Another piece of cake, take two

A recent study suggests that the lowest mortality in the US is found among those categorized as "overweight" by national guidelines and that body mass index has a much smaller influence on mortality than had been commonly believed and that being thin causes as many or more deaths than being too fat. This is obviously a major change in the way we think about obesity and health, so I've blogged a lot about it.

In part I, I reveiw the study itself
in part II, I ponder the question of what thin people die from
in part III, I criticize BMI as a measure of obesity and the current categories (normal, obese, etc.)
in part IV, I try to illustrate what I see as the most important implications of the study

(Boy this post adds a lot -ed; I wanted to organize it for Carnival of the Vanities; you'd better hope vanity doesn't increase mortality; guilty as charged)