How Much Accelerometry Data is Required to Reliably Estimate Preschool Children's Habitual Physical Activity?

Introduction There is no consistency in the literature as to the amount of accelerometry monitoring time required to reliably determine habitual physical activity behaviour in preschool children. This study aims to identify how many days, and what percentage of time each day, of monitoring is required.

Methods Data from ActiGraph GT1M accelerometers were collected over an eight-day period from 284 preschool children (aged 3-5 years) who participated in the Melbourne based Healthy Active Preschool Years (HAPPY) Study. After downloading, data were managed by a specially developed series of macros in Excel and syntax in Stata 11 which applied cutpoints [1], and identified wear time and time of day/day of the week. The primary outcome measure was the percentage of time spent in light-moderate vigorous physical activity during the child’s usual waking hours (parent reported). SPSS 17.0 was used to assess individual day reliability using ICC and repeated measures ANOVA against the criterion measure of 80% of possible wear time (equivalent to just over 10 hours). The Spearman-Brown Prophecy Formula [2] was used to determine the number of days required at each of the given ICC values to meet reliability of 0.8. Analyses were undertaken for wear time of 40%, 50%, 60% and 70% of usual waking hours (possible wear time).

Results When the criterion was set to 50% of possible wear time, ICCs for any individual weekday ranged from 0.62 to 0.69, and number of days to achieve 0.8 reliability ranged from 1.8 to 2.5. For weekend days, the ICC for an individual day was 0.74 to 0.86, and 0.7 (Saturday) to 1.4 (Sunday) days were required to achieve a reliability of 0.8.

Discussion and Conclusion A minimum of 50% of preschool children’s usual waking time on a minimum of three weekdays is required to accurately represent habitual weekday physical activity. Variability in weekend day data was higher and ideally data for 50% of the child’s usual waking hours on both weekend days would be included to ensure physical activity is reliably estimated. Future studies investigating physical activity in preschool children should include the minimum amount of accelerometry data identified here to ensure the findings are reliable.

References [1] Sirard, JR, Trost, SG, Pfeiffer, KA, Dowda, M, and Pate, RR, Calibration and Evaluation of an Objective Measure of Physical Activity in Preschool Children. JPAH, 2005; 2(3):345. [2] Stanley, JC, Reliability, in Educational Measurement, R.L. Thorndike, Editor. 1971, American Council of Education: Washington, DC. p. 395.