Validity of the ActiGraph Inclinometer Algorithm for Detecting Sitting and Standing Postures

Aim: To assess the criterion validity of the ActiGraph inclinometer algorithm for detecting sitting and standing postures when worn on the thigh and waist.

Methods: Thirty-nine adults (46% female, age: 31.0±8.8 years, BMI: 26.7±5.0 kg/m2) were directly observed whilst participating in a range of sedentary (typing on a computer, watching television, playing a hand held computer game, playing on a Wii) and non-sedentary (standing still, walking at 0.4, 0.6, 1, 1.3, 1.6, 1.9, 2.3, and 2.6 km/h) tasks, each lasting 10 minutes. Participants wore 2 ActiGraph GT3X+ accelerometers, one on the anterior aspect of their right thigh and one on their waist. Time recorded in each posture according to the inclinometer algorithm during each task was compared to direct observation for each wear location.

Results: The inclinometer algorithm of the thigh-worn ActiGraph was 100% accurate at detecting standing postures in all non-sedentary tasks (standing still and slow walking). During the sedentary tasks for this wear location, the error of the inclinometer algorithm ranged from 4.7% to 8.1%. The error of the inclinometer algorithm of the waist-worn ActiGraph ranged from 2.4% to 39.1% for detecting non-sedentary tasks, with the exception of walking at 2.6 km/h where is was 100% accurate in detecting standing. During the sedentary tasks for this wear location, the error of the inclinometer algorithm ranged from 16.7% to 30.6%.

Conclusion: The validity of the ActiGraph inclinometer algorithm for detecting sedentary and non-sedentary postures is improved when the device is worn on the anterior aspect of the thigh.