Validity of physical activity monitors for estimating energy expenditure during wheelchair propulsion

Background: It is unknown if activity monitors can detect the increased energy expenditure (EE) of wheelchair propulsion at different speeds or on different surfaces. Methods: Individuals who used manual wheelchairs (n = 14) performed 5 wheeling activities: on a level surface at 3 speeds, on a rubberized track at 1 fixed speed and on a sidewalk course at a self-selected speed. EE was measured using a portable indirect calorimetry system and estimated by an Actical (AC) worn on the wrist and a SenseWear (SW) activity monitor worn on the upper arm. Repeated- measures ANOVA was used to compare measured EE to the estimates from the standard AC prediction equation and SW using 2 different equations. Results: Repeated-measures ANOVA demonstrated a significant main effect between measured EE and estimated EE. There were no differences between the criterion method and the AC across the 5 activities. The SW overestimated EE when wheeling at 3 speeds on a level surface, and during sidewalk wheeling. The wheelchair-specific SW equation improved the EE prediction during low intensity activities, but error progressively increased during higher intensity activities. Conclusions: During manual wheelchair propulsion, the wrist-mounted AC provided valid estimates of EE, whereas the SW tended to overestimate EE.