Variation in validity of the FitBit Charge 2 in treadmill walking: with differing hand positions in the healthy adult population.

Abstract

Question: What is the validity of the FitBit Charge 2 at 1.2m/s for step count and does this vary with differing hand position? Design: Within-subject experimental validity study. Participants walked on a treadmill at 1.2m/s for each hand position, whilst wearing the FitBit on their non-dominant wrist. Participants: 27 healthy adults (18 female, 9 male) from within a university. Outcome Measures: The three hand positions investigated were; normal arm swing (NAS), hands in pocket (HIP) and holding phone (HP). FitBit recorded steps were compared to criterion measure. The criterion measure was defined as visual observed steps calculated from a lower limb video recording by two raters. Results: Data from 26 participants were included in analysis with one participant removed due to treadmill malfunction. Inter-rater reliability was excellent for VOS (all positions ICC ≥ 0.997). The mean difference of step count and standard deviation (SD) (FitBit – VOS) for NAS, HIP and HP was -12.92 ±23.40, 1.62 ±5.31 and -1.60 ±5.48 respectively. The MPE and SD that represents these differences was 5.61% ±10.10%, 0.77% ±2.39% and 0.65 % ±2.41% for each position respectively. Conclusion: The FitBit Charge 2 is a valid device in measuring step count in the healthy adult population. It is most accurate for HP, with least accuracy shown in NAS. Therefore despite changes in hand position during walking it can be used as a physical activity monitor for healthy adults, however it is suggested that caution is taken when using it.