The ice cream effect
29:21 Share

Ryan Reynolds here for, I guess, my 100th Mint commercial. No, no, no, no, no, no, no, no, no. I mean, honestly, when I started this, I thought I'd only have to do like four of these. I mean, it's unlimited premium wireless for $15 a month. How are there still people paying two or three times that much? I'm sorry, I shouldn't be victim blaming here. Give it a try at mintmobile.com slash save whenever you're ready. For

$45 upfront payment equivalent to $15 per month. New customers on first three-month plan only. Taxes and fees extra. Speeds lower above 40 gigabytes. See details.

The land down under has never been easier to reach. United Airlines has more flights between the U.S. and Australia than any other U.S. airline, so you can fly nonstop to destinations like Sydney, Melbourne, and Brisbane. Explore dazzling cities, savor the very best of Aussie cuisine, and get up close and personal with the wildlife. Who doesn't want to hold a koala? Go to united.com slash Australia to book your adventure.

Hey, Brian. Hello. I was reading this article recently. It was in The Atlantic. And it was one of those articles where every next paragraph was like, "Oh, I have a question. Oh, I have another question. This question is getting bigger. Now I'm really confused. Now it's making me question everything I thought was true." Was this an article on, like, the nature of consciousness or something super heady like that? It was actually a lot more simple. It was about one of my favorite things in the whole world, something I thought I knew a lot about.

which is ice cream. Oh, made you question everything? Well, sort of. I talked to the guy who wrote this article. His name is David Johns. He's a public health historian. And he told me that last summer he was doing what he normally does. He was talking to people, reading through scientific papers on nutrition. And he came across some really strange data. Ice cream intake, eating ice cream, consuming ice cream was somehow protective against diabetes.

Apparently, for diabetics, eating half a cup of ice cream every day leads to a lower risk of heart problems. Ice cream good? Apparently, maybe. Okay. I was like, there's no way this is real. It's too absurd to be real, right? Like, that goes against so much of what I would take to be just bedrock.

bedrock fact of what I'm supposed to eat. Yeah, ice cream is loaded with fat and sugar, and it's just super calorie dense. Very dense, yeah. And I wouldn't necessarily expect this based on at least the nutritional advice I grew up with. Exactly. So he dug into the data, and he found a reference to an earlier study from 2001, which

which was looking at what thousands of people had eaten over a couple decades. So I pulled up this paper and I was going through the data tables, and almost every dairy food they looked at was protective. Low-fat, high-fat, milk, cheese, you know, all of these dairy products seem to reduce the risk of diabetes.

But then he kept reading. And I looked at the bottom and they had actually looked at what in this study they had called, quote, dairy-based desserts. Dairy-based desserts, which the researcher behind the study told David mainly consisted of ice cream.

they seem particularly good at reducing the risk of diabetes, like even better than other kinds of dairy. - The largest effect by far, like way bigger than any other effects, was this ice cream effect. It was there. And when I saw the number, I was like, "Ah!" Like I just like literally screamed at my computer screen. I was like, "This is like a little too on the nose. Like I'm screaming for ice cream somehow." You know, just because it made me just like squawk.

You know what? I would say that finding is squawk worthy. I might squawk in the same position. It seems reasonable, right? Because it's not just dairy is good for you, but ice cream, which is delicious and fun and the stuff of childhood dreams, seems to be particularly better than just milk. Yeah. So almost two and a half times better than milk at reducing diabetes risk.

which seems really strange. - I was like, maybe this whole thing is just, I'm just like insane, right? - So he dove into this rabbit hole, dug into decades of data. - But then I like, I kept finding it and I was like, now I'm like preparing my yarn board on my wall connecting. Maybe I'm the conspiracy theorist. - And he really wanted to know whether this was just a fluke of the data or whether this was something real. Because if it's real, that raises a pretty big question.

What happens when researchers confront a finding that, like, goes against everything they've ever said before? So that's what I want to get into this week. Is ice cream good for you, at least in this one specific way? And more broadly, what is the fact that this question is so hard to answer? Tell us about the field of nutritional science as a whole. Like, do we know anything about ice cream or anything you eat? Do we know what we should eat?

I ask that question every single day. So once David heard about these strange ice cream results,

He had tons of questions. Yeah, where have they been all my life? Exactly. Why haven't they been shown to me every morning before I wake up? But the first obvious question here is just how? Good question. If we assume that this data is accurate, that ice cream is good for you,

How would that happen? Like what in the chocolate chip cookie dough or the rocky road would actually lower your risk of diabetes? Yeah. So the first basic answer here is just that ice cream is loaded with dairy.

This was, you know, the big takeaway from that 2001 study that dairy is good for you. But ice cream is not just dairy. It's dairy with a lot of sugar, fat and other stuff, eggs. Well, actually, believe it or not, one of the theories for why ice cream seems to lower diabetes risk, you know, more than other kinds of dairy is exactly all that, you know,

other stuff in the ice cream. At least that's what David heard from this professor at the Tufts School of Nutrition. He was like, you know, ice cream's got fat, it's got minerals, it's got protein. Yes, it's got a lot of sugar and saturated fat, but it's a relatively whole food. And the thing that really, like, made my jaw drop to the floor, he was like, it's better for you than bread. Better than bread?

Better than bread. Okay. And I was like, ice cream is better for you than bread? Like, that's, I mean, ice cream is like the canonical, you know, unhealthy food in a way. David said that, you know, he might have been referring specifically to white bread. Yeah, yeah. Super processed. Yeah, yeah, yeah. Not the sourdough I make at home.

Yeah, and ice cream might have things bread doesn't. There's a way that we look at foods in a kind of reductionist manner, and you think of ice cream as being like this sweet thing because it's got all the sugar in it, but it's a much more complex food than that. And maybe these potential...

benefits that it brings. You know, the protein, the minerals, the good fat, like maybe that could potentially outweigh the sugar, at least when it comes to something like diabetes risk. Okay. And then another explanation is that there's something called the milk fat globule membrane. Don't you call my mother that. Yeah.

I apologize. Just kidding. Anyway, it's basically this membrane that surrounds the fat in dairy products. There are certain ways of processing dairy, like in butter, that break this membrane and

And there are some studies that show that if this membrane is intact, it's better for you. Okay. But on the other hand, David told me that regular cream has an intact membrane and it doesn't lower the risk for diabetes. So it's possible something else is going on here besides the membrane thing. All right.

All right. So what is the something else that could explain this ice cream effect? Yeah. So David thought there might actually just be a problem with the study itself. My first reaction was, oh, this is obviously an artifact of the data. So in order to find out what people were eating, the study used questionnaires that people would fill out on their own.

But an obvious problem here is just making sure that people are honest about what they eat. It's not like you have like a researcher on your back who's like, you know, measuring and following every food that you eat and testing you precisely every single day so that they know the exact moment when you tip into pre-diabetes or get diabetes or whatever. And one of the researchers David spoke to said that some people might feel bad or self-conscious about eating ice cream, especially if they're at risk for diabetes, and

And they might just under-report how many desserts they eat. A lot of these questionnaire-type methods may have biases in them where people are reporting things that they want to believe about themselves or they would prefer to appear more virtuous. So the idea here is that maybe people at risk of diabetes are eating ice cream, but they're not being honest about it. So when scientists, you know, crunch the numbers, it looks like eating ice cream isn't actually correlated with diabetes.

That seems plausible. Yeah. And the other possible option here is something called reverse causation. Have you heard of this before? Okay. This is one of those topics that has come up in my science journalism career that I've read about, but it has never stuck in my brain. Yeah.

Perfect. So he actually gave me this really helpful example. There had been studies that had found a correlation between mouthwash and oral cancer. So there was concern. Is mouthwash somehow causing oral cancer? But it's possible the mouthwash wasn't causing the cancer. Oral cancer tends to give people bad breath. So the oral cancer may have been leading to the mouthwash use, not the other way around. Mm-hmm.

How is ice cream like mouthwash in this situation? So the idea is that maybe people who were eating less ice cream were less healthy to begin with. So maybe they had high blood pressure, elevated cholesterol, and either their doctors had told them not to eat ice cream or maybe they consciously tried to avoid it because of health problems.

So basically, the idea here is that the people who ate less ice cream might have been less healthy to begin with. So that's how this could have been reverse causation. So the ice cream just doesn't mean anything here. It's just a marker for, you know, how people think about their health. Potentially. Yeah. Like, if you're healthy, you might...

feel comfortable allowing yourself to have some ice cream once in a while. That makes sense. Like when I go do exercise, you know, that temptation of, oh, I exercise today. Yeah, treat yourself. I can treat myself a little bit. Yeah. And if you're someone that's pretty healthy and regularly has some ice cream for dessert, yeah.

That might lead to this ice cream effect. But there was a scientist that was able to control for some reverse causation by only focusing on unhealthy people. And the ice cream effect did shrink. But it was still there. And it was still statistically significant. And it was still like stronger than other findings that this research group had celebrated and publicized themselves. All right. So there still might be a there there. Yeah. So I just asked David like point blank it.

Is ice cream good for you? I mean, does it seem like ice cream is good for you? I mean, if I had to guess, I like so. So I think I told you where I started from in this, where I started from in this story, which was like, this is an artifact of the data somehow. And I was you know, I felt like that was obviously the most likely explanation. However, in talking to some of these researchers, they actually convinced me by the end that the better answer on my part was to say we don't know.

Essentially, these, you know, mechanical, chemical explanations for how the ice cream effect could work aren't perfect. But the argument that it's all due to flawed study design, that doesn't seem to solve the problem either.

So after all of this, somehow the ice cream effect still lives, which feels really strange. This is sort of a basic question, like, does ice cream lower the risk of diabetes? And scientists don't seem to be able to answer clearly yes or no, which kind of, I don't know, it introduces sort of a bigger question here, which is like, what does this say about nutritional science in general beyond ice cream?

Well, the ideas that underlie our assumptions about ice cream being bad, are those even correct? You know, or just is anything we know about nutrition right? Exactly. Like, how uncertain should we be about nutrition in general? And then also, how can we get better data? How can we do better research to get stronger answers so we can really know what we're supposed to eat? I'm here for that. Will you be here for that in a minute?

Can you stick around? I think I'm like employment obligated to stay. All right. But I'm also interested in hearing what you have to say. I'll see you in a minute. Okay.

How do you feel great on vacation? Like really good? Easy. You go to Aruba. You'll spend your time relaxing on cool white sand beaches and floating in healing blue water. You'll immerse yourself in natural wonder and find your center on an island where things move at your speed. You won't just feel great. You'll feel relaxed, renewed, and ready for life. That's the Aruba effect. Plan your trip at aruba.com.

Feel your max with Brooks running and the all new ghost max too. They're the shoes you deserve designed to streamline your stride and help protect your body. Treat yourself to feel good landings on an ultra high stack of super comfy nitrogen infused cushion that takes the edge off every step every day. The Brooks ghost max too. You know, technically they're a form of self-care Brooks. Let's run there. Head to brooksrunning.com to learn more.

I drink your milkshake. I drink it up. Brian, we're back. True.

So we've been talking about this ice cream effect, which is strange because we've got lots of reasons to believe that ice cream isn't that good for you. It's got sugar, saturated fat. It's packed with calories. And then on the other hand, we've got decades of research that seem to show that ice cream might lower diabetes risk, even though, you know, as we've basically been talking about this whole episode, scientists are still saying

Okay. So we're left with this question, which is that if data is popping up all over the place that seems to contradict some of the basic truths we take for granted about nutrition, do we question the studies? Do we question the things we take for granted? Or do we just try to rethink the whole thing? I would want to look at those kind of fundamental assumptions because...

I've always wondered if scientists really have the tools to forcefully answer a question like, is ice cream good for us? Is chocolate good for us? Is anything good for us? Or are there just fundamental limitations here? Yeah, I think that's a really good point because David, you know, this public health historian we've been talking to, one of the first things he told me was that this field, you know, nutrition science, it's so much harder to do research here than you'd think.

Like, just think about the way we normally do scientific studies. It's easier to do it with, you know, pharmaceuticals where it's like you can give one person the pharmaceutical and one person like a sugar pill. But with diet, if you're going to do like an ice cream trial, it's like, what's the placebo for ice cream, right?

Why can't you just give some people ice cream and some people no ice cream? Well, you can't just give the other person nothing, right? Like, then you're just testing the difference between food and no food. So you'd have to give them something...

that's kind of like ice cream. But what do you do? Do you give them the same amount of calories? Do you give them the same amount of sugar? Carbs? Do you make it something cold? Like, what is the thing that makes ice cream ice cream? And is it even just a good idea to do nutrition like this, like food by food? Like, ask, oh, is ice cream a problem? Or is chocolate a problem? Or is the sum of what we eat more important than the individual components of it? Like, do we even know that? Well...

Or it seemed to be hard to tease out part. It does seem hard to tease out because what you eat can actually impact what you eat next. So if you go to your refrigerator and you have a glass of milk or you eat some cheese or you eat some ice cream, what is it that you might have eaten otherwise? Right. Maybe you would open a bag of Fritos and you would eat the entire bag of Fritos instead of, you know, some cheese or milk or ice cream.

And if it fills you up, eating a small bit of ice cream might be better than eating that whole bag of Fritos. So it's really hard to study something like ice cream in a vacuum. So it may be that it's not so much that the dairy foods are good for you. It's that by eating the dairy foods, you're preventing yourself from eating things that are even worse for you. And all of these variables, the things that make it really hard to study nutrition, are

These are all before we even get to some of the more fundamental problems with questionnaires specifically, which is what researchers used in that 2001 study. You don't know if you can trust people to report accurately. You don't know about reverse causation. There's the issue with placebos we talked about.

And so they're super limited. And David said that when he talked to the researcher from that study... I mean, he said to me, I think we've probably reached about the limit of what we can learn from these methods. And I think we need to shift towards more exacting trials, you know, controlled experiments that may involve fewer people, but where we really think we have a handle on, you know, what people are eating, because it's very hard to measure. And that's what I'm curious about, like...

these questionnaires, are they just kind of useless? Or is there an argument in defense of them? Yeah, I mean, we definitely shouldn't throw out questionnaires. And David did tell me that we've gotten a lot of useful information there. Like, they helped us learn that trans fats were bad.

But then we went and, you know, backed up that knowledge that we kind of originally got from questionnaires with more experimental studies. So is the ideal doing some combination of questionnaires and then experiments? Yeah, but a lot of nutritional experiments can be really hard to do. Like there's this move to focus on something called metabolic chambers, which are basically these small rooms that...

that are so super controlled that scientists can get all of this really intense data. - Human beings are essentially confined like lab rats for a short period of time, and they are fed specific diets. And they actually measure like your CO2, you know, the movement of gases in and out of your body.

And scientists have learned all sorts of things from these experiments. Like, just one example is that exposing people to cold temperatures when they sleep actually allows them to burn more calories. Oh, interesting. So it's not just even about food. It's about your environment, too. Exactly. And that reverses when study participants sleep in warmer temperatures. So it's not like a long-term thing.

And I don't know, that kind of thing would be pretty hard to figure out from a questionnaire, right? Like figuring out exactly how many calories you burn relative to your exact sleeping temperature. Who knows that? Yeah, but just like with questionnaires, metabolic chambers have tons of limitations. If you put 24 people in a metabolic chamber for a month or two weeks, you're probably not going to get any diagnoses of heart disease or deaths, right?

With metabolic chambers, it's hard to get long-term data. It's hard to get these outcomes of which disease a person got and when. You just get, okay, this particular food is leading to this short-term change in your body. And if you're going to do a huge study with these, it would be really hard. Not just because it's hard to get long-term data and not just because it's a lot to ask of all these people to live in metabolic chambers. Mm-hmm.

But in general, it's hard to do nutritional research because of the way we've historically approached nutrition science in America. Nutrition science up until like the 40s, say, it was like in schools of like home economics or like agricultural schools. And frankly, it was regarded as, quote unquote, women's work because this was the kind of sexism that prevailed in, you know, the first part of the 20th century. And so like nutrition was kind of like barely thought of as a science.

And then even today, a lot of the organizations that are responsible for nutrition research, they also support producers. USDA has a dual mission of like both producing science related to nutrition, but also promoting the well-being of farmers and food producers, right? Yeah. There's sort of like mixed incentives to do nutrition research,

But, you know. Also prop up the dairy industry. Exactly. And then even one more step. There's so much incentive in the media to have sensationalized nutrition claims like this is the stuff that sells. People love to hear about the new diet, you know, the superfood.

There is this kind of like pipeline from nutrition science researchers to the university press offices to the journalists who write about this stuff because it's frankly of interest. It's always newsworthy. This was a staple on like the Dr. Oz show for years. Yeah. And it's this kind of self-fulfilling cycle that can prevent some real nuanced scientific work from getting done.

And this exact thing actually played out with David's article. Like so much of the reaction was just finally an excuse to eat all the ice cream I want. People like threw up their hands and they were like, I'm headed straight to the freezer. Like, I don't need to sift through the data. I don't need to like scrutinize thing. This is enough of an answer for me. It's like green light towards that tub of ice cream sitting in the freezer. But that's not the point he was trying to make. If it's this hard to figure out whether ice cream reduces the risk of diabetes, just a

Imagine how hard it is to figure out whether it's quote unquote good for you or not in larger ways. The whole sort of good, bad foods thing that we've kind of come to think about is like it's all on kind of a gradient because you have to eat something, right?

That just brings us back to that big question is like nutrition science really robust enough to answer these questions. I mean, yeah, at the end of my conversation with David, I was wondering the same thing. Like, do we just need to lean into the uncertainty here? So I wanted to see what he thought. Our show is really about kind of getting comfortable with uncertainty. And I'm wondering if that's sort of

what it's going to be? Are we always really going to be somewhat uncertain about nutrition and we just kind of have to get comfortable with that? I don't think we're going to always be uncertain about nutrition. I mean, you know, think about these like metabolic studies of the type I was describing. If you went out to, you know, the broader population and said, we want to put 500 people in metabolic chambers or we're going to do this most insanely ambitious study. We're going to pay you, but we're going to keep you in these metabolic chambers for six months. You might get some pretty interesting answers.

David's basically saying that it's theoretically possible to get the best of both worlds here. You know, you could use questionnaires and experimental studies. You get long-term data, short-term data. Yeah, we're not stuck with uncertainty. We can certainly do more work here. Yeah, and David didn't know this when he said that to me, but the NIH recently announced this huge, almost $200 million study that's going to track 10,000 people to figure out how different diets work for different kinds of people. Hmm.

People are going to fill out questionnaires, but they're also going to visit clinics. They're going to wear glucose monitors. They're going to wear activity trackers.

Some of them are even going to get these kind of camera sensor thingies for their glasses that'll detect when you're chewing and then automatically take pictures of the food when you eat it. So the memory problem is not an issue here. Exactly. No false reporting. And then, like, the thing that's really exciting is that 500 people are going to live at a clinic for weeks at a time and be intensely monitored in this kind of metabolic chambery sort of way. Mm-hmm.

And we might actually get a better sense of what food is good for us to eat. What does David think about this study? Well, he's kind of skeptical about whether this study is actually going to be able to answer all the questions it's laying out, like which specific diets are good for which individual people. But he does think it could come up with some interesting results. And, you know, when I look at this, I don't know, part of me is just hoping that

Like maybe we could learn somewhere here about the ice cream effect. Or maybe there's like a bit of data here that someone could look at and figure this out. Well, maybe if this is still being developed, this study, they can add an ice cream component. Yeah, what would help us? So David actually did talk to a researcher who said that he does want to do an actual targeted study on the ice cream effect.

But, you know, regardless, to me, this new study shows just how difficult it is to answer basic questions about nutrition. Like, it's just so hard to get even a simple understanding on the smallest bits of nutrition science. And if we want to push forward, you know, we have to invest tons of time and resources. And even then, we still might not get clear, satisfying answers. Well, I guess...

Call me back when this study is done and if they include ice cream data. Yeah. When is this going to finish? So, I mean, it looks like it's going to be completed around 2027. So a ways away. And who knows what we're going to get out of it. Probably, to be honest, not a definitive answer to the ice cream effect.

And yeah, I guess for now we're just going to have to wait. It just can take a mountain of data to say... A mountain of ice cream. A mountain of ice... It's just, you know, in the meantime, I can live with some ambiguity on this question. Yeah, and I guess we'll just have some ice cream while we wait. I have some in my freezer right now. Oh, what do you have? Mascarpone and this berry jam.

It's just like cheesecake flavored ice cream, but like less cream cheesy. It's just like a more melty, like a strawberry cheesecake. This episode was produced by me, Noam Hassenfeld. We had editing from Brian Resnick, with help from Meredith Hodnot, who also manages our team. Mixing and sound design from Christian Ayala. Fact checking from Serena Solon. Music from me. Manding Nguyen has been listening to Blood, Sweat & Tears' Spinning Wheel. And Bird Pinkerton was shocked.

What do I have to do with all of this? Why am I even involved with this bird-octopus war? The octopus looked at her and responded, Birds can't stand it when non-birds have beaks. But bird, you are the only bird without a beak. Yeah, I mean, I tend towards more the fruit-based or contained ice cream.

Thanks to the researchers that David spoke to in his reporting, which we referenced in this piece, Frank Hu, Dariush Mozaffarian, and Mark Pereira. For transcripts of our show, we've got a link in the show notes. And if you have thoughts about this episode or ideas for the show, please email us. We're at unexplainable at vox.com. And we'd also love it if you left us a review or a rating.

This podcast and all of Vox is free, in part because of gifts from our readers and listeners. You can go to vox.com slash give to give today. And if you do give and you're giving because of our show, maybe let Vox know that you enjoy Unexplainable when you give. It would mean a lot. Unexplainable is part of the Vox Media Podcast Network. We're off next week for Thanksgiving, but we'll be back in your feed on November 29th.

I really don't like super chocolate caramel, you know, just punch you in the face with. I kind of like that combination of creamy and sharp. So I definitely I'm definitely looking for some like strawberry. My pies are always sour. I always put some sour elements in the pie.

Life is full of complicated questions. I want to know how to tell if my dentist is scamming me. What age is it appropriate or legal to leave your kid at home? From the silly to the serious and even the controversial. Can I say something that will probably just get me canceled? I'm John Cullen Hill.

And I'm hosting a new podcast at Vox that'll be your go-to hotline for answers to the questions you don't know how to answer. Email a voice memo to askvox at vox.com or call 1-800-618-3545. I promise you it's better than asking ChatGPT.