Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·

Longitudinal association of neighborhood variables with body mass index in dutch school-age children: The KOALA birth cohort study

Publication files not online:

Author: Schmidt, S.C. · Sleddens, E.F.C. · Vries, S.I. de · Gubbels, J. · Thijs, C.
Source:Social Science and Medicine, 135, 99-108
Identifier: 525595
doi: doi:10.1016/j.socscimed.2015.04.018
Keywords: Health · Birth cohort · Child · Environment · Generalized estimating equations · Neighborhood · Netherlands · Obesity · Overweight · Body mass · Child health · Cohort analysis · Equation · Age distribution · Caloric intake · Controlled study · Energy expenditure · Female · Follow up · Human · Human experiment · Linear regression analysis · Longitudinal study · Male · Maternal behavior · Normal human · Preschool child · Questionnaire · School child · Social behavior · Social capital · Socioeconomics · Healthy for Life · Healthy Living · Life · LS - Life Style · ELSS - Earth, Life and Social Sciences


Changes in the neighborhood environment may explain part of the rapid increase in childhood overweight and obesity during the last decades. To date few theory-driven rather than data-driven studies have explored longitudinal associations between multiple neighborhood characteristics and child body weight development. We aimed to assess the relationship between physical, social and perceived safety related characteristics of the neighborhood and Body Mass Index (BMI) development in children during early school age, using a longitudinal design. We included an examination of moderating and confounding factors based on a conceptual model adapted from the EnRG framework (Environmental Research framework for weight Gain prevention) and empirical research. Analyses included 1887 children from the KOALA Birth Cohort Study followed from baseline age 4-5 years until 8-9 years. For children age 4-5 years, parents completed a questionnaire measuring characteristics of the neighborhood. Reliability and factor analyses were used to identify constructs for neighborhood characteristics. Linear regression analysis was performed to assess the relationship between neighborhood constructs and BMI z-scores cross-sectionally at age 4-5 years and longitudinally using Generalized Estimating Equations with BMI z-scores over 5 repeated measurements until age 8-9 years. Fourteen constructs were identified and grouped in three domains including perceived physical, social, or safety related characteristics of the neighborhood. Cross-sectionally, a lower BMI z-score was associated with higher perceived physical attractiveness of the neighborhood environment (standardized regression coefficient (β)-0.078, 95% CI-0.123 to-0.034) and a higher level of social capital (β-0.142,-0.264 to-0.019). Longitudinally, similar associations were observed with potentially even stronger regression coefficients. This study suggests that BMI in children is mainly related to the modifiable physical and social environment of the conceptual model and not related to safety as perceived by parents. © 2015 Elsevier Ltd.