Seth’s blog

As any reader of The Shangri-La Diet knows, I attribute the obesity epidemic to ditto foods — foods that taste exactly the same each time, such as factory food and fast food. We eat a lot more of these foods today than 50 years ago or even 20 years ago.

An alternative explanation of the obesity epidemic that many people believe is too little exercise. People who deal with childhood obesity, in particular, often say the problem is too much TV, too little playground.

If kids are fat due to lack of exercise, more exercise should be a good solution. A new study shows it isn’t. It turns out that giving kids more P.E. doesn’t cause weight loss:

In studies involving nearly 10,000 children, primarily in elementary schools, none demonstrated a reduction in BMI with those who were assigned to the most phys-ed time, compared to those who didn’t have as much.

Via Calorie Lab.

A recent issue of JAMA has an article titled “Comparison of Strategies for Sustaining Weight Loss: The Weight Loss Maintenance Randomized Controlled Trial”. It reports an experiment that compared three ways to keep from regaining weight you’ve lost.

If you want to lose weight it paints a discouraging picture. It was an very expensive study, 27 authors, five grants. About 1000 subjects. Four years just to collect the data. The whole thing might have taken seven years. Must have cost millions of dollars. Might have cost tens of millions of dollars.

Given the huge expense, surely the subjects got the best possible establishment-approved weight loss advice. They did lose 19 pounds in six months. Here’s how the advice was described in the article:

Intervention goals were for participants to reach 180 minutes per week of moderate physical activity (typically walking); reduce caloric intake; adopt the Dietary Approaches to Stop Hypertension dietary pattern . . . and lose approximately 1 to 2 lb per week. Participants were taught to keep food and physical activity self-monitoring records and to calculate caloric intake.

Shades of Marion Nestle’s “move more, eat less”! Aside from the DASH “dietary pattern,” which was meant to reduce blood pressure, not weight, this advice could have been given fifty years ago. Apparently, those who did the study and those who funded it — who are representative of the larger research establishment, I assume — believe there has been no theoretical or empirical progress since then.

Many fields haven’t progressed in 50 years. Fifty years ago, 2 + 2 equaled 4. The basic principles of thermodynamics and inorganic chemistry were the same then as they are now. Lack of progress in weight loss advice would be fine if the advice actually worked but the whole study derived from the fact that the advice is poor — the weight loss it produces cannot be sustained.

To help people sustain their weight loss, the study compared three methods: 1. Monthly contact. Usually a 10-minute phone call (”with an interventionist”), every 4th month a hour face-to-face visit. Although the article claims this treatment was “practical,” I suspect it is too expensive for widespread use. 2. Encouragement to visit an interactive website. The website helped you set goals, allowed you to graph your results, and had a bulletin board, plus several other features. This was the focus of the whole huge research project: the effect of this website. It could be offered to everyone practically free, except that if the subject didn’t log on after email reminders she got a phone call. 3. “A self-directed comparison condition in which participants got minimal intervention [that is, nothing].”
The personal contact condition was slightly better than nothing. By the end of the study, the website was no better than nothing. And nothing was bad. The subjects regained about two-thirds of the lost weight during the maintenance year and, looking at the weight-versus-time graph, were apparently going to regain the rest of the lost weight during the coming year. Subjects in all three conditions continued to regain the lost weight throughout the year of maintenance.

In other words, this exceedingly expensive study could be summed up like this: We tried something new, it didn’t work. The abstract didn’t face this truth squarely. It concluded: “The majority of individuals who successfully completed an initial behavioral weight loss program maintained a weight below their initial level.”

It’s a Catch-22: Without a good theory, it’s hard to find experimental effects. You’re just guessing. Most of what you will try will fail. Without strong experimental effects, it’s hard to build a good theory. I was in this situation with regard to early awakening. I had no idea what the cause was. It took me ten years of trying everything I could think of, dozens of possibilities, before I managed to find something that made a difference. From that I managed to build a little bit of a theory, which helped enormously in finding more experimental effects.

The people who did this study had no good theory about weight control. Nothing wrong with that, we all start off ignorant. The website they tested was just the usual common-sense stuff. What’s discouraging for anyone who wants to lose weight is how little progress was made for such a huge amount of time and money. If it takes seven years and ten million dollars and a small army of researchers to test one little point in a vast space of possibilities . . . you are unlikely to find anything useful during the lifetime of anyone now alive (or any of their children). The people behind the study also had a poor grasp of experimental design. With 300 people in the website group, it would have been easy to test many website design variations: weight-loss graph (yes or no), bulletin board (yes or no), etc., using factorial or fractional factorial designs. Their study merely showed that one particular website didn’t work. They learned nothing about all other possible websites. They might have been able to say: no likely website will work. They can’t because the study was badly designed. The study cost something like $10 million and that was the statistical advice they got!

The huge expense and the lack of progress in the last 50 years go together. The methodological dogmatism I discussed recently has bad consequences. It leads to studies that are more expensive and take longer. The proponents of the methodological rigidity say they are “better” not taking account of the cost: continued ignorance about health. A better research strategy would be to fund and encourage much cheaper ways of testing new ideas.

My exercise research suggests our brains work better when we walk. Here’s one way to combine walking and writing:

While working on a paper, which was most of the time, [Niels] Bohr would select an assistant from among the young physicists in Copenhagen. The assistant, affectionately dubbed the victim, was supposed to sit in place while Bohr paced around the room, constantly puffing away at his pip, working and reworking his ideas, talking aloud as the idea took shape, trying and retrying to dictate his sentences to the victim.

From Faust in Copenhagen: A Struggle for the Soul of Physics by Gino Segre.

The modern version may be use of word recognition software with the computer screen on a wall or large TV. You walk back and forth in front of it. I have spent a lot of time writing while walking on a treadmill but it was noisy and tiring. Moreover, it was hard to start and stop and it was monotonous.

As part of my digression into the effects of exercise, I tested the effect of a 15-minute walk (= on a treadmill at about 2.8 miles/hour) twice more. Here are the results:

Here is the result (posted earlier) of the first test:

Here is a test of a 40-minute walk:

What do I learn from all this? For my omega-3 experiments, which might cover 6 hours, I should keep the walking involved under 15 minutes. If I want to get some sort of mental benefit from walking, I should spend 40 minutes or more. Less obvious is this: I take these results to indicate the existence of a mechanism that “turns up” our brain when we are doing stuff and turns it “down” when we are inactive. This suggests what Stone-Age activity consisted of: more than 15 minutes of walking. This also suggests that whatever the benefits of exercise, they require more than 15 minutes of walking to obtain.

The practical question these results raise is how to use this effect to help me with what I do all day — most of which, such as writing, seems to be incompatible with walking. Walking breaks every few hours? What about running 10 minutes every few hours?

Gary Taubes’ new book on food and weight comes out today. Taubes agrees with what I say in The Shangri-La Diet: Exercise is a poor way to lose weight. The results above provide a different reason to exercise, of course. But the details should change. My impression is that most people focus on burning calories; whereas these results suggest choosing exercise that best produces this reaction-time-lowering effect.

Exercise reduces reaction time, I’ve found. What’s the threshold? I wondered — how little exercise do you need to get the effect? I wanted to know so that in my omega-3 experiments, I could be active — e.g., walk to a cafe — without distorting the results. Also, for practical reasons, I wanted to produce the effect as easily as possible.

To learn more about the threshold, I walked on my treadmill for 15 minutes at a comfortable speed (2.8 miles/hour). Here’s what happened:

If anything, the short walk increased reaction time. Thirty minutes of walking produced a clear (and repeatable) decrease, so the the effect appears to require between 15 and 30 minutes of walking.

I did this experiment three days ago. Self-experimentation is many times easier than conventional science; blogging is many times easier than conventional publishing. A powerful combination, I hope.