I came at this one assuming it would confirm my "BMI is bad" bias. It did not. It is a different story. It's a health economics story I think, and a statistics-is-hard story.<p><i>Estimates of the BMI-mortality association in the NHANES data are significantly affected by all three biases, and obesity-mortality associations adjusted for bias are substantively strong at all ages. The U.S. mortality consequences of overweight and obesity have likely been underestimated, especially at older ages.</i><p>So it's arguing to an under-estimation. This goes very quickly into health policy effects in diet and advertising, food regulation, education. It goes very quickly into "personal responsibility" minimisation of intervention which is the other side of the view on how to manage public health crisis (ie, by not managing it because poor people have personal responsibility for their lifestyle choices, in this world view)<p>Because aged care is a least-cost burden (as in people seek to implement it for the least cost), and providing high protein, nutritious diet to people at end of life is both expensive and hard, they often have carb-rich fatty sweet diets and for other reasons low exercise levels. So this pretty much goes to "old people die because we don't invest in helping them live" -But the mortality stats won't put "died because they were fat and we fed them badly" because thats not in the aged care facilities best (financial) interests. I guess thats one of the bias' they talk about in the paper at root.