Obesity Hype

The Robert Wood Johnson Foundation and Trust for America's Health just released a report entitled F as in Fat: How Obesity Threatens America's Future 2012.   The associated press release reports:

For the first time, the annual report includes an analysis that forecasts 2030 adult obesity rates in each state and the likely resulting rise in obesity-related disease rates and health care costs. 


If obesity rates continue on their current trajectories, by 2030, 13 states could have adult obesity rates above 60 percent, 39 states could have rates above 50 percent, and all 50 states could have rates above 44 percent.

Obesity is out of control!  Or is it?  If we look at the actual data, we can see that no such explosion in obesity rates are occurring.  One group of researchers writing in 2010 in the Journal of the American Medical Association reported that that the prevalence of obesity :    

may have entered another period of relative stability.

Another re-appraisal in 2012 by the same authors using new data in the Journal of the American Medical Association reported

Over the 12-year period from 1999 through 2010, obesity showed no significant increase among women overall

and only slight increases for men.

Here is the key graph from the JAMA paper:


As you can see, trends in BMI since 1990-2000 are basically flat.

So, how is it that the Robert Wood Johnson Foundation and Trust for America's Health can claim run-away trends?  I don't know, but it makes me suspicious of anything else they claim in their report not to mention their policy prescriptions.  

Indeed, it might have done the authors of the report some good  to read  or at least acknowledge what was written in the 2012 JAMA paper (footnotes omitted):

In part because we know relatively little about the precise causes of the trends previously observed, it is hard to predict the future trends in obesity. Several analyses have modeled increasing obesity prevalence as a function of calendar time and then projected future obesity prevalence from these models. These obesity predictions in effect assume that the causal factors for obesity will continue to rise with time or will have an increasing effect over time, and therefore calendar time itself is a reasonable predictor of future obesity prevalence. However, the results reported here and the apparent slowing of trends suggest these may not be valid assumptions and these predictions may be inaccurate.