First of all, many are calling the study proof of the "Obesity paradox," although the study clearly showed that obese people, that's people with a BMI above 30, had an overall increased risk of dying. Also this was an observational study, not a cause and effect one, which already makes the results somewhat questionable.
Second, if 67% or so of the population is overweight and 36% is obese, which means that 31% are overweight and not obese, that only leaves 33% of the population as either normal weight or underweight. Included in the 33% are people who have lost weight due to illnesses such as cancer, eating disorders, hyperthyroidism, malnutrition, etc.
The study does not account for why people fall into the normal and underweight category (healthy or unhealthy lower weight) given the preponderance of people in the overweight category.
Third, it doesn't account for people who have normal BMIs, but high waist to hip ratios which signifies excess white fat, the dangerous kind of fat. A recent Mayo Clinic study showed that among 14,000 study subjects, those with a normal BMI, but a high waist-to-hip ratio, were the most likely to die of heart-related disease during the 14-year follow up, even at higher rates than those in obese ranges.
Fourth, it is well established that people in advanced stages of cancer and diabetes do better with additional weight because as the body wastes away, it takes longer to do so. G-d forbid you should need excess weight for that reason.
Fifth, the study did not assess people who were underweight as it is known that such people have increased mortality. The underweight people were lumped together with the normal weight people to come up with a total death rate, so one can expect that the underweight inclusions skewed the numbers somewhat.
Sixth, the study did not address people who are overweight on the road to becoming obese, as opposed to those who are chronically overweight. By not knowing if recent illness is contributing to change in weight, we can't do an apples to apples comparison.
Seventh, the study underscores why BMI may not be the most reliable indicator of unhealthy weight because it does not take into account the muscle to fat ratio. Athletes, for example, could have increased muscle mass that will increase their BMI and put them into the overweight (even sometimes obese) category despite their otherwise good health.
Eighth, it is well know that excess weight, in and of itself, does not mean you are unhealthy if you don't have concomitant increases in cholesterol, blood pressure or blood sugar.
Ninth, is there a difference between a BMI of 25.1, which puts you into the overweight category and 29.9, which keeps you out of the obesity category? The study does not address gradations of being overweight.
Tenth, and finally, it not entirely clear to me what exactly is the point of this study. Is it to tell people that being overweight is not as bad as most people think? Should people who are overweight, but not obese be less concerned about their diet and lifestyle?
Of greatest concern, should people who are overweight, but not obese, make sure that they don't lose any weight as that may increase their risks for death?
I am not surprised that this is a government study as I find such studies to sometimes be the most generically misleading. Such a study would seem to offer comfort to the 67% of Americans who find themselves in the overweight category as you can rest assured that most people will not even bother to look past the headline that states that extra weight may be healthy. They will just take comfort in the notion that that extra piece of bread or serving of pasta may be health promoting.
G-d help us all if the government keeps doing these types of studies. Next thing we know they will tell us to cut out fats and eat more carbs. Oops, they did that already and we all know what happened.
Update, February 3, 2913: A recent study offers support for the notion that far more important than BMI is the waist to hip ratio. This measurement of "belly fat" is more indicative of overall health because the build-up of white fat cells at one's midsection is more reflective of overall inflammation within blood vessels and risks for heart disease. So those who took heart from the study that it's better to be heavier than be normal weight, please understand that this referred to well distributed weight and not weight concentrated around your gut.