The measurement, levels, and correlates of physical activity in a bi-ethnic population of young children
thesisposted on 19.06.2017, 15:38 authored by Daniel D. Bingham
Background: The first five years of life, called the early years is a period of rapid and vital physical, behavioural, emotional and social development. Physical activity (PA) is one of the behaviours which develop during the early years, and high levels of PA during the early years have been shown to be related to multiple health outcomes. The importance of PA of young children (children during the early years) has been highlighted by a number of national governments. In order to better inform future interventions and public health policies, a greater understanding of the correlates and determinants of young children s PA is vital. Previous research has been limited through measurement inconsistencies, and few studies have been conducted within multi-ethnic communities, where many young children in the United Kingdom are born. Thesis Aims: 1) to systematically review published research in order to establish currently known correlates and determinants of PA in the early years and identify gaps within the literature. 2) Calculate an accelerometer wear-time criteria to reliably measure young children s habitual PA. 3) Investigate the validity and test re-test reliability of a new parental proxy reported PA questionnaire. 4) Investigate the levels and correlates of moderate-to-vigorous PA (MVPA) in toddlers (1-3 year olds) from a predominately bi-ethnic and bi-linguistic population. Methods: Aim 1) nine electronic databases were searched to identify previous research which investigated associations between an exposure/variable, and a quantitative measure of PA. Correlates/determinants of total PA (TPA), MVPA and light-intensity PA (LPA) were reported using an ecologic model. Aim 2) to calculate a wear-time criteria for young children a simple stepped process was used whereby statistical tests were run to determine the minimum length of wear for one day, if there were any differences between weekdays and weekend days, and the presence of reactivity. Intra-class correlation models and the Spearman-Brown prophecy formula were used to calculate wear-time reliability. Aim 3) this study was undertaken by 196 parents completing a proxy report questionnaire on their child s PA and sedentary behaviour after seven days of their young child (mean age 3.2 ,SD: 0.8 years) wearing an accelerometer. A total of 156 (79.6%) questionnaires were completed in English and 40 (20.4%) were completed in transliterated Urdu. Of the 196 parents, 109 parents completed the EY-PAQ a second time seven days apart from the first EY-PAQ completion; this was to assess test re-test reliability of MVPA and sedentary time. Validity analysis used all data and data falling with specific proportion boundaries for MVPA (2%-41%) and sedentary time (30%-94%). Reliability was assessed using intra-class correlations (ICC) and validity by Bland Altman plots and rank correlation coefficients. Aim 4) this study was undertaken by conducting a cross-sectional analysis using 24 month olds and their mothers data collected as part of the Born in Bradford (BiB) birth cohort sub-sample study called BiB1000. The outcome variable was daily minutes of MVPA measured by the EY-PAQ. Numerous independent variables covering the layers of an ecological model were selected. Univariate linear regression models accounting for sex, age, language and season were conducted to examine the differences between White British and South Asian children s daily minutes of MVPA and each of the EY-PAQ s domains, and the proportion of time spent in MVPA within each of the EY-PAQ s domains. A series of univariate linear regression analyses were performed to examine and identify correlates of MVPA (for the whole sample, and separately for White British and South Asian children). Significant variables found in univariate analyses were then included in hierarchical multivariable regression models (based upon the ecological model), in order to examine the percentage of variance accounted for in daily minutes of MVPA. Results: Aim 1) The systematic review identified a large volume of published research. All studies took place in high income countries and few studies (6%) were of high quality. A small number of correlates and determinants of TPA were identified. The only correlate of MVPA was sex and no determinants of MVPA or LPA were found. PA correlates/ determinants were relatively consistent between objective and subjective PA measures and few studies investigated correlates of toddlers or between children with White and South Asian ethnicity. Aim 2) No differences in accelerometer-determined time in TPA, MVPA or sedentary time were observed between weekdays and weekend days within this sample of young children. Similarly, there was no evidence of reactivity to accelerometer use. For young children living in Bradford, an accelerometer wear-time of a minimum of six hours on any three days was shown to provide reliable estimates of accelerometer-determined time in TPA, MVPA, and sedentary time. Aim 3) The test re-test reliability of the EY-PAQ was moderate for sedentary time and fair for MVPA. The EY-PAQ had poor agreement with accelerometry with both sedentary time and MVPA before the application of boundaries. Post application of boundaries the EY-PAQ still had poor agreement with accelerometer-determined sedentary time but good agreement for MVPA. Limits of agreement were wide for all variables and language and ethnicity did not confound results. Aim 4) Bradford toddlers were found to be very active and no difference was observed between proxy-reported time spent in MVPA between White British and South Asian children. However, White British toddlers were found to have spent significantly more time in reported MVPA while walking for transport compared to South Asian toddlers; while South Asian toddlers reportedly spent significantly more time of MVPA in the home compared to White British toddlers. Correlate models were only statistically significant when multi-layers of the ecological model were included; and correlates differed for South Asian and White British children. Conclusions: There is a need for more high-quality studies exploring correlates/determinants across all layers of the ecologic model, and research investigating MVPA correlates/determinants of toddlers and between ethnicities is sparse. The work reported within this thesis has produced a reliable wear-time criterion for use to estimate accelerometer-determined PA and sedentary time in young children living in a bi-ethnic community. This criterion can now be used in future accelerometer studies (validation, observational and intervention) and the stepped-process offers researchers a method to derive sample-specific wear time criteria. The EY-PAQ is a promising habitual population-level measure of young children s MVPA from a bi-ethnic community. In situations when objective methods are not possible for measurement of young children s MVPA, the EY-PAQ may be a suitable alternative. Levels of toddlers MVPA did not differ by ethnicity but the contexts and correlates did. Therefore, future interventions should seek to maintain and maximise high levels of toddlers MVPA and tailor interventions by ethnicity. The research conducted within this thesis will inform the development of surveillance systems, interventions and public health polices to improve young children s PA levels, particularly children living in a bi-ethnic community.
Bradford Institute for Health Research.
- Sport, Exercise and Health Sciences