NBER: Obesity costs U.S. $168B annually
Using instrumental variables (IV) to estimate the impact of obesity on medical costs, researchers found that the causal effect of obesity on medical expenditures is four times higher than when non-IV models are used, indicating that previous literature has underestimated the impact of obesity on medical costs, according to a National Bureau of Economic Research working paper published this month.

The report estimates the annual cost of treating obesity-related illnesses annually in the U.S. is $168.4 billion, or 16.5 percent of medical spending. Previous studies have underestimated this cost and reported that obesity is only responsible for 9.1 percent of medical spending.

In 2007-2008, 33.8 percent of American adults were clinically obese, and this number has doubled in the past 30 years.

Because previous obesity studies focused on the correlation of obesity with medical care costs rather than the causal effect, John H. Cawley, PhD, of Cornell University in Ithaca, N.Y., and Chad D. Meyerhoefer, PhD, of Lehigh University in Bethlehem, Pa., used data from the Medical Expenditure Panel Survey (MEPS) for 2000 to 2005 to address both sides of the spectrum—endogeneity of weight and reporting error in weight.

To do so, Cawley and Meyerhoefer utilized IV models and controlled for: gender, race/ethnicity, respondent age, education level, census region, household composition, whether survey information was self-reported, employment, gender of the oldest child and age of the oldest child in months.

“The impact of obesity on medical expenditures may vary across the distribution of medical spending,” the authors wrote. “Furthermore, the marginal impact on medical expenditures of an increase in body weight is presumably greater for individuals of poorer health status with multiple chronic conditions.”

The researchers reported the average body mass index (BMI) of patients evaluated in the study to be 28.17 for men and 27.37 for women; the prevalence of obesity was 28 percent.

The sample was limited to adults with biological children, "as they are the only MEPS respondents for whom we can estimate the IV model," the authors wrote. The results showed that 79 percent of men and 88 percent of women in the MEPS database incurred some type of medical expenditures during the duration of the survey. For men, the unconditional average medical expenditure was $1,999, for women it was $2,617.

“Our results indicate that the effect of obesity on medical care costs is much greater than previously appreciated,” Cawley and Meyerhoefer wrote.

Pooled results showed that obesity (relative to having a BMI less than 30), when a non-IV model was used, was associated with a $676 total medical expenditure. These rates were $582 for men and $786  for women. When researchers weighed an additional unit of BMI beyond just the mean (IV model), the rate of medical expenditure was raised by $154 in the pooled sample, $82 for men and $178 for women.

Relative to being non-obese, obesity raised the medical expenditures by $2,826 for the pooled sample (23,689) and $1,171 for men (9,852) and $3,696 for women (13,837) when the IV model was utilized.

“A comparison of the results of the IV and non-IV models indicates that the causal effect of obesity on medical expenditures is four-times higher than its association, for both the pooled sample and for women; in other words, the previous literature has considerably underestimated the impact of obesity on medical costs,” the authors wrote.

Cawley and Meyerhoefer also reported that obesity-related medical costs are higher for patients who were uninsured (n=4,379) compared to those who held private insurance (n=16,475), $3,496 and $2,624, respectively. Additionally, the researchers found that the association between obesity and medical expenditures were twice as large for whites (n=12,575)  compared with non-whites (n=11,114), $827 versus $414, respectively.

“Despite the large point estimate of the impact of obesity on third-party medical expenditures for the Medicaid population ($3,647), it is not statistically significant, meaning that one cannot reject the null hypothesis that obesity does not raise medical costs in this large public health insurance program,” the authors wrote.

When comparing the medical expenditures of the obese and non-obese, the researchers found that obesity raised these costs by almost 50 percent, from $1,840 to $2,826. For women and men, obesity raised medical expenditures by 90 percent and almost 70 percent, respectively.

“The largest relative increase in expenditure occurs among the uninsured, for whom obesity raises medical expenditures by 490 percent, from $593 to $3,476,” the authors wrote.

“Our estimates indicate that the annual direct medical cost of obesity averaged $22.9 billion over the six-year period, of which 90 percent was borne by third-party payors. In the most recent data, from 2005, the cost of obesity in this population was $25.5 billion, of which 89 percent was borne by third-party payors,” the authors wrote.

“The results of this paper also indicate that insurance companies spend more treating obesity-related illness than was previously thought, and thus there may be a better business case for insurance companies to cover effective treatments for obesity (such as prescription drugs for weight loss, bariatric surgery and nutrition counseling) than was previously appreciated,” the authors wrote.

The authors noted that one limitation is the fact that the IV method was only implemented on adults with biological children, and noted that caution should be used when generalizing these results to the entire population.

“[T]his paper makes an important contribution by providing the first estimates of the causal impact of obesity on medical care costs, which are significantly higher than the estimates of the association that are published in the previous literature, and thus have important implications for insurance, government policy and estimates of the cost effectiveness of anti-obesity interventions.”