Physical activity and obesity in children: measurement, associations, and recommendations Research Completed


Physical activity and obesity in children: measurement, associations, and recommendations

Lead Author

Scott Duncan


Auckland University of Technology

Publication Year



Auckland University of Technology


Primary Supervisor:

Grant Schofield

Phone: +64 9 921 9169






Widespread increases in the prevalence of childhood obesity have raised the prospect of serious public health consequences in many countries. New Zealand is no exception; according to the most recent national estimates, approximately one in three children is overweight or obese. As a consequence, an understanding of the specific risk factors that predict this condition in children is becoming increasingly important. It is generally accepted that the promotion of physical activity is a key strategy for reducing the risk of childhood obesity. However, there is limited information describing physical activity and its relationship with body fatness in young New Zealanders. The overall aim of this thesis was to gain insight into the associations between excess fatness and physical activity in New Zealand children from a diverse range of socio-demographic groups. Three related studies were conducted to achieve this aim: a large descriptive survey of obesity and physical activity patterns in primary-aged children, and two preceding studies which develop the methodology for objective assessment of physical activity in this population. The first study provided the only validation data for the NL-2000 multiday memory (MDM) pedometer in children. In a sample of 85 participants aged 5-7 and 9-11 years, the NL-2000 offered similar accuracy and better precision than the widely used SW-200 pedometer (NL-2000: mean bias = -8.5 ± 13.3%; SW-200: mean bias = -8.6 ± 14.7%). The second study investigated reactivity to wearing pedometers over four 24-hour testing periods in 62 children aged 5-11 years. The sample was divided into two groups: one was given a full explanation of the function of the pedometer, while the other received no information prior to testing. The absence of significant differences in step counts between the first and last test periods indicated that there was no evidence of reactivity to this device for either preparation procedure. The central study presented in this thesis was the measurement of physical activity, body composition, and dietary patterns in 1,226 children aged 5-12 years, from which four chapters (4-7) were derived. The sample was ethnically diverse, with 46.8% European, 33.1% Polynesian, 15.9% Asian, and 4.1% from other ethnicities. Physical activity levels over three weekdays and two weekend days were assessed using NL 2000 pedometers. Percentage body fat (%BF) was determined using hand-to-foot bioelectrical impedance analysis with a prediction equation previously developed for New Zealand children. Waist and hip girths, height, and weight were measured using standard anthropometric techniques. Parent proxy questionnaires were used to assess demographic and lifestyle factors and pedometer compliance. The first reported analyses of this dataset (Chapter 4) examined the effect of weather conditions on children’s activity levels. In boys, a 10ºC rise in ambient temperature was associated with a 10.5% increase in weekday steps and a 26.4% increase in weekend steps. Equivalent temperature changes affected girls’ step counts on weekdays only (16.2% increase). Precipitation also had a substantial impact, with decreases in weekday and weekend step counts during moderate rainfall ranging from 8.3% to 16.3% across all sex, age, and socioeconomic (SES) groups. The aim of Chapter 5 was to understand the relationship between children’s step counts and their body mass index (BMI), waist circumference (WC), and %BF. Mean step counts for this sample were 16,133 ± 3,864 (boys) and 14,124 ± 3,286 (girls) on weekdays, and 12,702 ± 5,048 (boys) and 11,158 ± 4,309 (girls) on weekends. Significant associations were detected between and both WC and %BF, but not between and BMI. The findings in Chapter 6 extended these results by estimating the number of steps required to reduce the risk of excess adiposity in children (16,000 and 13,000 for boys and girls, respectively). Finally, the study described in Chapter 7 examined the associations between excess adiposity and a series of demographic and lifestyle variables, providing the first assessment of body fat correlates in young New Zealanders. Our results indicated that children aged 11-12 years were 15.4 times more likely to be overfat (boys, %BF ≥ 25%; girls, %BF ≥ 30%) than those aged 5-6 years. In addition, the odds of overfat were 1.8 times greater in Asian children than in European children, and 2.7 times greater in the low SES group when compared with the high SES group. Three modifiable behaviours related to fat status were also identified: low physical activity, skipping breakfast, and insufficient sleep on weekdays. Clustering of these risk factors resulted in a cumulative increase in the prevalence of overfat

Keywords:Public health; Body composition; Pedometers; Bioelectrical impedance analysis; Epidemiology; Step counts


How to access

Areas of Focus

Population Groups

Settings (location)

Provision (delivery type & infrastructure)





June 18, 2014