The Relationship Between Depression and Obesity Among Hispanic Adolescents in Three-Generational Households Using Structural Path Model

Schedule:
Thursday, January 15, 2015: 4:50 PM
Preservation Hall Studio 5, Second Floor (New Orleans Marriott)
* noted as presenting author
Hae-Nim Lee, MSW, Doctoral Student, Boston College, Chestnut Hill, MA
Background:Three-generation households have become more prevalent in the United States. In particular, Hispanic youth are more likely to co-reside with a grandparent than native-born youth. Depression and obesity have been recognized as major public health issues in Hispanic adolescents. Given those trends, I expect that incorporated adolescent health behaviors into the one comprehensive model would give us better understanding of the mechanism behind the relationship between depression and obesity in adolescents who belonged to two-parent households with and without a grandparent present.

Therefore, the study aims (a) to examine how adolescent depression and health behaviors are associated with obesity separately for Hispanic adolescents living with and without a grandparent in two parent households, and  (b) to determine the equality of structural paths between adolescents with a grandparent and without a grandparent in two-parent households.

Methods: The data come from the National Longitudinal Study of Adolescent Health in Wave 1 and the analytic sample consists of 2,089 adolescents who belonged to two-parent households with and without a grandparent present. The dependent variable is self-reported BMI scores in wave 1. Then, BMI was categorized into four levels: 1) underweight, 2) normal weight, 3) overweight, and 4) obese. The independent variable is depression measured using the modified version of CES-D scale. The mediators are adolescents’ health behaviors such as vigorous physical activity, sedentary activity, intake of fruits and vegetables, hours of sleep, and smoking.

First, the separate path model was used to evaluate how depression and health behaviors are associated with obesity for each group. Second, the multiple group path model was used to test whether or not the relationship of each health behavior variables with depression is equal between two groups. The LR test compares each model with constrains to the baseline model with no constraints between groups. 

Results: For the separate model for Hispanic youth living with a grandparent, physical activity (b=-0.06, p<0.05), intake of fruits and vegetables (b=-0.02, p<0.05), and hours of sleep (b=-0.04, p<0.05) are significantly associated with depression. However, no significant associations were found between health behaviors and obesity (X2=18.78, df=9, p-value=0.027, CFI=0.65, RMSEA=0.078). For the separate model for Hispanic youth not living with a grandparent, vigorous physical activity (b=-0.03, p<0.05), intake of fruits and vegetables (b=-0.02, p<0.05), hours of sleep (b=-0.03, p<0.05), and smoking (b=0.02, p<0.05) are significantly associated with depression. In addition, results showed that sedentary activity (b=0.02, p<0.05) and smoking (b=0.07, p<0.05) are significantly associated with obesity (X2=14.96, df=6, p-value=0.021, CFI=0.97, RMSEA=0.03). Findings for the multiple group analysis and the LR test showed that the relationship between depression and vigorous physical activity is negatively stronger for Hispanic youth living with a grandparent than those not living with a grandparent in two parent households.

Conclusions: These results suggest that adolescent depression plays a critical role in their health behaviors for both groups. However, smoking and sedentary activity moderate the relationship between depression and obesity only for Hispanic youth not living with a grandparent in two-parent households.