Time to Toss the BMI

One of the most widely reported science stories yesterday was about a group of scientists who've just shown that the standard measurement of obesity -- the Body Mass Index (BMI) -- is completely flawed. All the media coverage made it sound like this was a new and surprising finding, but in fact, the BMI has been known to be flawed for years.

I wrote about this last June, when a study from the CDC was widely misreported as saying that people who were slightly overweight lived longer than those who weren't overweight. People went nuts. In a column I was writing for Popular Science at the time, I explored the problems with that study, and the way it was being interpreted by the media (and therefore the public). The biggest problem, I wrote, was the BMI:

"The study’s most obvious limitation is its use of the unreliable “body mass index” (BMI)—a number determined by a person’s height and weight—to define “normal” and “overweight.” A BMI of between 18.5 and 24.9 is “normal,” between 25 and 29.9 “overweight,” and 30 or more “obese.” But BMI doesn’t take into account many important factors: physical activity, fat versus muscle, gender, diet. This means George W. Bush—a nearly-six-foot-tall 200-pound guy who eats well and works out regularly—has the same BMI as a six-foot-tall 200-pound guy who sits on the couch all day eating junk. With a BMI of 27.1, they’re both “overweight.” But President Bush has precisely the right amount of body fat for his age, and he’s in great cardiovascular health. I’d like to see the same study use some kind of body fat index. Bush’s percentage of body fat is 18.3, which is considered excellent for his age. Not the case for that out-of-shape guy on the couch."

When you scrap the BMI and determine obesity by looking at hip-to-waist ratio instead, which is what these authors did, you actually triple the number of people who qualify as being at risk for heart disease. Many will argue against this, but the way I see it, this Lancet study (free registration required), has finally proven that BMI is not the way to go. That always seemed pretty obvious via common sense. But common sense doesn't always apply when dealing with obesity, so apparently we need hard stats to back it up.