Validation and comparison of 3 accelerometers for measuring physical activity intensity during nonlocomotive activities and locomotive movements.

BACKGROUND The current study evaluated the validity of 3 commercially-available accelerometers to assess metabolic equivalent values (METs) during 12 activities. METHODS Thirty-three men and thirty-two women were enrolled in this study. The subjects performed 5 nonlocomotive activities and 7 locomotive movements. The Douglas bag method was used to gather expired air. The subjects also wore 3 hip accelerometers, a Lifecorder uniaxial accelerometer (LC), and 2 triaxial accelerometers (ActivTracer, AT; Actimarker, AM). RESULTS For nonlocomotive activities, the LC largely underestimated METs for all activities (20.3%-55.6%) except for desk work. The AT overestimated METs for desk work (11.3%) and hanging clothes (11.7%), but underestimated for vacuuming (2.3%). The AM underestimated METs for all nonlocomotive activities (8.0%-19.4%) except for hanging clothes (overestimated by 16.7%). The AT and AM errors were significant, but much smaller than the LC errors (23.2% for desk work and -22.3 to -55.6% for the other activities). For locomotive movements, the 3 accelerometers significantly underestimated METs for all activities except for climbing down stairs. CONCLUSIONS We conclude that there were significant differences for most activities in 3 accelerometers. However, the AT, which uses separate equations for nonlocomotive and locomotive activities, was more accurate for nonlocomotive activities than the LC.