Background: Identifying older adults with increased fall risk due to poor postural control on a large scale is only possible through omnipresent and low cost measuring devices such as the inertial measurement units (IMU) embedded in smartphones. However, the correlation between smartphone measures of postural stability and state-of-the-art force plate measures has never been assessed in a large sample allowing us to take into account age as a covariate. Research question: How reliably can postural stability be measured with a smartphone embedded IMU in comparison to a force plate? Methods: We assessed balance in 97 adults aged 50 to 90 years in four different conditions (eyes open, eyes closed, semi-tandem and dual-task) in the anterio-posterior and medio-lateral directions. We used six different parameters (root mean square and average absolute value of COP displacement, velocity and acceleration) for the force plate and two different parameters (root mean square and average absolute value of COM acceleration) for the smartphone. Results: Test-retest reliability was smaller for the smartphone than for the force plate (intra class correlation) but both devices could equally well detect differences between conditions (similar Cohen's d). Parameters from the smartphone and the force plate, with age regressed out, were moderately correlated (robust correlation coefficients of around 0.5). Significance: This study comprehensively documents test-retest reliability and effect sizes for stability measures obtained with a force plate and smartphone as well as correlations between force plate and smartphone measures based on a large sample of older adults. Our large sample size allowed us to reliably determine the strength of the correlations between force plate and smartphone measures. The most important practical implication of our results is that more repetitions or longer trials are required when using a smartphone instead of a force plate to assess balance.

中文翻译：

---

验证了用于测量老年人姿势稳定性的智能手机嵌入式惯性测量单元

背景：只有通过无所不在且低成本的测量设备（例如嵌入在智能手机中的惯性测量单元（IMU）），才能大规模识别由于姿势控制不佳而导致跌倒风险增加的老年人。但是，从未在大样本中评估智能手机的姿势稳定性测度和最新的测力板测度之间的相关性，这使我们可以将年龄作为协变量来考虑。研究问题：与压力板相比，使用智能手机嵌入式IMU测量姿势稳定性的可靠性如何？方法：我们评估了97名年龄在50至90岁之间的成年人在前后前后方向上的四种不同情况（睁眼，闭眼，半串联和双任务）的平衡。我们为压板使用了六个不同的参数（均方根和COP位移，速度和加速度的平均绝对值），为智能手机使用了两个不同的参数（均方根和COM加速度的平均绝对值）。结果：智能手机的重测可靠性比测力板小（类内相关性），但两种设备都可以很好地检测条件之间的差异（类似Cohen d）。随着年龄的增长，智能手机和测力板的参数会适度相关（稳健的相关系数约为0.5）。意义：这项研究全面记录了通过测力板和智能手机获得的稳定性测度的重测信度和效应大小，以及基于大量老年人的测力板和智能手机测度之间的相关性。我们的大样本量使我们能够可靠地确定测力板和智能手机测量值之间相关性的强度。我们的结果最重要的实际含义是，使用智能手机而非测力板评估平衡时，需要进行更多的重复或更长的试验。