Is There a Link Between Food Insecurity and Overweight Status in Children? Evidence from the Early Childhood Longitudinal Survey
Research Center: Department of Nutrition at the University of California, Davis
Investigator: Rose, Donald, and Nick Bodor
Institution: Tulane University
Department of Community Health Sciences
Tulane University School of Public Health and Tropical Medicine
1440 Canal Street, Suite 2301
New Orleans, LA 70112
Recent literature indicates that close to two-thirds of the U.S. adult population is
overweight or obese. Moreover, the prevalence of obesity has increased in the
United States, particularly among children. This research identified literature
linking obesity with increased risk of poor health outcomes, such as cardiovascular
disease, diabetes, and cancer.
While there is little controversy about the proximal determinants of overweight
status—imbalance between energy intake and expenditure—there is considerable
complexity in the framework of distal factors that give rise to this imbalance.
An emerging area of research exploring distal factors is concerned with
the relationship between household food insecurity and obesity.
Several empirical studies cited in the report have explored the food insecurityweight
status relationship. There is apparent agreement among studies of adult
women: most studies show a positive association between food insecurity and
the probability of being overweight. While a positive relationship may seem
paradoxical, several explanations are possible. Food insecurity could lead to
overweight status if individuals overcompensate for periods when food is
scarce, resulting in greater overall intake. Weight cycling could also make the
body more efficient in utilizing dietary energy, leading to weight gain over time.
Finally, energy-dense foods are often less expensive, so that food-insecure
households who cannot afford to eat balanced meals or who must rely on a few
kinds of low-cost foods may have an overall greater energy intake.
No clear pattern has emerged regarding the food insecurity-overweight link in
children. Some authors have suggested that the issue in children is unresolved
because of limitations (including sample size) of previous datasets. Literature
cited in the report has found that food insecurity or hunger is associated with
negative academic outcomes and poor psychosocial functioning at school,
adverse health outcomes, and poor mental health. The objective of this research
is to test the hypothesis that household food insecurity is positively associated
with overweight status in children.
The study analyzes data collected in the Department of Education’s Early
Childhood Longitudinal Survey-Kindergarten Cohort (ECLS-K). The ECLS-K
is a large nationally representative survey of children which began with the
kindergarten class of 1998-99. The survey collected measures of children’s
heights and weights twice per year in the kindergarten and first grade, the full
18-item USDA food insecurity module in the spring of 1999, and a rich set of
variables on the home and school environments of these children. Algorithms
from the Centers for Disease Control and Prevention (CDC) were used to assign
Body Mass Index (BMI)-for-age percentiles to each child’s measurements.
Children with a BMI greater than or equal to the 95th percentile of their sexspecific
BMI-for-age chart were considered overweight. In addition to this indicator,
the study calculated a dichotomous variable indicating “risk of
overweight,” a CDC term for children with a BMI greater than or equal to the
85th percentile of their BMI-for-age chart.
Weight status is affected by a number of biological and socio-economic factors.
To control for potentially confounding variables, the research developed multivariate
logistic regression models in which the dependent variable was a
dichotomous indicator of overweight status. Independent variables included a
measure of household food insecurity and a full set of control variables,
including: age, sex, and birth weight of the children; maternal educational
attainment; income, region, and urbanization of the household; as well as family
meal patterns and child activity patterns. All analyses used ECLS-K weighting
variables, and accounted for the clustered nature of the sample by using jackknife
replicate methods to estimate standard errors.
The primary study finding is that household food insecurity, when modeled with
appropriate controls, is not associated with a higher prevalence of overweight
among young school children. If anything, household food insecurity seems
inversely associated with weight status. The finding is relatively robust, since
similar results were demonstrated across a range of different models. The study
applied dichotomous (food secure/insecure) and trichotomous (food secure/insecure
without hunger/insecure with hunger) expressions of the household food
security variable and also used dichotomous and trichotomous expressions for
child food insecurity. Models using different expressions of the dependent variable
were performed, using “risk of overweight” as an indicator in one model,
and simply BMI in continuous form in another. A cross-section analysis based
on data collected in the spring of the children's kindergarten year was run. It
should be noted that parents reported on household status in the 12 months prior
to the interview, so in effect a food-insecure condition would have preceded the
children's weight status. The research further tested whether household food
insecurity in the spring of 1999 was predictive of overweight status a year later
and found that it was not. Additionally, the study tested whether household food
insecurity in 1999 was predictive of a high weight gain over the next year and
found an inverse association.
These findings may mean that food insecurity is less relevant for those
whose main concern is childhood obesity than for those focused on
academic and psychosocial outcomes or physical and mental health. In addition,
study estimates indicated that while 10 percent of the sample overweight
children came from food insecure households, 24 percent came from
households in poverty. Thus, targeting overweight prevention might be more
focused on a general population of the poor than on the food-insecure population.
Moreover, social marketing techniques or environmental change
strategies that affect large groups of people may cost less than strategies that
must first identify food-insecure households.