Stability reliability of habitual physical activity with the Actical® activity monitor in a youth population

Abstract

The purpose of this study was to examine and compare stability reliability of hip-and wrist-worn activity monitor (AM) physical activity (PA) outcome variables for adolescents (12-17 years) under free-living conditions. Twenty-one adolescents (11 girls, 10 boys) wore AMs at hip and wrist locations for a 14 consecutive day monitoring period. The dependent variables were total time (minutes/day) occurring at sedentary (TS), light (T L) and > or = moderate (T MV) intensity PA, and the sum of activity energy expenditure (AEE; kcals/day) occurring at a light (AEE L), > or = moderate intensity (AEE MV) and from total AEE for a monitoring day (AEE T). Intraclass correlation coefficients for each monitoring day (ICC DAY) were calculated to examine reliability of all dependent variables for both AM locations, over bout durations of 1-, 5-, and 10-minutes. Excluding T L measured by the wrist AM location, ICC DAY values for all variables decreased as bout duration increased from 1- to 10-minutes. The number of monitoring days to reach an ICC DAY > or = 0.80 for AEE L, AEE MV, AEE T, T L and T MV across all three bout durations was higher for hip- versus wrist-worn AM, and ranged from 2-26 versus 2-10 days, respectively. Conversely, then number of monitoring days to reliably measure TS was lower for hip AM (4-5 days) versus the wrist AM (6-8 days) locations. Among all the variables, T L had the largest range of necessary monitoring days to reach an ICC DAY > or = 0.80, 7-10 and 6-26 days across all bout durations for the wrist and hip, respectively. Further comparison of the dependent variables by day type revealed that only T L measured at the wrist had significantly different means on weekdays versus weekends (pvalue < or = 0.001) across all bout durations. While not significant, mean values for T MV, AEE MV, and AEE T measured by both hip- and wrist-worn AMs, were higher on weekdays than on weekends. In summary, reliability of time and AEE PA variables was dependent on bout duration and AM location. Generally, to reliability measure free-living PA in adolescents wrist-worn AMs required fewer monitoring days than hip-worn AMs.