Department of Orthopaedic Surgery and Rehabilitation, Oregon Health and Science University, Portland, Oregon, U.S.A.
Department of Neurology, Oregon Health and Science University, Portland, Oregon, U.S.A.
Arthroscopy. 2017 Dec;33(12):2110-2116. doi: 10.1016/j.arthro.2017.06.042. Epub 2017 Aug 31.
To develop a model using wearable inertial sensors to assess the performance of orthopaedic residents while performing a diagnostic knee arthroscopy.
Fourteen subjects performed a diagnostic arthroscopy on a cadaveric right knee. Participants were divided into novices (5 postgraduate year 3 residents), intermediates (5 postgraduate year 4 residents), and experts (4 faculty) based on experience. Arm movement data were collected by inertial measurement units (Opal sensors) by securing 2 sensors to each upper extremity (dorsal forearm and lateral arm) and 2 sensors to the trunk (sternum and lumbar spine). Kinematics of the elbow and shoulder joints were calculated from the inertial data by biomechanical modeling based on a sequence of links connected by joints. Range of motion required to complete the procedure was calculated for each group. Histograms were used to compare the distribution of joint positions for an expert, intermediate, and novice.
For both the right and left upper extremities, skill level corresponded well with shoulder abduction-adduction and elbow prono-supination. Novices required on average 17.2° more motion in the right shoulder abduction-adduction plane than experts to complete the diagnostic arthroscopy (P = .03). For right elbow prono-supination (probe hand), novices required on average 23.7° more motion than experts to complete the procedure (P = .03). Histogram data showed novices had markedly more variability in shoulder abduction-adduction and elbow prono-supination compared with the other groups.
Our data show wearable inertial sensors can measure joint kinematics during diagnostic knee arthroscopy. Range-of-motion data in the shoulder and elbow correlated inversely with arthroscopic experience. Motion pattern-based analysis shows promise as a metric of resident skill acquisition and development in arthroscopy.
Wearable inertial sensors show promise as metrics of arthroscopic skill acquisition among residents.
使用可穿戴式惯性传感器开发一种模型，以评估骨科住院医师在进行膝关节诊断性关节镜检查时的表现。
14 名受试者对一具尸体右膝进行了诊断性关节镜检查。参与者根据经验分为新手（3 年级 5 名住院医师）、中级（4 年级 5 名住院医师）和专家（4 名教员）。通过将 2 个传感器固定在上肢（背侧前臂和外侧臂）和躯干（胸骨和腰椎）上的每个部位（2 个传感器），来收集惯性测量单元（Opal 传感器）的手臂运动数据。通过基于关节连接的一系列链接的生物力学建模，从惯性数据中计算出肘关节和肩关节的运动学。为每个组计算完成该程序所需的运动范围。使用直方图比较专家、中级和新手的关节位置分布。
对于右上肢和左上肢，技能水平与肩外展-内收和肘旋前-旋后密切相关。新手在右肩外展-内收平面上完成诊断性关节镜检查所需的运动平均比专家多 17.2°（P =.03）。对于右肘旋前-旋后（探针手），新手完成该手术所需的运动平均比专家多 23.7°（P =.03）。直方图数据显示，与其他组相比，新手在肩外展-内收和肘旋前-旋后方面的变化明显更大。
我们的数据表明，可穿戴式惯性传感器可以测量诊断性膝关节关节镜检查过程中的关节运动学。肩部和肘部的运动范围数据与关节镜经验呈反比。基于运动模式的分析有望成为关节镜手术中住院医师技能获取和发展的指标。
可穿戴式惯性传感器有望成为衡量住院医师关节镜技能获取的指标。
