Test-retest reliability of the ActiGraph® accelerometer in children and adolescents with congenital hemiplegia: How much more activity is needed?

Abstract

Objective: The reliability of accelerometers in individuals with cerebral palsy has not been determined. This study aimed to establish the reliability of the ActiGraph®GT3X+ tri-axial accelerometer in children and adolescents with congenital hemiplegia. Design: Test-retest measurement study. Method: Thirty children (n=16 males; n=13 right hemiplegia) with spastic type congenital hemiplegia aged between 8 and 17 years (mean 11yr 11mo, SD 2yr 7mo) classified at GMFCS levels I (n=16) or II (n=14) were recruited from a tertiary hospital. Participants completed 43 minutes of standardized tasks on two consecutive days wearing an ActiGraph®GT3X+ tri-axial accelerometer at the hip on the unimpaired side, concurrently measuring heart rate (HR) using a PolarRS400® monitor and video recording. Testing protocol comprised 5 minutes seated rest period (REST) followed by 5 minutes walking at light (LW: 0.53m/sec), moderate (MW: 1.16m/sec) and vigorous (VW: 6min at 1.53m/s) pace on a 10 m linear track, and 3 minutes rapid stepping on/off a 20 cm step (STEP), interspersed with 5 minutes seated rest. ActiGraph® counts were synchronized with HR and standardized task then time coded into 5 seconds epochs. Two minutes of synchronized data were extracted from each task when HR reached steady state. Y-axis activity counts were converted to intensity using Evenson cut points. Reliability was calculated using intraclass- correlation coefficients (ICC) between days for each task (two-factor, mixed-effects, agreement model). Percentage agreement between ActiGraph® calculated and task intensity, and minimum detectable difference (MD=[SD√1-ICC] x1.96x√2) were calculated. Activity counts are counts/epoch, data is mean (SD). Results: Activity counts and HR increased with increasing task intensity. Mean activity counts during REST: 0.5 (5.3), LW: 52.9 (61.5), MW: 125.2 (82.4), VW: 355.9 (107.1), STEP: 286.0 (110.6); rise in HR from REST (bpm): LW: 17.7 (8.0), MW: 22.6 (8.8), VW: 51.9 (15.4), STEP: 51.8 (10.2). Agreement between activity intensity was 80-98% during REST, LW, VW and STEP tasks, but only 17% during MW. Testretest reliability was strong for LW (ICC=0.80), MW (ICC=0.80) and VW (ICC=0.70) but moderate for STEP tasks (ICC=0.67). MD was 15 counts/epoch for SED, 77 for LW, 101 for MW, 162 for VW and 178 for STEP tasks, meaning a difference of 23 345 counts or 10 minutes of moderate to vigorous physical activity (MVPA) between 12 hours wear periods is a statistically important change. Conclusion: The ActiGraph®GT3X+ tri-axial accelerometer is reliable during standardised walking and stepping tasks in independently ambulant children and adolescents with congenital hemiplegia.L716833842014-11-18 <br />