Precision and accuracy of an ankle-worn accelerometer-based pedometer in step counting and energy expenditure

BACKGROUND: Walking is a widely used approach to increase physical activity levels in obese patients. In this paper, we investigate the precision and accuracy of an ankle-worn dual-axis accelerometer (Stepwatch) and investigate its potential application as a predictor of energy expenditure. METHODS: Twenty healthy subjects (10 lean, 10 obese) wore spring-levered (Accusplit), piezoelectric (Omron HF-100), and Stepwatch pedometers. Subjects walked on a treadmill at 1, 2, and 3 mph and in a hallway at 1 and 1.85 mph, during which energy expenditure was measured. RESULTS: The Stepwatch counted 99.7 +/- 0.67% (mean +/- SEM) of the manual counts. In comparison, the Omron pedometer counted 61 +/- 3.3% and the Accusplit counted 26 +/- 2.8% of the manual counts at 1 mph although all pedometers were accurate (> 98% of counts) at 3 mph. In repeated measures, the Stepwatch produced negligible variance (SD = 0.36) over all speed whereas the other pedometers showed a large amount of variance at all speed (SD = 4-13). Stepwatch counts were predictive of walking energy expenditure corrected by weight (r2 > 0.8). CONCLUSION: The counts from the Stepwatch were virtually identical to the manual counts from a trained investigator and provided a reliable predictor of walking energy expenditure.