A large body of literature finds strong links between obesity and health status and health care use and expenditures. Depression is similarly linked to chronic health conditions such as heart disease and diabetes but the causal pathways remain uncertain. A few studies (Goodman and Whitaker, 2002) also find associations between obesity and depression, but is little known about the relationship between obesity, depression, and mental health treatment, especially among adults.

We examine the association of obesity with depression and treatment of depression using nationally representative data for adults aged 18 and older from the 2004-2007 Medical Expenditure Panel Survey. We focus in particular on the question of whether the relationship between obesity and depression and treatment use differ by gender and racial and ethnic groups. Previous studies find differences in obesity-related behaviors across racial and ethnic groups (Harris et al., 1991). Moreover, there is strong evidence of differences in perceptions of mental health treatment need and treatment seeking behavior by gender and racial and ethnic groups (Freiman and Zuvekas, 1998; Zuvekas and Fleishman, 2008).

In logistic regression models, we find a positive association between obesity and a two-item PHQ-2 depression screener scale (Kroenke, Spitzer and Williams 2003) among white females but not among other groups defined by gender and race/ethnicity. However, obesity is not associated with perceived mental health, conditional on PHQ2 depression status. We find obesity is positively associated with use of antidepressants among white females and white males without a current indication of depression (PHQ2), while obesity is negatively associated with antidepressant use among depressed black males. We also find that obesity is positively associated with ambulatory visits for depression among white females, depressed or not. No statistically significant associations between obesity, depression, and depression treatment were observed among Hispanic males and females. Future work will explore the use of instrumental variables and panel data methods (Meyerhoefer and Zuvekas 2008, 2010) to better sort out the causal pathways between obesity and depression and depression treatment.