Variability in bimanual wheelchair propulsion: consistency of two instrumented wheels during handrim wheelchair propulsion on a motor driven treadmill

BACKGROUND: Handrim wheelchair propulsion is a complex bimanual motor task. The bimanually applied forces on the rims determine the speed and direction of locomotion. Measurements of forces and torques on the handrim are important to study status and change of propulsion technique (and consequently mechanical strain) due to processes of learning, training or the wheelchair configuration. The purpose of this study was to compare the simultaneous outcomes of two different measurement-wheels attached to the different sides of the wheelchair, to determine measurement consistency within and between these wheels given the expected inter- and intra-limb variability as a consequence of motor control. METHODS: Nine able-bodied subjects received a three-week low-intensity handrim wheelchair practice intervention. They then performed three four-minute trials of wheelchair propulsion in an instrumented hand rim wheelchair on a motor-driven treadmill at a fixed belt speed. The two measurement-wheels on each side of the wheelchair measured forces and torques of one of the two upper limbs, which simultaneously perform the push action over time. The resulting data were compared as direct output using cross-correlation on the torque around the wheel-axle. Calculated push characteristics such as power production and speed were compared using an intra-class correlation. RESULTS: Measured torque around the wheel axle of the two measurement-wheels had a high average cross-correlation of 0.98 (std=0.01). Unilateral mean power output over a minute was found to have an intra-class correlation of 0.89 between the wheels. Although the difference over the pushes between left and right power output had a high variability, the mean difference between the measurement-wheels was low at 0.03 W (std=1.60). Other push characteristics showed even higher ICC’s (>0.9). CONCLUSIONS: A good agreement between both measurement-wheels was found at the level of the power output. This indicates a high comparability of the measurement-wheels for the different propulsion parameters. Data from both wheels seem suitable to be used together or interchangeably in experiments on motor control and wheelchair propulsion performance. A high variability in forces and timing between the left and right side were found during the execution of this bimanual task, reflecting the human motor control process.