Extrapolating

I was driving around in my car yesterday, minding my own business, when a story came onto my local NPR station that caught my attention.  Kevin Hall PhD was on the air talking about his new study comparing how the body responds to low fat vs. low carb diets.  He was careful to state that the study compared people in carefully controlled environments as opposed to relying on food journals for information on food journals (because you KNOW us fat folks can’t be counted on to be honest about those things).  He said that the study participants were brought in for two, two-week inpatient hospital stays where every morsel of food was accounted for.  In fact, every person was fed the exact same breakfast, lunch and dinner every day for two weeks, were expected to eat all of it and were not allowed to have anything different.  (I think the study creators took their lives into their hands with this one, because if you fed me the exact same, low-calorie, diet crud for two weeks running, I’m quite sure I’d become pretty stabby.  And I don’t know what could possibly compel me to return for the second phase of that torturous crud.  But I digress.)

Kevin said all of the subjects “had obesity” which is a funny, disease-language sort of way to say they were fat.  But none of them had diabetes.  And according to Kevin, the study indicated that at a metabolic level there was very little difference in weight loss results from the low fat vs. the low carb diets.  Although, the low carb diet had a greater impact on blood glucose levels than the low fat diet.  And the low fat diet apparently led to slightly greater fat loss within the two week period than the low carb diet.  Okay.  Got it.

But then the report started to go off the rails for me.  Kevin suggested that based on this study we could extrapolate that the low-fat diet would lead to greater weight loss over the long term than the low-carb diet.  And that’s when I had to pull the car over because I was shouting at the radio.

“How can you extrapolate anything over the long term other than the notion that most of these people won’t keep the weight off or will even gain weight after this experiment,” I said.  “This study has a lot to do with metabolism but you can’t really expect to transfer the results of this study to real life!”  I tried to call in to express my opinion, but by time I got through the story was over.  So I resolved to hit the radio station website, find the study and take a look for myself.

I did just that.  And the data extrapolation situation was so much worse than I thought.  The study contained a grand total of 19 subjects.  During the first two-week stay, the subjects ate a eucaloric baseline diet for 5 days, followed by 6 days on either a reduced fat diet or a reduced carbohydrate diet.  Then there was a 2-4 week “washout period” when they went home.  After that, they returned and ate the eucaloric baseline diet for 5 days followed by a reduced fat or reduced carbohydrate diet.  Each subject switched diets for the second two week stretch. So if they did the low fat diet in the first stretch, they did the low carb diet in the second stretch and vice versa.

Okay.  So there’s some good stuff here.  The study carefully tries to control for various factors.  The food is the same each day, so there won’t be wild fluctuations there.  All subjects are, um, subjected to both diets so that helps eliminate some of the variables.  And everybody eats exactly the same thing.  Subjects spent a significant amount of time in a metabolic chamber to very carefully measure the way the calories were metabolized.  I see the point of needing this level of precision to control the variables of the experiment.

But my big, BIG problem came during the radio interview when Kevin started talking about how this could apply to real world fatties and what it would mean for them.  He suggested that over the long term (so far the study doesn’t even have 6 months worth of follow up data) the minute differences in weight loss from the two diets would accumulate and you would see significantly greater weight loss from the low-fat diet vs. the low carb diet.  This despite the overwhelming evidence outside of this study that indicates that over the true long term (2-5 years) the most likely outcome for the vast majority of the subjects is either insignificant weight loss or even weight gain.

Don’t get me wrong, I love it when science people are science-ing.  But when you extrapolate from 19 subjects over 22 days in an extremely controlled environment that is virtually impossible to replicate in the real world to what effect these diets will have over the long term in the real world, I’ve got a serious problem with that.  I have talked about it before, and I will talk about it again.  If you want to understand science, possibly the last thing on earth you should do is listen to science reporting in the mainstream media.

And I have to wonder about the millions of people who are going to hear this report and hop on the diet/shame/regain cycle all over again because science people say they should.  Because when you take science and simplify and stretch it beyond all limits in order to get a popular news story to help bump up your funding, you are behaving irresponsibly and people, especially fat people will wind up getting hurt.

Love,

Jeanette DePatie (AKA The Fat Chick)

P.S. Want me to come and sort out some science reporting for your group?  Click here to learn more about hiring me as a speaker!

6 Comments. Leave new

  • Iiiiiiiiiiiiiiiiiiiiinteresting. I heard about this study, but I haven’t gone digging myself. Thank you for reading it and writing this. I think the study design could be useful in some ways, as you point out, but it does get frustrating when, over and over again, extremely short-term, or very low-n studies are used to make wild extrapolations about long-term weight loss for great numbers of people. It’s just endemic in the science at this point, and it’s annoying and it contributes to an overall hostile environment, because people are led to believe (by the reporting on these studies) that weight loss dieting has been “proven to work!” BY SCIENCE!!!! and anyone who is fat just hasn’t tried hard enough.

  • “How can you extrapolate anything over the long term other than the notion that most of these people won’t keep the weight off or will even gain weight after this experiment.”

    It’s called a “model,” and it’s pretty accurate. Here’s the guy explaining it (this isn’t the first study supporting this model, there is a ton of other data as well): https://www.youtube.com/watch?v=hPi1LQHBWBk

    PS, the people were only on the restricted portion of the diet for 6 days, and the model predicted about 430 grams of fat loss on the RF diet, and about 245g for the RC. The actual results were the RF diet resulted in 463 ± 37 g of fat loss compared to 245 ± 21 g of fat loss with the RC diet . Note the very low variance between subjects. Increasing the number of people in the study won’t change the numbers much.

    • Considering the poor results of every single intentional weight loss protocol over the long term, I believe it’s okay for people to want actual long-term data for any new model being presented.

    • Hi Tetondon,

      Thanks, I understand very well what a model is. And yes a model can be very accurate. However, the vast majority of available evidence indicates that weight loss does not continue unabated forever. In fact, it would be impossible because people who lived long enough would weigh zero pounds. The research outside of this study suggests that for most people, there is an initial weight loss phase which becomes slower and shallower over time, followed by weight regain for almost all subjects in the long term (I’m talking at the 2 to 5 year point). So while I don’t have a problem with the methods in this study in exploring certain very specific things, I do have a problem with the study’s author stating in an interview that “in the end” low fat is going to lead to more long term weight loss than low carb. I don’t think the study provides enough information to tell us what will happen in the end, and the research that we do have about the long term (2-5 years) predict that the vast majority of subjects won’t maintain weight loss regardless of which weight loss regime they follow.

  • Fat Chick, you are right on the money! I am the scientist cited in the article you link to (cite 5 in that article) stating that within 2-5 years, the average dieter regains most of the lost weight. That result came from combining the findings of every single diet study that followed the dieters for at least two years (most diet studies don’t). We simply cannot extrapolate from short-term studies. The pattern is clear: short term weight loss followed by long term weight gain. I explain all of this, plus why the weight returns, in my book: Secrets from the Eating Lab: The science of weight loss, the myth of self-control, and why you should never diet again. You and your readers may find it of interest: http://www.amazon.com/Secrets-Eating-Lab-Science-Willpower/dp/0062329235

    Keep up the important work you are doing!

  • Looking at the actual numbers in the study (http://www.cell.com/cell-metabolism/fulltext/S1550-4131%2815%2900350-2) I see some interesting issues that make it further difficult to extrapolate. First of all, I find it highly intriguing that the reduced fat diet provides more protein than the reduced carb diet. I don’t think that would ever be how it would pan out in the real world.. Actually, I am INCREDIBLY curious now what kind of diet would provide 350 carbs, 17 grams of fat, 21 grams of fiber and 105 grams of protein in 1,918 calories. I would guess lots of fat-free cottage cheese, chicken breast, and rice..? Fat reduction, or carb restriction, isn’t exactly the only thing being tested here.

You must be logged in to post a comment.
keyboard_arrow_up