[Transcript] – How “Machine Learning” Can Predict Your Blood, Urine, Stool, Saliva & More! (Special Release: Podcast & BONUS Article!)

Affiliate Disclosure


Podcast from:  https://bengreenfieldfitness.com/podcast/self-quantification-podcasts/how-machine-learning-can-predict-your-blood-urine-stool-saliva-more/

[00:00] Introduction/BioTropic Labs

[03:38] Dr. Tommy Wood

[07:37] The Idea Behind The Test

[12:57] The First Most Common Performance Killer

[15:28] What Do They Do With The Questionnaire

[16:50] The Second Most Common Performance Killer

[20:56] The Third Top Performance Killer In Ben's Test

[23:36] Number Four of Ben's Top Five

[26:01] Using Metformin

[29:16] The Fifth Performance Killer Tommy's Lab Tests Look At

[31:13] The Accuracy of The Machine-Based Learning Test

[35:58] Tommy's Morning Routine

[39:18] Tommy's Workout Routine

[40:59] Tommy's Blood Flow Restriction Training

[44:10] Other Key Things Tommy Would Not Want To Leave Out

[46:27] Is It Free?

[52:36] End of Podcast

Ben:  Hey, it's Ben Greenfield.  And as it seems I am prone to do these days, I'm releasing another podcast episode on a day that I don't usually release podcast episodes because I'm just rife with amazing content.  I'm surrounded by all these smart cookies who have amazing things that I want to tell you about.  So today is no exception.  You're going to learn about blood biochemistry, machine learning, artificial intelligence, and how all of this can blend together to diagnose and to fix your body.

This podcast is brought to you by one sponsor today, BioTropic Labs.  BioTropic is basically a company that produces some of the most potent oxygen enhancing and oxygen utilizing herbs and supplements that you'll find.  You may recall the podcasts that I've done in the past with this guy named Craig Dinkel.  And Craig is a climber and a swimmer who is really into figuring out what makes the body tick, specifically when it comes to blood oxygen, nitric oxide, erythropoeitin, which is your body's precursors for red blood cell production, ATP, and he has these supplements that are combinations of very, very weird fringe things, but they actually work.  Like liver extract mixed with chlorella, mixed with nitric oxide precursors.  You got to try this stuff before workout.  It's really amazing.

So you can go to bengreenfieldfitness.com/biotropiclabs, just like it sounds, BioTropic, B-I-O-T-R-O-P-I-C.  bengreenfieldfitness.com/biotropiclabs and use code benpod10.  That's a mouthful.  benpod10 to save 10% and plus you get to enter to win one of five amazing prize packs that they have for you.  And I'm going to announce the winner August 29th on my Facebook page over at facebook.com/bgfitness.  So check that out, bengreenfieldfitness.com/biotropiclabs, use code benpod10.  Alright, let's go join Tommy.

In this episode of the Ben Greenfield Fitness Show:

“We often know that it is really important to change how you interact with your work environment.  So if somebody is getting stressed at work, work stresses them out, that's the perfect place to do some of these meditation and mindfulness practices.” “If you can do some diagnosis that doesn't require a physician to be there, those guys could really, really benefit.  However in the Western world, we know that faster diagnosis isn't necessarily the issue.  We have a lot of diseases where we're not really digging down into the root cause and stopping them much earlier.”

Ben: Hey, folks.  It's Ben Greenfield.  And literally about 45 minutes ago, I took this thing called an EPA, elite performance analysis tool.  What this tool promised to do for me was to use something called “machine learning”, which I guess is kind of like artificial intelligence, although today's guest will be able to fill us in thoroughly, machine learning to predict biochemical results, like what's going on with your blood, and your urine, and your stool without you necessarily bleeding, and peeing, and pooping.  And what this tool is supposed to do is it's supposed to predict different patterns of performance killers that you would regularly see in your body or in your brain, like blood sugar dysregulation, or low sex hormones, or gut dysbiosis, or circadian rhythm issues, or low oxygen availability, those type of things.  And this entire algorithm was developed by a guy who's actually been on the show before.  There was a show back in the day called “The Underground Test That Shows You How to Legally Upgrade Your EPO, Increase Your Oxygen Levels, Boost Your Red Blood Cells, and Build Double Digit Percentages in Power and Endurance”.  I know it's a mouthful.  But this guy came on that show, and in that show we talked about oxygen deliverability, something that will probably come up today as well.  And I will link to that original show if you go to bengreenfieldfitness.com/machinelearning.

But now, he's back.  His name is Dr. Tommy Wood.  And Tommy is a senior fellow at the University of Washington, not too far from me, on the other side of the state.  He's the chief medical officer of Nourish Balance Thrive, which is this company that uses advanced biochemical testing to optimize performance, particularly in athletes.  They work with thousands of athletes, like Olympians and world champions, and they use all this data to develop what they're now doing, which are machine-based learning algorithms to predict issues that you might have.  And so Tommy himself has a bachelor's in biochemistry, he has a medical degree from the University of Oxford, he has a PhD in neuroscience and physiology from the University of Oslo.  The dude is one smart cookie.  And he has published and spoken on everything from like multiple sclerosis, to insulin resistance, to ancestral approaches to health.  And he also competes in rowing, CrossFit, powerlifting, and ultra-endurance racing, so he's kind of an underachiever.  Tommy, welcome back to the show, dude.

Tommy:  Thanks, Ben.  Thanks for having me again.  It's a real pleasure to be here.

Ben:  Yeah.  Word, man.  And by the way, I know that you and I are both going to be talking together over in Iceland, right?

Tommy:  Yeah! So the Icelandic Health Symposium, you're our star guest.  And I'll be hosting the conference, and there'll be some other guys talking, as well as you, about how to improve longevity, which is something that most people are really in interested now.  So, yeah.  I'm really excited to do that and see you there.

Ben:  The Who Wants To Live Forever Conference.  I don't know who came up with that name.  It must sound sexier in Icelandic language, Who Wants To Live Forever.

Tommy:  Well I guess it's based on the Queen song, Freddie Mercury.  So, yeah.  The Icelandic guys came up with a name, but that's where it came from.

Ben:  Okay.  I'm out of the loop on that one.  September 8th through 11th.  I'll put a link to that over at bengreenfieldfitness.com/machinelearning too if you want to come hang out in Iceland.  Me and Tommy'll both be there.  So Tommy, you sent me over this evaluation that I took and I have to admit, I always have my eyebrow raised when I'm answering questions online and it's telling if I might have gut dysbiosis or low oxygen deliverability, but walk me through kind of like the idea behind this test.

Tommy:  Yeah.  So one of the things that we want to do, as well as make athletes perform better, is try and find a way to make the stuff that we do accessible to more people.  We do a lot of testing, gut testing through the stool and urine, and then blood tests, and this is stuff that's expensive and not necessarily everybody has access to both because they're in a country where they don't have it or 'cause they don't have the money.  So one thing…

Ben:  Or they live in New York.

Tommy:  Or they live in New York! Exactly!

Ben:  Some places where you can't do this stuff.

Tommy:  Yeah.  And so one of the things we want to do is try and find a way to get some similar results as we would get on these tests, but maybe using a different approach.  Machine learning is really the best way to do this.  It's going to rapidly improve our ability to diagnose disease, hopefully give us much better insights into the causes of disease, and it's really good at stuff like pattern recognition.  So machine learning is already out performing radiologists in terms of ability to detect cancers on x-rays.  It predicts longevity from a CT scan.  So just by training an algorithm with examples, it will then get some data from a new person and can then predict something based on what it seemed previously.  So we have data from a thousand athletes, and this includes all that stuff you talked about.  So advanced blood panels, stool tests, urine tests.  And then at the same time, we ask them a questionnaire, which is this the things that you did, and it covers stuff like how well you sleep, your sex drive, how well you recover, whether you feel lonely, whether you feel angry, and all these kind of things that we know are sort of simmering under the surface in a lot of athletes and they don't necessarily know how to express or how to deal with it.

And we originally thought that maybe we could predict some fancy tests, like the urine test from a blood test 'cause most people can get a blood test, but then we moved one step further away and we're now, based just on the answers to a questionnaire which is a questionnaire we sent to everybody when they first turn up and work with us, but then also over time to sort of track how they're doing, based on those test results, I can tell you with a very high degree of accuracy whether, say, you have a yeast overgrowth or whether you have some kind of blood glucose, probably how high your fasting blood glucose is.

Ben:  And that's just because a whole bunch of other people who have replied to this question who have also done blood testing, or urine testing, or gut testing for example, have had displayed the same issues?

Tommy:  Yeah, exactly.  The good thing about the machine, if you want to call it that, is that the things that I might think are important in terms of, say, blood sugar control aren't necessarily what the machine picks up.  So on some of the things that we've analyzed, the algorithm decided that what was really important was the time of day that somebody's answering the questionnaire.  So maybe it means something that you're more likely to do the questionnaire in the morning than you are in the evening, and then that figures into the algorithm to then predict what's happening later on.  So there are lots of things and interactions that you can pick up.  And if you think about, so you did 53 questions, there's five answers for each, it's just an analog scale from “None at all” to “Lots” essentially, and you have five to the power of 53 different combinations there, which is a lot of combinations, I don't know the number in my head, plus all this other stuff that's being analyzed at the same time.  So it can pick up associations between symptoms that a human just would never be able to see.

Ben:  Okay.  Are you able to send me these results? ‘Cause we're going to through 'em.  But in case people want to download them, is there a PDF that folks could look at to kind of go along with these results as we look through what it's telling me?

Tommy:  Yeah, definitely.  If anybody does, when people do the analysis, you'll get a ranking.  So we have five most common problems that we see in people.  So there's blood sugar dysregulation, low sex hormones, sub-optimal hemoglobin levels, something that we call low oxygen deliverability, a gut dysbiosis, and then also a circadian, or a HPA axis dysregulation.   Those are the top five things, and you'll get a ranking in terms of what's most important for you personally and then sort of gives you an idea of what you can focus on.  We have data on the back end which sort of breaks down all those predictions and gives percentages and things like that.  So we can definitely send that for you and then people can look through it as we talk about it.

Ben:  Okay.  So go through explaining what each of these are with me, and I want to do this in a way that kind of outlines to people what these top five things are that tend to be the performance killers that you see regularly in your athletes so that they can wrap their heads around them.  So it looks like the first one, well I really don't care which one you want to start with, but what I'd like to do is just launch into the first performance killer and then get into how that is determined and then what someone would commonly do about something like that.

Tommy:  Yeah, sure.  So we can go for the thing that came top.  Your top ranking was a gut dysbiosis.  So that's what it predicted.  And sort of your overall probability was fairly low.  So it was 30%.  So we're not saying you definitely have a dysbiosis, but if you have one of these issues, then that's your most likely one.  And of those, I then have the sub-data, which suggests to me that if you have a dysbiosis, it's most likely to be a yeast overgrowth and that's what we'd see on gut testing.  So we might see it on the stool test.  Or you can see some evidence of fungal overgrowth markers, yeast overgrowth markers on a urine or organic acids test, which I know you've done before as well.  So say somebody has, they predict the gut dysbiosis, and then we want to look for potential causes.  So you know it could be something in your environment, in athletes particularly, overtraining, pushing the gut too hard.

Training really, really intensely, and then eating right afterwards or eating inappropriate foods right after, particularly more processed-type foods, that can definitely drive drive dysbiosis, drive something like a yeast overgrowth.  So for somebody like that, you might want to look at how the training is structured, they might want to look at their diet.  We might want to look at an elimination diet or some probiotics potentially.  And if somebody does come back and their top problem is likely a gut dysbiosis, still at this point we'd still do a formal stool test 'cause we want to make sure that we know exactly what we're treating.

Ben:  So you're doing the questionnaire?

Tommy:  Yeah.

Ben:  And that's allowing you to say, “Okay.  Well if you're going to spend a thousand bucks, these are the tests that you should get based on your questionnaire.”  So it's almost like you're pre-screening someone to make sure they're not getting a bunch of tests that they don't necessarily need to be getting because they're asymptomatic for a certain issue? So in my case, for example, this ranked high, let me pull it up here, the gut dysbiosis ranked high, you'd say, “Okay.  Well you should get an organic amino acids test, like a urinary test,” and by the way, me and your buddy over there at Nourish Balance Thrive, Chris, we actually did a whole podcast on the Nourish Balance Thrive, or on the organic amino acids test.  I'll link to it in the show notes here.  But you would say, “Okay.  Go get this test because that's going to help us determine if this really was the case with you.”  Or do you just go straight to fixing somebody before testing based on the results or the questionnaire?

Tommy:  Yeah.  That's absolutely the goal is getting to a point where we can predict stuff with a high enough accuracy such that we can skip the expensive testing and we can try an empirical protocol.  At the moment, particularly for the gut dysbiosis, just because there are a wide range of different pathogens that could be causing issues, we will still do a stool test.  However, it allows us, the analysis allows us, like you said, to just focus on the test that somebody really needs.  So it gives us a way to reduce the cost to the user potentially, but also allows us to start much sooner.  So as soon as somebody comes to us, we know that gut dysbiosis is their main issue, we can still look at some of the other stuff like their training and their diet, and we can really focus in on that because we know that that's where the issue is.  And then when the test results come back, then we can start to treat.  The long-term goal is that you know we can then skip the testing altogether and we can start treating just based on the results of the algorithm.  And that's something that we're definitely moving towards.  We just we released the algorithm, we were so excited about it that we thought we'd come and talk to you about it because that's something that you and your listeners would really enjoy.

Ben:  Yeah.  It's super interesting.  So we've got gut dysbiosis.  What would be the second most common thing that either you're seeing in me that you want to talk about or that would be like the next perfect performance killer? ‘Cause there's five total, right?

Tommy:  Yeah.  So again, let's just work down your list.  So the next most likely one was low oxygen deliverability.  And that's something that we look at, it's basically lower-than-optimal hemoglobin.  So it's not the same as an anemia.  So when somebody has very low hemoglobin, they'd be considered to be anemic.  And we have higher cutoffs than that.  So if people are familiar with the values, it would be something like less than 14.5 grams per deciliter of hemoglobin in men and less than 13 in women.  So we have different cutoffs in men and women.  And the reason why this is super important is because we know that something like power at a given intensity, so say threshold power, functional threshold power increases linearly with hemoglobin.  So that's why blood doping is so successful in endurance sports because it makes you faster and it makes you more powerful.  So if we can identify things that might be causing you to have a lower hemoglobin than you might like, then absolutely that's going to result in better performance almost immediately as that hemoglobin just starts to come up.

So we're basically looking at anything that's going to be causing inflammation and oxidative stress.  So that could definitely be overtraining.  We know there's an interaction between training volume, hexodine and iron regulation, which can then lead to an anemia, particularly in endurance athletes.  We would then again go to the gut dysbiosis.  So if you had both predicted low oxygen deliverability and gut dysbiosis, then we know there's probably an interaction there.  Iron overload is actually something that we see associated with anemias, which is slightly counterintuitive, but we've definitely had some people with excess iron accumulation.  And this is what we talked about partially when we were on the podcast the last time.  And if they do something like a therapeutic phlebotomy, they donate some blood, we've actually seen hemoglobin go up despite the fact that they're donating blood.  So there's definitely an interesting interaction there.  And then some other toxin exposures.  So we do, if we think that somebody has a likely toxic exposure, either heavy metals or we also do and non-heavy metal tox tests through Great Plains, which looks at some interesting organic pollutants, and those can also interact with things like general oxidative stress interact with iron metabolism.  So all of those things are something, that's where we're really going to narrow down and focus if low oxygen deliverability is your top issue.

Ben:  Okay.  Gotcha.  I'm actually looking over my most recent lab results to see if this questionnaire is actually accurate when it comes to hemoglobin.  I'm trying to find my hemoglobin results here to see if I can hunt them down. ‘Cause what value did you say you'd like to see for hemoglobin in men, for example?

Tommy:  At least 14.5.

Ben:  At least 14.5.  Okay.  Let's see if I can see…

Tommy:  And it's worth saying that I don't think that anybody that we put into the algorithm is quite as super human as you are, Ben.  So we'll actually use your data, we have some of your biochemical data from previously, we now have your results from this test, we will then put that back into the algorithm so we can continuously update it.

Ben:  Oh, my hemoglobin! It's at 14.4.

Tommy:  Oh! You just managed it! See?

Ben:  There you go.  You did something! You know, I have done that urine test, that organic amino acids test, and I also did that along with a small intestine bacterial overgrowth test I think it was.  And it did show that there were some issues with, not a yeast overgrowth, but I believe it was almost like a little bit of a bacterial overgrowth, like a small intestine bacterial overgrowth.  So that might also go hand in hand with agreeing with some of the results of your gut dysbiosis prediction.  So, yeah.  You're on track so far in terms of your ability to seance my performance killers.  Okay, so what would number three be?

Tommy:  So number three on your list was hormone imbalance, and that's essentially low testosterone in males.  That's sort of our main indicator.  And the way that we normally measure that is using a DUTCH test, which I know that you've done before.  And for you, by the time we're working our way down this list, your likelihood of having low testosterone is pretty low.  That's just this is where it ranked.  But when we're looking at low testosterone, that's probably something that people are going to be really interested in because they know that that's associated with health, well-being, definitely performance.  And this is something where we can really dig in and get you started on a lot of stuff very early even if we don't have an absolute test telling you exactly how high your testosterone is because we know we can do stuff like work on your light exposure.  Ben, are you straddling your Joovv right now?

Ben:  Am I straddling it?

Tommy:  Your light on your balls?

Ben:  I'm not.  Here.  Dude, I'll flip it on.  Now I am!

Tommy:  That's one of your favorite hacks, I know.  But getting light at the right time, bright light in the morning is really good for improving testosterone production, proper sleep.  We'll look at your strength training, maybe get you out chopping some wood, cold thermogenesis.  We'll look at your calorie intake, so people who aren't eating enough calories or eating enough carbs, that can definitely reduce testosterone levels, maybe think about your micronutrient status.  And then again, you might think of some toxin exposure, and if necessary we could do some detox stuff.  And so really early on, there were loads of things that we know that we can implement to try and get your testosterone back up, and that's just based on the results of, based on this questionnaire.

Ben:  Chopping wood is an interesting one.  You know what I do now, because I do chop wood but not in the middle of the summer, but I've got one of these maces, one of these Onnit maces and I actually slam it against a tire, and I include that as part of my workout, and you do kind of feel a little bit more macho after you do it.

Tommy:  Yeah.  So my version is I'm clearing out some stuff in the back of my yard, so I have a pickaxe.  So this morning I was out in the sun with my pickaxe, digging up some weeds, and rubble, and stuff.  So anything like that, I think, is going to give you that feeling and hopefully boost some testosterone too.

Ben:  Okay.  Got it.  Number one is gut dysbiosis is one thing that this thing can predict.  And then we've got oxygen deliverability issues, and then low sex hormones or some kind of a hormonal imbalance.  What would be number four here?

Tommy:  So number four for you, and again I think very unlikely that you have this, but glucose intolerance, and that's something that we're seeing even in, maybe because all of the amount of exercise that athletes are doing, we're seeing higher fasting blood glucose than we'd like.  So the cutoff that we use is 88 milligrams per deciliter, higher HbA1c.  So we look at above 5.3 as an HbA1c target, or high fasting insulin.  So a fasting insulin greater than five.  All of those we know, as that starts to creep up, that's going to be associated with poorer forms in long-term, but especially worse health associated with cardiovascular disease and various types of cancer, and dementia, and all that kind of stuff.  So you know we really focus on getting that blood sugar sorted, and that's actually fairly common, particularly in a lot of hard charging athletes.

So if somebody came to us and blood sugar dysregulation was their top issue, we definitely start by tracking fasting blood glucose.  You can do that with a cheap blood glucose monitor and finger prick.  Or you really want to dive into it, you can do a continuous glucose monitor.  And then depending on what we see, we might look at some time-restricted eating or intermittent fasting.  Again, we talk about strength training and walking.  We want to try and get a greater glucose demands inside of your cells.  Really important to look at circadian rhythm and sleep timing.  And then we might look at something like a low carb diet or an elimination diet if there's something that you're eating that's maybe causing an issue and then that's causing a high blood glucose response.  Sometimes we might see lower blood glucose than we'd like, and then we'd look at stuff like cortisol, autonomic functions.  So look at HRV, again we look at the urinary organic acids test, things like mitochondrial function which can be altered in people with a poor energy status or very low blood glucose.  So depending on how that goes, we can approach that in different ways.  But again you can see that a lot of the stuff, we're finding issues in people, and a lot of the stuff is something we can implement straight away for free without any fancy testing.

Ben:  By the way, speaking of blood glucose, are you one of those guys who's a fan of this metformin that everybody's using these days, it seems like in the anti-aging sector?

Tommy:  Yeah.  It's such an interesting question because in people who have dysregulated blood glucose, so say somebody who is a type II diabetic, when you put them on metformin, their cancer risk decreases, their mortality risk decreases.  There definitely appears to be a longevity benefit.  It's also been done in a number of animal models.  However, it does give me kind of a bit of pause for somebody who's healthy to take it because we know it is a mitochondrial poison, it affects complex 1 particularly.  We talked so much about mitochondrial function, and if you're a hard charging athlete who wants their mitochondria to be working as well and as hard as possible and then also recover properly, then I do kind of wonder if it's something we should be recommending to everybody.  And it's not something that, at least until I've seen a study in healthy people, it's not something I'm going to take personally.

Ben:  When you say that it's a mitochondrial poison, can you explain what you mean by that?

Tommy:  Yeah.  That's probably a severe way to state it.  But part of the mechanism of action of metformin is to inhibit the first complex of the electron transport chain, complex 1 in the mitochondria.  So that has some feed-on effects that then can activate things like AMPK, and you're sort of activating some of those pathways that you know you'll activate with exercise too.  So AMPK is activated.  It's like the energy sensor in the cell activated when you do endurance training, but also intense resistance training.  And some of the benefits are coming from activating AMPK, and that's associated with the way that it interferes with complex 1.  So I think it's definitely possible to get the same benefits or similar benefits by doing something else, say like strength training or some aerobic training, or maybe you take some polyphenols that can activate AMPK or NRF2.  They can sort of activate those pathways without actually affecting the mitochondria at the same time.  So that's where I'm leaning in terms of the evidence and how best to approach that.

Ben:  Yeah.  I just think there's a lot better options out there personally.  ‘Cause I've messed around a little bit with post-prandial blood glucose measurements and I've found things like apple cider vinegar and ceylon cinnamon for example to work pretty dang well at controlling blood glucose post-carbohydrate rich meal versus this metformin stuff.

Tommy:  Yeah.  I definitely agree with doing those stuff.

Ben:  Yeah.  Plus you can still get a lot of the calorie mimicking anti-aging and longevity effects without necessarily taking, again, what might be kind of an extreme definition, but mitochondrial poison in a way.  Plus I have actually seen some evidence too that metformin may interfere a bit with sleep and deep sleep cycles.  So that can be another issue when it comes to that particular drug.  I raise an eyebrow at how many people are freaking using that stuff.  Okay.  So we've got gut dysbiosis, we've got sub-optimal hemoglobin or low oxygen deliverability, low sex hormones, blood sugar and dysregulation.  And then what's the fifth parameter, or the fifth performance killer that you guys look at?

Tommy:  So the fifth one is circadian rhythm or HPA axis dysregulation.  And the reason that those are kind of lumped together is because we're basically looking at cortisol timing and the cortisol pattern over the course of the day.  So that's what we look at.  We use the DUTCH test.  Originally, a long time ago we did the cortisol saliva test, but the DUTCH test is just so much better than that.  So if you have cortisol outside of the normal range at any one particular point in time, that's going to tell us that you have some kind of circadian or HPA axis dysregulation.  And then we're going to look at stuff, again we talked about light, getting light at the right time early in the day, not at night.  Getting food at the right time, so particularly eating when it's light outside.  But then also stuff like stress management if appropriate, and then something that we talk about a lot because it's so important is social isolation or social connectivity, like how many friends you have, the quality of your relationships, whether you work stresses just don't make it, just aren't worth handling and you should try and do something else.  All of those things, and again it's something you talk about too, having a greater goal in life is super important.  And all of that can translate to seeing these changes on a cortisol test.  So depending on who you are and what we see, particularly on the responses to your questions, 'cause we ask questions about social isolation, and stress, and anger, and all that kind of stuff, then we can sort of start to pinpoint where maybe you need to focus.

Ben:  Okay.  So have you actually gone in and studied whether this is true or is this all just blue sky that if I take this test and it's suggesting that I have an issue, like the top on the count is gut dysbiosis, have you then taken a group of people who gave these responses, tested them, and found out these tests were accurate, this machine-based learning test was accurate?

Tommy:  I mean that's a great question, right? Because we can say all this stuff, but we want it to work in real life.  And we're currently running the first set of guys who are getting the prediction done and then we'll get the test done afterwards.  So we haven't done the full suite of testing afterwards, but what we have done, and we've actually been really pleasantly surprised by is when we've got this analysis out there, we've been approached by a number of new clients who have known issues with certain things.  And honestly they don't tell us about that up front, but they do the analysis tool, and then it predicts it right off the bat.  So we've had guys with low testosterone, or low free testosterone, libido issues, and then that's their number one problem.  We've had guys with problems with blood sugar, type II diabetes or pre-diabetes, and that comes out top.  We've also had some people who they know there's some issues in their gut, and that's come out top.  So even though we're still working on the next step to get people tested after they do the analysis, the people who've come in fresh, we've never seen them before, and they have a known issue, that's being predicted straight away.  And it's worth mentioning that when we train the algorithm, we split up the data.  So algorithm was trained with 80% of our data, 75% of our data in fact, and then the other 25% we used to test how well the algorithm predicted things.  So then I know how well the algorithm is predicting stuff because we've already tested it out in 25% of our patients or clients.

Ben:  Gotcha.  Very interesting.  So in terms of machine-based learning, is this something that is taking place in other areas of medicine at this point?

Tommy:  Yeah.  Definitely.  And people might have heard of IBM Watson.  That's sort of IBM's artificial intelligence machine learning algorithm and platform, and a lot of big hospitals are kind of employing this.  The idea is that we're going to ways, sort of machine learning is going to assist the physician or assist the health care provider in some way 'cause it can sort of boost their decision making.  We know that it's unlike a human, it's completely objective.  I said it can find associations that maybe we wouldn't be able to see otherwise, it can reduce cost, increase speed, and doesn't require any training unlike a physician who might require years of training to get to the point where they're as good as they can be.  So there's a load of interest in the healthcare field.  What we're hoping to do is kind of introduce a new way of thinking about it.  And the reason why I think that it's important is because the rapid diagnosis is something that's really important in places where they don't have enough doctors or they don't have any doctors.  So you think about places in the third world.  If you can do some diagnosis that doesn't require a physician to be there, those guys could really, really benefit.

However, in the Western world, we know that faster diagnosis isn't necessarily the issue.  We have a lot of diseases where we're not really digging down into the root cause and stopping them much earlier.  So if we're just applying machine learning to loads and loads of patient data once they already have [0:34:28] ______, that's not necessarily going to tell us anything about how the disease happened in the first place and that's what we need to know.  So we're kind of trying to move the process up much earlier so that we can get even subjective data, just like how somebody feels, and then try and find something out and intervene much sooner than, say, trying to do some complex diagnostic task, which is maybe not what most people in the Western world need from a machine learning type healthcare platform.

Ben:  Interesting.  Yeah.  It's really interesting.  So knowing what you know about these top five things that you see cropping up over and over again, especially in active people, blood sugar dysregulation, and low sex hormone, sub-optimal hemoglobin or low oxygen deliverability, gut dysbiosis, and circadian rhythm or hypothalamic pituitary adrenal axis dysregulation, that's a mouthful, I'm curious about you, Tommy.  Because you're a pretty smart guy, like I mentioned.  You're well-educated.  I'm kind of curious how you structure your day personally to set yourself up to be optimized on a lot of these levels.  What kind of routines, and what kind of nutrients, and what kind of food groups, and what kind of exercise that you personally do or you endorse based on what you know, based on what you've learned.  I'm just curious how it is that you're setting up your day.  So for example, if I could ask you a few questions, you don't mind do you?

Tommy:   No, absolutely not.

Ben:  Do you have like a morning routine or anything like that?

Tommy:  Yeah, definitely.  And it's something that I've sort of managed to develop while I've been in the US.  Because when I was doing my PhD, it's just kind of, every day is just triage, trying to get it done, and get things finished, and get out.  So now that I'm here, my day, first thing, I get up, I get outside.  I have a yard here, so I'm usually outside, barefoot with my dog, or dogs now.  So we're out running around outside.  And then I usually come in, I'll make myself, I'll have a cup of coffee in the morning.  I make a smoothie, which just sort of sets me up for the day with lots of nutrients.  It has various plant powders.  I get mine from Four Sigmatic and Dr. Cowan's Garden, not that I have any affiliation with them.  I just like what they make.

Ben:  I love Dr. Cowan's Garden, dude.  That's some good stuff.  And for those of you who are listening in, I interviewed Dr. Cowan about these delicious, nutritious vegetable powders that he makes, and we have a whole cupboard full of them.  And every time mom makes dinner, it's like an insult to mom.  I tell the kids, “Go grab all of Dr. Cowan's Garden powders,” and usually they'll choose two or three and they bring them, I just dump 'em all over all the food.  For those of your listening, I'll put a link to my interview with Dr. Cowan in the show notes so you can learn more about these superfood plants that he grows and powders.  But you're putting those in your smoothies along with the Four Sigmatic mushrooms, Tommy?

Tommy:  Yeah.  Absolutely.  And this is like a vegetable-type base.  So whenever I'm cooking in the kitchen, I'll always end up with ends of vegetables and things like that.  They get thrown in a bag in the freezer, that becomes the base of the smoothie with some berries, probably some coconut milk too, and then some powders.  And that's my breakfast for the day.

Ben:  So you just keep the whole bag of stuff in the freezer and then dump it into the blender? So when you're collecting berries, and plants, and stuff like that during the day, you're actually freezing them and then taking them out and blending them?

Tommy:  Yeah.

Ben:  That's interesting.  What I do is just like a boatload of ice and then some ascorbic acid, like a squeeze of lemon or lime, and then I'm typically just picking 'em fresh and putting them in there.  But that's a good idea, to freeze some of it.  I hadn't thought of that.  Now what I do freeze are the, I've been putting broccoli sprouts in my morning smoothie.  And so I freeze those because from what I understand, you actually get a little bit more activity of the sulforaphane in the broccoli sprouts when you freeze them first.

Tommy:  Yeah.  Definitely.  I definitely want some sulforaphane too.  I have a slightly different way of getting mine.  I developed a tea product for Nourish Balance Thrive, and one of the main ingredients is broccoli seeds, which…

Ben:  Oh, yeah.  That, what's it called?

Tommy:  Hormetea.

Ben:  Hormetea! Yeah.  You sent me some.  It's good stuff.

Tommy:  Thanks! But yeah, so actually the highest concentration is in the seeds.  And then you activate them with some warm water, and then you can sort of top it up as a tea, or I sort of activate my tea and then I just pour into my smoothie.  So I get all that good stuff.  And there's some other things like some berry anthocyanins and some green tea extract, and I sort of get that all in there, so all those nutrients sort of come at the same time.

Ben:  Nice.  I like it.  Okay.  So you've got your morning smoothie, you've got your cup of coffee.  What does your exercise routine look like, Tommy, when it comes to striking a balance between avoiding excessive inflammation or gut dysbiosis from exercise, but also enhancing longevity with movement? What are you doing?

Tommy:  So I just moved house and I'm in the process of building my home gym.  But in the meantime, the last few months, what I've been doing is three or four times a week, I'm out, like I said, I'm with my pickaxe for maybe an hour.  There's quite a lot of stuff I have to clear.  So that's kind of my warm-up.  And then I've been doing, I have a weights vest and some rings in the area that I'm going to build my gym.  So I'll do some light gymnastics and maybe I'll do some lunges or pushups with the weights vest.  And then I have some blood flow restriction bands.  So I'm doing a bit of blood flow restriction training.  And then I just have some rubber bands with handles so I can do some other, some of the exercises you might do with dumbbells.  I can do those with the resistance bands.  And if you adds some blood flow restriction, then you can sort of get some of that, you get that metabolic benefit without adding a huge amount of stress.  So that kind of gives me, I work the muscles in those sort of the more strength-type domain if I've got a bit of, add some weight to myself.  Or I do some blood flow restriction training, and that gets the metabolic benefit.  Sort of overall, the total stress is much lower than maybe what I was exposing myself to when I was training for ultra-endurance marathons.

Ben:  Yeah.  I have blood flow restriction training bands.  And the way that I use those is I wrap them like a tourniquet around my arms and my legs, and this would be in a day where I don't want to beat my body up too much in terms of eccentric muscle tissue damage, but I'll do exercises like body weight squats or pushups wearing these bands.  What type of workout do you do when you're wearing a blood flow restriction band on the arms or the legs?

Tommy:  So I'll mainly focus on a lot of pushups, and then maybe also if I'm doing a bit of an arm workout, some bicep curls and tricep extensions with some resistance bands.  Those are great with the blood flow restriction bands.  And then one of my favorite things actually is just to put the bands around the top of my thighs and then just go out lunging in the garden.  And I'll just sort of lunge up and down for a few minutes.  If anybody thought that a few lunges wasn't a hard workout, as soon as you put those bands on, it becomes a lot of fun.

Ben:  Yeah.  I actually had a guy who designed some of these bands come to my house.  He happened to be here for an entrepreneurship mastermind group that I have in my house every year, and he took me through a workout.  He was doing sets of 30, 20, and 10, which I guess is one of the more efficacious ways to train with these bands on as you'll do 30 reps of an exercise, and then introduce a whole bunch of blood and lactic acid, and then a short recovery period, and then 20 reps of the same exercise, and then a short recovery period, and then 10.  And he was having me on many of the exercises pyramid up in weight as I went.  So just in those very short breaks in between, I'd pyramid up in the actual weight that I would use.  But yeah, you get a good pump, as Arnold would say, when using those things.

Tommy:  Yeah.  Definitely.

Ben:  Okay.  So you've got blood flow restriction training and pickaxing.  You've got your superfood smoothie with your Hormetea and your broccoli seeds, or your broccoli sprouts, your cup of coffee, your Four Sigmatic mushroom extracts.  What about for your spirit? Do you have a meditation, or a yoga, or a gratitude routine, or anything like that that you do?

Tommy:  Yeah.  That's something that I've been bringing in more and more recently.  It was something that you always know that you should do but haven't been doing.  So I mainly use, I use a combination of the Headspace app, which I'm sure people have heard of, and then I also do some of the guided meditations on brain.fm, that's the other app.  So I generally do that towards the end of the day.  We often know that it's really important to change how you interact with your work environment.  So if somebody is getting stressed at work and work stresses them out, that's the perfect place to do some of these meditation or mindfulness practices.  So if you can then associate work with some of the more calming meditation, then you can actually change the way that you respond to stresses at work.  So if work's getting really tough, then I'll sit down for 10 or 15 minutes and then I'll do some of that meditation too.  So that's how I do mine.

Ben:  Okay.  Gotcha.  So you like the Headspace app.  And did you say you do any journaling or anything like that?

Tommy:  So I do not do any journaling, but that is definitely on my list of stuff to start.  Absolutely.  Guys like you…

Ben:  I'm jaded, now that I've created a gratitude journal, but I still, three years now, I've barely missed day.  I swear by the morning journaling, man.

Tommy:  Yeah.  I was around at Mike [0:43:55] ______ house recently, doing some podcasts with him, and he said the same thing.  So I need to get started on that.

Ben:  Yeah.  What are some other key things that you do each day that you wouldn't miss from a health standpoint or a longevity standpoint? Or just a happiness standpoint, really.

Tommy:  So the importance of gratitude is very important to me.  And though I don't journal it, it's really important to me to always be thankful to others.  And I think about, I'm conscious about that.  Whenever somebody does something for me or in a way that benefits me or anything like that, I always try and make sure that I'm thankful to them, and this is something that I learned, it's an interesting place to learn it but when I was working as a doctor in hospital, one thing you learn is that if you're nice to your co-workers, especially the nurses, as a doctor, your life is just so much better.  And then also the patient gets better care.  So being nice to people, thankful to people, that's super, super important to me.  Then I try and move every day.  So I have stand-up desks, I try and walk, I don't necessarily work out everyday but try and move everyday, get exposure to light every morning, particularly.  And then I do all that stuff at night that we know is important.  I have blue blocking glasses, I have flux on my computer if I'm exposing myself to a screen, which I do less and less now at night.  So all that kind of stuff to bookend my day and make sure that I'm getting the right light exposure.  I think that's really important too.

Ben:  Interesting.  Are you a fan of any of this idea of hormetic stressors of cold or heat? Are you one of those guys who's doing barefoot walks up Mt. Kilimanjaro, or at least jumping in a cold bathtub every week, or doing sauna treatments, or things like this?

Tommy:  Yeah.  I have Scandinavian genes.  I'm half Icelandic.  So that stuff is kind of build into me.  Where I am at the moment, I'm in the process of building a cold plunge and will then get a sauna once I've built my home gym.  So it's not something I'm doing everyday currently.  But when I've had access to it, so when I worked in Norway before I moved to the US, so I had access to a sauna and cold plunge in the gym.  So I was doing that.  So definitely something that I agree we should be doing more of, just not something I have in my house right now.

Ben:  Gotcha.  Okay.  Cool.  I'm going to come full circle back to this elite performance analysis tool.  This is something that someone, is it something someone would purchase on your site or is this just something someone takes for free?

Tommy:  So the analysis is free.  You can come on to the website, so people listening to this podcast can go to the URL ben.nbt.ai, then go there.  We know that they're coming from this podcast and then they can do the analysis.  And what that will give them is then a ranking based on where those five issues are in terms of what's their number one, particularly.  So then when we see what their number one issue is, they'll also get, you'll get some follow-up.  So you'll get some podcasts where Chris or I have talked about that particular issue, and strategies that you can do to try and overcome it.  A lot of people will then want to dig in a lot deeper.  So this is now the first part of working with us.  So somebody does the analysis, they decide they want to work with us as coaches and maybe dig in to their performance and get them performing as well as possible.  And this could be anybody with a performance goal.  Your performance goal could just be playing with your kids for the next 10 years.  It doesn't have to be that you want to qualify for the Boston Marathon or Kona.  Whatever your performance goal is, we can definitely help you with that.

And so somebody comes in, they do the analysis, then they can jump on a free call with us.  We'll talk about what the analysis found, what they can maybe start doing.  We'll get them started straight away on the most important things they can fix, then we'll do the necessary testing, we'll tweak the program.  And then over time, we have what we call sprint.  So we'll talk to you on a call, we'll set up a countdown board, so it's kind of like a To-Do list, so we can always see what we've suggested you do, we can all see what you're doing 'cause it's like an interactive board.  So we can always see who's doing what and if you have any questions, you let us know.  And so we'll set some tasks, you do them for two weeks, you meet up with us again, and we reiterate that over the next few months and see you improve, and then we'll reuse the analysis over time so we can see how you're doing.  We don't necessarily need to redo all the tests because you can redo the analysis and say, “Okay, great.  My likelihood of the gut dysbiosis is coming down,” or, “My blood sugar control is improving.  My circadian dysregulation is improving.”  So we can then use that to track over time, and then if necessary, we can also retest.  But it's kind of allowing us to get started with people much faster, and get a much better idea of where they need to focus, and then allow us to track over time.

Ben:  Got it.  Cool.  I like it.  And by the way, for those of you who have difficulty remembering all these URLs and everything, just go to bengreenfieldfitness.com/machinelearning and I'll put a link to all this over there.  I also will upload my own results so you can kind of see what this simple little algorithm looks like.  And there's some cool cute little videos that walk you through each of these things too, like the gut dysbiosis and the oxygen issues and everything else.  But it really is interesting, these five issues that you see pop up over and over again, 'cause I definitely see these same things in a lot of people and it's kind of cool that you guys have identified them and then figured out a way via a simple questionnaire for people to figure out whether they actually have some of these issues.  I like it.

Also, for those of you listening in, when you visit the show notes, I will indeed put a link to that conference in Iceland because it promises to be a really great one when it comes to anti-aging and longevity.  And I'll be speaking there.  I'll also be leading a bicycle ride through Iceland for those of who want to join me there.  That'll be fun as well for all fitness levels.  So check all that out, and again that's going to be at bengreenfieldfitness.com/machinelearning.  Tommy, was there anything else you wanted to fill people on when it comes to the elite performance analysis tool, Nourish Balance Thrive, Iceland, your broccoli sprouts, or anything else?

Tommy:  Well, yeah.  If people are interested in tea, they can go to hormetea.com, which is HORMETEA.com and have a look and try it out.  But the main thing is that if somebody's listening to this podcast and they can come to Iceland on the main days on the 8th of September and if you sign up to work with us on our elite performance program and you can come to our Iceland, we will buy you a ticket to the conference and you can come and have your first consultation with us in a hot tub in Iceland and we can talk about all things health-related.  So if you're interested in that, we'll definitely help get you there and we'd love to see you in person because that's what really drives us and we really enjoy just meeting everybody we work with and getting them the best possible results.

Ben:  In a hot tub? Wow.  That's quite an offer.  You can't turn that down.  Alright.  Well cool, folks.  I'll put a link to Iceland, the EPA tool, everything else that Tommy and I talked, about as well as the previous podcast that I've done with Tommy if you go to bengreenfieldfitness.com/machinelearning.  Thanks for listening in and have a healthy week.



I recently took a test that uses machine learning to predict biochemical test results (like blood, urine and stool) – a test called the “Elite Performance Analysis (EPA) tool”.

Over the last three years, the folks at Nourish Balance Thrive (Dr. Tommy Wood and Chris Kelly, both former podcast guests) who designed this test have worked with over 1,000 athletes, averaging over 100 biochemical markers collected per athlete, including:

  • Blood biochemistry
  • Urine tests (DUTCH and organic acids)
  • Stool tests (PCR and culture)
  • Subjective quality of life questions (a Health Assessment Questionnaire, or HAQ), scored on an analog scale (1-5)

As well as working to optimize the performance of athletes at every level, another goal of Nourish Balance Thrive is to give more people access to the type of work they do by increasing speed of access and reducing cost. Machine learning provides for a very good way to do this. By training an algorithm based on historical HAQ and biochemical test data, they can predict five common patterns of performance killers that they regularly see in their clients, including:

  1. Blood sugar dysregulation (high/low fasting blood sugar and HbA1c, or high fasting insulin)
  2. Low sex hormones (testosterone in men and oestrogen in women)
  3. Suboptimal hemoglobin (“low oxygen deliverability” <14.5 g/dl in men and <13 in women)
    Gut dysbiosis (including clostridial and yeast overgrowths, parasites, and H. pylori)
  4. Circadian/HPA axis dysregulation (total/free cortisol outside of the normal pattern during the day)

Dr. Tommy Wood, who I interview on today’s show, is a Senior Fellow at the University of Washington, and Chief Medical Officer of Nourish Balance Thrive, an online-based company using advanced biochemical testing to optimize performance in athletes. Close to 1,000 athletes have been through the Nourish Balance Thrive coaching program, from weekend warriors to Olympians and world champions, with hundreds of performance biomarkers available for each client. Based on their accumulated data, the team is using machine learning algorithms to rapidly and reliably predict common problems in athletes including low hemoglobin, low sex hormones, and gut dysbiosis.

Tommy has a bachelor’s degree in biochemistry from the University of Cambridge, a medical degree from the University of Oxford, and a PhD in physiology and neuroscience from the University of Oslo. He is President-elect of Physicians for Ancestral Health, and is on the scientific advisory board of Hintsa Performance. Alongside his career in medicine and research, Tommy has published and spoken on multiple topics surrounding functional and ancestral approaches to health, including examining the root causes of multiple sclerosis (paper, talk for the public, talk for physicians) and insulin resistance (published article). Tommy has also coached and competed in multiple sports including rowing, CrossFit, powerlifting, and ultra-endurance racing.

During our discussion, you’ll discover:

How a “machine algorithm” can predict gut dysbiosis issues such as yeast overgrowth and a urine organic acids test…[7:30]

The surprising fixes for low testosterone and hormonal imbalances in both men and women…[21:30]

Why Tommy is such a fan of chopping wood…[22:44]

The three powerful nutrients that can lower your blood glucose…[29:30]

How personal relationships and friends can affect your HPA axis and cortisol…[30:05]

Tommy’s morning routine, including the one superfood he swears by…[35:30 & 38:25]

How Tommy uses blood flow restriction training…[39:51]

What Tommy’s spiritual practice consists of…[42:30]

And much more!

Resources from this episode:

Elite Performance Analysis (EPA) tool

The results of my machine-based learning test

Elite Performance Program (click here to join, and you’ll automatically get free tickets to the “Who Wants To Live Forever” Sep 8-11 Iceland conference and a free in-person consultation with Tommy in Iceland!)

My original podcast with Tommy

The Underground Test That Shows You How To Legally Upgrade Your EPO, Increase Your Oxygen Levels, Boost Your Red Blood Cells & Build Double-Digit Percentages In Power And Endurance.

Podcast: What Is An Organic Acids Test?

Podcast: Dr. Cowan’s vegetable powder

Nourish Balance Thrive coaching program (follow this link to get a complementary one-hour 1-on-1 video conference call with Tommy)

Blood flow restriction bands

Christian Gratitude Journal

September 8-11, 2017: “Who Wants To Live Forever?” conference in Reykjavik, Iceland. Most of us not only want to have a long lifespan, but also a long health span; to be fit and healthy throughout the course of our lives. As we move into this unprecedented era of human history, a question arises: how far can the human health span be extended, and what are the most effective ways to achieve longevity? Join me and other health experts as we dive deep into this subject in the country with strongest and healthiest humans on the planet! Click here to register.*

*There is a Cycle Workshop as part of the “Who Wants To Forever” conference: In this workshop you’ll cycle in and around the Reykjavik area with one of the most influential people in health and fitness (ME!), gaining direct access to my wealth of knowledge on how to optimize your performance. Among the things you will learn along the ride are: Breathing techniques to enhance your longevity and health. Modern and ancient techniques to harness the natural magnetic fields of the planet we live on. How to use heat and hypoxia to increase blood flow and maximize your response to training. Basic techniques for cold thermogenesis, how to use both cold and warm water exposure effectively. And much, much more! Sign up now – we’re sure you don’t want to miss this! Click here for the bike event only.

Podcast Sponsor:

BioTropic Labs – Go to bengreenfieldfitness.com/biotropiclabs and use code “benpod10” to save 10 per cent PLUS enter to win 1 of 5 awesome prize packs now! Winner will be announced August 29 on our Facebook page.

But the podcast ain’t all…

…in addition to the podcast, Tommy, Chris and Megan Roberts MSc of Nourish Balance Thrive put together this comprehensive article all about the top five performance killers and the Nourish Balance Thrive (NBT) 7-Minute Elite Performance Analysis Tool.


Ask Ben a Podcast Question

Leave a Reply

Your email address will not be published. Required fields are marked *