Assessment of physical activity in children and adolescents

Abstract

The objectives of The Identification and Prevention of Dietary and Lifestyle Induced Health Effects in Children and Infants study (child sample) and the Kenya Adolescent Physical Activity study (adolescent sample) conducted in this thesis was firstly, to assess the impact of methodological and practical decisions such as the appropriate epoch length and cutpoints to use in accelerometry studies involving children and adolescents across the physical activity continuum and the reliability of these accelerometer outcomes in predicting habitual physical activity. Secondly, the performance of uniaxial ActiTrainer accelerometry with heart rate (HR) monitoring was compared to triaxial GT3X accelerometry against indirect calorimetry during structured activities in the predominantly active Kenyan adolescent cohort. Similarly, the performance of uniaxial ActiTrainer accelerometry with HR monitoring vs. triaxial 3DNX accelerometry was compared against DLW under free living conditions in both children and adolescent cohorts. Finally, the validated uniaxial ActiTrainer was used to assess the impact of physical activity and the environment on energy expenditure and indices of adiposity in the two cohorts. The main findings of the thesis were: (a) that 15 s epoch reports significantly higher engagement in physical activity compared to a 60 s epoch in both the children and adolescents cohorts (b) choice of cutpoints significantly affected classification of physical activity and sedentary behaviour in both cohorts (c) a minimum of 6 h for 7 - 9 days in the cohort involving children and a minimum of 6 h for 4 - 5 days in the adolescents were required to reliably measure physical activity (d) triaxial accelerometry reported better predictive validity compared to uniaxial accelerometry during structured activities. In addition, HR monitoring did not improve the predictive validity of either accelerometer during structured activities (e) during free living activities, however, uniaxial and triaxial accelerometry reported comparable predictive validity. The addition of HR monitoring improved the predictive validity of uniaxial accelerometry by approximately 4% in both cohorts (f) in both cohorts, total volume of physical activity and patterns (time engaged in light and moderate to vigorous physical activity) were significantly associated with energy expenditure (g) physical activity and sedentary behaviour were significantly related to indices of adiposity in both cohorts (h) the environment was a significant predictor of physical activity and indices of adiposity in the adolescents but not children. The findings of this thesis have important implications on strategies to standardise accelerometry field protocols and future studies on the validation of accelerometers and the association between physical activity, the environment and health.