← The Neuro Experience
Loading summary
Transcript94 lines
- [00:00]
Renee Dehan
I love wearable data because it's a little bit different from blood. Blood you measure probably once every couple of months or twice a year. You know, however much, however often you want to do, do that. And you have to, you know, go to a phlebotomist and have your blood drawn or maybe they come to your house or something like that. But it's a process. Of course, fitness trackers are wonderful because you just wear them and they passively collect information. Women between 40 and 50 with LDL cholesterol levels of under 110, they are better off when it comes to cardiovascular disease risk or something like that. Right?
- [00:33]
Louisa
Yeah.
- [00:34]
Renee Dehan
If we think about machine learning, what that requires is a fair amount of data. So that's why we hear a lot about it right now, because we are, hands down in the data age, data collection age. We are absolutely drowning and swimming in data. Very good at collecting it, not very good at making sense of it. But machine learning helps. If there aren't enough women included in the study to really, truly evaluate how that drug works in women, you know, then that's why women are more likely to have adverse events. Like, that's horrible.
- [01:04]
Louisa
Extremely underrepresented women. And not just that. I think from a scientific standpoint, it may cause people may be like, well, why are women unrepresented? Like in terms of this. It's sometimes it's because women menstruate and you have to get them at timestamps. It's like we have to get them at this time and it's just very hard to track them on a. So it's just easier to get a guy. Really? That's the premise of it.
- [01:32]
Renee Dehan
Absolutely, absolutely. And that's, that's a big, that's a big factor. Is like women have a confounding effect of cycling.
- [01:39]
Podcast Host
Before we move further, this is a quick word regarding InsideTracker. So InsideTracker is a personalized nutrition platform that analyzes data from your blood and DNA to help you better understand your body and help you reach your health goals. I've long been a believer in getting regular blood work done for. For the simple reason, and that is that many of the factors that impact your immediate and long term health can only be assessed with a quality blood test. What's unique about InsideTracker is that while there are a lot of different tests out there for hormones and metabolic factors, InsideTracker actually does something really different. You get the numbers back in terms of your levels and they also give you a very clear directive in terms of lifestyle nutrition and supplementation that can help you bring those values into the ranges that are best for you and your health goals. And that's very different to a lot of other programs out there where you get all the information, but they don't really help you understand that information. Whereas InsideTracker makes it very easy to understand and very actionable based on the very easy to use dashboard. So if you'd like to try InsideTracker, you can visit InsideTracker.com Louisa to get 20% off any of the InsideTracker plans. I just had my blood work done. I had over 70 different biomarkers tested. So if you want to do this, just use the code Louisa20 at checkout for 20% off.
- [03:26]
Louisa
Renee Dehan, welcome to the Neuro Experience podcast. I'm very excited to have you here because we are talking all about AI and health and you are the best person to be talking to this subject about. You are the vice president of Science and Artificial Intelligence at Inside Tracker where you're building, correct me if I'm wrong, the world's largest data set of blood and DNA to create evidence based solutions that are both simple and actionable. So let's dive right in by first understanding what are data sets, what is blood, biomarkers and DNA.
- [04:05]
Renee Dehan
Awesome. Thank you. It is a delight to be here. I really appreciate the invite and I'm very happy to talk about these subjects. So InsideTracker can be boiled down into kind of two main steps. And the first step is measurement, right? We want to take an individual and really assess where their health state is at that moment in time. And then the second step is analysis of that to give personalized insights and recommendations about your health and ways in which you can improve your health. And our goal here is really we want to be able to do this for everyone. It's about improving and optimizing health span even more so than longevity. Of course, not everybody wants to live a long time if those years aren't great. You know this, I know you talk about this a lot as well. So what we're trying to do is optimize out of all of the years you're going to live. You want those years to be healthy. You want to be able to do the things that you love with your friends and your family and walk your dog and whatever it is. So we are aiming to do that through essentially techniques that will prevent disease, prevent these diseases of aging like metabolic disorders, diabetes, cardiovascular disorders, Alzheimer's, things like that. And the other piece here is we want to give people the tools to take control over their own life and their own health state, to be able to improve it. So we don't make recommendations about what statin you should be on. You have the medical system for that. But where we fall is somewhere else. So if you're, let's say you're not, you have elevated ldl, it's not super elevated. Your doctor doesn't think you should be on a statin, but you do want to get it down, then what are the exact ways in which you can change your life to try and actually reduce your LDL cholesterol levels? So if I unpack both of those two steps, the measurement and the analysis step. Measurement means we look at actually lots of different kinds of biomarkers. This could be. It started all with blood. The company started with blood. Right now we look at up to 48 different blood based biomarkers. These blood based biomarkers in some instances are things that are standard, routine lab values that you get done at the doctor's office. In other cases, they are things that a doctor may order but is much less likely to order. So for example, we added APOB and insulin recently, which is great. I've never had those measured.
- [06:51]
Louisa
Well, APOB is a much stronger predictor than LDL of cardiovascular disease risk.
- [07:00]
Renee Dehan
Yes, that was a huge one. And we're kind of waiting for the medical field to catch up. Yeah, yeah. And I can say that I have been, I've actually seen a women's cardiologist and they recommended having the APOB test, but it wasn't really done as kind of a normal, standard, routine thing. But it should be like, I want to see PCP offices off offer this because of exactly what you said it is. It's a more direct measure of the number of atherogenic particles that you have in your blood. And so it's a great complement to ldl, hdl, triglycerides, things like that. We measure all different blood biomarkers that look at cardiovascular health, metabolic health, gut health, cognitive health, sleep health, anything like that, liver enzymes, et cetera. But in addition to that, we also will measure DNA. You can have your genome sequenced. We will understand what is your baseline, what are you showing up to the equation with that you've been given by your parents. And then we will also measure data from fitness trackers. So if folks have a fitness tracker, then we'll take in data from that, which is awesome. I love fitness tracker data.
- [08:20]
Louisa
Like Oura Ring, for example.
- [08:21]
Renee Dehan
Exactly.
- [08:22]
Louisa
Oh, wow, that's amazing.
- [08:24]
Renee Dehan
Yes. And we just Added OURA relatively recently, but currently we'll incorporate data from Fitbit, Apple Health and Garmin. And we'll add more types of wearables as we go. I love wearable data because it's a little bit different from blood. Blood you measure probably once every couple of months or twice a year. However often you want to do that and you have to go to a phlebotomist and have your blood drawn or maybe they come to your house or something like that. But it's a process. Of course, fitness trackers are wonderful because you just wear them and they passively collect information. Then the last piece of data that we collect currently is something that we just call user generated health data. And that might be survey questions, right? Like how often, how many times a week do you eat vegetables? Or what is your smoking status? Basically anything that is provided by the patient or the customer about their health or their life. It could even be information like what is your zip code? You know, because that might tell you on average how much sunlight you are exposed to in a year.
- [09:31]
Louisa
Right?
- [09:32]
Renee Dehan
So we also take in all that information as well and incorporate it. And what that does is you do get that high definition picture of your health through all of these different methods of measuring elements of your health. And then you can take those measurements over time. Because of course, as you age, things change. As you exercise, let's say, take on a new exercise program, things change. You want to capture that. If you make changes to your dietary habits, you want to capture that as well. It's not a one and done. Genomics always is. If you have your DNA measured, that's not really going to change. You only have to have that done once. But we measure all of the other things throughout time. Then we essentially, at the very highest level, we take all of those measurements and we put them into a machine. This machine does a couple of different things. So first of all, it will analyze that data and say, okay, well what happened today compared to three months ago? Or can we potentially detect changes in your sleeping or resting heart rate patterns or something like that? But we also, in that machine have a lot of knowledge about different lifestyle interventions that can affect the biomarkers that you have. So we look at your biomarkers and we say, okay, how many of them are optimized for you for your sex and your age? How many are optimized and for also.
- [11:07]
Louisa
What you're trying to do? Because I can imagine if a woman who is wanting to optimize for pregnancy, for example, maybe in a different position to a woman who is not trying to optimize for fertility and pregnancy.
- [11:20]
Renee Dehan
Absolutely, absolutely. And actually currently we do not support really any pregnancy specific or fertility specific recommendations. Although we talk about it. Yeah, yeah, but absolutely. Like it's. We would love to be able to do that because there are different needs if you're trying to conceive than if you're not. There are different needs if you are in the perimenopausal state or if you are post menopause compared to earlier on, if you are trying to run an ultramarathon and you have low iron. Right. Like really? Yes, absolutely. So your goals and what you're trying to achieve out of it matter. But we look at all of these biomarkers and identify which ones are optimized, which ones are in a clinically normal range. They may not be optimized, but a clinically normal range. Because inside tracker within a clinically normal range for biomarker will say, okay, that's great, but you could do even better and be in this optimal zone because it's been shown that women between 40 and 50 with LDL cholesterol levels of under 110, they are better off when it comes to cardiovascular disease risk or something like that. Then you may be in, you know, outside of the clinically normal range. So we'll take a look at especially the ones that are outside the clinically normal range and then the ones that are just not optimized and say, what can we do to help this person get into the optimal range for as many biomarkers as possible? And what that does is then it searches this big knowledge base of information that has been curated and entered into our system by subject matter experts like registered dietitians and exercise physiologists and people who really truly understand the literature and the articles that they're reading and all of that. What comes out the other side are interpretations of your data, like, what does it mean if this biomarker is out of range? What does this mean for me? What does it mean for disease risk? What does it mean? And then we say, what should you do about it? Is there any science that will show me how I can lower my cholesterol without using drugs? Is there a fitness program that I can undertake? What is that fitness program? What is the dosage of that fitness program? So the goal here is to use science and evidence, look at everybody's bio, look at your individual biomarkers, and then provide personalized recommendations for you that have been shown to be effective and published in Other clinical studies.
- [14:05]
Louisa
Yes. And that's probably the biggest and probably I can imagine something that must be very hard for you is like, what to do with the data once you get it. Cause I imagine at scale you're getting thousands of data sets from individuals all over the states currently. And it's like, okay, we've got all this blood work, we've got all this data, what do we do with it?
- [14:29]
Renee Dehan
How do we. Yes, how do we, how do we make sense of it? And yeah, we take, you know, almost 50 blood biomarkers. The data that you capture from physio trackers is huge, right? I mean, you're getting resting heart rate measurements on the second VO2 max every time you go out for a run. Sleep data broken down by sleep cycles, REM versus deep and. Yes. What do you do with all of that information? And what we're essentially trying to do is recreate if you had a team of people around you supporting you, let's say a trainer and a nutritionist and somebody with experience in sleep and meditation, we'll call recovery and a physician. If you had that team available to look at your data every time you got a blood test or every time your fitness tracker captured something and interpret it for you, that would be amazing.
- [15:31]
Louisa
That is like true, like optimization.
- [15:34]
Renee Dehan
Yeah.
- [15:34]
Louisa
Like, that's like having a board of directors for your health. Really?
- [15:39]
Renee Dehan
That's exactly what it is. And wouldn't that be awesome?
- [15:41]
Louisa
Oh, that's my dream. I used to think, I used to say to my mum, like years ago, like 10 years ago, I'm like, mom, how does the queen stay so healthy? And she used to say to me, well, because. I'm like, why? Why in the royal family, how are they always living to like 100 and beyond? And she's like, well, she's like, honey, they've probably got like a board of directors. And I used to think, even 20 years ago, I used to, that's my dream. Like, I was thinking about this years and years ago. Yeah, that's like the ultimate, right?
- [16:08]
Renee Dehan
It is. And that would be. And you reminded me, I once worked with this company that did exactly that for high net worth cancer patients. They literally, these are people that had kind of very rare cancers. Right. And also a lot of money, which is important here. And they would assemble these tumor boards of doctors from all over the world, they would fly them in and then we would do kind of like a research grade analysis of, you know, their rna, which wasn't normally done at that time. And it blew my mind.
- [16:43]
Louisa
Does that company still exist.
- [16:45]
Renee Dehan
I think that they have morphed. I don't think that they do that program anymore. But I do think, I mean, obviously you have to be super wealthy to afford that, right? So that's, that's great. But not everybody can afford that, obviously. So really, like what, you know, what I've tried to do my whole career is to take something amazing like that and build a technology that can try to think like that group. Right. You know, try to do that at scale for less money. Right. And that's essentially what InsideTracker is trying to do with the, with the system and the platform that we have that I cannot take credit for building. I only got there a year and a half ago. So it was built by the founder, Gil Blander and a lot of other really wonderful, fantastic scientists along the way that built a very beautiful system to be able to kind of take all of that complex data and come out with interpretations on the other end of it.
- [17:43]
Louisa
Can you describe why it's so important for us to marry blood work with DNA? Because from my understanding, and this is what I see at neuro athletics, if we have somebody and they've got, let's just say we just do blood work and we see that our 42 year old male has an elevated homocysteine. And by elevated, I'm saying maybe he's coming back with a reading of 12 in our books that's elevated to us. We try and actually get a below 8. But let's just say we work on. And by the way, for everybody listening, it's just a protein that builds up and it can be correlated with cardiovascular disease and dementia. So let's just say this just does not, you know, we try and, you know, we get him to take B vitamins, for example, to bring it down and nothing is happening. But we still don't know. We can then get his DNA checked and if he has the MTF HR gene. Did I say it right? Then we know, okay, there is a reason why his homocysteine is very high. That's one reason why I believe, you know, we need to marry the two. But can you describe to me why it's just so important to include both of them?
- [18:56]
Renee Dehan
Yeah, no, I love that and I love that you're doing that for, you know, for your client. It makes a lot of sense. Right? You want to understand what, what are you bringing to your, to the table. So that might mean if we take VO2 max, for example, the upper end of your VO2 max is, is probably in part designated by your genetics. Some people are just going to have a better VO2 max than somebody else, even given all the same exposures. Right. Because their genes create a system that. That does so. Right. That doesn't mean that you can't change your VO2 max if yours is low, but it does mean that eventually you may hit a ceiling somewhere down the line, I think. Another. Another example. And, you know, this is rather personal to me, but my. My cholesterol has always been kind of high ever since I was in my early 20s. Right. You know that. Like, borderline high. And then, you know, you walk around and you talk to people and they're like, yeah, my cholesterol is like, 80. Wow. How. I have no concept of that.
- [20:07]
Louisa
Just to be clear, when you say your cholesterol has been high, you know, you're talking specifically ldl.
- [20:12]
Renee Dehan
Yes, yes, yes. Okay. Thank you for. For that. Yes. Specifically my LDL cholesterol has always been in that kind of borderline high range and pretty resistant to lifestyle changes, actually. But it kind of blew my mind. So recently I was looking at my dad's. My dad's labs from the doctors, and his cholesterol just. It's so much better. Like, all of his cholesterol parameters, like HDL and LDL and triglycerides. Actually, triglycerides are about the same, but in his diet is so much worse than mine. And, you know, I just was like, I don't think that this is all me. Right? I don't think it's all the choices that I'm making. And, you know, my mom always had high cholesterol. My grandmother always had high cholesterol. My grandfather had. So I inherited my mom's pattern of cholesterol. Synthesis, production, you know, metabolism, whatever it is. You know, I don't know exactly what it is, but if you look at my DNA, there are little secrets in there that say, yeah, you are at an elevated risk of having high cholesterol. And why did that make me feel better?
- [21:27]
Louisa
A.
- [21:28]
Renee Dehan
It made me feel better because I went on a statin. Then I was like, okay, I can't eat my way out of this problem. I cannot exercise my way out of this problem. I know that I am really going to benefit from taking a medication in this case. And it helped me kind of like, release actually a lot of guilt, like, oh, I should be working harder, or I should lower my saturated fat intake. And, you know, sure, I should do all of those things, but I knew that in order to really dramatically change it, I'D have to make some very unpleasant lifestyle changes that I didn't want to do and weren't even necessarily something that I could sustain for the rest of my life. So it was good to know that it, you know, in my case, my high cholesterol, My high LDL cholesterol was due, in part, and my elevated apob. I should say, was due, in part to genetics. Yes. And not all of it, but some of it. So it's just. It's good to know that because we don't want somebody bashing their head against the wall trying to do something that, frankly, it's not going to pan out for them. At the same time, let's say I did not have an elevated risk of having elevated LDL cholesterol. Let's say that was the case. Then I would know more concretely that dietary and lifestyle interventions absolutely would be more pliable for me.
- [22:46]
Louisa
Yeah. And one of the biggest things that we're finding across our board is stress is extremely correlated with ldl, which you never think you're like, okay, let's just cut the fat. Really, people, let's just cut the fat. We know that exercise interventions don't really have an effect on serum levels. So it's like, can we please just manage the stress in your life? What? I.
- [23:10]
Renee Dehan
How do you do that?
- [23:11]
Louisa
Yeah, Well, I mean, if you're following me on Instagram, you probably know what has occurred in my apartment right now. And I'll tell you, because you may not be following me, I recently got a cold plunge in my apartment. It was not meant to be in my apartment. It was meant to be outside. I have a beautiful balcony, and management was like, yeah, that's not going outside on the balcony. I'm like, all right. Next to the couch, it goes. So it now is like a water feature in the bedroom. So now I'm forced to have a cold bath every single day because it's there. So, you know, that's really doing a lot for my. I would say stress, or it's doing a lot for my cortisol and inflammation, if you will. Yeah, I love it. So I'm directly impacting stress from the cold standpoint. Okay, I want to shift gears now and talk about AI and how AI is transforming the way blood work is analyzed. So can you talk to us about AI? Because when I. When I talk about AI, people are probably thinking, oh, chatgpt, you know, but we're not really. So I'd love to talk about AI. And if you're utilizing that.
- [24:18]
Renee Dehan
Yeah, yeah, Absolutely. And you know, chat. ChatGPT is one fun, interesting, maybe scary type of AI, but probably the other types of AI that we use are. They're more heavily utilized for research, the type of research that we do or implemented, at least in the InsideTracker platform. So I bucket the two AI types into different categories. One is data driven AI, and that's something like machine learning, Right. So that is predicting what movie you should watch based on your prior history. Another type of AI is called knowledge representation and reasoning. It's a lot of syllables, but basically that is another type of AI that nobody talks about. Everybody talks about machine learning. You don't hear as much about knowledge representation and reasoning. But both of these AI subtypes have been around for decades and decades and decades. And the second type of is really what we were talking about, where we might, let's say, gather all of the important results that have ever been published about the impact of dietary oatmeal on LDL cholesterol or APOB levels or something like that, where you extract that from a large body of literature and you put it into a computable format. And that's the knowledge representation part. Then the reasoning part says, all right, given all of this stuff that we just put in there, how do you make sense of it? Right. So that's really kind of teaching the AI how to think like a subject matter expert would. So a lot of logic engines, which is the type of engine, one type of engine that we have at InsideTracker falls under that category. So they're both used for different things, sometimes overlapping. But if we think about machine learning, what that requires is a fair amount of data. That's why we hear a lot about it right now, because we are, hands down, in the data age, data collection age. We are absolutely drowning and swimming in data. Very good at collecting it, not very good at making sense of it. But machine learning helps. I think the classic example is an AI that you can feed it a picture of a cat or a dog and it will tell you if it's a cat or a dog. Right.
- [26:50]
Louisa
So I'm waiting for ChatGPT to be able to do that. I don't know if it can't do it.
- [26:55]
Renee Dehan
Yeah, interesting.
- [26:56]
Louisa
Yeah, yeah.
- [26:58]
Renee Dehan
I guess you can't upload a. I.
- [26:59]
Louisa
Don'T think you can upload right now. I don't think you can upload a photo and say, describe what this image is, because I could see that that would have massive implications for students. You know, I can imagine if I was back at School. And I'd be like, just, hey, chat cheap. Like, describe this diagram to me and what it means and write a synopsis on it and argue this point. That would be amazing.
- [27:21]
Renee Dehan
Ye.
- [27:22]
Louisa
I'm sure we're going in that direction for sure, which is pretty much going to create dumber people, right?
- [27:27]
Renee Dehan
It's hilarious. It's funny. Say that, you know, movie Idiocracy.
- [27:31]
Louisa
Yeah, yeah, yeah, yeah, totally.
- [27:33]
Renee Dehan
I feel like the first time I tried TikTok, I was like, this is the road to Idiocracy. It's, like, paved in gold and glitter. And I was hooked and, like, did not get off my phone for seven hours when I first got on TikTok. It's like, it knows me.
- [27:46]
Louisa
Yes.
- [27:47]
Renee Dehan
But, yeah, that's. That is an algorithm that's actually figuring out where my attention is going, and it is optimizing for exactly that. So it shows me a bunch of different videos and figures out, like, okay, well, Renee scrolled way, you know, right by, you know, video A, but she really kind of paused on video B, and then she actually hearted it, so she really must like it. And then it keeps track of all of that information. And then we'll show you more of things like video bio. So that's exactly what it's doing. And it needs to look at a lot of people's behavior, their user behavior, for example, to make those accurate predictions. And so the way that we can use this type of technology in, let's say, health, wellness, and medicine is we can ask it to make predictions based on a lot of data. So an example might be to figure out whether or not what your risk is of having a cardiovascular event in the next 10 years. You can give it. Let's say it measures 100 different blood biomarkers in, say, 1,000 people. And I'll say, okay, we have 100 biomarkers measured in 1,000 people 10 years ago. Then for the next 10 years, they measure whether or not anybody in that group has a cardiovascular event. If they do, they get a one. If they don't, they get a zero. Let's say 300 people have a cardiovascular event. 700 people do not have a cardiovascular event. Machine learning algorithm is going to say, all right, well, what's different about the 300 people compared to the group of 700 without the event? And it'll look at those hundred different biomarkers and it will figure. It'll parse out really truly what is different about those individuals. And then you can give them another person. You can roll in there, give them your 100 biomarkers. And the algorithm will say, okay, yeah, you have this likelihood of having a cardiovascular event in the next 10 years. So it learned from those thousand people and their information and their data. And so it can handle kind of a new scenario or a new person or a new patient walking in. So we can do things.
- [30:12]
Louisa
Extremely intelligent.
- [30:14]
Renee Dehan
Yeah, yeah. It's very. It's just very good at pattern finding. Right. Like, that's what it's essentially doing is it's deeper and more sophisticated ways of pattern finding.
- [30:24]
Louisa
Yeah. And how many people. I mean, I don't know if you can share this information, but, like, how many data sets do you have currently that you're working off to?
- [30:32]
Renee Dehan
So our. Yeah, I actually don't know if technically I can share it, but we have tens and tens of thousands of people like, that have been collected over the last 14 years or so. Wow, that's a lot of information.
- [30:47]
Louisa
That's great then. So in terms of, like, what you get, because just for everybody listening, I'm getting my blood work checked with InsideTracker, which I'm so excited about, by the way, because I'm also getting DNA, which. Which I've actually never done. And I speak about APOE risk, you know, in terms of Alzheimer's disease and dementia. So I think it'd be very interesting for me to see and marry the two up as well and go through the process. I'm extremely. I wouldn't say reliant, but I am, like, married to my Oura ring. I love Oura too. Yeah. I have a great resting heart rate. 44, 45.
- [31:24]
Renee Dehan
Wow.
- [31:24]
Louisa
Yeah. And I have a. My HRV is around 180, so.
- [31:29]
Renee Dehan
My God.
- [31:30]
Louisa
Yeah. If I have one glass of wine, it plummets to like, 70. It's if, like, it's like, you know, and my resting heart rate goes up to about like 54. So it's like signaling like. Louise, are you dying? But the thing is, and I want to know your opinion on this, sometimes I get stressed or anxious about all of the data that I do have coming in. As you can see on the table here, I've got my apple, I've got the Ultra watch, because I am a runner, so, you know, of course I had to have the Ultrawatch, but I've also. I'm an ex triathlete, so I was raised on Garmin, so I also wear Garmin. And then I'm sleeping on a performance mattress, which every morning it's giving me different. It's giving me a different reading of my REM and deep sleep compared to everything else, and I'm like, which one do I trust? And so I get stressed and anxious about it.
- [32:24]
Renee Dehan
And you are actually, you're hitting on one of the big technical challenges that we with, let's just say, like, extant wearable devices will include the mattress in that equation. Everybody has a slightly different algorithm for how they calculate resting heart rate, how they calculate VO2 max estimates, how they measure sleep, and do the estimates for REM and deep. And it's completely. You know, they're different. So I, too, I've definitely worn Aura, an Apple watch, and my Garmin and gotten three very different answers for what happened in my sleep the night before. I mean, for sure, what I would say is that the important thing is kind of looking at the overall trends for all of that. And actually, I would love. I love doing this with VO2 Max, too. The VO2 Max estimates across the.
- [33:19]
Louisa
You said VO2 Max about 15 times in this episode. Now everyone listening knows how much I love VO2 max. We actually have a much better way of measuring it. We do it with the actual device. So with every person that comes in. We don't take it from a wearable. We actually do the, like, MetCard testing. We do the entire thing. We get metabolic rate as well. We get. We get everything. So we can try and find your true zone two, zone five, and your true VO2 max.
- [33:48]
Renee Dehan
Oh, God, that's so.
- [33:49]
Louisa
We'll have to do it on you.
- [33:50]
Renee Dehan
That's so important. I love it. I love it. I love VO2 max.
- [33:53]
Louisa
We love VO2 max.
- [33:54]
Renee Dehan
Yeah, I know. We could just talk about VO2 max all day. No, I totally do. And I. I do get my VO2 max measured at the lab every once in a while. Because I do. I started running a couple years ago, and it was really important to me to find the right zone, too.
- [34:10]
Louisa
Yes.
- [34:11]
Renee Dehan
Because I just knew that I was going to go at it too hard unless I had some, like, really accurate data to say. Like, this is what easy is. And I like RPE as well. Of course.
- [34:20]
Louisa
Yeah.
- [34:21]
Renee Dehan
But I'm like, it's great. Both. I want both. Like, I want my own estimates and the actual data. But Apple is wild. Like, how is your VO2 max on Apple compared to Garmin?
- [34:31]
Louisa
Oh, no, it's. Garmin is a much better predictor of my VO2 max.
- [34:35]
Renee Dehan
Absolutely.
- [34:36]
Louisa
Same what I think everything that's been coming up right now with what you're talking about in terms of future projections. Let's. Let's Take a magic pill and jump 20 years into the future. This is what I see because I think this would be amazing, right? We've got the three levels, you've got the questionnaire, you've got the blood work, you've got the DNA. Imagine marrying that with MRI, CT, everything. Like looking at your CAC score, looking further beyond. Do you think maybe we'd be including that in the future?
- [35:07]
Renee Dehan
Absolutely. And you know, we're just, we're. I was going to use an inappropriate word, but like, we will do despicable things for data. Like the more the better. Right. And you know, as, you know, measuring different things, like you said, bone density, it's all like, they all provide different information. And we're a huge network. We're a huge network in our bodies. We're a huge network with the things that live in our bodies, like our microbiome or a huge network with our environment around us. And, you know, the more of that we can capture, the more of a high definition picture we're going to have of what's actually happening. Right. And the better we're going to be at predicting things that we can't measure. Right. The more data you can capture, the more you can kind of like impute or make guesses about the data that's missing. But yes, I want to look at microbiome, I want to look at epigenetics, I would love to look at metabolomic sequencing, all of these types of things and incorporate tons of information that you might get from electronic health records. CAC scores would be great. All of these things really help you dial in on what your actual disease risk is, which is really health span. Like health span is just sort of like another way of thinking about collective disease risk. So I absolutely think that that's going to be the case. I think it's going to be a lot easier for people to get measurements. I hope more and more of these measurements can come passively. Like, I know lots of people wear CGMs, right. So those are easy. I want the lactate meter in the cgm. Like, I want that so badly. Right.
- [36:51]
Louisa
My, my biggest hope is to find a continuous blood pressure monitor. Yeah. So, yeah. And I speak about it often because my dad in 2019 had a stroke. Very unpredictable. And I think, imagine if we had him with a, you know.
- [37:06]
Renee Dehan
Yeah.
- [37:06]
Louisa
And we can't do that yet. So I think maybe 20, 30 years, maybe we'll have a device for that.
- [37:12]
Renee Dehan
Yeah.
- [37:13]
Louisa
So one question that has been on my mind and a question I actually get asked Quite often is, what are some of the drawbacks in utilizing AI and blood work?
- [37:25]
Renee Dehan
Absolutely. This is a super important question and one that I'm pretty passionate about personally. But. But the first thing that comes to mind is bias, right? So if you are trying to build a machine learning model that will predict your risk of something or whether or not you have a disease or condition, you do require training data, and that data is subject to bias. So if you think about it, the majority of research that has been done out there and in clinical studies have been done on white men. They are the best represented in, let's say, the collective training set of health care data that we have. So if you're a woman, if you are trans, if you are a person of color, if you are very old or very young, there have been systems in place to either keep you out of clinical trials and clinical research, or you don't want to be part of it. Because we've absolutely abused it in the past, and we know that things like the Tuskegee study. But it's this garbage in, garbage out scenario where if you build a model that's tested only really in white men, but you're not a white man, then the effectiveness of that model being applied to you is decreased. And this can be really harmful. So, for example, women are actually much more likely to experience an adverse event from a drug. And the reason why is because the majority of clinical trials are done on men and not women. And it's not just that. The joke in the skiing world is shrink them and pink them, but women aren't just small men. So you can't just change the dosage and hope everything works out okay. The actual mechanics of how women metabolize a drug are different. Right. So all these pharmacokinetic parameters that people measure, you know, if there aren't enough women included in the study to really, truly evaluate how that drug works in women, you know, then that's why women are more likely to have adverse events. Like, that's horrible.
- [39:41]
Louisa
Extremely underrepresented women. And not just that. I think from a scientific standpoint, it may. Because people may be like, well, why are women unrepresented? Like, in terms of this? It's sometimes it's because women menstruate and you have to get them at timestamps. It's like, we have to get them at this time, and it's just very hard to track them on. So it's just easier to get a guy.
- [40:06]
Renee Dehan
Yeah.
- [40:06]
Louisa
Really? That's the premise of it.
- [40:08]
Renee Dehan
Absolutely. Absolutely. And That's a big factor, is like, women have a confounding effect of cycling. I would also say, though, that it's pretty wild. Like, the circadian people are obviously, they're banging this drum a lot. But even the time of day that we collect data is not necessarily standardized in a clinical study because it's just operationally really difficult. And I get that. But your body is doing very different things at 9:00am than 4:00pm or before you ate or after you ate or whatever it is. There's a lot of other confounding effects there.
- [40:40]
Louisa
My body's definitely doing things differently when I talk to somebody that upsets me compared to when I don't talk to somebody that upsets me. So, yeah, there is definitely many confounders there. One other question that I want to ask you is. So I, you know, I told Instagram, actually, hey, guys, I'm getting my DNA done. I'm just so excited about it. And an influx of messages saying, aren't you afraid of what they're going to do with your data? I'm like, no. So could you describe to me why people are so afraid?
- [41:11]
Renee Dehan
Yeah. And data privacy in AI is really important because we are collecting phenomenal amounts of data about people. And I think the main worry is that by looking at your DNA, we'll be able to figure out that it's you. And maybe there's something hiding in your DNA that you didn't know about. Like, let's say you had the 44 variant for Apoe and you have an elevated risk of developing Alzheimer's and you don't want to tell your employer that because there could be ramifications that you don't know about or your insurance company. So I'd say any company that is collecting sensitive data like DNA sequencing information, you know, you should do your homework and make sure that they're taking this very seriously. And they are. You know, they have policies and procedures in place in order to protect your data and that your data is not going to just randomly be sold to a third party or something like that, where it's really out of your control and the control of the entity. So I personally, I think my DNA is like smattered all over the world, like in multiple places. So I personally am not super concerned about it, but I definitely have a lot of friends who are just like, absolutely not. No way. I don't want to expose myself to that risk.
- [42:35]
Louisa
But on that as well, there's also the other end of the spectrum where people also are too Scared to see what their DNA is. They're like, well, hey, Louisa, if I'm going to die, I'm going to die. It's like, but that's not what your DNA, your DNA is not going to tell you, hey, you will die tomorrow. It's basically saying to you, hey, for the last 10 years of your life, pretty much whenever that may be, how can we ensure that you're not going to be bedridden for the last 10 years? That's what I, the way I try and describe it is, let's just understand what's happening because if it is apple, like at least we know instead of you starting to lose your short term memory, which is the first thing that goes, we will know why you won't just think that maybe it's because you're tired or you're, you're angry. Yes. Or you hate your husband or your wife or it's menopause and it's, we can find out because sometimes maybe it is, maybe you just are sleep deprived. But if we have a full clear picture, then we can just optimize for it.
- [43:30]
Renee Dehan
Personally, I fully agree. And every time I meet with a doctor, I'm like, I'm an over tester. I want to know. Like, don't hesitate. I tell me everything, tell me everything. And I would just, I would rather, I would rather know than not know. And I will tell you. Like, I have had my DNA done in many different types of capacities and it's amazing what you can find in there. Like a lot of things just jump out and start to make sense. Like little risks here and there for different autoimmune conditions. We're like, oh, that makes sense. I have a relative with RA and one with ulcerative colitis and asthma. And all of a sudden this kind of constellation of factors that you see in your broader kind of family history, they show up there, number one. And number two, I think about this for folks who are adopted and have no idea of their family history. So I would say that if you're adopted and you're comfortable with having your DNA done, that can be a great way to start to get information about your genetic lineage, essentially.
- [44:30]
Louisa
Okay, last but not least, I really have to ask this question because I'm very big on education. I think we've got so much education right now, but there is a gap in terms of how to acquire the right education. And so my question is, what role can AI play in making blood work more accessible and understandable to the general public? And how might this then empower individuals to take ownership of their own health.
- [44:57]
Renee Dehan
Yeah, that's another great question. And I always just immediately jump to the caveats, like what about when ChatGPT gives you wildly false answers because it cannot cite its sources? And I know that it's getting better and that's really wonderful and it will certainly get there. But I think tools like ChatGPT are a great first line resource for folks where you don't have to. Maybe you can kind of get the TLDR on a given topic. That might be fairly complex. I mean, I always joke that ChatGPT kind of reads like a sanctimonious LinkedIn article, but it does give you the like here are the five bullets that you need to know why you should measure APOB instead of ldl, for example, or something like that. So I do think that that certainly is is a way to make things more accessible. I think it is really important, especially in today's social media age, for people to understand and try and figure out what is a trustworthy source versus a not trustworthy source of information. That one can be really difficult. And I think actually probably us as science educators should be putting out more material on how to find out who to trust. It doesn't have to be me, but you should look for people with these types of criteria. They should say that they don't know something, they should be able to cite their sources. But the way that Inside Tracker works, and it was a reason why I wanted to go to InsideTracker outside of the big data set, which is always wonderful to get to have access to. But as a researcher, but what I wanted was a way to be able to parse out all of the scientific information that we have into what is useful. So, for example, I am not a nutritionist, right? I have not read every clinical study that has ever been published about the impact of dietary olive oil on your health. I haven't. I mean, there's got to be hundreds. Hundreds, right? But you know, the system that we have, this knowledge representation and reasoning does that. It basically does a meta analysis for whatever scientific question you're asking it, like what is the impact of dietary olive oil on LDL cholesterol levels or triglyceride levels or you know, I don't know anything else really. So. So I don't have to read those 500 papers because some nutritionist who has way more expertise in this field already has and they've curated and took all of that information and they have put it into a system that will remember it Right. And then that system will look and say, like, okay, for people like me, a woman in her 40s with the following characteristics. Here are the 10 studies that actually looked at people like you and looked at dietary olive oil intake and then showed that there was actually a positive effect or a negative effect or no effect. And it will then do that, a meta analysis, and say, okay, there's more, you know, there's more evidence to the positive impact of olive oil on LDL than, you know, no evidence or negative evidence. And that whole process, just for that one question, can take weeks. And I don't have that time. And I don't have that time to do that for every blood biomarker or every lifestyle intervention. But AI enables you to take that and capture it and store it and then reuse it over and over and over again for me, for you, for whomever's coming to the table. That is an aspect that I love. Because you don't need to hire that doctor. Right. Like, there's. I mean, you actually, you should. I. I love doctors. I have many of them, but I understand. Yeah, yeah. When you're not looking, you're. When you're not. When you're a relatively healthy person and you're just looking to get healthier, then how do you parse all of that information? And you have a technological solution driven by AI that can help. So that part is great.
- [49:17]
Louisa
God love the future and God love AI. Really do. I think it's going to make some amazing advancements in the healthcare system. But, Renee, thank you so much for being part of the neuroexperience podcast. I want to know where people can find out more about you and how they can potentially get their own blood measured.
- [49:35]
Renee Dehan
Yeah, sure. So definitely you can go to InsideTracker.com and InsideTracker is they're on Twitter and Instagram and Facebook and all the social medias. And that's where folks can go and get your own blood DNA tested, hook up your fitness tracker and identify actionable solutions for you. I myself, I think I lurk on Instagram, but I'm occasionally on Twitter making snarky comments.
- [50:03]
Louisa
We will connect everything below in the show Notes. Thank you.
- [50:08]
Renee Dehan
Thank you.