medical professionals or be able to advocate for themselves more. You know, obviously living healthier lives. I definitely get moving more because every couple of, you know, minutes I get a stand up notification, right? I hate that stuff. Yeah, yeah, no, I hate it too, but I'm also acutely aware of how much I'm moving my body in a day, you know, based on my wearable.

[00:09:43] And then I can go back and track and see, okay, over the course of a year, here's how often I worked out. Here's how often, you know, how many steps I got. Here's what my resting heart rate is. Here's what my sleep patterns, you know, looked like. I think it's really important to know your trends. Yeah, so that's good. But back to your earlier question, it's all of the above. I think we need people to take control of their health. The physicians today, they're very, very well-meaning, but they only have 15 minutes to spend with you. And that's not good enough for you. But is this going to make us healthier?

[00:10:13] Is more of the information going to make us healthier? Is it going to lead to actual behavior modification? It will because we're all going to have our own personal AI agents who will be tracking our health. And that's the key in today's medicine. You actually don't get measured very often. And nobody's following your trajectories. Even in the current medical system, they measure you. They see whether you're in normal range or not. And if you are, you're fine. But you can, we have a good example in the group of people we've been studying.

[00:10:43] We had one person who was at the low end of normal range, always very consistently. And then at one of his visits, he jumps up to doubles his value, but he's still in the normal range. No physician would, in fact, physicians are supposed to look over the data. No one flagged anything. But he brought it to me and I said, well, that's strange. Yes. Why don't you do another measurement, which he did, and then he was out of range. So it's a trajectory that counts.

[00:11:09] And then he took corrective action, actually, just by lifestyle change, fixed it up. So I think we need to bring trajectories into this whole thing, which is not currently done. And that's way, way more important than absolute values. So, yeah. Well, it's interesting because the continuous monitoring, in some ways, I'm like, okay, this could drive you crazy if something spikes for a second and maybe goes back down. But then I could see the value. And I know on a week-to-week basis or a month-to-month basis versus maybe every six months

[00:11:39] when I'm going to my doctors or something is wrong, you know, to the earlier part of us opening this conversation of I'm only going when something is wrong versus I can nip it in the bud or at least start asking questions a lot sooner. You know, but to your point in terms of getting a bit more... Can I pick on that? Yeah, yeah, yeah. The first thing they ask you when you go to the doctor, too, is when did it start? Yeah. If they had your wearable data, they'd know exactly when it started because your heart rate would have jumped up and your heart rate variability would have dropped. They wouldn't even have to ask you and you'd probably tell them wrong anyway.

[00:12:09] Well, I think I noticed it two weeks ago or something like that. But you could just get exactly what that value is from your wearable data. So more potential accuracy there in the monitoring. Yeah, and it's going to be better diagnostics, too. I think you're going to combine all these data in very useful ways. So I have a question about kind of the noise that's generated by wearables. You know, sometimes you just have like bad sleep. You had a stressful day.

[00:12:34] There's kind of just life stuff that sometimes interferes with our health and it's not necessarily permanent, right? We use smartwatches and rings for, you know, infectious disease detection, respiratory viruses, in my case, also Lyme disease, and the whole red alerting system if things are off. But we tuned it in a way where you get red alerts about every six weeks. We did surveys. People were comfortable with that amount of alerting. And it's a balance between having too many alerts and not being sensitive enough.

[00:13:04] You want to be sensitive enough to catch when you're ill, but not, you know, have this thing going off every day. If you run a marathon, your heart rate's going to be up for the next several days. You're going to get red alerts from that. And so that's easy. Just ignore it. And that's the way a lot of these are. But if you're just sitting around listening to boring Mike Snyder and you're getting red alerts, you're either physically or mentally stressed, hopefully not from Mike Snyder.

[00:13:28] But the ones where it's not obvious what's going on, you might think about that and, you know, adjust accordingly. But it takes a longer alert, by the way. It takes more than just a few hours to get a red alert. It would have to be very, very high for, say, half a day or lower for, say, a day and a half kind of thing to get a red alert. So they have to be a substantial signal. And again, the easy to contextualize ones, you basically ignore.

[00:13:58] Even we think mental health is a great area. We're actually pushing on this very hard with our wearables now. We don't quantify mental health very well. We use surveys, right? Yeah. Which is pretty also archaic when you get right down to it. Because, you know, when you give someone medicine, how do you tell if they're improving? Well, you give them surveys every few months, right? That's a terrible. Yeah. That's the extent of the measurement. Yeah. It's subjective. It takes forever.

[00:14:26] And so what you want are really quantitative measurements. And I think biometrics like wearables. And we're also bringing in microsampling for that to be able to get good quantitative biomarkers, digital or biochemical, for mental health. And I think that's a great way to, A, diagnose it because there are also going to be subtypes of depression. We already know there are. Anxiety, same thing. PTSD.

[00:14:50] All these things we should be able to get biomarkers for, quantify it, subtype it with these biomarkers, and then treat it. And I think this is where mental health field, more than any, needs proper biomarkers. The potential is clearly infinite and growing. And especially as I think more folks start to embrace wearables, like I feel like majority of my friends and colleagues have the rings, right?

[00:15:19] Like I have not advanced beyond the watch myself. But I'm pretty sure I'll be peer pressured into get a ring at some point. Don't forget to see Jim. I really think they're super valuable. Okay. You're going to have to hold me to that. So there was this New York Times reporter who, you know, tried a CGM on and called Mepi and said, Mike, I thought I was eating the healthiest lunch every day. I had salmon on salad. What could be healthier than that? Then he puts on a CGM and it turns out it spikes through the roof.

[00:15:49] And guess what it was due to? The salad dressing? You got it. So in hindsight, and this happens all the time when you wear a CGM. In hindsight, it should have been obvious and probably could have tasted it too. But what you're doing at the time, you don't necessarily realize it. And you wear those things like, for me, it was a pulled pork sandwich. Okay. I ate a pulled pork sandwich. I think I was diagnosed with diabetes. But my glucose shot through the roof in the, you know, 300s and stuff.

[00:16:19] It was really, really high. And I'm watching this in real time, but somebody's standing next to me. He said, wow, look at my glucose go up. I just ate this pulled pork sandwich. He goes, Mike, everybody knows pulled pork sandwiches have sugar. I didn't know it. Again, a lot of the stuff is obvious. Yeah, the barbecue sauce. Yeah. Well, it didn't even have barbecue sauce, but it did have something in there that was definitely sugary. Yeah. And so, bingo. Now, the thing that I'm...

[00:16:46] In real time, and that's what's powerful about CGM, because they, and same with the wearables, they're giving you back this data in real time. So, you can see your, literally, eat a McDonald's shake. Yeah. And you'll watch your glucose go up in real time. Well, I'm going to assume if I have a McDonald's shake, it's going to go up, for sure. I'm going to bet on that, you know? Well, it's a safe bet. I'm just going to have some kale juice after, and it'll, you know, we'll balance it out.

[00:17:19] Self-monitoring gives you this sense of empowerment and agency. But there's also this side of data ownership and understanding who now owns the data, because there's obviously wide open opportunity for research, for making all aspects of this industry better and smarter. But at the same time, I don't want my data necessarily used in ways that I haven't given permission to. Right. So, a couple comments there.

[00:17:48] You own your own data. Now, whether you can always get access to it is not so easy, especially for EHR data. But for wearable data, you can get it. Now, often you do have to sign that you give permission to use data for research to the companies. I am a believer we should be sharing data, because if we all share data, then we can actually build better health management systems from the shared data. It's hard to do that just on one person.

[00:18:15] So, I am a believer in data sharing. And I think we should have rules in place. And another comment is to actually protect you. So, I think, you know, we're in a wealthy society. So, my opinion, the purpose of government is to do things for citizens they don't do for themselves. And one basic, you know, right, in my opinion, is to have some minimal level of health care. And I think we should all have that in a wealthy society.

[00:18:42] And so, you shouldn't be discriminated against by your own data. And it's true for employment and for health plans. That's true for genetic data. We should have that in plan for other things as well, I would argue, so that we can better manage people's health. But I'm going to flip this whole thing around, which is that we actually, you should get paid if you share your data is one way to look at this. And quite frankly, you should get a discount in your health plan if you wear a smartwatch

[00:19:12] because you're going to take better care of yourself. You should get a discount if you get your genome sequenced because you're going to take better care of yourself. You're going to know what you're risked for. And then you'll get surveyed for those things. You know, if you have a BRCA mutation, you have a couple options. One is you could do prophylaxis surgery or you could just get monitored more. You should do one of at least those things. Right, right. And in BRCA meaning the breast cancer gene. Yeah, the BRCA genes put women at high risk for breast and ovarian cancer.

[00:19:41] But men are at high risk for breast cancer too. That may not always be realized. And so with that information, you should get screened more and then you'd catch that early. But if you catch it late, it's very hard to reverse. So I want to think about like 20 years in the future, right? Like we've solved some of the systemic challenges of health care access and insurance is paying for wearables, right? And insurance is paying you to like get the tests done or what have you. So we're going to pretend like the barriers have been reduced significantly.

[00:20:10] That's the number one barrier now. Who pays, right? Nobody pays to keep you healthy. We've got to solve that. So let's say in 20 years, what does a doctor's visit potentially look like? And does the annual physical exam even exist anymore? Is it still relevant? I think a lot of this could all be done at home. So the wearables are the first things that'll move into this remote monitoring, I think. And the blood sampling too, the micro sampling or other versions thereof. That can all be done at home.

[00:20:38] You collect at home, mail it in, get back your results. And so the power of this is, you know, I don't know how often you go to the doctor when you're healthy. Most people rarely go, as we talked about. And, you know, they make recommendations when you hit 40, go every two years or something like this. And even still, people have a hard time going. Every two years is enough. If you have aggressive cancer, you could be dead in two years. So we actually need to do much more frequent health monitoring.

[00:21:08] The nice thing about the wearables, it's all passive. Meaning you put this on, the data come in, just have to keep it charged. Pretty easy. And I'm comparing it, but I'm a researcher. But as a user, I would just have one device or maybe two at tops. Let the data come in. I'll have a dashboard. My AI agent will do a readout. And this isn't 10, 20 years away. This is the wearable stuff can all be there, quite frankly, now.

[00:21:39] And the AI agents are emerging now too. So this is not 20 years away. Let's say two years tops. And what are the AI agents doing? It's basically taking all your wearable data. If it's built properly, it's watching it. So that when you do go off, like I mentioned for the real-time alerting for infectious disease, when it does go off or it's an AFib event, it flags it. And it flags it just for me? Or is it also flagging it to my doctor?

[00:22:09] That's up to you to decide. I believe, yes, it should share with your doctor. Because that's their job is to help make you, keep you healthy. This information, it's powerful because it's coming all the time. And then we can really use this in beneficial ways. But it's up to you to share. And again, I think it can keep us healthy because it is following your trajectories.

[00:22:31] And I do find it to be just interesting to learn a little bit more about yourself and your own health and then be able to at least make some kind of informed decision. Or at least when I go to the doctor's office. I want to come back to that first point. Right now, you only have 15 minutes with your doctor. And that's just not enough. That's why you have to take agency for your own health. Because you're going to learn more.

[00:22:53] Or, you know, just a fun example, even before all the wearables and things, when people had things go wrong in my family, they would call me up and say, Mike, so-and-so has this. What do you think is going on? What do I know? I'm a PhD. I don't know any of this stuff. Now, I do have access to good doctors. I call them up. But these days, I just put this stuff in some of the AI algorithms we have. And it's kind of interesting what they come up with. And you have to view these things. In fact, I want to go back to this point.

[00:23:21] As what they're doing is they're flagging things that might be off. And that's why I don't think it's a big deal. You don't drive your car around without a dashboard, right? Why would you? You'd run out of gas or you'd speed or whatever. And then your, or your injury would break down. But you have a dashboard to tell you what's going on and when things are off. Even if the check engine light doesn't tell you exactly what's off, you know something's off. And then you bring your car in. And that's what we should be doing with health. These flags are just flags.

[00:23:50] They're not necessarily, oh, that was, maybe it was you had some transient thing that happened and it went away. But that's okay. Get it checked. Yeah, absolutely. Especially because you mentioned the 15 minutes with the doctor. And for me, I always make the assumption that the efficiencies that technology should provide us should give us more time. And so, sure, I want the AI agent to be additive to what my medical professional can do, right?

[00:24:20] And maybe they have more of my data on the back end so they have a longer history. And they can point to these very specific moments or spikes or what have you. But then once I get into that exam room, I want to be able to sit with my doctor longer and know that my doctor now has more time because a lot of the things have already kind of been either addressed or documented. And so now they're coming in. They have all this knowledge. They have some potential recommendations. Maybe the agent has kind of put forth some ideas.

[00:24:49] And now they can kind of fully focus and service me because they're not working on just like all the paperwork. I'm with you 100%. Yeah. All the things a nurse does before the doctor comes in, most of that can be done before you even show up at the door. So from your wearable data, your other data, it's all there. Yeah. Even your blood pressure and things. And your AI is just poised, right? It's done a first pass, like you say. And then you get maybe some additional measurements from the physician.

[00:25:15] And so they actually could think of that 30-minute appointment at least 25 minutes with you now instead of 15. Absolutely. You've given me so much to think about. I'm excited for our listeners to hear your talk and to continue to follow your work. Michael, thank you so much for joining us. Thanks for having me. Here's what stays with me from my conversation with Michael.

[00:25:38] 15 minutes with a doctor once a year is not an effective way to understand the human body. I'm fascinated by Michael's thoughts on health data too. He sees the positive potential of sharing your data. And I'm glad he started thinking of ways to encourage people to learn about the data generated by our own bodies. Get a discount on your health plan if you wear a smartwatch. Get paid for your data.

[00:26:05] Don't get discriminated against for it. The last one isn't a feature. It's a hope. Right now, protection against genetic discrimination in this country is narrow. It doesn't cover life insurance. It doesn't cover long-term care. And wearable data isn't genetic data. It's a different category. And the rules for it are barely written. So when Michael says you should get a discount for wearing a watch, I think of the next question.

[00:26:33] What happens to the people who don't? A discount for some is often a penalty for everyone else. And the people least likely to have the ring, the patch, the eight devices, those are the same people the system already fails. I'm also thinking about the infrastructure needed to support this vision. Michael has a lab, a team, and access to some of the best physicians in the country. When his data flags something, he has somewhere to take it.

[00:27:02] But a lot of people get the alert and nothing else. A red light on a dashboard is only useful if there is a mechanic. None of this means he's wrong. In fact, he might be early. But early is exactly where the rules get written. And right now, the devices are shipping faster than the protections. Michael wants you to take agency over your own health. And I want that too. The question is how much agency wearables can truly grant you.

[00:27:31] And what we need to do to make it truly equitable for everyone. TED Tech is a podcast from TED. This episode was produced by Rahima Nasa. Our editor is Alejandra Salazar. And the show is fact-checked by Julia Dickerson. Special thanks to Constanza, Gallardo, Daniela, Belarreso, Maria Ladias,

[00:27:59] Tanzika Sangmanivan, and Roxanne Hilash. If you're enjoying the show, make sure to subscribe and leave us a review so other people can find us too. I'm Sherelle Dorsey. Let's keep digging into the future. Join me next week for more.