The medical costs of obesity are twice as high as previously reported, according to the first study to estimate the causal effects of obesity on U.S. medical costs.

Using innovative methods, Cornell health economist John Cawley found that the annual estimated cost of treating obesity is $168 billion -- 16.5 percent of the country's total medical care costs.

The higher costs are not incurred so much by the nearly two-thirds of American adults who are overweight or obese, said Cawley, associate professor of policy analysis and management. Rather, they are driven by the skyrocketing costs of medical care for a small percentage of highly obese individuals, for example the 4.8 percent of Americans who are classified as "morbidly obese."

The study, conducted with Chad Meyerhoefer of Lehigh University, was recently published as a working paper (No. 16467) by the National Bureau of Economic Research, a nonprofit, nonpartisan research organization.

Previous studies have reported that obesity is associated with roughly $1,400 higher annual medical care costs per person. The new study reports that it is more in the range of $2,800. The cost of a case of obesity among recipients of Medicaid, a taxpayer-supported program, is some $3,378 annually in 2005 dollars.

"Given the difference between our findings and previous studies," Cawley said, "we checked our estimates in many ways and found that the findings are very robust. Our estimates are higher in part because we correct for reporting error in weight, and partly because we measure the causal effect of obesity rather than just its correlation with medical costs."

The researchers used a federal survey of 24,000 non-elderly U.S. adults, their doctors and other medical providers from 2000 through 2005, the same survey that other researchers have used. However, rather than correlating weight with health care costs, which would not be able to take into account obesity caused by medical conditions (such as surgery or a bad back triggering inactivity and weight gain), the researchers measured the causal effect of obesity using one's inherited propensity to obesity. Using the weight of relatives, they could determine if obesity itself triggered medical costs.

"Our method builds on previous studies that show that the similarity in weight among family members is due to genetics, not shared environments," Cawley explained.

They also statistically adjusted the subjects' weights to eliminate the influence of reporting error. "Weight is notoriously underreported," Cawley said.

Cawley and Meyerhoefer scoured the medical literature on obesity: "It is striking that so many medical conditions are aggravated by obesity," Cawley said. "Obesity increases the risks of cancer, heart disease, asthma, stroke, degenerative joint disease and so on."

The researchers conclude that their findings not only suggest that previous studies have underestimated the estimated annual cost of treating obesity but also have resulted in underestimates of "the cost effectiveness of anti-obesity interventions and the economic rationale for government intervention to reduce obesity-related externalities."