Disentangling the Combined Effects of 24-Hour Time-Use Behaviours on Childhood Obesity: A Compositional Data Analysis

Hedayatrad, Leila
Duncan, Scott
Stewart, Tom
Item type
Degree name
Doctor of Philosophy
Journal Title
Journal ISSN
Volume Title
Auckland University of Technology

Obesity in children is a worldwide health problem, and New Zealand is no exception. One in three New Zealand children is overweight or obese. Thus, identifying modifiable determinants of obesity in children is essential to inform future interventions. Evidence shows that time-use behaviours, including physical activity, sedentary behaviour, and sleep, are related to obesity in children. However, the associations between these individual time-use behaviours and obesity shown in past research have been criticised for failing to appropriately adjust for time spent in each of the other behaviours. The emerging research field of time-use epidemiology suggests that interactions among these time-use behaviours may impact health in ways that cannot be explained by studying these behaviours in isolation (e.g., increasing one behaviour displaces another). This has prompted a global shift in behavioural epidemiology research where an integrated approach focusing on complete (24-hour) days is now a research priority. The overall aim of this PhD thesis was to advance the area of time-use research in children through four studies, guided by the Viable Integrative Research in Time-Use Epidemiology (VIRTUE) framework. Study 1 investigated the concurrent validity of two accelerometers (i.e., ActiGraph GT3X+ and Axivity AX3) for measuring children’s time-use behaviours against direct observation. Both accelerometers reached 65% to 97% balanced accuracy for detecting various postures and physical activity intensities, with the AX3 offering slightly better accuracy than the GT3X+ accelerometer. These findings showed that the AX3 device could effectively measure activity type and intensity in child populations. Studies 2¬–4 utilised data from a sample of children who participated in the 8-year wave of the Growing Up in New Zealand cohort study. Study 2 investigated if the 24-hour AX3-measured time-use behaviours (measured from activity intensity and activity type perspectives) and reallocation of time across these behaviours were associated with obesity-related outcomes using compositional data analysis. The 24-hour time-use composition was significantly associated with body mass index (BMI). More time spent in light-intensity physical activity (LPA) and walking (relative to the other behaviours) was associated with lower BMI. Study 3 examined how these children clustered based on their lifestyle behaviours (including time-use behaviours and diet) and associations between cluster membership and obesity. Three unique clusters were identified, with children in the healthiest cluster (lowest sedentary time and sitting time, healthiest diet) had the lowest BMI compared to other clusters. Study 4 examined which sociodemographic factors were associated with adherence to the New Zealand 24-hour Movement Guidelines. Only a small number of children met these guidelines, and child gender, ethnicity, mother’s education, and household area (urban vs rural) were associated with guideline adherences. Overall, the studies in this thesis have made significant contributions to time-use research by providing insight into the structure of time-use behaviours and their relationship with measures of obesity in New Zealand children. It is hoped that these findings will assist in developing and tailoring future interventions to improve child health.

Publisher's version
Rights statement