Home 🏠 🔎 Search


Bad Transcripts
for the
Beat Your Genes Podcast & More

Chef AJ: The Evidence on Nuts | Interview with Dr Doug Lisle
an auto-generated transcript


To get a shareable link to a certain place in the audio,
hover your mouse over the relevent text,
right click, and "copy link address"
(mobile: long press & copy link address)
 


well good afternoon everybody thank you so much for being here and welcome to healthy living live I'm chef AJ and my guest today is dr. Doug Lyall he is not only the co-author of the pleasure trap but the psychologists at both the trimmer and health center and MacDougal Living Program both conveniently located in Santa Rosa California he has the star of the weekly podcast beat your jeans GE and es which you can listen to live every Wednesday evening at 8:30 p.m. Pacific time and even call in and ask questions there's over 140 episodes now that you can access on blog talk radio and iTunes so please welcome dr. Doug Lyall hi Jay hi we love having you here you're like our resident psychologist and even Bailey decided to come stick her head up to see what you're going to talk about today so it's funny because there's a microphone right here it looks like she's actually talking so we are actually going to talk about nuts and cardiovascular disease today and the reason is some of you may know my weight loss story that I told it the MacDougall Advanced Study weekend is called from fat vegan to skinny you can find it on me dr. MacDougall's website as well as on my youtube page and it it talked about how I've worked with dr. Lisle to lose these 50 pounds which I've kept off now for about seven years and one of the strategies I used was to decrease the amount of fat in my diet and I found that for me I had to go to no added fat for my weight to move I was not cheating dr. Lisle knows this I was eating an SOS free diet but even eating about an ounce of nuts a day I just aged I couldn't lose weight and so it's been almost seven years now since I've added any fat over back to my diet not nuts Nazis not avocado I eat a lot of greens I had a lot of purse line and I also get my omega-3 fatty acid levels checked every single year and they're not only excellent the doctors are like I can't believe how high your numbers are but very often especially when I speak at conferences there are other wonderful plant-based doctors who insist that we need to eat nuts specifically for the prevention of cardiovascular disease which runs in my family and I got really concerned and I asked dr. Lisle if he could please talk about this because he was a statistics professor at Stanford and when I look at the research I I don't know what it means I know that sometimes it says it was funded by the nut industry but I have no idea what it means so I thought we would trust our resident psychology to actually look at the research and and tell us do we have to eat nuts to be healthy do we have to eat nuts for longevity or to prevent cardiovascular disease or perhaps reverse it so take it away dr. Lyle all right EJ well I heard about this issue you know I've had the echoes of it have come to my attention but I haven't really bothered to investigate it when I have investigated in the past I didn't see anything that was of any significant interest but you know if you're worried and you got people that are worried about this we we should take a look at it and so let's uh the way we do this in science is we we look at members so i got it i got i'm sorry i did just give you an opinion i you're gonna have to we have to look at numbers and that's what that's what statistical people do so we're gonna do let's let's go find ourselves the best some some really excellent data so hold on a second so let's punch in nuts and cardiovascular disease into your computer okay sorry everybody people didn't know that I needed to wear reading glasses but yeah alright so what comes up first for me is a first study is net consumption and risk of cardiovascular disease it's published through the NIH it looks good let's see what it is 2017 okay it's a guy by the name or a person by the name of couch da looks like pretty fancy internationally Journal of American College of Cardiology 2017 November 14 so if anybody's following along at home this is where we're gonna look so the the background of the study was to look at the association between specific types of nuts and specifically peanuts and walnuts and look at cardiovascular disease and whether or not there's a association between the intake of total and specific types of nuts and cardiovascular disease coronary heart disease and stroke lists okay so this looks good and now we're going to look through at what they used looks like they used the big Nurses Health Study which was data collected from 1980 to 2012 seventy six thousand three hundred sixty four women and then there's a second Nurses Health Study that looks like it was started a little bit later 1991 through 2013 that's a ninety two thousand nine hundred forty six one so these are these are huge cohorts that obviously were very carefully done so this is this looks like very good data so if you're a scientist looking at this you're going to be interested in finding out what these people found the next thing is forty one thousand five hundred and twenty six men so we now balance it out for some men that health professionals follow up steady 1986 2012 so twenty-six years so and they started with people everybody in these studies was free of cancer heart disease and stroke at the baseline and so so basically looks to me way we're going to do this we're gonna do this like I would do it because we are going to do it so what I wanted to do AJ is I want you to take those three numbers and seventy six thousand three sixty four ninety two thousand nine forty six and forty one thousand five twenty six and I want you to add those up so while I continue here and that we get a total of how many subjects are in the investigation okay the over two hundred thousand people which is it's just going to be spectacular okay I got the first one seven six to me six four there the next one is ninety two thousand nine hundred forty six and the third one is forty one thousand five hundred and twenty six so we're going to add all those up and come up with a number of the total subjects that were in these studies okay I got I got two thousand two one zero eight three six two hundred and ten thousand and eight hundred and what thirty sixteen three six okay that's very good I'll I'll check you later we wouldn't want to make a mistake today I actually used a calculator yeah okay all right so then what we're going to do is it looks like what they're studying the big thing we're interested in is I'm not sure if we care whether we die of a heart attack or a stroke we're really concerned about whether whether there's anything that we can do other than things that we already know like get rid of smoking probably reducing reducing animal food and diet exercising sleep okay is there anything else like for example specifically these people are interested they're taking a look at this huge database and they're going to take their their razor and they're going to narrow it down and nuts they're going to see it did nuts have anything to do with reducing cardiovascular disease so this is a very interesting question and so we're going to look at this now the so it turns out that out of these two hundred and ten thousand people if they looked at for about 25 years the there was fourteen thousand one hundred and thirty six cardiovascular cases some have more heart disease some of them were strokes so we're not too concerned about which is which here were concerned about the fact that we have fourteen thousand one hundred and thirty six cases now they're not all fatalities but that that's okay so we we don't want I don't want to stroke and then live I just should avoid it so we're just going to say these are bad things and we we don't want these to happen to us okay so that's uh so now what we're going to do is we're going to look at whether or not now they look at nut consumption so they looked at to find out whether or not if you if you ate nuts a lot in this case it was a serving and nuts five times or more per week compared to people that almost never ate nuts so it's people that over 25 years ate nuts five times a week as opposed to people that did not do this this is pretty good while we're gonna look at two hundred thousand people and out of those people there's going to be a subgroup and I don't know how big the subgroup is because they don't tell us there but we looked down into the original data we would be able to find it but it doesn't matter the so let's suppose they've got a couple of thousand people or ten thousand people that that check the box that as they they actually followed these people up with questionnaires every four years so they've got huge amount of data they've got on average five different or six different self reports across people's lives saying that you know I'm a kind of person that I eat nuts you know pretty much every day expect what this is so as opposed to people that never read that's so we got a nice distinction there and it's going to turn out but they're now gonna run data up on these fourteen thousand one hundred thirty six cardiovascular cases so people that either have a heart attack or a stroke or they you know that they need to live or they died but whatever it was this was bad okay so we got fourteen thousand one hundred thirty six cases and now we're going to find out that if you are in the unit that I never ate nuts category we're gonna find out what your odds were of having one of these diseases and if you were in the high net intake category I eat five times a week for twenty five years whether or not you know what your odds were okay so let's just say further before we get started to what what would be exciting what would be exciting is yep if or if these guys over here had some amount of disease that they got and that if you ate nuts your per capita or per hundred people there was less of those people that they ate all these nuts that they have less of the heart attacks or strokes that would be a good thing so let's suppose dad nuts that's were really good and for every hundred of these people that have a heart attack or explode then there was like 50 people that ain't nuts that would be a you know a big reduction it would be very interesting to see that alright so we're gonna look and see what they found and what they found was that the hazard ratio which is like did it go from a hundred people you know they had strokes to the 50 Anheuser the ratio is 0.8 6aj so that means that if a group of people a hundred of these people have heart attacks or strokes that a high net group eighty six people with that so there's a bit of a reduction I mean it's it's interesting that it's there so now now we're going to try to see how valuable nut consumption would be you know to an individual as they look at this problem okay so here's what we're going to do we're gonna take the first thing we must do is we need to see how much risk there is for the people in the study so we're gonna take the fact that we've got 210 would you say now what was your never the total number two one zero eight three six okay so now we've got 14,000 136 people that had a disease process so I want you to put 14,000 136 and divide that by 210 836 that will give us the statistical likelihood of having the cardiovascular or stroke crisis that either kills you or names you or it's just scares the daylights out of you okay I think what's that one six point oh six seven okay so your odds in this in this group of people on average your odds of having a heart attack or a stroke or 66.7% okay now that I immediately my statistician brain turns on and I say okay that's kind that's relatively low but remember this was a group of people that started out that didn't that were free of cancer heart disease and strokes so we're already looking at people that are reasonably you know healthy and now we're watching them for 25 years and then we're seeing what happens so we're not taking people that that we're a bunch of smokers that had already had a stent put in their heart and then following that so we're following people that are disease free at some point in their lives probably in their 30s or 40s or 50s whenever they signed up for this study and then we're following over them for 25 years and this is what we get so 6.7% the people wind up with cardiovascular disease okay it doesn't surprise us any other Americans there's nothing about this study that says that these people who'd be getting sore vegetarians are particularly health just or anything else so so this is what we got you got 6.7 percent of these people wind up in in some serious spot water okay now what we're going to do now is we're gonna find out how valuable the nuts were for these people for the people in this study that ate nuts five times a week for 25 years we're going to see if the nuts did something to help them or it or might help so what we're going to need to do is we're going to need to multiply the odds of having a crisis which is 0.067 and then we're going to multiply that by five point eight six which is the hazard ratio if you're a nut eater so if we do that what do you get if we go point oh six seven times point eight six I got point oh five seven point oh five seven okay so what that means is that means that the the odds of having a heart attack or a stroke if you were a nut eater was five point seven percent and if the odds if you were not on that meter Peter was six point seven percent okay so there's a 1% difference though now in a very large study like this a 1% difference uh we don't kind of have time to explain statistics at this level today but when you have that that's a very small effect size there was so it's there that means that if you have hundred people in each group if you take a hundred people that were just the regular leaders in the group and a hundred people that were the nut eaters in the group that at the end of 25 years that in the general people there would be a ninety three point three percent chance that you wouldn't have a problem or that there would be call it six point seven percent chance that you have disease call it seven people okay so seven people out of those hundred would have had a vascular or cardiovascular crisis of some kind in the nut group it would be ninety four percent would be disease-free so six people would have had a heart attack or stroke so the difference between the two groups the nut eaters are have reduced their odds from instead of ninety three out of a hundred being successful they've gone up to 94 I've done so so they've they've helped themselves by one percent so that's the that's what the research actually shows on this topic now the now statisticians so I'm gonna look now what they say it's in in the conclusions of the study in three large prospective cohort studies higher consumption of total and specific types of nuts was this associated with the reduction of total cardiovascular disease okay so the one percent reduction that that they're seeing here is it is statistically significant and so you will hear people say this so if you were if you're a layperson that was looking at this and you didn't really have confidence in yourself to actually go down through the arithmetic the way we just did then you would said you would read the conclusion and you would say wow well there it is and that would be true so that people are pushing the truth there you if you also were maybe a little bit more sophisticated and we're looking at the hazard ratios of 0.86 you would say wow that's a 14% reduction so that seems like a big deal but remember we're going to have to compute that against the overall risks that the disease is associated with in order to get any feel for the value of the issue so if somebody tells me that they can reduce my odds of I don't know kala hawa syndrome in coma hawa so they can you know that I eat a thousand raspberries a day that I can reduce my odds of Kyle how a syndrome by 50% and they've got research to that effect but if Halle Halle syndrome only impacts one in seven billion people hello out there alright so so what we want to do is we have to take a look at the overall risk of the disease process we're looking at and we have to then compute to see the value of a change in hazard ratio in this case it's a hazard ratio reduction of 14% and then we which makes us point eight six or you're eighty percent eighty six percent just as likely to have the disease process if you're the net eater whooshes but not not eager and then we wind up then we can look it at it in terms that you know a grade schooler could understand this is out of 100 okay so out of a hundred people the ninety ninety three of the people that didn't eat so still alive and disease-free and at the eighth and that's--it's 94 so that that's what that is now if I have I have anybody really sharp and pay attention to this tune to this they're gonna understand that this point oh six seven four then where the non leaders of nuts is not quite accurate and that's because remember the nut eaters are in that group so when we take them out the point oh six seven is going to go up a little bit it's probably going to go up to point those six a to point it was six nine so there's going to be so we might have to adjust this in Contessa Mele but it's almost perfectly accurate okay so four so that so now you know AJ that if you like let's back back the camera and really take a look at what it is that we've learned from this it's interesting yeah we're gonna we're gonna we're not going to stop at one study even though this is an extraordinarily comprehensive study because we've got three cohorts for over 20 over 20 years per individual in the study measured multiple times and 200,000 people so in other words when you look at numbers that large that are that immense then what you're going to see is that the ratios that you're observing should be extremely stable so you're not going to have a new study from somebody else that's going to say oh no the effect is twice as large if you got a spell it's like no you're not going to find that I mean you may find a small study with 2,000 people in it it shows an effect size twice as large but that's because there's something biased about that study it's some way that the data was collected it is giving us a ungenerous able observation okay when you get a study this massive this carefully done these are excellent estimates as to what these effect sizes are and it also tells me something as a statistical person that that there's a real correlation coefficient down so it is true that the people that that need nuts are having a little bit lower risk factors so I we now have to say okay that's pretty well-established now we don't know why so we don't know that it's the nuts that are actually causing this little bit of the defect it could be what we call in science the third variable puffle okay so you and I are going to look at that in a minute this is this is what's known as a correlational study so did I I told you what I did for a living as a kid no that is a kid I know you're a statistics professor at Stanford as an adult but I don't know what you did is I was a statistics professor at Stanford before that I mowed lawns and I need more money mowing lawns then you do now working at the True North health center that's right so I thought statistics for many years at three different universities including Stanford and so I spent years sort of taking students through problems just like this so students would come in and they'd say doctor while actually those days they called me professor but hey you know that's how that works no no more the so they'd say well professor a while this is this you know this is what this says you know what about this and I would say well let's take a look at it and we get out can you get out the original study so we wouldn't read what was in the magazine a little synopsis we wouldn't read what the Mayo Clinic little bulletin says what we did was we actually go to the original study like you and I just did okay and we would fearless leap and confidently take out our calculator and we would trust our minds ability to step through the basic logic and then we would see what we would see okay so that's what we've seen today the now-now so I want to point out that the on the basis of this evidence which looks to be highly critical I have a credible is stable it would appear that the 1% difference is the maximum possible effect size there could be possibly attributed to to net consumption in other words if there's a third variable that's actually correlated with the chantin with the nut consumption if there's anything about these people that eat nuts that is also helping their cardiovascular disease that has nothing to do with the nuts then the effect that we're observing on the nuts is reduced so this is telling us that the maximum possible effect nut consumption on human cardiovascular disease reduces the odds by 1% over 25 years okay so now what we're going to do just for our own entertainment is you might say to me well I don't know what that means like well should I be worried about 1% and you know and I believe I'd be saying well you know AJ 1% 1% I don't know I don't know how else to tell it to you the weekend we could use a another statistics term to give you a feel for this which is what we call the expected value of the function and so we would say okay well let's look at let's look at these vascular incidents that you have a 1 percent greater chance if you happen to be in the group that's not eating nuts every day so if we did that the big bad thing that could happen to these people and did to some of them was they could die of a heart attack or stroke so now in this case these people didn't all die but let's just suppose they did so let's suppose we have the worst case scenario and these 14,000 136 people died of a heart attack or a stroke and if they happen to be in the nuts condition they have a slightly reduced likelihood that that was going to happen in fact it's about 1% so now what we're going to do is we're going to say ok well what is the amount of life that is lost by the heart attack of the stroke so it's going to turn out that if we were to compute this we're going to find that the average length of loss of life on the order of about 10 years so the average person is not having heart attacks you know at 42 or 52 or at 62 okay so we're gonna find that particularly if we take smoking out of the equation we're gonna find that the average person is having their heart attacks you know yeah you know yeah and dying of heart attacks in there around their 70th year or so so it's going to be essentially the estimated cost of a heart attack or stroke that's fatal it's going to be about ten years a lot so what we're going to do is we're going to multiply that ten years which is 12 months that's 120 months now we're going to multiply that hundred 20 months by 1% which is the advantage that that the nut consumers demonstrated in the study so it's a hundred and twenty times one percent which is AJ no - 2010 it's one percent what's one percent of a hundred one what's then one percent of a hundred 2012 one point - that's why we're doing this okay this is exactly why we're doing this today okay what's going to one point two months so one point two months is about 36 days 37 days so that would mean that the people that ate the nuts five times a week for 25 years how to have an outcome if all those people died from their disease which they didn't but if they did the worst case scenario the total protective effect of nut consumption in this study would have been somewhere around 36 days okay so now you have a way of gauging what the value would be from taking this risk of not eating nuts these high levels so that's that's what the scientific evidence tells us at this point so but now what we're going to do is we're going to do something smart and that is when whenever we see one study and we look at it and as in as impressive as this study is which I have to tell you it's very impressive it's a tremendous amount of work went into this the now what we're going to do is we're going to look on the internet for another study that looks like it but it's not the same people but it has the same kind of methods that they is the same kind of questions that they're asking we're going to find out did they have results that look similar if the results look similar then gives us confidence in everybody's results if the results are vastly different then that's then we're sort of back to the drawing board or not quite sure what to do so let's look at again net consumption and cardiovascular disease and what we're gonna find let's go do it again nah may I ask you something though I think you sent out some of them okay great well there's there's a few comments that I want to share with you if you don't mind what people keep asking what kind of nuts I'm not sure that's relevant we can also link to the study but ironic was saying well if you're that one person that lived then maybe it's relevant to you but I guess what I'm saying is when they do these studies are they taking people that are healthy vegans that are eating an A+ TrueNorth diet like me without nuts and comparing do you know what I'm saying so silly good question AJ the truth is obviously the six or seven percent of people that have vascular disease were the worst people in the study for diet health problems so we're talking about the F students that's obviously who that is so so the question is let's suppose that and but let's be fair that so over ninety percent of the people don't have any problems in fact it's more like ninety five percent of people it's the bottom five percent of the class that's flunking there now so the question is you know is there is there a protective effect if you're in the middle of the curve let's say you're just average eater and health or health person in this study the answer would be that the effect that were observing and that 1% would be vastly reduced because your risk factors are a great deal less than the people that are that are getting hit people that are getting yet or at the bottom of the pie out there okay so that's a very very good question that the study the risk factors and the numbers that you're seeing are not on healthy vegans we would only expect the people that would live more healthfully than the average person these studies would have very significantly attenuated numbers so that good very good question it's a good question that came from peanut gallery okay really did like how do you know it's the nuts that made them live longer what if they just said did blueberries instead of nuts I mean it just seems like good that's what we're gonna go so we're gonna know but before are we gonna go to that we've got time to do this today so we're gonna we're gonna take our time make sure that we get this right is what I want people to to have a really good look at look at this question because I know that if people are anxious I probably tried to answer this question a hundred times or 50 times in the last five years of that people coming up to me at conferences and tapped me on the shoulder and asking me they ask me all kinds of things should I take vitamin C when I get a cold I'm sorry they they ask all constant exam one of the things they do is they asked about this and and I said I've said up to now I don't see I haven't seen any compelling evidence and and that now remains to be true today okay now so now what we're going to do is we're going to look at we're gonna look again we're going to look for a very similar study to this one that is with a different group of people okay we also want something that looks that looks fancy but we don't here's a British Medical Journal that's really fancy okay 2018 so it's this year and it's Susanna Larsen and sweetie okay so it's a major study sixty 1364 Swedes were followed for 17 years really good okay so now this is extremely similar to the studies that we just cited it from the United States so this is a very large cohort study they're looking at them for 17 years and what they're doing in the background is that they said that consumption has been found to be inversely associated with cardiovascular disease mortality ie nuts have been found to help reduce mortality that's how scientists talk universally associated okay that means that consumption goes up cardiovascular disease goes down so people have found that just in the study that we just talked about which was found but the association between net consumption and the incidence of specific cardiovascular diseases is unclear in other words there's congestive heart failure there's heart attacks there's you know no ventricular fibrillation there's different kinds of issues so we're going to now look these guys are going to like take the take the magnifying glass out and they're going to see if there's you know if the nut consumption the effect that we that has been demonstrated is holds up across different kinds of problems so the problems that they're looking at for our heart attack heart failure atrial fibrillation and abdominal aortic aneurysm I don't even know what that is okay okay so they're gonna look at four different problems with their very large cohort and let's see what they found okay they found that that there was very similar kinds of things the people in see compared so that the the met the people that consumed a bunch of nuts as opposed to not a bunch of nuts were three percent less likely to have atrial fibrillation though that's actually very very very tiny effects on stuff okay the let's see but for heart failure let's see the corresponding hazard ratios were between 0.87 and 0.98 so that means that it's a round point we would have point eight six in our in our study okay the previous one this looks like theirs isn't as good and so the the hazard ratios are not as useful that was not associated with the aortic valve stenosis she mixed row or intracerebral hemorrhage in other words that consumption was not found to be helpful at all with the aortic valve stenosis ischemic strokes or intra sea bream hemorrhages those are out the only thing that they were able to find was a very very tiny effect for atrial fibrillation and also an extremely small effect for heart failure okay the the heart failure was significantly less let's see now nut consumption one to three times a month one to two times a week three times a week okay heart failure a charter school okay it was actually unbelievably small for heart failure so the they make sure I'm reading this right AJ compared to the consumption of no nuts the hazard ratios for atrial fib four point nine seven point eight eight point one two times a week and 0.82 the three times a week oh now they're a little bit better so it's a little bit of effect there that the heart failure it was point eight seven point eight point nine eight okay so the people that consume nuts huh great at more than three times a week had 98% as many heart failures as people that didn't consume any vegetable in other words was almost exactly the same strangely enough the people that consumed nuts one two three times a month we're at 87 percent which is very similar the 86 percent that we found in our study and 0.80 for people in the middle one to two times a week so what you get is a thing that says it goes down a little bit if you rarely it goes down a little more if you have one to two times a week and then then your risk goes back up if you're eating nuts more than three times a week it goes back basically to the very same ratios if you didn't have any nuts at all okay here's a clue in science when you find nonlinear relationships of a variable so a little bit of that seems to up a little bit a little bit more a little bit more than more uh-oh gets worse when you have a pattern like that you know that your variable is worthless okay are basically worthless it's a it's a death to scientists on that question it's like Oh God so now we're getting down to oh wow we're gonna reduce your risk a little bit if you nuts two times a week but if you go over three who then your risk goes back ups like God means you got nothing okay and all through this what we see is we see extremely similar numbers to the numbers that we saw in the United States so let's look at the conclusions conclusions of the study is that net consumption or your this is beautiful because we did not see this in the other study as part of inclusion so this is beautiful writing these findings suggest that nut consumption or factors associated with this nutritional behavior so there are that the nuts may not have anything to do with it but it may be something associated with people that eat nuts very important caveat good scientists ok these findings suggest that nut consumption or factors associated with this nutritional behavior may play a role in reducing the risk of atrial fibrillation may and possibly heart failure I need it's not looking so good for heart failure because it weird up at three times a week not suddenly we're back up to 98% the risk factors okay so what do we find from this we find that two different investigators looking at four different cohorts with an average of twenty years of follow-up on up each of these cohorts looking at over a quarter of a million people in two different societies find almost exactly the same numbers it's good that tells us that the question is basically settled scientifically and all points have it aren't settled but the effect size and of whatever nuts do poor cardiovascular disease we can see that the effects are extremely small okay the they are associated with something like at most a few days of life so that's where we are with this and that's fine it shows you that if you take out a telescope powerful enough you can see a footprint on the moon so we have taken out a statistical telescope unbelievable precision so we've looked at over 250,000 observations and so we we can we can grind this and see if there's something here there is and then there's something here that keeps showing up it's an extremely small relationship so the question is though once we've established the relationship between two variables and we have a feel for what it is if we were interested because we're academics and we said you know what there appears to be a tiny little thing there hangnail in our theory and it says you know what looks like more nuts as something about that is associated with the need I need it better outcomes I don't we want to know more now one person might say hey looks like nuts are protective let's eat nuts but actually folks that that would be that's one that's one person's opinion but it actually doesn't have anything to do with science that's just that's just taking a stab at something so what we have what we have established here is a correlation coefficient or Association we have no idea what the cause of that association is now one person might say well I've seen that if you eat nuts you know that's how good things in them and improves the lipid profile so therefore we think that nuts are definitely doing something usen't you got nothing okay so we're all we have is a correlation coefficient and it's an interesting question so we're gonna spend a few more minutes today AJ like academics just like academics we do bagus I was in my office back to the University I would take out my calculator take out a piece of paper and I would start thinking through what might be what's called the third variable so this is known as the third variable problem this is a classic problem in science and in statistics so we're gonna walk through the third variable problem so that people can understand it better and then we're going to I'm going to think about whether or not there's any possible third variable that could be associated with this effect the the suppose that you and I were political scientists God God help us AJ but let's suppose we work now make it economist political economists that's good that's better just kill ourselves now all right so suppose the we were political scientists and we were very interested in the wealth of nations and so we were interested in what was causing some nations to be successful their people to do well economically and other nations to be poor this is actually an unbelievably important problem in the world it's probably more more brains or trying to solve that problem than probably any other problem so there's a lot of people looking at that problem and trying to figure it out and so what they do is they study what we're going to call variables and so the variable of interest here is going to be per capita income in a country that's like heart disease you either have it or you don't where you got more disease or less disease that's an outcome variable what we're going to try to do is get a variable that's a predictive variable that's associated with it that we think could be controlling it that's like net consumption so if we got a heart disease over here some people have it some people down we've gotten that consumption over here some people have a lot of nuts some people know are these two things associated with each other okay so the first step in science that you know on some certain types of problems is we got to first see if there's an association okay so so now we're going to be a political scientist and we're gonna say okay we're gonna look at all the countries of the world we're gonna get numbers on how much money the per capita income nets and then we're gonna say to ourselves what do we think could be associated with it so we might say things like how much education in people have we put that as a variable how many years of education that's a variable to predict a variable okay you see well more education does that lead to higher incomes maybe does maybe it doesn't okay we're gonna run that number and find out a third so soon you could also say things like how far from the equator is the center of the country that's and so if somebody runs that and they say wow you know actually we're gonna make it north in the equator how far north in the equator is the country turns out there's an association the high of the further north you are from the equator the higher the income there isn't in the society it's like oh wow look about that maybe there's something about being north of the Equator and out of all the hot Sun helps people as economically denied or know somebody else says you know I think it's English speaking let's let's face the facts America's really rich people speak English England is pretty Reds people speak English Australia that's in the southern hemisphere forget about your theory about the equator there in the southern hemisphere they speak English they're pretty rich by Singapore now there are some very successful I think there is an english-speaking country okay I think we got something here I think English speaking is the variable right now so somebody else says yeah I don't think English speaking is the real deal I can see that you have an association but I don't believe that Association I think that there's a third variable in the thing it isn't English speaking causing people to be rich it's English speaking is associated with the fact that you have a democratic representative limited constitutional government with free markets that you're going to see in Australia and the United States and in England and if you're Venezuela you don't have that so we can't go to Venezuela it's sort of teaching everybody English and expecting them to look like San Diego and tenures because we don't think with the variable that's causing english-speaking countries to be rich is English we think it's a third variable that we're missing okay obviously that's true but it doesn't get so obvious when it comes to other things so the third variable problem gets to be interesting somebody might say for example the heart attacks they believe are caused by stress they don't think they're caused by God and it turns out that they think a big thing that causes stress is traffic and so they think that if you run a correlation coefficient between how many cars the country and how many heart attacks that have they think that they've got it and you know what they would the more the Morgan Motor Vehicles per country there's a country house the more heart attacks that they have the correlation is pretty good it's a hell of a lot better than nut consumption and heart disease it's a very good correlation so now some some psychologist says I don't care about all you people with your mumbo jumbo with cigarette smoking and diet and exercise and sleep anything else I think you're all crazy I think the issue is traffic stress and I've got a correlation coefficient here that is very strong and so you guys can just go pound sand because I believe that what people need is that they need to have less cars and less stress because when we go to the Blue Zones and we find out that people are living longer we're going to turn out but a lot of those people live in societies that less cars so I think it's frickin traffic now I know that when I'm in traffic I freaking feel really stressful I like feel like shooting people and I get short of rap and I'm yelling at people the car next to me so don't tell me that that's not a variable okay totally reasonable so what we're going to do well we've got to do really careful science we're going to look at everybody's patterns of Association and then we're going to try to figure out is there a plausible third variable if the third variable is actually a correlates with the outcome variable even better than this other variable over here then this set variable over here may not be so important so watch English democratic free-market government and well it's going to turn out the democratic free market government correlates better worldwide with well then does English so English is correlated with free market governments so that association between English and outcome it's there and we can see it it's a it's a correlation coefficient it's an association that is statistically significant but we see that there's another that's better and it actually explains what we're seeing more more crisply and has a lot of reason for us to believe that that's the issue so Switzerland has a democratic constitutional government they are wealthy just like the United States they don't speak English okay and so we start to see ah Luxembourg ah I see how this goes okay so this is going to be helpful so this is the third variable problem so we have a weak but statistically observable relationship between high note consumption and and very slightly lowered risk of cardiovascular disease interesting one person they say oh it's the nuts it's that I'm telling ya the guy with the idea that traffic stresses causing heart disease is frickin idiot okay he doesn't understand he doesn't know that all he's looking at is the countries that are wealthier have more cars cars people the countries that are wealthy or eat rich or food you know and as a result about the rich animal food that they eat is causing a heart disease he doesn't understand that the causal chain isn't what he thinks he owes but I believe that I know but I when I when I see this little association between nuts and reduced cardiovascular disease that in fact I believe that there's this chemical and not sort of a protective or anti-inflammatory etcetera etc okay so you know I've got a big story question is is there a third variable in the equation that's associated with net consumption that could be responsible for the outcome that is actually a better correlation than the nut is to the outcome well people aren't looking very carefully at that but when the Swedish study looked at that they would find that when they controlled for other cardiovascular variables the effect of nuts on outcomes went very low almost nothing so this starts to as the per people said in their conclusion nut consumption or behaviors associated with that with people that do that well let's think about something suppose and remember we're not none of these studies are on vegans or vegetarians or anything else on the Sun so we're talking about a few percentage of the people running into trouble we think we know who they are if we were to ask any intelligent informed cardio scholar about what they think the cause was if these people are not smoking I'm not sure if they will I think maybe the world I don't know if they were let's suppose that those two hundred ten thousand people were non-smokers and we've got six and point seven percent of them or having bastard disease let's take our first guess that what we think would be responsible for that vascular disease I know what my first guess would be astronomy and cheese sandwiches with mayonnaise for 25 years that's what I think the causes okay so when we look at them as 6.7 percent of people and we see that this fraction of people wanted the trouble and you're asking me why do I think they went rent ran into trouble it's because I think that their diet was very rich now with animal food now it turns out that there is 6.7% chance of a stir disease out of that group and I think I know who they were there were people that are eating cheese in patronise and which is some french fries and cheeseburgers that's who I think they were that's my hypothesis okay now now what we're going to do is we're gonna say okay but if out of that group of people that was in trouble I mean the net people they had 5.7 percent of those people well who do I think those are that ate nuts and also almost everybody today might nuts you know they wound up in the same boat exactly this would be able they didn't eat the nuts I think that the tiny effect size is likely to be responsible in the following way then it isn't that the nuts have some pretty protective lip soon enough instead think about this if I need nuts five times a week for 25 years we're gonna call that call that a peanutbutter sandwich right so if I'm eating five peanut butter sandwiches a week what am I not eating pastrami I'm not eating pastrami five times a week I'm eating peanut butter five days a week so I'm hardly surprised that if read peanut butter five days a week instead of something else that that's something else I don't think was freaking rice and potatoes or any other healthy food I think it was some rich food okay and now we say well now way to something if we subtract it out the rich food if we control for that and instead don't worry about whether it was nuts or not but if there's a dip in the rich food yeah in the rich animal food by X percent ie by with the equivalent of five doses of nuts a week let's run the correlation coefficients on people that ate whatever that is a couple of thousand calories less a week of animal food will we get that corresponding 1% reduction of cardiovascular disease I believe in well okay now I can't prove it I'm just a scientist with a hypothesis I believe that the third variable in the equation is that we are in is that when the increase of net consumption for that group that cohort has led to a reduction in the amount of animal food that they eat and therefore that change which is a modest change in the people's diet because you're only talking about you know probably you know thousands over 2,000 cars a week out of you know 15,000 calories so you get a change in their behavior and we wind up with a little modest change in their health okay not shocked so hopefully well I took a whole hour sorry AJ that's a bit this is a sort of a comprehensive look and you know whenever anybody comes to me you know sort of worried and worrying about what some study says somebody says they'll happy I don't know razzle berries here to reduce my odds of colon cancer and you know there's a correlation coefficient the doc says if this is really important it's like you know I never addressed none of these correlation coefficients like this whenever I hear anything like this we usually make much sense to me and the reason this is true is that it's actually relatively inconsistent with the theory of evolution it doesn't really make sense there's some magic food out there that is is going to stop some specific disease heart disease and cardiovascular diseases are caused by toxic processes as a result of dietary patterns that are inconsistent with the natural history organism the it is always possible that some healthy food or healthy food extract in concentration perhaps or some medicine could actually have a medicinal or pharmacological effect that might be useful in any given case that's always possible but is a general process I don't think that green beans are better than oranges and I don't think that oranges are better than blueberries and I don't think blueberries are better than kale and I don't think the kale is better than the potato and I don't think potatoes are better than nuts I don't think nuts are better than rice I think natural food is natural food and that whenever anybody starts looking for correlation coefficients between some specific disease and some natural food that are including Laura I think they're actually chasing or I think they're looking for health in all the wrong places in case anybody is interested I wrote an article by that name for this specific problem 20 years ago in health science magazine and it then became the chapter by that title in the pleasure truck okay so we morphed it into it's a Sherlock Holmes story about how Sherlock Ned's kind of look at the right place to find the answer to the mystery the people have a strong tendency to look in the wrong they seized on a correlation that looks exciting I understand this correlation or Association is what they're talking about this is where we begin a question and so when we hear that there's an association in science we say interesting how has it been replicated the answer in this case yes yeah is the is the Association strong not in this case it's not strong it's actually very weak okay is it if it is it isn't stable yes well then that's interesting scientifically may not be that interesting practically but it may be interesting scientifically so I believe that the Association nuts to vaster disease is scientifically interesting I believe that it is it is a very little practical interest but it is an interesting one and I think that some grad students somewhere may write a PhD thesis where they go back into this database and the databases that are available and they they may determine some vegan grad student maybe that I might be inspired by this discussion may be able to go into these databases and find out that the actual reason for that reduction and vaster disease turns out to be a reduction of the animal food as a result of the inclusion of medicine and I okay so I don't I'm speculating now that that could be the case and it would be kind of cool to us but for those of us that are in this arena the message is this don't eat don't get clot in a lot of 5 or 6% of the distribution where you wind up with heart attack or stroke day those people are were irresponsible with their lives and they ate okay and it turned out that a bunch of a made a bunch of nuts it didn't help Martin laid off and wound up with basically the same amount of pathology either way we know that there's people that made well and did intelligent things and didn't wind up with the pathology in fact most people ain't 3/4 Lea didn't want apology daddy but we know as healthy eating plant-based people our restrict we are that we should be doing an excellent job and we should not be in the bottom 10% of his class and put ourselves at risk okay from one last look at this to help you understand the statistical code synthesis if you were in the bottom 10% we are taking the 10% losers out of a hundred that the worst of the worst the genetic worst and the worst behaved people of the non net eaters seven of those ten wound up with a heart attack or a stroke in the net eaters the guys that ate a bunch of nuts that were in the bottom 10% six of them wound up with a heart attack or a stroke okay you can tell me whether or not or anybody else whether or not you think that's right and that's important and then you need to be really careful that's that judgment is up to you I know what I think and I hope that I've made that clear well there's rythmic is watching and he's a college student he said he wished he had a statistics professor like you I love teaching stats AJ I have to tell you I loved it so much that yeah it's kind of every now and then in this life I think God I can't believe that I'm not teaching stats because you know there's jobs that you have in life like I don't know that there may have been some jobs that you just absolutely loved and there's two jobs that I loved I loved delivering sprouts for dr. Alan Goldhamer in Portland High was his I was his driver and he paid me six dollars an hour now nothing has changed I loved that job but it was finally time for me to move on my life to go to school I loved with driving Sprott truck for perfect sprats they're the cute girls who would say oh it's the perfect sprung iron I'm like yeah that's me nothing good happened but it was fun and the other or the fun another funnest job that I ever had was teaching statistics and so now now obviously we do new things but sometimes sometimes it's it's it's fun to look back on those days that was great well thank you so the take-home message everyone is saying they understand is don't worry about having arugula versus artichokes or kale versus chards don't eat you heard it today from that line that's what I said perfect thank you so much dr. Lyle and thanks all of you for watching another episode of healthy living live I'm chef AJ I make healthy tastes delicious but dr. Lyle makes it understandable take care everyone
Back to the top
🏃     👖




Artist