Gains in income during early childhood are associated with decreases in BMI z scores among children in the United States

Abstract

Evidence suggests that changes in family income are an important determinant of children's body mass index (BMI). However, few studies have leveraged longitudinal data to investigate the association of changes in family income on changes in BMI z score. This study aimed to assess whether gains in family income are associated with changes in BMI z score among children in the United States by using the nationally representative Early Childhood Longitudinal Survey Birth Cohort (ECLS-B). We used longitudinal data from the ECLS-B to assess whether gains in family income, assessed by using the poverty to income ratio (PIR), were associated with changes in BMI z score among children aged 2-6 y. Child anthropometric characteristics and family income were assessed at 2-y, 4-y, 5-y, and 6-y visits. Sex-stratified, individual fixed-effects linear regression models compared children with themselves over time to control for time-invariant measured and unmeasured confounding factors. Models also controlled for time-varying confounders, including number of siblings, household structure (2 parents, one parent, or unrelated guardian), age, and age squared. Children (n = ~9200) had a mean ± SE change in BMI z score of 0.12 ± 0.022, and family income increased by ~$3361 ± $536 during the 4-y period of observation (2003-2007). The association between increased PIR and change in BMI z score varied by sex but not by race-ethnicity. Among girls, an increase in PIR was associated with a statistically significant decrease in BMI z score (βPIR = -0.022; 95% CI: -0.042, -0.0016). There was a statistically significant association between PIR and BMI z score among preterm boys (βPIR + β PIRXpreterm = -0.067; 95% CI: -0.12, -0.018), but the relation was not statistically significant among boys born at term (βPIR = -0.0049; 95% CI: -0.024, 0.014). By comparing children with 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.