Regional variation in lifestyle patterns and BMI in young children: the GECKO Drenthe cohort

Abstract

BackgroundA better understanding of lifestyle behaviours of children < 7 years and the relation with childhood overweight is needed. The aim of our prospective study was to examine how lifestyle patterns in young children are associated with the development of childhood overweight. As ecological models suggest focusing on not only the child as an individual, but also their environment, we also considered the role of socio-economic status (SES) and spatial clustering of lifestyle and body mass index (BMI).MethodsIn 1792 children (aged 3–6 years) participating in the GECKO Drenthe cohort, diet, screen time, outdoor play and sleep were assessed by questionnaires and moderate-to-vigorous physical activity and sedentary time by accelerometry (Actigraph GT3X). At 10–11 years, height and weight were measured to calculate age- and sex-specific standardized BMI z-scores (zBMI). Lifestyle patterns were identified using principal component analysis. To assess spatial clustering for the lifestyle patterns and zBMI, we calculated the Global Moran’s I statistic. Linear- and logistic regression models, taking into account SES, were performed to examine the association between the lifestyle patterns and the development of overweight. For the spatial analyses, we added spatial terms for the determinants, the outcome, and the error term.ResultsThree lifestyle patterns were identified: (1) ‘high activity’, (2) ‘low screen time, high sleep and healthy diet’, and (3) ‘high outdoor play’. No associations were observed between the ‘high activity’ or ‘high outdoor play’ patterns at young age with the development of childhood overweight (all p > 0.05). In contrast, children who adhered to the ‘low screen time, high sleep and healthy diet’ pattern had lower odds to become overweight and a lower zBMI at 10–11 years (odds ratio [95% CI] = 0.766 [0.65; 0.90]). These findings remained similar after taking SES into account. Regarding the spatial analyses, we found spatial clustering of zBMI, but no spatial clustering of the lifestyle patterns.ConclusionsLow screen time, high sleep duration and a healthy diet cluster into a pattern that seems favourable in the prevention of childhood overweight, independent of individual SES. The spatial analyses suggest that there are likely other neighbourhood factors that contribute to the spatial clustering of childhood overweight.