Influence of Accelerometer Type and Placement on Physical Activity Energy Expenditure Prediction During WheelChair

Abstract

An understanding of the impact of physical activity on the health of manual wheelchair users (MWUs) could be improved with the development of an accurate and unobtrusive monitoring tool. PURPOSE: (i) to identify the most appropriate anatomical location and accelerometer-based device to predict physical activity energy expenditure (PAEE) in MWUs and (ii) to develop and test associated regression equations to predict PAEE. METHODS: Seventeen manual wheelchair users (36 ± 10 yrs, 72 ± 11 kg) completed ten activities; resting, folding clothes, propulsion on a 1% gradient (3,4,5,6 and 7 km.hr-1) and propulsion at 4km.hr-1 (with an additional 8% body mass, 2% and 3% gradient) on a motorised wheelchair treadmill. GT3X+ and GENEA accelerometers were worn on the right wrist and upper arm. Linear regression analysis was conducted between outputs from each accelerometer and criterion PAEE, measured using indirect calorimetry. Subsequent error statistics were calculated for the developed regression equations for devices at the wrist using a leave-one-out cross-validation analysis. RESULTS: Accelerometer outputs at each anatomical location were significantly (p < .01) associated with PAEE (Upper arm, GT3X+; r = 0.68 and GENEA r = 0.87. Wrist, GT3X+; r = 0.82 and GENEA; r = 0.88). Average PAEE estimation error for all activities using new regression equations at the wrist was 13.8 ± 49.6% and 3.8 ± 25.7% for GT3X+ and GENEA, respectively. Absolute PAEE estimation errors for devices worn on the wrist varied from 17.1 to 122.0% for the GT3X+ and from 16.5 to 32.1% for the GENEA. The larger overall error associated with the GT3X+ worn on the wrist was mostly accounted for by a 1.2 Kcal.min-1 over-prediction of PAEE during the folding clothes activity. CONCLUSION: Of the two accelerometers and anatomical locations considered in this study, the results indicate that the GENEA worn on the wrist provides the most valid prediction of PAEE in MWUs. Variation in error statistics between the two devices is thought to be associated with inherent differences in internal components and on-board filtering processes of each device. The GT3X+ when worn on the wrist appears to lack sensitivity to accurately quantify the PAEE during low intensity activities, with low frequency components, which could lead to erroneous PAEE measurement during free-living monitoring.