Why no diet wins (what 40 years of nutrition research shows)

Why No Diet Wins (What 40 Years of Nutrition Research Shows) | Dr. Christopher Gardner & Mike Haney

In a recent episode of A Whole New Level, Levels editorial director Mike Haney sits down with Dr. Christopher Gardner, professor of medicine at Stanford University and director of nutrition studies at the Stanford Prevention Research Center. Gardner has spent more than three decades designing and running nutritional intervention trials — randomized controlled studies in which real people change their diets so researchers can measure what actually happens in the body. His work spans landmark studies on competing weight-loss diets, plant-based proteins, fermented foods, and the surprising limits of what any single study can tell us. He has served on the Dietary Guidelines Advisory Committee and is one of nutrition science's most prominent voices for honest, humble communication about what we actually know.

The conversation covers why randomized controlled trials are both the gold standard and deeply limited, what decades of diet research actually show about low-fat versus low-carb eating, why food quality matters more than macronutrient ratios, what the science really says about ultra-processed foods, and why — despite all the complexity — the core message of good nutrition has barely changed in 30 years.

At one level it couldn't be more simple — and at another level it's very personal, cultural, physiological. There are some tweaky things around the margins.

— Dr. Christopher Gardner

What a nutritional interventionist actually does — and why it's hard

Mike Haney: Dr. Chris Gardner, thanks for joining us.

Christopher Gardner: Pleasure to be here, Mike. Let's do it.

Mike Haney: So, by way of introduction, I've heard you refer to yourself as a nutritional interventionist and that that is kind of your superpower. What is a nutritional interventionist?

Christopher Gardner: Yeah. So, in the world of nutrition, you can study mice and you can study test tubes and you can be an observational epidemiologist, which means you have a 100,000 people or 200,000 and you every once in a while email them and track them for 40 years. There's all different kinds of angles for trying to get at nutrition questions. And one of the I think least frequent is somebody who actually asks people to sign up and change their diet, track them over time, see if they really did change their diet, to what extent they did, bleed them, take their poop, find something that might have changed because they changed their diet. So that's the nutrition interventionist part. It's a really — you have to have a really thick skin. People don't want to sign up. People don't want to stick to it. People aren't always completely honest with you. Cost craploads of money to do it. So not that many people do this thing that I do.

Mike Haney: And what are the knocks against it? I feel like you know — you do randomized control trials which I think even to the lay person they hear "randomized control trial" and they sort of have some sense that that's the gold standard. But as you say there are other forms of research that have their own purpose. And you've certainly taken guff about some of — I think every big study you've done — there are people have criticisms of it. What are the valid knocks against RCTs, or what are the shortcomings of RCTs and in particular intervention RCTs?

Christopher Gardner: Yeah. No, absolutely. So the biggest one is how narrow the questions are if you're going to intervene in the first place. So let me differentiate also sort of dietary supplements, which is kind of like drugs, versus diet and foods and things. So if all you wanted to know is fish oil or something that could come in a pill and nobody had to change their diet, they just got assigned to two pills, a placebo pill or an active agent — you don't have to change your diet. Those can run for a long time. But if you wanted somebody to say, "I really am curious how a vegan or a keto or a paleo diet will work" — so Mike, in order to understand coffee, to see if it will prevent cancer or heart disease, I need you to agree to be randomly assigned to either drink coffee or avoid coffee for the rest of your life. No one will do that.

So in the kinds of studies I do, I say, "Look, I think in four weeks or eight weeks or six months or a year, I could get this health thing to change. So let me see if I can put you on a vegan versus a keto diet, or let me see if I can give you more garlic or soy, and please agree to come in for some blood samples, and please agree to let our dietitians call you to see to what extent you're actually following our advice." And then I'll publish a paper and I'll say when we randomly assigned this one group of people to this thing — and usually we have to pick one group. Are these healthy people? Are these people with high cholesterol? Are these older? Are they younger? Are they athletes? I usually don't get to pick everybody. And then I pick a dose and then I pick an outcome.

So if I could give you my favorite example, which I actually have to talk about for another talk tomorrow. One of the best studies I ever did was a garlic study. We got 200 people to either have one of two types of pills or a sandwich that had a condiment that had a clove of garlic in it. The clove of garlic was in the condiment and it was to see if garlic would lower cholesterol, and it was NIH funded, and it took me four years of my life, and neither of the garlic pills or the California Early Harvest garlic — none of them had any impact on LDL cholesterol.

I finished the study and I said, "Wow, this is pretty disappointing. There's like a putative active agent here that could have done this. I'm sorry it didn't work." And all of a sudden my email was flooded by supplement companies who said, "How do you know it doesn't work with my supplement? I have a different supplement." I said, "Well, that's fair." I had a garlic farmer call me and say, "I grow 90 varieties of garlic. How many of the varieties did you try?" I said, "One. California early. That's the most common garlic consumed in the US." He said, "Well, then you don't know anything about my other 89 varieties of garlic." I had somebody say, "What if you had started with people with a higher cholesterol level? You had people with a moderately elevated LDL cholesterol." I said, "But that would take doing another study." Somebody said, "Blood pressure — what happened to the blood pressure?" It was a six-month study. Somebody said, "What if you've done it for a year or two?"

Just being incredibly modest here — two years, 200 people. We tried two types of supplements and real garlic. We had excellent retention. It was six months, not four weeks. We did so much more than anybody had ever done in this field. And all their criticisms were correct. The question I asked was one dose in one population for one outcome over one period of time. And 400 newspaper articles covered this, because Stanford was tracking how often we get in the news and they shared them with me, and one of them said "Stanford finds garlic isn't good for you." That was the headline. Like, okay — it wasn't Stanford, it was me and our lab group, so don't blame Stanford. And it was the most narrow of questions. Like, did anybody want to ask me if it cures vampires and keeps vampires away? It still surely does that. I know there are published papers on that.

But all those other questions are very reasonable separate questions that would have to be answered another way. If you were doing observational epidemiology, you would have tracked hundreds of thousands of people — they would have been older, younger, some with high cholesterol, some with low cholesterol, some who ate a lot of garlic, some who ate a little garlic — you could actually address almost all the questions they asked in one study. But it would be association, not causation. So really the best idea, Mike, is to take our study and try to match it up with others for a more complete picture. And that's what we always do. None of my studies are ever definitive for the whole field.

The art of study design: dose, duration, and population

Mike Haney: So how do you think then about taking that narrowness of what you do and generalizing into advice? Because you do a lot of public education, your studies because they are very big and very well-run get a lot of coverage. And I get the sense just from listening to you that you are genuinely interested in helping people figure out what to eat. Your passion feels like it's less a very narrow mechanistic thing. There are scientists who just study one particle all their lives and just want to sort of advance their narrow thing. Your mission feels broader. So it feels like you want to be able to land at some place of being able to tell people something we have learned about what to eat. And yet, as you say, these RCTs by definition have this narrowness to them. How do you square that?

Christopher Gardner: Okay, great question. And so to be honest, most of physiology is pretty similar. Is there a male-female difference? If it is, it's really modest. Is there a young-old difference? There can be, but it's really modest. Is there a dose difference? Sure. And I actually labor over trying to figure out what that dose is. Like, if it's too small, nobody would expect anything. And if it's too large a dose, everybody will say that's nuts, nobody could ever eat that much. So you kind of try to find some very reasonable level. And I feel like in a lot of our studies we pick something that's probably generalizable to a lot of people. It isn't all the questions, but it at least addresses it more than the mouse study or the test tube study or the observational epidemiological study. It's a pretty reasonable contribution.

And another thing we often do, Mike, is we say we are going to cut out like the five or ten percent of the population that is the sickest and the third of the population that's super healthy. So we have some inclusion and exclusion criteria. So we would like to help people in need of help. The ones that are really sick probably need to see their doctor and more extensive intervention is involved. But throughout my life, most of our studies have usually been generally healthy people who have modestly elevated cholesterol or blood pressure or insulin or weight, and they would be interested in not going on drugs someday and doing it with lifestyle. So this is going to be relevant to a huge swath of the general population. That's kind of how we decide what to do. And then try to be as humble as we can and say, but it's limited to the narrow thing that we did.

Mike Haney: Is dialing in those variables — dose, duration, population characteristics, population size — is that the thing that you do particularly well? What have you gotten good at over 30 years of doing this?

Christopher Gardner: Well, I'm not sure if I'm good at it, but it's what's really fun. Like get a bunch of people in the room and what would be the most reasonable dose to use? One clove a day of garlic was what we did. We used Beyond Meat products to do a plant-based meat versus animal-based meat and it was two servings a day. We did a fermented food study for the microbiome. We picked six servings a day as a goal and people were aghast — six servings? I got to eat six meals a day of fermented food? No, actually a serving's pretty small. Like a serving of kimchi or sauerkraut is like 50 calories and it's half a cup. And a bottle of kombucha is two servings that you'd buy in the store. So it was actually — you could get six servings of fermented food with 300 calories. It would actually be a relatively small part of your diet.

Duration is one of the biggest deals. So if you were to do a trial where we're going to try to lower cholesterol or blood pressure or insulin or glucose or something like that with diet — if you had a stable diet A and you switch to a stable diet B immediately, and sometimes we do this by having food delivered so they don't have to learn how to shop and cook and buy — if you do that and they do it consistently day after day, most of those metabolic things, the changes start to plateau in two weeks and then by four weeks they've really plateaued. So if you wanted to see what that physiologic metabolic impact would be, four weeks is usually fine. We sometimes make it eight just to get the point across that it was stable at some time and we did hit that plateau. But when people say why didn't you do it longer, why didn't you do it for six months or a year — the higher the participant burden, and one of those big factors is duration, the less likely they'll stick with it and quit. The more they quit, the lower my retention is, the lower my retention is, the more I lack internal validity. So it's a balance.

Can I tell you one of my most fun ones was lactose intolerance — giving people raw milk versus pasteurized milk? And you had to have lactose intolerance because otherwise there was no room for benefit. And I said, "How much would I have to pay you or how long would you be willing to do this?" And it was really fun working out this thing where we said, "Okay, we're going to start with four ounces and on a daily basis we're going to go to four to eight to twelve to sixteen to twenty to twenty-four ounces of milk." The study period is only seven days long. As soon as you reach that point where you are intolerant of the symptoms, you can stop. That's actually part of the study — to see how long it takes you to become intolerant. But seven days is enough. And in that particular case, Mike — it depends on the outcome. How long does it take for you to experience a difference in the outcome? Well, if you're lactose intolerant, that same day — if you didn't tolerate, you've got diarrhea and you've got abdominal sounds in a day. So really a day is enough, but we did it for a week. To me that is really part of the fun of science — balancing out the dose, duration, participant starting point to get at the most generalizable finding you can.

To me that is really part of the fun of science — balancing out the dose, duration, participant starting point to get at the most generalizable finding you can.

— Dr. Christopher Gardner

Equipoise: why both diets have to be good to get a real answer

Mike Haney: One of the things that's going to come up a lot in this conversation is food quality. And one of the ways that you address the idea of food quality in your study design is equipoise. Can you talk about the concept of equipoise?

Christopher Gardner: Yeah, that's — I'm not sure if I'm using that word correctly, but I'm saying it's one of my new favorite words. If you look at the literature, when the public gets confused, it can easily be because one day somebody got a headline that said diet A is better than diet B and the next day it says B is better than A. And if you're not a scientist, you say, "Wow, that's super frustrating. I guess there's no real answer here." And the closer you look, what I often find in my field is that as much as we'd love our scientists to be unbiased and completely objective, sometimes if they really want to push their diet forward, it's pretty easy to make a kick-butt diet A and a crappy diet B and have diet A win, and flip it — and the next day have the proponent of B have a kick-butt diet B and a crappy diet A. And then it sounds like it's an opposite conclusion. When you look more closely, it's like, "Oh my god, that wasn't really fair. Those were very different contrasts, even though the headline said A versus B."

So what you're referring to is at least a couple of studies that we've done where we really paid attention. One was having a really good quality low-fat versus low-carb diet for weight loss. One was a Mediterranean versus a ketogenic diet for glycosylated hemoglobin control for people with diabetes. One was vegans versus omnivores where one group went vegan and the other group stayed omnivore. And in all three of those cases, we really paid attention to having both diets be good representations of what somebody would say, "Wow, I'm a low-carb fan" or "I'm a low-fat fan — Gardner gave our diet the best chance." People have said, "Oh here's a Gardner study — he's a prominent vegan, he's probably just trying to push his thing." Well, we actually improved the diets of the omnivores in the comparison. So we can say we really tried to give omnivores the best chance they could, which to be honest diminishes my chance of seeing a difference, right? So if there was to be a difference, it would be easier if I had a great A and a crappy B. But if you make a great A and a great B and say I put them head-to-head and there was nothing there, then that's really a more reasonable conclusion.

I'm looking for more consensus in the field. Controversy is clickbait that sells, but it messes with the overall nutrition message and I think we lose out in the end.

Mike Haney: I want to come back to a bunch of those specific findings and also the controversy versus consensus point, but a little bit more on the idea of how we conduct and then communicate research. One of the things that I struggle with — because I'm on the journalism side, so I'm reporting on and communicating out study results to people — is maybe we just shouldn't write about individual studies, because of the confusion thing that you just mentioned. Even when you have a really well-done study, any study is one learning, and knowledge — what we actually know — is the accumulation of lots of stuff, good and bad, various quality, and some sort of synthesis of that over some long period of time. But the thing that gets communicated is "garlic is bad for you," "blueberries will save your life," "red wine will make you live longer." And then the thing that changed my mind about it a little bit was listening to you talk about your experience with the Netflix documentary. You did the twin study — it was omnivore versus vegan — and it was the most covered thing you've ever done because it was paired with this extremely well-viewed Netflix documentary. How do you think about the value of communicating out the results of a single study?

Christopher Gardner: Sure. That study that got turned into a Netflix docu-series was not the best, most clever study that we ever did. It was almost a no-brainer. The main outcome was LDL cholesterol. That's affected by lowering your saturated fat and increasing your fiber. A vegan diet has to be higher in fiber and lower in saturated fat than an omnivorous diet. So, oh my god, we lowered LDL cholesterol. But what was really fascinating was putting that narrative into the lives of real people struggling with it — some are having an easier time than others trying to be vegan — pairing it with stories about animal rights and welfare and climate change and making a whole narrative out of it and talking to people starting businesses. Chido Hoyos, who did this, who produced and directed this, did a marvelous job of tying a whole narrative around it to make the narrative engaging.

One of the more rewarding social media posts I got was somebody named Lizzy, PhD, who wrote, "This is the best science communication I've seen in a decade." Part of the reason it was the best is because people didn't know they were getting science — they were just entertained. And I went in to look at the rest of her post and figure out who she was. She had been one of the investigators at Harvard that got the COVID vaccine made and out rapidly. And so what I'm reading between the lines is her frustration at getting out the vaccine that cut short the pandemic faster than anybody thought possible, and having people not take the vaccine because of a distrust of science.

Some people had criticized our study and the portrayal of our study in the Netflix docu-series because we had dumbed down the science. And that was an epiphany for me — that's actually the wrong way to phrase this. We have made the science so elitist and with so much jargon that the average person can't relate to it. So yeah, let's try harder to make it more engaging.

I had a really fun comment the other day from somebody watching me, in response to something, one of those little comments that follows an article, saying "Wow, Christopher always seems to be smiling when he's talking about science." I noticed that he's always got a big grin on. And I love science. I love the challenge of designing a question and then trying to narrate the story and make it engaging. When you're super dry-faced and objective — "here's the facts, here's this" — it puts you to sleep. So this idea of making it more accessible, which journalists have tried to do forever — that's your job, to make my study engaging — I think there's a lot to learn from making it more interesting and engaging, despite the fact that probably the biggest conclusion to come out of the next decade is going to be "eat more vegetables and beans." That's so boring. Yeah. But here's why, and here's the farmer's lives who make them, and oh my god it's related to the environment — we can weave a whole engaging story about that, and that is really fun to talk about.

On swearing, smiling, and making science relatable

Mike Haney: I think it's what also makes — you've got the credentials of having done this a long time, you've got a fancy university behind you, and you do these gold-standard RCT studies. But also as you're trying to communicate to people what to eat, you don't sound punitive. Just as somebody who listens to a lot of researchers talk about this stuff, your sort of joy in this — I think there is something to that. I think if you were teaching your colleagues how to communicate about this stuff, asking them to smile more and talk about why this is fun — I think that is a bit of an unlock.

Christopher Gardner: Yeah. I have another funny thing to say. I started swearing more. So I remember in a talk at least three or four years ago, I said, you know, it's a shitty job, but somebody has to do it. Something like that. And I got a good laugh from the audience. And the laughter keeps them awake as they're listening. And I remember a post-doc that came up to me afterward and said, "Oh my god, thank you so much for your talk. I relish the day when I'm senior enough that I can swear in front of an audience." And I thought, what a bizarre comment — you aspire to use foul language in front of a medical audience. But every once in a while now I will just say, you know, use an f-bomb or something else and say, "Ah god, we were trying to do this and this thing happened. Damn, that sucks." And it makes it a little more — I hope a lot more — relatable. It's like, oh, I'm not calling anyone names or using bad language, but somehow it softens up my delivery. It's like, oh, this is someone who's actually not just talking science-speak, but he's talking to us as an audience and trying to keep us engaged as he gives his scientific message.

Industry funding and how to read research critically

Mike Haney: Well, that kind of transparency that you use when you talk about this stuff — being willing to say when a result sucked or a study — I want to hit on a couple of things before we leave this topic of studies in general. One is industry funding. You've taken industry funding to do research. As a journalist, that catches my eye, but I'm also very skeptical of the outright dismissal of industry funding. How should a person who's reading a study — or more likely the press coverage of a study — who hears that it is industry funded, how should they interpret that?

Christopher Gardner: Yeah. So if you can figure this out — most of the industry funding that I'm aware of doesn't give the investigator any money personally. They're not taking extra vacation trips from this. They couldn't get it funded by the NIH or other groups. So they're looking for funding and they got it here and they still did it objectively. Do they have stock in it? Are they hawking something? If you don't hear any of that, you can probably be a little more open-minded.

I have to say that we found a positive result of Beyond Meat versus red meat in our study called Swap Meat. And I was preparing for an onslaught of conflict of interest. And I pointed out that at least six times I've taken money from industry and published null findings. So I was citing my track record of failure. Okay, it wasn't really failure — I got null findings and I published the null findings, because those are just as important as positive findings. You don't want anybody to have to repeat this because you did it and it didn't work.

So I think one possibility is looking at a track record. And I tried to make it very publicly available — here are A, B, C, D, E, F where we got funding from industry and it didn't work and we said it did not work. I guess you've got to be looking for someone who is clearly expressing the limitations — all of them, not just conflict of interest, but "oh, we had high dropout," "oh, we never met our recruitment goal." There are so many things that get in the way of a perfect study that I think the more you could hear that honesty from a scientist, the more you'd be less likely to dismiss it and more likely to say, "It looks like a pretty well-conducted trial and they are pouring out all the limitations. Okay, it's reasonable. I should wait to see if somebody else replicates it, but I should be cautious." You should definitely be cautious.

The dietary guidelines: a more rigorous process than you think

Mike Haney: A related topic that I want to hit on — you actually served on the board that worked on the revision of the dietary guidelines, the every-five-year revision. That's another space where I think there was some real reflexive dismissal or at least a lot of skepticism or griping. What do people need to know about that process and the utility of those guidelines that maybe gets lost in some of the communication about them?

Christopher Gardner: Sure. So a huge step back: for the last however many sessions of this, on that every-five-year basis, a group of scientists gets selected to look at all the literature that's come out since the last one and provide an advisory report to the secretaries of USDA and Health and Human Services. And then we let it go — and the secretaries don't have to take any of our advice. The dietary guidelines that come out could have been reshaped by that report, but it's not up to the dietary guidelines advisory committee. All we get to do is give advice. And I think there's a misunderstanding that we rewrite them.

So first of all, we don't. And it's well known that the secretaries historically have often dismissed the suggestions. Let's just take added sugar: for quite a while, the US, unlike the rest of the world, had an upper limit of 25% of calories from added sugar. No more than 25%. WTF — 25% of calories from added sugar. That is obscene. So at one point they lowered it to 10%, and that was reassuring. And the next dietary guidelines advisory committee said it should probably be even lower than 10%. And the secretaries did not accept that — they left it at 10%, surely because of some political blowback from the food industry that didn't want it to be that low in sugar.

So people need to know that this group volunteers to do this. You have to be kind of nuts to do this. When they call you up, they say, "Okay, you're going to be a government contract employee, but you're not going to get paid. You have to work with us for two years — anticipate ten hours a week for two years of your time." Over the last several rounds they've made it more and more methodological and rigorous, so no one can stack the deck. You have to get all the search criteria ready and then you hand it off to a group of federal staff members who find all the papers that exist on that topic. Then they help you extract the data, and a couple of people read it as their primary question and the rest read it as something else, and the whole group has to agree on your conclusion by the time you're done. After you come to a conclusion, there are five criteria that force you to say whether this is a strong conclusion, a moderate conclusion, or a limited conclusion. And most of the things that we looked at — we looked at 80 questions, many of them had sub-questions — most were limited conclusions, and some of them were "no data available."

My assigned role was food sources of saturated fat. So the main question was, how should we look at food sources of saturated fat? Mike, there were 31 sub-questions to that question. So we had to potentially come up with 31 conclusions, have a directionality, say if there was enough evidence, say if it was strong, moderate, or limited, and then present it to the whole group and have them sign off on it. Most of them were limited conclusions, and some of them were no data available. And it couldn't be more dry and boring and objective and rigorous. And in the case in particular of ultra-processed foods, we kicked the can down the road and people were mad. "I can't believe the dietary guidelines advisory committee only had one limited statement on ultra-processed foods — that's a huge issue." Yeah. But according to this method, there weren't enough data to answer the questions that were being posed the way they were being posed. And so the conclusion was "more data needed." We can't say more than that. It's set out for us — the method doesn't allow us to say that.

So it is really objective, really thorough, really impressive. I myself had cynically mocked some of the past conclusions of the dietary guidelines advisory committee until I served on it. And I was totally impressed, absolutely impressed — not only with the whole process, but with these twenty different scientists from all over the country and many different disciplines, how well we worked together. When the evidence is in front of you and you're surrounded by colleagues, you can't get away with anything. It's like, "You're right, there isn't enough." Or, "There's a lot of evidence and it's pretty consistent and you really can't look the other way. Yep, that's what it is."

I myself had cynically mocked some of the past conclusions of the dietary guidelines advisory committee until I served on it. And I was totally impressed.

— Dr. Christopher Gardner

Ultra-processed foods: a new name for an old problem

Mike Haney: Well, unburdened by those structures around how we evaluate this, let's talk about ultra-processed foods. If there's one theme that I've seen come out of a lot of the work you've done, it's that the real lesson is it doesn't really matter what the diet is, just eat real food. And the opposite of real food is ultra-processed foods. Why are ultra-processed foods bad for us?

Christopher Gardner: And I almost might push back — they're no worse than foods high in added sugar, sodium, and saturated fat. There's tons of data on that, and those have been messages from the health community for decades. The ultra-processed thing — Carlos Monteiro, who came up with the NOVA classification, has four classes. NOVA 4 is ultra-processed. David Kessler, the ex-FDA commissioner who nailed the tobacco companies for nicotine and food addiction a long time ago, has turned his attention to this and he thinks "ultra-formulated" is a better word for it. The reason being that there are two important terms here: processed and formulated. Processed, I think, means chopped up, cut up, extracted — something physical or chemical. And formulated is what you added. Did you add colors? Did you add flavors? Did you add emulsifiers? Did you add gelling agents, anti-glazing agents, something?

The NOVA that Carlos Monteiro came up with, I think very interestingly, is that there may be something beyond salt, sugar, and fat — what he calls "cosmetic additives." I think it's a brilliant term because cosmetic makes it sound like, "Oh, it's just for looks — it's not for nutrition, it's not for health, it's just so I'll buy it when I go to the grocery store." His point is there's something beyond the salt, sugar, and fat that is cosmetic. And at the end of the day, he's kind of right — except almost all the foods that are ultra-processed are high in salt, sugar, or fat, or a combination. So it's kind of like another name for junk food.

And one of the ones on his list is turmeric. Turmeric is sometimes added to make the food yellow. And yes, turmeric is supposed to have curcumin in it and be this cool thing that might help you metabolically. But if the purpose was cosmetic, then it's ultra-processed. And if the purpose was to spice it, it's not. So that seems kind of arbitrary. I'm not so keen on turmeric being a defining factor of ultra-processed food.

But here's what kind of cracks me up: we've been saying "have less added sugar, sodium, saturated fat" for decades and Americans haven't really done much of that. And if all of a sudden you say, "Yeah, but the colorants and the flavorants and the emulsifiers" — and suddenly people go, "Oh my god, thank you for pointing that out, I'm going to have much less of those now" — it would have been almost all the same foods we were saying were on the naughty list for the last several decades. If messaging it differently leads to a bigger response from people, I'm all in. I would say, "Wow, that's kind of funny." I don't really think it's the colorants and the flavorants as much as the bad nutrients that are in there and the lack of fiber and the lack of vitamins and minerals and the lack of being whole foods. But wow, if that's a way to message it and have it work — as a behaviorally-oriented person, I really want people to eat better food. So sure, I'll buy that.

Hey America, you should be mad. You're being manipulated. Companies are doing this and marketing it in a way that it's not your willpower. The GLP-1 drugs have been pretty fascinating in this regard — if you take these GLP-1 drugs, you're not hungry, you're just not thinking about it anymore. It's like, "Wow, you've developed willpower." No, you didn't. You took a drug that just cut off all that marketing and cosmetic stuff that was hooking you before. Now you're just not thinking about eating that much.

Mike Haney: The danger of the ultra-processed foods framing — or the way maybe people think about this — is less about going in and scrutinizing the ingredient list to look for something that you can't pronounce or don't recognize. Maybe that's useful as some kind of model. But the danger is not that thing that you haven't heard of that just sounds a little bit scary, or that maybe some influencer told you was scary, and maybe there's a rat study that shows that it causes cancer at twenty times the dose you would ever take. That's not really the concern. The concern is that these things are all being put together to make that food ultra-palatable and very attractive and shelf stable and cheap — all the things that would make you want to buy it — but that food is also lacking in all the nutrients, the fiber, the natural micronutrients, the things that you might want, and it's loaded with the stuff you don't want. One of the things you've said which I like is: people don't shop for nutrients, they shop for food.

Christopher Gardner: Absolutely. And that's not clickbait. That's not so fun to hear. That's not, "Oh my god, this is so fascinating — there's a superfood, a super ingredient, and if I just take this, the rest of my diet can be crap. As long as I do this one thing." No, it's really day after day making good choices, eating whole foods. It's just not very exciting to tell it that way.

Simple or complex? How decades of research land on Michael Pollan's seven words

Mike Haney: Well, that leads to the real reason I've called you here today. There is one question that I really want answered. I find myself saying to people a lot now that nutrition advice is actually pretty simple — it's kind of the Michael Pollan thing of just eat real food, not a lot, mostly plants. And if you just do those things, you're probably going to be fine. At the same time, the other thing we find ourselves saying a lot as communicators and scientists is nutrition is really complex — everybody's really individual, look at this range of outcomes in this study, there's the microbiome and the genetic differences. How do you think about nutrition being basically simple versus incredibly complex and individual?

Christopher Gardner: Oh, this is going to be so selfish and self-serving. I need grant money and I need to publish and I need to keep my job. So if it is complex, you need me.

Let me tell you what an Italian nutrition scientist told me in 1993, right after I got my PhD and my post-doc job at Stanford. He said, "Wow, welcome to Stanford. Good job. It's odd that you picked this time to get your PhD because we actually now know everything about nutrition. So there's not much left to know. But good for you. Good luck with that job." I thought, seriously — I just got the thing and I was going to make a career out of this. And looking back 33 years, I feel like he was right in a sense. Except at that time, we didn't call them seed oils or the hateful eight. And so somebody needs to tease that apart and explain what that means. We really didn't have any plant-based meats. And people had made lentil burgers, but they hadn't really made Beyond Burger and Impossible Burger. And so a lot of things have come up over time — somebody found a new way to eat. "Is this really good for us or really bad for us?" That part's really complicated. But I am so enthralled by and annoyed by Michael Pollan. He nailed it. Seven words and he nailed it. That was like the best comprehensive statement ever made. And did you notice how many citations he has after that? None. It's just seven words. I can't get away with that as a PhD — I have to have 150 citations after those seven words.

So Mike, one of the things I do like to say — I was asked a couple of years ago when I was chair of the American Heart Association's nutrition committee to take advantage of the new 2021 American Heart guidance and all the popular diets out there and say: we now have these ten domains of advice from the American Heart Association. Christopher, we don't need you to show all the evidence for whether one diet pattern or another works — can you just descriptively say which components of the different popular diets align with the American Heart's ten domains of advice? And it was a really fun exercise because I didn't have to prove that they worked or not. I just had to sort of describe them. And it was really fun because not everybody describes Mediterranean the same way, and people follow ketogenic in different ways. So it was really fun to sort of say: okay, this is paleo, this is keto, this is vegan, very low-fat, and here are the guiding principles of those. And as we ran the full spectrum — if I can point out the two extremes, one would be low-fat vegan and the other one would be ketogenic. I can't think of anything more polar opposite than those and then eight other patterns in the middle. All ten that we came up with said: more vegetables, more whole foods, less added sugar, and less refined grain. All ten diet patterns said this.

Now, this would be super boring if we said everybody does that and people eat lots of whole foods and plenty of vegetables and they avoid added sugar and they avoid refined grain. Except it's the opposite — those are all the things that Americans do wrong in a huge way. And so if we could just have a kumbaya moment here around simplicity. Yes, we can argue about omega-3s and omega-6s and inflammation and curcumin and spices and insulin resistance — there's some really cool physiology things to discuss and maybe tailor from one person to the next. But it is amazing how much consensus there is around things like Michael Pollan's seven words that very few people follow. Why? Why is that, when there's so much consensus around the core foundational diet components that everyone agrees on, that we don't do it?

Do lentils have lectins in them? Oh, I wonder how many lectins I'll get if I have that chickpea. No, no, no. That should not be at the top of your list. It should be the whole foods, the vegetables, the added sugars, and the refined grains. And the next category — the next ring around those that I always talk about — almost full agreement on: lots of beans, peas, and lentils, and fruits and nuts, eggs are fine, fish are fine, vegetarians — even though they have a face — and maybe not all of dairy, but yogurt. That would be my next six after the core four. That would be like 75% of most people's diets right there, for which there's like full scientific agreement, and people don't do them. Boring.

Mike Haney: Why don't people do them? What have you learned, thinking about this and doing studies around it?

Christopher Gardner: Oh, we're so good at marketing and we're so good at marketing to kids, and we're so bad at funding schools. I mean, one of the places a lot of folks set some of their lifelong food habits is in school. I think America more than anyone else is afraid of inconvenience. So they want it to be convenient. They want it to be low cost and they want it to taste good. The food industry has figured out — and David Kessler wrote in his first book, The End of Overeating, this capacity of layering on salt, sugar, and fat in different ways. He has one chapter that's salt and sugar, and another one that's sugar and fat, and another one that points out that you can make a food where you layer these things one on top of the other to make them hyper-palatable. And we can make fun colors and we can do great marketing and it's America and it's capitalism and it's convenient and it's low cost. So there's just a ton of factors in there that are making it super easy and low cost to eat poorly and have dopamine reward pathways fire in your brain. And we're not very good at being long-term thinkers. It's short-term gratification.

I'm not sure if I got them all right or all in the right order, but it's pretty easy to point to a lot of different things that are all coalescing around — it's really hard to do this. If I could sort of quote David Kessler's first book that he wrote — I remember his opening, I'm not sure if I remember this correctly, I think it was 2009. He said, "Look, I have an MD. I've been the dean of two medical schools. I have a law degree. I have a business certificate. There's a chocolate chip cookie in front of me right now. I am smart enough to know that that chocolate chip cookie is probably not good for my health given all my medical background." The cookie is gone. He ate it already. The point I took home was: I'm not stupid. Probably there are very few people as smart as he is in the whole country, but he ate the cookie. It's not an intelligence issue. It's the level of stress and time pressure and interest and satisfaction and flavor profiles. It's all been set up against you and it works, and a lot of people make money when it works.

Mike Haney: Can I throw another one in there that sort of indicts my profession? Because I also wonder sometimes if part of the problem is that when people go to eat healthy or do what they think is going to be healthy, they end up following one of these diets. And as you've shown and as we're going to talk about more, most of these diets either don't work or the diet itself is not the thing that's going to make it work — they're very hard to stick to. Or people are seeing conflicting headlines: keto's the best, vegan's the best. And they just sort of throw up their hands and go — if people understood that consensus, and didn't have that sort of almost nihilism of "well, nobody knows so I'm just going to eat the cookie." How much do you think that's a factor?

Christopher Gardner: Absolutely. And that is super frustrating for me — that people make it out to be more confusing than it is. But you're right, you and I are both to blame. Just to be super simplistic here — if I want to get promoted and I did a study and it got funded and it got published, and my media folks at Stanford are excited, you know, I'm feeling a little egotistical today. Maybe my findings are a little more important than I'm stressing right now. And then Mike, you have to publish this — your editor says, "That headline isn't very catchy. I think it would be catchier if we said this refutes everything we thought we knew for the last 50 years." And between the two of us — well, I want to keep my job and I want to make it interesting. And then the consumers are confused and they throw up their hands, and it's not serving us well in the long run. So I'll take some responsibility, and you already have.

A to Z and DIETFITS: what 600 people and a year of data actually show

Mike Haney: Let's get into some of the studies that you've done. So let's talk about DIETFITS. I'll do the 30-second summary and then you can expand. 600 people, year-long, low-carb versus low-fat. And what you were really trying to figure out — if I understand this — was does pre-existing insulin resistance affect whether low-carb or low-fat will work better for you? And does a very particular set of genetic SNPs affect whether one diet will work better or not for you? And what you found is: nope, on both. And you saw this huge range — you saw very little difference between low-carb and low-fat, but within the groups you saw a huge range. Someone gained 15 pounds, someone lost 50 pounds. Why?

Christopher Gardner: Yeah. So let me go back just a minute because we did another big study called A to Z before that, where there was very little difference. We picked four diets that had a huge range of fat and carb, and the differences were much less at the end of the year than I expected on average. But one of my colleagues a long time ago taught me to present results not just as an average and a variance bar — which is how scientists really do this — but to make what's called a waterfall plot and show every single person in the study for the amount of weight they lost. And the patterns of weight loss were stunningly identical across the four groups. In all four diets, somebody lost a crap ton of weight, but somebody also gained weight — with the same advice. Wow, there must be some predisposing factors.

And then an important thing we did in DIETFITS was the equipoise thing. We said: hey low-fatters, I know that added sugar and refined grain is low-fat, but don't eat that. Hey low-carbers, I know butter and lard is low-carb, but don't eat butter and lard. Eat a healthy low-carb and a healthy low-fat. And we had a bunch of literature behind us to show this predisposition — people had done low-carb and low-fat and mixed up the insulin resistance. Those were short studies with small numbers of people, and this was 600 people for a year, and we didn't replicate any of it.

And some people said, "How did you get it wrong?" I said, "Why did you think we got it wrong? The other studies before us were smaller and shorter and didn't do this equipoise thing that we took really seriously. Maybe the other ones were wrong because we've done this bigger study now." And now that we have this in our hands and we are seeing, oh my gosh, just as he said — we replicated this same amazingly huge variability within each group. Not an average, but within. God, we have a responsibility now to keep working with the data.

So Mike, you saw one paper where a post-doc at the time — her name is Michelle Hower, she's now the president of the American College of Lifestyle Medicine, she gives a teaching kitchen to medical students where she's teaching medical students how to cook — she looked at the quality aspects of the diet and showed that quality explained part of the difference. Another post-doc looked at gender and actually there are some male-female differences that were small. Another undergraduate actually found this by mistake — there's something called respiratory quotient. Your respiratory quotient can be found if you sit on a bed completely still early in the morning before metabolism's really kicked in, and it measures how much oxygen and CO2 you give off. It can tell you how much fat versus carb you burn just sitting still. And people burn slightly different ratios. That explained part of it. We had another psych student look at some of the — what are these things called — self-confidence and self-efficacy and determination, all kinds of things. Found nothing.

We've published 30 or 40 papers on DIETFITS and we've found a small microbiome thing and a small gender thing and a small quality thing — and if you added them all up together, it would explain a part of that variance, but nothing explained a lot of it. Which was very frustrating, to be honest. What probably made the biggest difference — and we haven't really looked at it this way — is: did you have a happy marital relationship? Did you lose your job? Did you get in an accident? Did you get a promotion? There are things in life that undermine our attempts to follow a diet. We had a couple of people who had disastrous things happen to them over the course of the year and said, "This is so disastrous I have to drop out of this study because I'm overwhelmed with the rest of life." And we had a couple of other people who said, "I am not dropping out of this study. This study is actually keeping me sane. Having this rigid thing that I'm supposed to follow for the sake of science is sort of giving me a little sense of purpose at a time when my life is falling apart." To me, that was one of the most amazing examples of — wow, the same disaster completely derailed this person and refocused another. Oh, we humans are so confusing.

We had people who said, 'This study is actually keeping me sane. Having this rigid thing that I'm supposed to follow for the sake of science is sort of giving me a little sense of purpose at a time when my life is falling apart.'

— Dr. Christopher Gardner

Mike Haney: Well, getting back to our point about the basics versus the individuality — that's what I found a little confounding. Okay, you've done the equipoise thing here. You've created high-quality diets. So quality should be a constant — ideally — across these two different formulations. How does quality then become something that explains the difference? I guess what I'm asking is: you're telling people this is what a quality low-fat or a quality low-carb looks like. Is there still just a decent range of how people are going to actually live that over the course of a year?

Christopher Gardner: Oh yes. So let me tell you another side paper that probably nobody saw. Luchia Aronica had this idea — she was really interested in the most versus the least adherent people. So the study was year-long and we collected data at three months, six months, and twelve. And at three months she found the ten percent of the people who were the lowest in carb on the low-carb diet and the lowest in fat on the low-fat diet. And they actually lost the most weight of anybody, and no difference between them — no difference between the most successful low-fatters and the most successful low-carbers. And we tracked them out to twelve months and they were still among the lowest, but they weren't able to maintain the same levels they achieved at three months when they were a little more excited about it. But they were still pretty good at keeping up with it.

And to me the most shocking finding there was: we had told the low-fat people that a quality diet doesn't have added sugar or refined grain in it. And at twelve months, the people who remained the lowest in fat and had lost a bunch of weight and kept it off had double the amount of added sugar and refined grain in their diet than they did at the beginning. We had told them not to have that. But the way they achieved that low a fat diet was to have more added sugar and refined grain and just say, "Well, I'm focused on the low-fat thing, but I'm going to have to have another baguette." And we didn't see quite that in the low-carb — it wasn't quite as obvious, but if we'd looked a little harder, I bet we would. So the quality range within the participants was just as wide as the range in how low-carb and low-fat they got.

The limbo-titrate-quality framework: finding a diet you can actually live with

Mike Haney: So what I'm hearing in both how life affects their ability to stick with this, and what you just discussed, is: the best diet you can sustain that focuses on quality food is the best diet for you. And maybe you could talk about the study you did where you had people drop to these extreme levels of carbs and fat and then titrate up to a level they felt like they could sustain — and kind of self-dial in something that felt like it was having an effect but also that they could stick to when not just reverting all the way back to whatever their standard American diet was before. Is that somewhere in the right ballpark in terms of a lesson that's come out of this?

Christopher Gardner: That should be the conclusion or the title of everything we do today. There are a lot of dietitians who will say: the best diet for you is the one you can stick with. Now, that's not entirely right, because Americans — a lot of Americans — stick with a super crappy diet. So that's not something to aspire to.

Sure, but you captured what we told these people. So, first we did a pilot study to try to define this — we were getting ready to start a big study for $8 million and we thought, wow, we're actually going to try something that nobody's ever done before. There was another weight loss study called Pounds Lost where they said, "Okay, here is the percent protein, the percent carb, and the percent fat that we want you on." And when you looked at it, you could say, "Wow, hardly anybody actually hit those numbers." That's even hard for dietitians to do.

So we came up with a term for this: limbo-titrate-quality. Limbo meant you're on low-carb or low-fat — limbo down as low as you can get as fast as you can. We don't even expect you to be able to stay there. But psychologically, you might be able to anchor yourself into thinking, "I am in a study, I'm going to try to get rid of as many carbs or fats as I can. Okay, I've done that." And we said: to be honest, at really extreme rates, very few people can stick with that. So once the limbo phase is over and you've found the lowest point you could get, just as you said — titrate back up. Add a little more and add a little more and do this to the point where maybe weight is starting to come back on and you kind of liked where you were a step below. So go back, titrate back down, maybe titrate back up, and find this space. And what we said to every single one of the 600 participants was: look us in the eye at some point — we don't even care if it's at two months, three months, four months, five months. Our goal is for you to look us in the eye and say, "You know, culturally, family-wise, economically, if this diet works for me, I could stick on this diet for the rest of my life." And then the last one was quality — like, don't do this with gamification. Don't find the low-fat brownies and the low-carb cupcakes. Eat food. Eat quality food.

In the study, Mike, we actually had four and a half different dietitians working with the 600 people and everybody taught both a low-carb and a low-fat class. And we actually had to say to them professionally: we want you to sleep at night. So we don't want you telling one group something that you don't believe in at all and telling the other group something that you think is the right answer. We want you to be super open-minded. If you teach the best low-carb possible and the best low-fat possible, America is waiting for the answer to this question, and we want you to sleep well at night professionally. And it was actually sort of career-changing for several of them — "Wow, I'm more open-minded than I've been before. I'm seeing successful people on both diets. And I'm seeing people really try hard and hit all the markers we want them to, and yet failing from a weight perspective. God, this is mind-blowing as a health professional."

Sorry, probably too long of an explanation to say exactly what you said was right. Like get to the point where it is sort of the healthiest set of options you can do for the persuasion that you're more drawn to and successful at — low-carb or low-fat — and try to stick with that and then tell us what happens. And so they did lose collectively 6,000 pounds at the end of a year. 600 people had on average lost ten to twelve pounds. Some sixty, some had gained fifteen. But we learned a lot and there wasn't any one thing that did it.

Insulin resistance, carbs, and why quality keeps winning

Mike Haney: There's a couple of other things I really want your take on. One is on this point of insulin resistance — you did this study in which you thought maybe somebody's insulin sensitivity state is going to affect whether they do well on a low-carb diet. You found no — that didn't seem to have an effect. How do you think about why that's the case? Mechanistically, the carb-insulin model — Gary Taubes and others push it and it makes sense, right? Insulin helps you store fat, eat a lot of glucose, you're going to spike your insulin, you've got high insulin, your body's going to want to store fat. Why didn't that work in real life, or what do you think are the limitations of that model as a way to explain obesity?

Christopher Gardner: I think when we asked everybody to try to eat a high-quality diet, that worked. In fact, Gary came up to me as the study was finishing and he said, "Christopher, I just realized you screwed up the study." I said, "It's almost done — why, Gary? You helped fund it." And he said, "You told the low-fatters not to eat added sugar and refined grain." I said, "Well, yes I did — just like I told the low-carbers not to have butter and lard." He said, "Well, the reason we see a difference all the time in insulin resistance is when people do low-fat and they do it with added sugar and refined grain."

And so I think that's what was going on there. I actually do think low-carbers have an edge — it didn't really come out in our study, so it can't be a big edge. But when you tell somebody to go low-carb, they are less likely to cheat on the added sugars and refined grain than the low-fat people. Because as you go on low-fat and your cheat is added sugar and refined grain, you're still low-fat. If you go on low-carb and you have added sugar and refined grain, it's a double cheat — it wasn't quality and it was the opposite of what you're supposed to take. So I think it's the added sugar and refined grain where everybody — kumbaya moment — needs to get on board. And then it could be a low-carb or a low-fat diet that's really super low in added sugar and refined grain. That would be a win for everybody in my mind.

Mike Haney: Why do you think those two things are so damaging? Does the carb-insulin model mechanistically explain it?

Christopher Gardner: So my quick answer is it's the proportion of the diet. There's a slide I show all the time that shows not just protein, carbs, and fats, but the types of each one. It shows saturated, mono, and poly, each contributing about 10% of calories to the diet. It shows animal and plant protein, each contributing about 10%. And for carbs, about 10% of the American diet comes from good quality carbs — from fiber-rich vegetables and chickpeas and things like that — and 40% from carbs from refined grains and added sugars. So I think everything else is ten. It's like ten, ten, ten, ten, ten, ten — and 40% comes from crappy carbs. And I think that is the biggest thing that has changed over the last century — the quantity of refined grain and added sugar that we have in the diet. That could be the easiest place to make a difference. But it's also, from an agricultural point of view — we grow a lot of corn and soy and wheat, and we make a lot of grain-based, wheat-based in particular, products that taste good in our mouth, they're inexpensive, they store really well, great shelf stability, and you throw that added sugar on there for that palatability, and it's almost hard to refuse.

What 40 years of nutrition research actually tells us

Mike Haney: So maybe last question here. What do you think are three things you could tell people you think are going to hold true for the next ten years — that the next twenty, fifty, one hundred studies aren't going to refute? Three basic truths about how to eat.

Christopher Gardner: So I'd almost like to give you ten. I worked with a bunch of chefs who worked on the 24 principles of change for the Menus of Change collaborative. And one of the things is: people want joy and pleasure from their food. And health professionals sort of give the anti-message — avoid this, don't eat that, don't have so much. I love working with chefs who say, "You've got to keep joy and pleasure there." So that would be one of my fundamental truths forever and ever.

The next one is: look at the food and see if it looks insanely modified versus "I can understand most of those ingredients — that looks like a whole food." I'm going to stop at two, because I think if it brought joy and pleasure and it was really pretty obviously a whole food that was accessible and worked for your culture, you are nine-tenths of the way there.

Mike Haney: I'm going to give you a chance to give the last tenth, which I know what you're going to say — which is beans.

Christopher Gardner: Yeah. Yeah. Yeah. So the thing that's under-consumed, which is just mind-boggling — I mean, we have this insane protein craze right now. All you've got to do is Google "Gardner rant protein" and like 40 things will come up. Beans are the best source of protein and a fantastic source of fiber and a good source of unsaturated fat. And the dietary guidelines suggest — I might be missing this a little bit — a cup and a half of beans a week. And Americans don't come close to a cup and a half a week. So divide a cup and a half by seven days and you can see it's not even half a cup a day. And there are so many kinds of beans.

Americans, please start to use the word legumes. Let me tell you what a legume is. The legume is the botanical hierarchical head of the family of soybeans and peanuts and fresh beans and peas and pulses — which include dry beans and peas and chickpeas and lentils. That's like the family. So the easiest word, if you only wanted to use one, is legumes. All of those are high protein, unsaturated fat, great sources of fiber. All of them are leguminous plants that fix nitrogen in the soil and replace synthetic nitrogen from the existential crisis of environmental concerns. And there are — oh, it's insane how many different kinds of beans there are. The Asian culture eats soy, the Mediterranean culture eats chickpeas and hummus, the African culture eats cowpeas, the Latin American culture eats beans and rice all the time, and the Indian culture eats lentils and dal. So much good. Yeah, I want to be part of the leguminati — this group that's pushing legumes. May we put that as an addendum on Pollan's phrase — we'll add three words: and more beans.

Beans are the best source of protein and a fantastic source of fiber and a good source of unsaturated fat. And Americans don't come close to a cup and a half a week.

— Dr. Christopher Gardner

Mike Haney: And more beans. Yeah. Or legumes. Okay. But if you don't say legumes, then beans.

Christopher Gardner: Sure.

Mike Haney: Well, I could definitely talk to you a lot longer. And maybe after I've chatted with Stuart Phillips in a few weeks, we can have you back on and we can have the protein debate that wasn't really a debate. And I do encourage people to go find your rants about protein, which I think are useful. This has been super illuminating. Thank you so much for your time today.

Christopher Gardner: That was really fun. I really loved all the stuff that you set up and all the homework you did ahead of time. A pleasure talking to you — and let's do it again.

This article is based on insights from Dr. Christopher Gardner, PhD, a professor of medicine at Stanford University and director of nutrition studies at the Stanford Prevention Research Center. He has spent over 40 years designing and running dietary intervention trials.