Validation Of The Actigraph (GT3X) Inclinometer Function: 2045: Board #174 June 3 8:00 AM - 9:30 AM

McMahon, Gregory C.; Brychta, Robert J.; Chen, Kong Y.

Medicine & Science in Sports & Exercise: May 2010 - Volume 42 - Issue 5 - p 489
doi: 10.1249/01.MSS.0000385098.02949.38
C-26 Free Communication/Poster - Objective Measures: Accelerometry and Pedometry: JUNE 3, 2010 7:30 AM - 12:30 PM: ROOM: Hall C

NIDDK, Bethesda, MD.


(No disclosure reported)

INTRODUCTION: Postural detection from portable activity monitors may facilitate activity assessment and energy expenditure prediction. However, traditional accelerometers lack postural information. The newly developed Actigraph GT3X monitors have added a new function for postural detection, which has not yet been validated.

PURPOSE: The primary purpose of this investigation was to compare the GT3X inclinometer function with direct observation. A secondary purpose was to determine the accuracy of the inclinometer output at different locations around the waist.

METHODS: Ten healthy participants, seven female and three male, BMI= 27 + 6.38 kg/m2 wore three GT3X's set at an epoch length of one second, placed around their waists on a single elastic waist band. The monitors were located on the right hip, above the right leg on the anterior spine of the iliac crest, and on the small of the back. Each subject randomly completed a series of twelve different postures, including sitting, standing, and lying down in both standardized and free-living positions.

RESULTS: At the hip (manufactures-recommended site), the inclinometer accurately classified 40% and 96.5% of the time spent in standardized and free-sitting respectively. For standing accuracy was 59.4% and 53.7% for standardized and free-standing respectively. For laying accuracy was 80% and 30% for standardized and free-laying respectively. At the back location, detection of standing improved to 83.8% and 90% for standardized standing and free-standing respectively, but accuracy for sitting and lying down was worse compared to the hip. At the right leg location, the detection of posture was lower than that of the hip for all activities.

Summary: The current postural detection algorithm from Actigraph GT3X has limited validity and is location-dependent.

© 2010 American College of Sports Medicine