Department of Neurology, University Hospital Schleswig-Holstein, Arnold-Heller-Strasse 3, Haus D, 24105 Kiel, Germany.
Digital Signal Processing and System Theory, Institute of Electrical and Information Engineering, Kiel University, Kaiserstrasse 2, 24143 Kiel, Germany.
Sensors (Basel). 2022 Mar 16;22(6):2304. doi: 10.3390/s22062304.
Evaluating gait is part of every neurological movement disorder assessment. Generally, the physician assesses the patient based on their experience, but nowadays inertial measurement units (IMUs) are also often integrated in the assessment. Instrumented gait analysis has a longstanding tradition and temporal parameters are used to compare patient groups or trace disease progression over time. However, the day-to-day variability needs to be considered especially in specific patient cohorts. The aim of the study was to examine day-to-day variability of temporal gait parameters of two experimental conditions in a cohort of neurogeriatric patients using data extracted from a lower back-worn IMU. We recruited 49 participants (24 women (age: 78 years ± 6 years, BMI = 25.1 kg/m and 25 men (age: 77 years ± 6 years, BMI = 26.5 kg/m)) from the neurogeriatric ward. Two gait distances (4 m and 20 m) were performed during the first session and repeated the following day. To evaluate reliability, the Intraclass Correlation Coefficient (ICC2,k) and minimal detectable change (MDC) were calculated for the number of steps, step time, stride time, stance time, swing time, double limb support time, double limb support time variability, stride time variability and stride time asymmetry. The temporal gait parameters showed poor to moderate reliability with mean ICC and mean MDC95% values of 0.57 ± 0.18 and 52% ± 53%, respectively. Overall, only four out of the nine computed temporal gait parameters showed high relative reliability and good absolute reliability values. The reliability increased with walking distance. When only investigating steady-state walking during the 20 m walking condition, the relative and absolute reliability improved again. The most reliable parameters were swing time, stride time, step time and stance time. Study results demonstrate that reliability is an important factor to consider when working with IMU derived gait parameters in specific patient cohorts. This advocates for a careful parameter selection as not all parameters seem to be suitable when assessing gait in neurogeriatric patients.
评估步态是每一项神经运动障碍评估的一部分。通常，医生根据自己的经验评估患者，但现在惯性测量单元（IMU）也经常被整合到评估中。仪器化步态分析有着悠久的传统，时间参数用于比较患者组或跟踪疾病随时间的进展。然而，尤其是在特定的患者群体中，需要考虑日常的变异性。本研究的目的是使用从背部佩戴的 IMU 中提取的数据，检查神经老年病房的一组神经老年患者在两种实验条件下的时间步态参数的日常变异性。我们招募了 49 名参与者（24 名女性（年龄：78 岁±6 岁，BMI=25.1kg/m2和 25 名男性（年龄：77 岁±6 岁，BMI=26.5kg/m2））来自神经老年病房。在第一次会议上进行了两次步态距离（4m 和 20m），并于次日重复进行。为了评估可靠性，计算了步数、步时、步长时间、站立时间、摆动时间、双肢支撑时间、双肢支撑时间变异性、步长时间变异性和步长时间不对称性的组内相关系数（ICC2，k）和最小可检测变化（MDC）。时间步态参数的可靠性较差，平均 ICC 和平均 MDC95%值分别为 0.57±0.18 和 52%±53%。总体而言，只有九个计算出的时间步态参数中的四个显示出高相对可靠性和良好的绝对可靠性值。随着步行距离的增加，可靠性也随之增加。当仅在 20m 步行条件下研究稳态步行时，相对和绝对可靠性再次提高。最可靠的参数是摆动时间、步长时间、步时和站立时间。研究结果表明，在特定患者群体中使用 IMU 衍生的步态参数时，可靠性是一个重要的考虑因素。这提倡谨慎选择参数，因为并非所有参数在评估神经老年患者的步态时都适用。
