Body Mass Index and Accelerometer Measurement Issues For Use in the Evaluation of Pedometer-based Physical Interventions in Children

Participation in physical activity (PA) of at least moderate intensity may yield important
health benefits for children. A popular behavioural tool used to promote increased PA
is the pedometer. There is however limited evidence regarding pedometer-based
strategies in children. This thesis reports on a series of anthropometric and
accelerometer-measurement issue studies which inform the methods used to address
the primary aim of this thesis- to determine the effectiveness of goal-setting, selfmonitoring
and step-feedback pedometer-based interventions for increasing PA in 10-
11-year-old children. In addition, each study in their own right provides an original
contribution to knowledge within their specific area of scholarship. The first objective
of this thesis was therefore to determine diurnal variation of height and weight and
the combined effect upon body mass index (BMI) weight status in children via a field
based study. Next, the reliability of the Actiwatch 4 (AW4) accelerometer was tested in
a mechanical laboratory experiment. Following this laboratory trial a second field
based study examined the impact of placement site upon AW4 output, and the validity
of a regression equation to predict hip-derived AW4 data from wrist-derived data.
Finally, a brief intervention mapping approach was used to develop goal-setting, selfmonitoring
and step-feedback pedometer-based interventions, the effectiveness of
which was evaluated in a small scale controlled trial involving two primary schools. The
main findings of this thesis were a) that diurnal variation in height (and in girls alone,
weight) impact upon increased BMI and BMI percentile in afternoon versus morning
measurements b) AW4 activity counts exhibit acceptable reliability statistics
(comparable to other accelerometer models), which improve when raw activity counts
are reduced into derived activity intensity variables c) wrist and hip derived AW4 data
are not comparable, and the derived regression equation may not be suitable for
group level prediction due to inaccurate individual level prediction and the large
standard error of the estimate observed d) pilot testing pedometer wear and
intervention materials may highlight practical pedometer issues (i.e. pedometer
attachment, wearing compliance and acceptability of instruction sheets) that inform intervention implementation and e) pedometer-based goal-setting, self-monitoring
and step-feedback interventions did not increase PA in 10-11-year-old children.
However, individual-standardised goal setting may be more promising as this appeared
to mitigate any decline in moderate-to-vigorous physical activity (MVPA) in moreactive
children, and increased MVPA in less-active. To summarise, the findings of this
thesis highlight important issues for physical activity scientists to consider when using
BMI-determined weight status as a grouping variable and accelerometers as an
outcome measure, when evaluating physical activity interventions in children. With
regard to the primary aim of this thesis, future researchers should further examine the
effectiveness of the individual-standardised against the group-standardised goal type
in a longer-duration intervention and using a larger sample of children, which may
permit sub-group analyses to be conducted. Of primary importance is future
clarification on the effectiveness of goal setting, self-monitoring and step-feedback
pedometer-based interventions per se for changing PA in children.