Resumen de Process and Treatment of Pedometer Data Collection for Youth: The Canadian Physical Activity Levels among Youth Study

Cora L. Craig, Catrine Tudor-Locke, Sue Cragg, Christine Cameron

• Background: Pedometry methods for collecting data in young populations are advancing, but it is unclear how many days of data are enough for population monitoring.

  Methods: Using random-digit dialing, 11,669 5- to 19-yr-olds were recruited into the Canadian Physical Activity Levels among Youth study and mailed a data collection package. Pedometers were worn for 7 d, and steps counts were logged daily. Reactivity was assessed by examining estimates from the pattern of pedometer data across days (arranged from first day of collection to last) using a repeated-measures ANOVA. Intraclass correlations (ICC) were computed for the first day and consecutive additional days (compared with the criterion estimate based on the whole week) to determine the minimal number of days required to achieve a reliability ICC of 0.70, 0.80, and 0.90.

  Results: Most children (>90%) wore the pedometer for 7 d. Mean steps per day differed across consecutive days (F = 52.7, P = 0.000); however, no difference occurred between the first and the second day of monitoring. Furthermore, no difference was observed between the first and either the third or the fourth day when monitoring commenced on a Monday or a Tuesday. Therefore, there was no clear evidence of reactivity. The first day provided a good representation of steps per day relative to the whole week in terms of both reliability (ICC = 0.79) and validity (relative absolute percent error [APE] = 2.5%), and these improved with additional days (2 d, ICC >0.85; >=3 d, ICC >0.90; and >=3 d, APE <1%).

  Conclusions: The Canadian Physical Activity Levels among Youth demonstrates the feasibility of national surveillance of physical activity using pedometers. Two days are sufficient to determine steps per day, and a single day appears defensible in terms of population monitoring if minimal standards for reliability are acceptable