Let's not exaggerate obesity revisionism

All the coverage has yelled "gotcha," following a new study by CDC scientists, who reported a lowered estimate of how many annual deaths are caused by obesity. The economists at Marginal Revolution and Steven Levitt's weblog both pursue this story line, although both posts cite secondary sources rather than the scientific article in question. The revisionism seems overstated.

Scientists have excellent estimates of trends in obesity. These trends are astonishingly rapid for a genetically stable population, and reflect major changes in our environment. Scientists also have a wealth of information showing the risks presented by obesity. Even the new article by the CDC scientists cited what scientists know on this topic:

A recent population-based study has found that overweight and obesity have a strong and deleterious impact on important components of health status, including morbidity, disability, and quality of life, and this impact is disproportionately borne by younger adults.

By contrast, it is difficult for scientists to estimate a key statistic that has special policy punch: the number of deaths attributable to obesity. Don't let your public health judgement blow too lightly in the back-and-forth winds of research on such a difficult-to-measure number.

The reason the new study lowered that estimate is not because the increase in obesity is in doubt, nor because obesity is found to be harmless. Rather, it seems to be because medical technology may have made obesity less fatal:

[T]he association of obesity with total mortality may have decreased over time, perhaps because of improvements in public health or medical care for obesity-related conditions.

Even if heart bypass surgery has become more effective at prolonging life, that hardly means the health hazards of obesity should be dismissed. I remember reading a quip by a vegetarian author, when told that the vegetarian diet seemed Draconian: "And what's your view of open heart surgery?"

Furthermore, the new study still found a strong association between official "obesity" (body mass index greater than 30) and estimated a low mortality impact only for official "overweight" (body mass index between 25 and 30). If this latest article leads you to worry about BMI>30 and to dismiss mild overweight as just fine (which I think is premature except for muscular athletes), then you may be interested in trends in the more severe category of obesity. Here is the 2003 graphic for rates of BMI>30, and you can follow the link for the full set of CDC slides for the time series.