New Years Weight Loss Edition 2: On Unfounded "Diets Don't Work" Claims, Weird Phrasing, and Science Bubbles

Meta-note: Judging by the feedback I got, the last article in this series was probably a little out of my depth in terms of what I’m able to explain in a clear and at least somewhat entertaining way. I’m trying something different with the format of this article to make a bit more readable, but because of that I might at a later time re-format the article again if things don’t work out.

If you’ve been following along, my last post attempted to point out problems with how science generally regards the role of a particular hormone (leptin) in regulating weight. I think looking at the actual evidence shows that the generally-accepted role of leptin is overstated based on the evidence provided. However, even if you accept that I’m more or less correct on leptin we are admittedly only looking at a very small slice of what science and pop-science reporting generally accept as true about weight loss.

If we look at how science thinks about dieting more generally, we find there’s a not-insignificant amount of people who hold that dieting itself is useless or even counterproductive. Take this quote from the abstract of this review:

These studies show that one third to two thirds of dieters regain more weight than they lost on their diets, and these studies likely underestimate the extent to which dieting is counterproductive because of several methodological problems, all of which bias the studies toward showing successful weight loss maintenance. In addition, the studies do not provide consistent evidence that dieting results in significant health improvements, regardless of weight change. In sum, there is little support for the notion that diets lead to lasting weight loss or health benefits.

Here’s a discouraging quote from the text of the full article:

It is clear that dieting does not lead to sustained weight loss in the majority of individuals, and additional studies of the effects of dieting on weight are not needed. A call for more rigorous diet studies seems unwarranted as it has been noted that among diet studies, “greater methodological rigor seems to be associated with poorer results” (Kramer et al., 1989, p. 126). We do not think further study of existing diets will lead to a different assessment, nor do we think a new diet formulation will appear that leads to more favorable outcomes.

I can’t overstate this: claims like these are very, very important. If it’s true that dieting doesn’t do anything to help with weight loss, then a huge amount of people are wasting an enormous amount of time on it. If it’s also true that dieting like this is unhealthy, then it’s not just the wasted time; we now have a situation where a lot of people are hurting themselves pursuing goals that their methods can’t help them obtain. On the flip side, obesity is a large enough detriment to health that if diets do work, then discouraging people from dieting is equivalent to killing some amount of them.

Despite this being important, I think there’s a tendency to present claims like the quote from the study above in a way that casually reads as something very broad and important (diets don’t work!) instead of the very narrow conclusions that the studies do support. In that spirit, I’m going to address what I imagine an average person would ask if they read some of the general anti-dieting information going around and then do my best to summarize what the data we have does and doesn’t show. Note that this won’t prove diets work - that’s well above my pay grade - but I hope by the time we are done to have neutralized some of the discouragement.

---

I’ve heard that studies have shown that diets never work.

The quick answer here is “kind of”, but this answer varies wildly based on what you mean by dieting, and what you think dieting studies are able to show.

When you or I talk casually about dieting, we usually mean something like “changing my diet in some way to lose weight”. We don’t necessarily mean something short term - some people think of diets as very temporary things, but generally most people allow that long-term results require long-term changes. We also think about “going on a diet” as something you actually do; we wouldn’t really think of someone who had thought about going on a diet but did not end up following through as a failure of the diet.

This is not always the case for the anti-dieting crowd. In the review linked to above, the three studies they weight the most heavily don’t necessarily deal with interventions. The first study provided counselling, free diet food, and monetary rewards in various combinations and tracked results. This study sent out some newsletters and gave some access to counseling and aerobics dance classes. The third study provided a lot of in-person counselling.

What none of the studies did is confirm that people actually followed the advice (beyond asking if people did, which isn’t particularly reliable) or, if they followed it, how well or for how long they did so. What each study measures is how well a particular attempt to get people to diet worked to get them to lose weight, which is a very different thing than finding out how well particular diets work if followed.

This isn’t the study authors’ fault, of course - keeping track of people’s attempts to diet is a notoriously difficult thing to do. People lie or are mistaken about how much they eat and exercise. People often state they are willing to make efforts they aren’t willing to make, or say they are making the efforts when they aren’t. Studies like this are then necessarily very muddy and only show very specific things, much more like “how well does various diet advice work” than “how well do sincere attempts to lose weight work?”.

This, of course, doesn’t stop the reviewers from making the claim that not only does no diet exist that works, but indeed that no diet could exist that had any potential of working:

It is clear that dieting does not lead to sustained weight loss in the majority of individuals, and additional studies of the effects of dieting on weight are not needed. A call for more rigorous diet studies seems unwarranted as it has been noted that among diet studies, “greater methodological rigor seems to be associated with poorer results” (Kramer et al., 1989, p. 126). We do not think further study of existing diets will lead to a different assessment, nor do we think a new diet formulation will appear that leads to more favorable outcomes.

The reviewers here are intentionally or unintentionally playing a pretty complex word game to make their point. “Diets” don’t work, but diets are in this context often just advice to go on a diet, rather than an actual attempt by an individual to lose weight. “Don’t work” is used abnormally here, where weight regained after the dieter stops the diet invalidates the weight lost when they adhered to it, much like arguing that a car accident death caused by prematurely removing a seatbelt proves that seatbelts don’t work.

The summary of all this is that science has pretty good evidence that diets don’t work in the long term, so long as “diet” means “a scientist told me to diet at least once”, and so long as we don’t know if you dieted or how long you dieted if you did. But otherwise, there’s nothing stopping you from making an attempt.

---

Let’s say diets might work; why have I heard science can’t find anybody who successfully lost weight and kept it off?

This particular claim was everywhere for a while and persists until today. Oddly enough, I think a lot of the blame for it belongs to an absurdly-popular-in-its-time David Wong/Cracked.com article, which made claims that basically nobody ever loses weight and keeps it off:

And when I say "no one," I mean those cases are so obscenely rare that they don't even appear on the chart. They can't even find enough such people to include in the studies. It's like trying to study people who have survived falling out of planes. Without surgery, obesity is effectively incurable. Every study shows it, and no one will admit it.

When he’s talking about this issue, he’s referencing literature analysis like this. And within the confines of the paper, he’s right; those reviews don’t show very good results. The people in them tend to gain back the weight they’ve lost, and then some. Judging only by those papers, the world seems a very hopeless place in terms of weight loss.

Luckily for us those papers don’t generalize to the purpose he wants very well because they deal with a very specific subset of people. For the most part, the dieters in these studies are obese/morbidly obese people recruited from patient populations - i.e. a self-selected group of difficult cases who are in situations where their doctors thought it a good idea to refer them to relatively extreme medically run diet programs.

The problem with doing this is you’ve now singled out a population who are obese, but have self-selected into a group who are particularly unlikely to succeed at diets in general. Think of it this way: If you are overweight, your first attempt at losing weight might be to eat less; if that didn’t work, you might add exercise. If those attempts failed, you might try reading up on various diets and methods of diets. If those failed, you might try structured for-pay weight loss methods (gym memberships, diet programs). It’s only after all these have failed (or failed to be attempted in the first place), and failed long enough that you have the kind of obesity-related medical problems that make you visible to these programs that you find yourself showing up in a review of this kind.

What these reviews don’t and can’t see is anybody for whom anything short of a medical intervention like these studies actually worked. A person who is very significantly overweight and starts spending a lot of time in the gym doesn’t show up in these numbers, and can’t - if he’s successful, he’s not eligible or thought of for the program. So you end up with numbers that are biased towards failure, because they ignore successes and select for those who have failed to lose weight to the point where it’s causing long-term medical problems.

The National Weight Control Registry is one program that attempts to capture these success stories, but it’s generally verboten for researchers to use their data for the same reasons that are ignored in the literature reviews above. The NWCR data is made up of people who voluntarily report their successful weight loss and thus is a non-generalizable self-selected group, so their data is mostly ignored.

While none of this proves that any individual diet is effective or really helps us understand how often people lose weight and keep it off, the reverse is also true - we have no idea how often committed, serious dieters succeed. This is good news, though - there’s no reason here, either, to accept that you can’t be successful.

---

Is my body going to slow down my metabolism and keep me overweight?

We covered this to a great degree in the last article, but I think it’s worth recapping in a more readable form. You are going to see quotes like the following in weight loss articles everywhere:

“As you lose weight, your metabolism fights back against you and makes it harder to continue with that downward trend,” said Sharon Zarabi, RD, CDN, CPT, bariatric program director, Lenox Hill Hospital, New York.

This, like everything in this post, is true when looked at from a certain angle. A larger person has more body to maintain and spends more energy moving around; as they get smaller, their energy needs tend to drop. So we know there’s some slowing of metabolism involved, and that has to be factored in by further reducing food intake or just setting the goal calories lower to begin with.

But what Sharon Zarabi is saying above isn’t that - she’s implying that some loss disproportionate to what you’d expect is happening, that your metabolism is dropping more than your reduced size implies. This is an accepted concept by nearly 100% of pop-science, but turns out the question is a lot more open than represented.

There’s two things happening here that muddy the waters. The first is that since different people’s metabolisms vary, researchers calculate a person’s projected “new” metabolism at their lower weight rather than by comparing them to the mean or average of those at a lower weight. The second aspect is for how long those weight-preserving metabolic adaptations last, if they exist at all; many studies that show them to exist during weight loss also show them disappearing as soon as weight stabilizes.

Again, the upshot here is that there’s nothing here that’s actually backed up with solid evidence in a way that should convince you not to make attempts to improve your health.

---

You’re being nice about it, but you are basically accusing all these researchers of lying. Why would they do that?

It’s true that I’m saying that an awful lot of researchers are overstating their results and findings here. If I’m right, why would they do this? I think the normal publish-or-perish incentives are in play here - nobody is excited to publish a study that implies that people could be losing weight, and doubly so something that implies they could do so with unexciting methods like reducing food intake / exercising more until they get to a target weight, and then continuing at a somewhat less-intense level to maintain it.

I think there’s more going on here - I think attempted niceness and social aspects are in play here. It’s not nice to believe that eating less and exercising more will produce weight loss, and it’s not nice to believe that continuing to do those things will maintain it. The reason it’s not nice to believe these things in certain circles is that it implies a certain level of responsibility of the person who is overweight - i.e. that they could be at a healthy weight, but aren’t exercising available options to get there.

For some people, saying that someone is responsible for not losing weight sounds a lot like saying they are at fault for doing so. And people are often mean to people who they believe to be at fault for an undesirable trait where they wouldn’t be to someone with a trait beyond their control. It’s possible we are looking at researchers letting their motivation to keep the overweight from being further stigmatized and abused, and letting this affect how they represent their findings. This is the kind of belief that builds up in a bubble - if we send out this data as-is, won’t the bad, lesser people use it for wrong in some way? Don’t we have a duty to make sure that doesn’t happen?

I don’t think that someone who thinks this way is doing something bad, so long as their motivation is considered in isolation. But taken as a whole, the last ten or 15 years of dietary science has used this motivation to shift towards a number of stances that taken cumulatively seem very likely to discourage people to make any attempt to lose weight at all. After all, why would you waste your time and deny yourself if there’s no chance of success?

I’m retreading a lot of the same ground as the last article, but let me encourage you again: there’s nothing in the literature that proves weight loss is impossible, and much less than you’d think that indicates that it’s any harder than we’d expect. That doesn’t mean there aren’t challenges - pretty much everyone agrees that long-term benefit in this arena takes serious, long-term effort.

I’m on record as saying that people should be able to pursue behaviors that are risky to themselves (drinking, smoking, etc.) if they know the risks and find it’s worth it to them. The same applies here - I’ll never make anyone feel bad for being overweight, because it’s legitimately tough to lose large amounts of weight and requires large, substantial sacrifice. I get it. But if it’s something you want to do - if weight loss is important to you - I’ll never discourage you from trying. Without adequate justification in the data, science shouldn’t either.