Purpose: Patterns of physical activity (PA) and sedentary time, including volume of bouted activity, are important health indicators. However, the effect of accelerometer epoch length on measurement of these patterns and associations with health outcomes in children remain unknown.
Method: We measured activity patterns in 308 children (52% girls, age range = 8-11 years) using ActiGraph GT1M accelerometers with 15-s epochs and reintegrated to 60-s epochs. We calculated the volume (minutes per day) of moderate-to-vigorous PA (MVPA), sedentary time, light, moderate, and vigorous PA, as well as bouted MVPA and sedentary time (0-5 min, 5-10 min, 10-20 min, and > 20 min).
Results: The difference between 15-s and 60-s epochs was statistically significant for all outcomes; however, effect sizes were small or negligible in 30% of comparisons. Bias ranged from 1.9 min/day (total MVPA) to 102.7 min/day (0-5 min sedentary bouts). Regression-based estimates of bias and 95% limits of agreement illustrated that the magnitude, and in some cases, the direction, of between-epoch differences varied with activity level. Correlations with body mass index and cardiovascular fitness were similar for 15-s (r = -.19 to .20) and 60-s (r = -.16 to .29) epochs. Estimated 15-s data (predicted from 60-s) were similar to measured data and had similar relationships with health outcomes.
Conclusion: Epoch length influences measurement of PA and sedentary patterns and the effect is modified by activity level. However, associations with health outcomes were similar and epoch differences can be adjusted. Future research should clarify the accuracy of different epoch lengths for measuring bouted activity and evaluate whether epoch length alters relationships with additional health outcomes.
Keywords: Accelerometry; bouts; measurement.