Clustering of diet, physical activity and sedentary behaviour among Australian children: cross-sectional and longitudinal associations with overweight and obesity

• Published in: International Journal of Obesity 2015-07, Vol.39 (7), p.1079-1085
• Main Authors: Leech, R M, McNaughton, S A, Timperio, A
• Format: Article
• Language: English
• Subjects: 631/443/319/1488; 692/499; 692/699/2743/393; 692/699/75; 692/700/2814; 692/700/478/174; Accelerometers; Accelerometry; Adolescent; Australia - epidemiology; Body height; Body Mass Index; Body size; Body weight; Care and treatment; Child; Child Behavior; Child Nutritional Physiological Phenomena; Childhood obesity; Children; Cluster Analysis; Clustering; Confidence intervals; Consumers; Cross-Sectional Studies; Diagnosis; Diet; Epidemiology; Exercise; Feeding Behavior; Female; Food consumption; Health aspects; Health Behavior; Health Promotion and Disease Prevention; Humans; Internal Medicine; Male; Medicine; Medicine & Public Health; Metabolic Diseases; Motor Activity; Nutrition research; Obesity; Overweight; Pediatric Obesity - etiology; Pediatric Obesity - prevention & control; pediatric-original-article; Physical activity; Public Health; Regression analysis; Regression models; Sedentary behavior; Sedentary Lifestyle; Socioeconomic Factors; Stability analysis; Statistical analysis; Statistical methods; Surveys and Questionnaires; Tracking; Weight gain
• ISSN: 0307-0565; 1476-5497
• DOI: 10.1038/ijo.2015.66

Background/Objectives: Evidence suggests diet, physical activity (PA) and sedentary behaviour cluster together in children, but research supporting an association with overweight/obesity is equivocal. Furthermore, the stability of clusters over time is unknown. The aim of this study was to examine the clustering of diet, PA and sedentary behaviour in Australian children and cross-sectional and longitudinal associations with overweight/obesity. Stability of obesity-related clusters over 3 years was also examined. Subjects/Methods: Data were drawn from the baseline (T1: 2002/2003) and follow-up waves (T2: 2005/2006) of the Health Eating and Play Study. Parents of Australian children aged 5–6 ( n =87) and 10–12 years ( n =123) completed questionnaires. Children wore accelerometers and height and weight were measured. Obesity-related clusters were determined using K-medians cluster analysis. Multivariate regression models assessed cross-sectional and longitudinal associations between cluster membership, and body mass index (BMI) Z -score and weight status. Kappa statistics assessed cluster stability over time. Results: Three clusters, labelled ‘most healthy’, ‘energy-dense (ED) consumers who watch TV’ and ‘high sedentary behaviour/low moderate-to-vigorous PA’ were identified at baseline and at follow-up. No cross-sectional associations were found between cluster membership, and BMI Z -score or weight status at baseline. Longitudinally, children in the ‘ED consumers who watch TV’ cluster had a higher odds of being overweight/obese at follow-up (odds ratio=2.8; 95% confidence interval: 1.1, 6.9; P