Lots of times I don't get around to finishing a series of posts that I had planned. Well, this is the third and probably final post on the popular study where energy expenditure was measured while folks tried various weight maintenance diets.
Lots has been said about it all over, and I was really waiting for Hyperlipid to say more about it, but that will probably go on for weeks. I hope people are reminded that I am not a research biochemist. I can analyze data, but I don't really KNOW anything, and so rely on folks like Peter to do some of the heavy lifting on those parts.
I have read lots about how small the sample size used in this experiment, and I find those arguments to be sub-optimal, especially when significant differences were noted. My take on this is that the researchers already sort of knew what they were looking for and had picked the sample size so they could detect minor but statistically significant differences between diets. The sample size problem would really be more of a problem if they had declared there to be no differences between any of the diets. This is a favorite trick.
People sure like to hone in on the sample size, and it is important to get that right, depending on the cost of the study and other factors. Another favorite trick is to bowl them over with lots of n's, and with lots of meaningless charts and graphs, hoping that nobody will really read anything and discover under the thin-veneer of a high-sample overly-charted masterpiece, everything underneath is rotten. Getting crappy data on 100,000 people, with a crappy food questionnaire, a notable lack of doubly-labeled water, doing a crappy epidemiological study and then doing a crappy job of "adjusting" for all sorts of variables that should have been controlled in the first place gives one, in the end, a bunch of crap. I would rather have a smaller experiment well-done. For this experiment, they kept pretty good track of the people in it, retention was high, parts of it were in-house and more carefully controlled.
There are a couple of issues I do have with this study. First, there is lots of talk that this is a test of the Atkins diet. I wish people would quit invoking Atkins when it isn't Atkins. The Atkins plan recommends VLC for a few weeks, and then on to a higher-carb plan, and when the weight is lost and they go on maintenance, which usually involves even more carbs. Additionally, the Atkins diet never attempts to be isocaloric. The plan recommends you eat when you are hungry and when you are full, stop. There is no prescription for it to be isocaloric. In fact, calorie-counting is discouraged.
I don't know a single person in the whole wide world who lost a significant amount of weight on a conventional CW diet and then decided to go on a VLC ketogenic diet for maintenance. Nope. Nada. Just doesn't happen. So, all the media attempts to say that this experiment somehow is representative of anything that actual dieters will ever actually experience is, well, an attempt and that attempt will hopefully not be successful.
My second beef with this study is that it has been reported as a comparison of weight-loss diets. Many people reported this experiment as a test of three different weight loss plans. It was not. It was a test of three different maintenance plans, after the participants lost weight successfully on the CW prescription. If they were not successful in this first phase of the experiment, they did not progress to the second part and did not get to supp the doubly-labeled water. So, unless you want to settle a university playground brawl, who cares? We really all just want to figure out how to lose weight. We don't really want to find a diet that will allow us to eat more, we want to find a diet that will allow us to not want to eat ourselves silly all day long, and not allow us to store all of the excesses we desperately crave and cave for. As for myself, I was too old to be enrolled in this study, and I would have failed the first part for sure, and would not be allowed to supp the doubly-labeled water either.
My third beef with this study is the crossover aspect. Very few reviews of this study seemed to notice that each diet was followed for a time by each person. Now I understand why they wanted to do it this way, because it minimizes the between-subject error due to individual metabolic differences, but SHEESH!, each diet was supposed to be followed for a few weeks, and then the participants were to magically adjust to the next plan?
I tried to comment about this same phenom on a Ned Kock blogpost, and he just didn't get it. These young guys and their fixation with the theory of balance They are like Tim Ferris and all their little boy-friends who try this or that for a day or so and then everything settles down back to normal. Meanwhile, I can try a few tweaks to my diet or exercise regimen and then try to get back into balance after WEEKS or MONTHS. Things just don't work that balance way. Anyway...... I was going to give Ned the benefit of the doubt for awhile, and he just never did rescue himself from his own confusion, despite my hints. At least he removed the half-naked sunburned paleo-required mugshot from his blog. Maybe there is hope for him yet.
...Anyway...oh, where was I?.....Oh, I would love to see this study done with the three plans tested after a weight-loss program modeled after the Atkinsarterycloggingsaturatedfat plan, and then start to supp the doubly-labeled water, and/or a longer time period on each plan and/or a longer wash-out period between plans (which of course, for people like me, would create a whole host of new problems). I don't right now remember reading if there was any transition time, but I am sure that whatever it was, someone like me would need more time, or perhaps it would be infinity.
Gosh, wouldn't it be cool to do this with people who have totally stalled out on either CW or the Atkinsarterycloggingsaturatedfat plan, and then finally put to bed for good that meme that we all fail because we can't use a measuring cup and we lie on our food logs? Now, that would be a cool study.