Comparison of Actical and activPAL measures of sedentary behaviour in preschool children

Abstract

Objectives: The purpose of this study was to examine the convergent validity of the Actical and activPAL to measure sedentary behaviour (SB) and non-SB in preschoolers in a free-living environment.

Design: A convenience sample of 49 preschoolers (22 boys; 4.0 ± 0.5 years) from six early childhood centres in Auckland, New Zealand were included in data analysis.

Methods: Participants wore a hip-mounted Actical and a thigh-mounted activPAL accelerometer simultaneously during centre attendance for one day and data were collected in 15s epochs. Bland-Altman tests were used to assess differences in group mean minutes and percentage of time in (non-)SB between both monitors. Agreement between binary coded (SB vs. non-SB) 15s-by-15s Actical and activPAL data was evaluated by calculating percentage agreement and κ statistic.

Results: The monitors were worn on average for 294.8 ± 46.3 min resulting in a total of 57,780 15s epochs. Bland-Altman tests suggested a small group mean difference in (non-)SB (1.3 min; 0.1%) and a wide prediction interval (121.3 min; 39.2%). No obvious systematic bias was observed in the Bland-Altman plot. Percentage agreement between the 15s-by-15s Actical and activPAL data of all participants was 73.0% (inter-child range: 36.8-93.8%). The κ statistic showed moderate agreement with a value of 0.46 (95% CI: 0.45-0.47).

Conclusions: Although the group mean estimate of (non-)SB was similar between the Actical and activPAL, the output of both monitors cannot be considered convergent as meaningful random disagreement was found between both monitors.