Skip Navigation

Global Obesity Prevention Center

Obesity Society Winning Poster

oddoDoctoral student Vanessa M. Oddo wins First Place, Epidemiology Section

"Gains in income during early childhood are associated with decreases in body mass index z-scores among children in the United States"

Faculty Mentor: Jessica C. Jones-­Smith



Background: Children with low family income in the U.S. are disproportionately burdened by overweight and obesity compared to those with high family income. However, few studies have leveraged longitudinal data to investigate the impact of changes in family income on changes in children’s body mass index (BMI).

Methods: We used longitudinal data from the nationally representative Early Childhood  Longitudinal  Survey  Birth  Cohort  to  assess  whether  gains  in  family income were associated with changes in BMI z-­‐score among 2-­‐6 year olds. Child anthropometrics and family income were assessed at 2-­‐year, 4-­‐year, 5-­‐year and 6-­‐ year visits. Gender-­‐stratified, fixed effects linear regression models compared children to themselves over time in order to control for time-­‐invariant  measured and unmeasured confounding factors. Models additionally controlled for time-­‐ varying confounders including number of siblings, household structure (two parent, one parent, unrelated guardian), age, and age squared.

Results: Children  (n=4,979)  had  an  average  BMI  z-­‐score  (standard  error)  of  0.42 (0.26) at 2-­‐years and an average change of +0.24 (0.026) over the study period. On average, family income increased by approximately $11,000 ($496). The association between gains in family income and change in BMI z-­‐score varied by gender, but not by race/ethnicity. Among girls, each additional $10,000 gained was associated with a  decrease in BMI z-­‐score  (β=-­‐0.018;  95% CI: -­‐0.033,  -­‐0.004).  Among boys, this association was not significant (β=-­‐0.003; 95% CI: -­‐0.017, 0.01).  

Conclusion: By comparing  children  to  themselves  over  time,  we  overcome  many barriers that typically impede causal inference in observational studies. In this way, our study provides stronger evidence that gains in income during early childhood may promote healthy weight outcomes among girls.  

To see the award winning poster, click here.