Clustering of Sedentary Behaviours and Physical Activity Among Youth

Abstract

Introduction: Physically inactive lifestyles increase the risk of premature mortality and morbidity. Reducing the time young people spend being inactive is particularly important because epidemiologic data suggest physical inactivity tracks better than physical activity from childhood to early adulthood (Malina, 1996) and small reallocations of sedentary time in favour of more active behaviour can be of clinical significance (Blair et al., 1992). Few studies have examined the many ways young people can be inactive and there is conflicting evidence over the impact of sedentary behaviour on levels of physical activity. The purpose of the present study was to examine the prevalence and interrelationships among sedentary behaviours and physical activity in a cross-national sample of youth ages 11–15. Methods: Participants were 2,501 middle school students from the USA (n = 1757, mean age = 12.9 ± 0.92 yr, 59% female) and UK (n = 744, mean age = 13.0 ± 0.94 yr, 85% female). Using a cross-sectional design, a modified version of the Self-Administered Physical Activity Checklist (SAPAC; Sallis et al., 1996) was used to assess levels of physical activity and sedentary behaviour. Participants were asked to recall the frequency and duration of participation (outside of school) in selected physical (n = 32) and sedentary (n = 7) activities during the past seven days. Cluster analysis was used to identify groups of young people who share similar sedentary and physical activity patterns. Results: A three-cluster solution emerged as valid and reliable for both boys and girls. One cluster contained young people who were physically active and reported more sedentary behaviour than their same-sex peers. For boys in this cluster, sedentary time was dominated by technology-based entertainment (labeled ‘techno-actives'). For girls in this cluster, sedentary time was spent socialising with friends (‘sociable actives'). A second cluster was characterised by low levels of sedentary socialising behaviour but equally high levels of physical activity (‘non-socialising actives'). A final cluster contained young people who reported little or no physical activity and lower levels of sedentary behaviour (‘uninvolved inactives'). Girls in the inactive cluster were more likely to be from the USA and to be older than girls in the active clusters. Chi-square analyses revealed all other clusters to be invariant with regard to age, self-reported body mass index and ethnicity. Discussion: Duration estimates of multiple sedentary behaviours highlight the many ways young people are inactive. Low intercorrelations between sedentary behaviours suggest youth sedentariness is multifaceted and cannot be accurately represented by any one behaviour such as TV viewing. However, because some sedentary behaviours appear inter-related there exists the possibility for shared determinants and the potential for interventions to shift the distribution of multiple inactive behaviours. There were no instances of negative correlations between physical activity and sedentary behaviour which argues against the assumption that physical activity and sedentary behaviour share an inverse and causal relationship. Further study should examine the health-related outcomes associated with sedentary behaviour that may be independent of physical activity and the modifiable determinants of sedentary behaviour among young people. Acknowledgements: Data collection for the USA sample was supported by grant HL54564, National Institutes of Health. Data collection for the UK sample was supported by a grant from NIKE, Inc.