Tuesday, May 15, 2012

A Case of Physics Envy?

With the airing this week of the HBO obesity documentary "The Weight of the Nation," reporting on the obesity epidemic has reached a fever pitch. I wanted to take a break from my usual programming to mention an article I saw today in the New York Times entitled "A Mathematical Challenge to Obesity."

Until now, mathematical hubris has been mostly the province of economic forecasters and financial risk modelers. As we have seen many times, if you do not fully understand the system that you are modeling, your model will be junk even if you think it's the greatest thing in the world. Now it seems the modeling bug is looking to infect the nutrition field.

The article is an interview with a mathematician who works for the National Institute of Diabetes and Digestive and Kidney Diseases, a division of NIH that I never knew existed. The mathematician, in an act of physics envy, claims to have created a model of a human being based on a single equation. I thought I would just re-print the email I wrote to Claudia Dreifus, who is the Times writer who conducted the interview. Text in brackets are for context if you have not yet read the article (they were not in my email to Ms. Dreifus).
"I read your interview with Carson Chow and had a few questions. 
1. Mr. Carson claims to have proven that increased food availability "caused" the obesity epidemic. He says that if there is extra food, people eat more. Yet animals typically do not become obese under "ad libitum" conditions (i.e. unlimited  food availability). Why are humans different and how does Mr. Carson's mathematical model "prove" that?
2. How does he know the increase in food availability is a cause and not a result of the obesity epidemic? The alternative hypothesis is that people are hungrier than they were before, and since they want to eat more, the food system produces more.
3. Why does Mr. Carson say that all diets work, when the clinical studies on diets almost invariably show that they don't work over the long term? If Mr. Carson has not done any experiments, how can his work prove that dieting works at all?
[Carson Chow said "it's so easy for someone to go out and eat 6,000 calories a day."]
4. Has Mr. Carson tried eating 6,000 calories a day for more than a couple of days? I don't think it is so easy, and professional bodybuilders and powerlifters often say that eating that much is the hardest part of their sport.
5. I tried the bodyweight simulator on the NIDDK site. Mr. Carson's mathematical model says that if I go from a very low carbohydrate diet to to a 6,000 calorie diet consisting of 100% carbohydrate, I would be able to gain hundreds of pounds within a week while maintaining a single-digit body fat percentage. While I agree that carbohydrates can cause weight gain for some people, that result is a little hard to swallow. Did the New York Times check the calculations to confirm they produced sensible results before interviewing Mr. Carson?"
I will update the blog if I get a response.

That last point is the most interesting to me. I believe the sports world will change forever once athletes realize they can gain nearly 750 pounds of lean mass in a week, without exercising. It is amazing nobody has taken advantage of this phenomenon before, and the first football team fielding multiple 900+ pound players is sure to be a favorite next season.

Here is a screenshot from the simulator showing this absurd result. The mistake probably stems from having "baseline calories from carbs" alone in a denominator somewhere (the app at least will not let you set this value to zero, so they caught that part of the bug).