Session: Cost of Obesity
Room: Hollister B14
Time: Mon 08:30-10:00

Presenter: Jeroen van Meijgaard (University of California, Los Angeles. Health Services)

Discussant: Tianyan HuLehigh University

Abstract

Research Objective:
Physical activity and obesity are both strong predictors of morbidity and mortality. We hypothesize that physical activity is an independent predictor of medical expenditures and confounds the relation between overweight/obesity and medical expenditures. The objective of this study is to evaluate the independent impacts of both physical activity and obesity on expenditures and how the relative impacts have changed from 1998 to 2006.

Study Design:
To estimate the independent impacts of physical activity and obesity on expenditures, two-part models were estimated for the under-65 and over-65 populations. The first part is a logit specification, which provides relative risks for having any medical expenditures, relative to the reference categories: sufficient physical activity, expressed in METhrs/wk, and normal BMI (18.5-25.0 kg/m2). The second part is a GLM specification with a log link and Gaussian distribution family. The study controlled for age, gender, race/ethnicity, income, insurance status, education, health behaviors and health status variables that may confound the relations between physical activity and obesity, and medical expenditures. The predicted marginal mean expenditures were calculated based on physical activity behavior and level of BMI. Confidence intervals of the means and of the differences across categories were calculated using simulation methods.

Population Studied:
The data used in this study is a representative sample of adults in the United States from the Medical Expenditure Panel Survey (MEPS), years 1998-2006, linked to the National Health Interview Survey (NHIS), years 1997-2005. The analysis is stratified by gender and age: (1) 18-64 years old and (2) 65 years and over.

Principal Findings:
We find that overweight and obesity significantly increase medical expenditures, by as much as 50%. However, decreased levels of physical activity do not increase medical expenditures; in both age groups, medical expenditures of sedentary individuals are not significantly different from those of physically active individuals. Furthermore, for women 65 years and over expenditures attributable to obesity rose more rapidly from 1998-2002 to 2003-2006 as compared to expenditures not attributable to obesity, by as much as 29%.

Conclusions:
The results confirm previous work that higher levels of overweight and obesity significantly increase medical expenditures. The impact is substantial and increasing over time. Currently as much as 6% of total medical expenditures may be attributed to increased BMI. No support was found for the impact of physical activity on the level of expenditures. Thus, although physical activity may lead to lower BMI, and indirectly lower medical expenditures, there is no direct impact of physical activity on medical expenditures. This also suggests that physical activity is not a strong confounder in the analysis of obesity and medical expenditures.

Key Terms

medical expenditures, health behaviors, obesity

Authors:

Jeroen van Meijgaard (UCLA School of Public Health. Health Services) , Gerald Kominski (UCLA School of Public Health. Health Services) and Frederick Zimmerman (UCLA School of Public Health. Health Services)