A Causal Estimate of Long-Term Health Care Spending Attributable to Body Mass Index Among Adults

Abstract

While high body mass index (BMI) is believed to be a major driver of poor health, there is little evidence about whether it leads to higher health care spending. Understanding the causal contribution of BMI to health care spending is necessary to estimate the returns to investment in weight loss efforts. We exploit genetic variation in BMI across siblings as a natural experiment to estimate the impact of BMI on cumulative third party and out-of-pocket health care spending among adults using the Panel Study of Income Dynamics data from 1999 through 2011. We estimate a two-stage residual inclusion model with a generalized linear model. We find a $611.60 increase in cumulative insurer spending for each one-unit increase in BMI. This amounts to $130.49 in mean annual spending, and is two times higher than the non-causal estimate. We find no difference in out-of-pocket spending by BMI. These findings suggest that having a higher BMI in young/middle adulthood leads to significantly higher insurer health expenditures over the life course, which can help to inform public and private insurer policies on BMI reduction and control.