Jaworski John, Makowski Brian, Weaver Michael, Seils Michael, Scheid Jennifer L
Department of Physical Therapy, Daemen University, Amherst, NY 14226, USA.
Department of Rehabilitation Science, University at Buffalo, Buffalo, NY 14214, USA.
Int J Environ Res Public Health. 2025 Jul 12;22(7):1100. doi: 10.3390/ijerph22071100.
Our study explored the accuracy of Fitbit recorded step count during the use of an assistive device (two-wheeled walker and standard cane) at various body positions (wrists, hips, and ankles). Participants ( = 11) ambulated an obstacle course (50 m total, including turns and a step up/down) a total of three times (two-wheeled walker, standard cane, and a deviceless control trial). Fitbit generated step counts (at the wrists, hips, and ankles) were then compared to the individual's "actual" step count captured through video analysis. During the cane trial and the deviceless trial, all positions were significantly correlated (r = 0.764 to 0.984; < 0.006) with the actual step count. However, increased variability (demonstrated by increased limits of agreement) was observed when the Fitbit was worn on the wrist (compared to the hips or ankles). During the walker trial, the step count was significantly correlated to the actual step count at the ankle and hip positions (r = 0.669 to 0.888; < 0.017) with an average error of 1.5%, while it was not statistically correlated at the wrist with a 31.2% average error. Our study suggests that Fitbits are a good predictor of actual step count, with the caveat that the location of the Fitbit should be considered if an assistive device (e.g., two-wheeled rolling walker and single-point cane) is being used.
我们的研究探讨了在使用辅助设备（两轮助行器和标准手杖）时，Fitbit记录的步数在不同身体部位（手腕、臀部和脚踝）的准确性。参与者（n = 11）总共三次走完一条障碍路线（总长50米，包括转弯和一个上下台阶）（两轮助行器、标准手杖和无辅助设备的对照试验）。然后将Fitbit生成的步数（在手腕、臀部和脚踝处）与通过视频分析获取的个体“实际”步数进行比较。在使用手杖试验和无辅助设备试验期间，所有部位与实际步数均显著相关（r = 0.764至0.984；P < 0.006）。然而，当Fitbit戴在手腕上时（与臀部或脚踝相比），观察到变异性增加（通过一致性界限增加来证明）。在使用助行器试验期间，脚踝和臀部位置的步数与实际步数显著相关（r = 0.669至0.888；P < 0.017），平均误差为1.5%，而在手腕处无统计学相关性，平均误差为31.2%。我们的研究表明，Fitbit是实际步数的良好预测指标，但需要注意的是，如果正在使用辅助设备（例如，两轮滚动助行器和单点手杖），则应考虑Fitbit的佩戴位置。
