Analyses examining the relationships of physical activity and obesity with total health expenditures

Introduction: Estimating health expenditures associated with physical inactivity and obesity is crucial to obtaining support for policies seeking to reduce the prevalence of physical inactivity and obesity. This study investigated whether changes in body mass index (BMI) and physical activity status are significantly associated with changes in health expenditures. Methods: Data included 18,715 observations from the 2004 Medical Expenditure Panel Survey, a national survey weighted to generalize to the US population. Ordinary least squares regression was used to examine associations between total health expenditures and the explanatory variables BMI and physical activity. BMI was splined into four categories: underweight (BMI<18.5), normal weight (18.5≤BMI≤24.9), overweight (25.0≤BMI ≤29.9) and obese (30.0≤ BMI). Physical activity was defined as moderate/vigorous activity > or = 30 minutes > or = 3 days per week. A Boxcox test indicated the dependent variable, total health expenditures, should remain untransformed. The model controlled for obesity-related co-morbidities, demographics, insurance and smoking status. Results: Cross sectional analysis show per unit changes in BMI categories underweight, overweight, and obese are not associated with changes in health expenditures (p=.05). A one unit increase in BMI in the normal weight category is associated with a $138.27 decrease in health expenditures (p=.05). Engaging in physical activity is associated with a $1,116.65 decrease in health expenditures (p=.001). Longitudinal panel data analyses from the 2003-2004 panel will examine these relationships over two data collection periods. Conclusions: Policies aimed at improving health through physical activity promotion have the potential to reduce total health expenditures while improving health.