Introduction: Accelerometers are used to quantify energy expenditure in field research. The ActiGraph™ GT1M (ActiGraph™) is a commonly used accelerometer for research. The FitBit ® Ultra (FitBit ®) is a low-cost alternative to the ActiGraph™; however, there is limited research on the validity of this device. Purpose: The pilot study compares the FitBit ® against the ActiGraph™ and metabolic cart for measurement of energy expenditure and step counts during treadmill walking. Methods: Thirty-two (25 female) adults, mean age 22±2 years, performed two thirty-minute phases of walking (slow and brisk) on a treadmill while concurrently wearing the FitBit ® and the ActiGraph™. Energy expenditure estimates were compared against energy expenditure measured by a metabolic cart. The Pearson's correlation and t-tests determine the linear association and similarity between the accelerometers. Results: Energy expenditure estimate is moderately correlated between the two accelerometers during slow walking (r=0.584, p=0.011) and strongly correlated during brisk walking (r=0.910, p<0.001). Step count is strongly correlated between the accelerometers during slow (r=0.974, p<0.001) and brisk (r=0.996, p<0.001) walking. The FitBit ® significantly underestimated energy expenditure during brisk walking compared to metabolic cart data. There is no difference between the slow and brisk phases' step counts using either accelerometer. Conclusion: The results of this pilot study suggest that the FitBit ® and the ActiGraph™ can be used interchangeably to measure steps, but not to measure kilocalories. Furthermore, the FitBit ® underestimates energy expenditure, compared to a metabolic cart, as exercise intensity increases. This limits its ability to accurately measure energy expenditure in active populations.

To assess the accuracy and reliability of reflectance photometers in estimating blood glucose concentrations, two were assessed: the Ames Glucometer 3 (Bayer Diagnostics) with Glucofilm Test strips; and the Reflolux S (Boehringer Mannheim) with BM Test Glycemia 20-800 strips. These were compared with laboratory blood glucose estimations in 100 assays (50 comparisons for each machine, measuring the difference (d) between the glucose reading and the mean of the reflectance photometer and the laboratory value). The Ames Glucometer 3 (mean d=+0-7 mmol/l, (SD 1.1) mmol/l) was less accurate than the Boehringer Reflolux S (mean d=0*2 mmoi/l, (SD (0.7) mmolfl).The range of error of both machines is wide (Ames 2 SD range +2*9 mmoUl to -1-5 mmol/ true readings; Boehringer + 1-8 mmol/l to -1l2 mmol/l of true readings). Because of this, any reflectance photometer readings that are even slightly low should be checked with laboratory estimations. The clinical value of such machines is limited in infants with low blood glucose concentrations.