Compliance of adolescent girls to repeated deployments of wrist-worn accelerometers

PURPOSE: To determine the cross-sectional and cumulative compliance of adolescent girls to accelerometer wear at three deployment points and to identify variables associated with compliance. METHODS: Girls from 20 secondary schools were recruited: 10 schools were participating in the 'Girls Active' intervention and 10 were control schools. Physical activity was measured using the GENEActiv accelerometer worn on the non-dominant wrist 24 hours/day for up to 7-days at baseline, 7-months and 14-months. Demographic and anthropometric characteristics were recorded. RESULTS: Seven valid days (≥16 hours) of accelerometer wear were obtained from 83%, 77% and 68% of girls at baseline (n = 1734), 7-months (n = 1381) and 14-months (n = 1326), respectively. 68% provided 7-valid days for both baseline and 7-months, 59% for baseline and 14-months and 52% for all three deployment points. Estimates of physical activity level from 3-days of measurement could be considered equivalent to a 7-day measure (i.e. they fell within a ±5% equivalence zone). Cross-sectionally, 3-valid days were obtained from at least 91% of girls; cumulatively, this was obtained from ≥88% of girls across any two deployment points and 84% of girls across all three deployment points. When controlling for clustering at school level and other potential predictors, physical activity level, being South Asian, being in the intervention group and prior compliance were positively associated with monitor wear. CONCLUSION: Compliance reduced across deployment points, with the reduction increasing as the deployment points got further apart. High prior compliance and high physical activity level were associated with the most additional wear-time.