Differences in the estimation of time spent in sedentary behaviour, light and moderate-to-vigorous physical activity in toddlers, according to different epoch durations

Publication Summary

PURPOSE: To investigate time spent in sedentary behaviour (SB), light and moderate‐to‐vigorous physical activity (MVPA) in 2‐3 year old South Asian (SA) and White British (WB) children and both parents, and the relationship between parental and offspring activity behaviours. METHOD: Eighteen toddlers (2.86±0.60 years; 10 girls) were video‐recorded during Semi-structured activity sessions. Videos were coded second-by‐second according to the Children’s Activity Rating Scale (CARS) direct observation system, resulting activity scores were averaged for each epoch duration, and time periods with complete matching 5-, 10‐ and 15-second epochs were selected for analyses. Bland‐Altman plots and paired t‐tests were used to test differences in estimated SB, light and MVPA time (according to CARS) between each epoch duration. Bonferroni corrections were applied for multiple comparisons. RESULT(S): In total, 12,915 seconds were included in the analyses. In relation to 5-second epochs, 10- and 15-second epochs significantly underestimated both time spent in SB (mean difference: ‐29.44 and ‐49.17 seconds respectively; p<0.0001) and time spent in MVPA (mean difference: ‐20 and ‐29.17 seconds respectively; p<0.0001). Time spent in light PA was significantly overestimated by 10‐ and 15‐second epochs (mean difference: +49.17 and +78.33 seconds respectively; p<0.0001) in relation to 5‐second epochs. CONCLUSION(S): Considering the differences found, accelerometry cut‐points derived for longer epochs will tend to overestimate light PA, and underestimate time spent in SB and MVPA in 2-3 year olds. Future research with such young children should use epochs ≤5 seconds to get a more accurate representation of their activity behaviours