Comparison of Shoulder Motion Measurements by Visual Estimate, Goniometer and Motion Capture

Abstract

Background: Brachial plexus birth injuries (BPBI) can result in lasting impairments of external rotation and cross-body adduction (CBA) that disrupt functional activities such as dressing, grooming, or throwing a ball. The purpose of this study was to compare the quantification of shoulder humerothoracic (HT) external rotation (ER), and glenohumeral (GH) CBA by 3 methodsphysician visual estimate, goniometer measurement by an occupational therapist, and motion capture. Methods: Twenty-six patients with BPBI (average age of 9.9 ± 3.2 y) participated in this study. Mallet scores and visual estimates of passive HT ER and GH CBA were recorded by a physician. The passive measures were repeated by an occupational therapist using a goniometer while motion capture measures were simultaneously collected. Active HT ER was also measured by motion capture. The passive measures were compared with analyses of variance with repeated measures, intraclass correlations, and Bland-Altman plots. External rotation Mallet scores determined by motion capture and by the physician were compared. Results: The measures of GH CBA were not statistically different and demonstrated good agreement, but substantial variation. For HT ER, all measures were significantly different and demonstrated poor agreement and substantial variation. When the joint angles measured by motion capture were used to determine the Mallet score, 79% of external rotation Mallet scores assigned by the physician were incorrectly categorized, with the physician always scoring the participant higher than predicted motion capture Mallet score.Conclusions: Both GH CBA and HT ER measures demonstrated substantial variability between measurement types, but only HT ER joint angles were significantly different. In addition, more than three-quarters of external rotation Mallet scores were misclassified by the physician. Motion capture measurements offer the benefit of less susceptibility to patients' compensatory and/or out-of-plane movements and should be considered for clinical assessment of shoulder range of motion in children with BPBI. If motion capture is unavailable, the use of a goniometer provides more accurate clinical measures of shoulder motion than visual estimates and care should be taken to minimize and account for compensatory movement strategies.

Validation of Angle Estimation Based on Body Tracking Data from RGB-D and RGB Cameras for Biomechanical Assessment

“…Furthermore, MediaPipe’s average error of 8.57 is below the error of tools commonly used in physical therapy clinics for joint angle assessment. Additionally, compared to the results obtained by Russo [ 15 ], the human error for angular measurements is 12.78 degrees. Thus, MediaPipe scored excellent for all data and lower limbs and good metric for upper limbs.…”

“…In Table 4 , it is possible to see the error for all movements and split for upper and lower limbs movements. In addition, it is possible to analyze its classification in relation to human error performed in clinical practice (12.78 ± 7.44 degrees) [ 15 ].…”

“…The absolute error was classified as excellent, good, moderate, or poor according to its median value and standard deviation concerning the clinical reference (CR), estimated by the human error performed in the clinical environment (12.78 ± 7.44 degrees) [ 15 ]. Thus, the final classification was as follows: (i) great, when error plus its standard deviation is below CR; (ii) good, if the error is below the CR; (iii) moderate, if it crosses CR until five degrees; and (iv) weak, if the error is more than 5 degrees higher than CR.…”

The Philadelphia Shriners Hospital Approach to Brachial Plexus Birth Injury

“…reachable workspace) and can differentiate glenohumeral and scapulothoracic joint motion. 4 5 Motion analysis yields a more accurate and nuanced assessment of shoulder function than can be obtained from a Mallet score. 5…”

“…4 5 Motion analysis yields a more accurate and nuanced assessment of shoulder function than can be obtained from a Mallet score. 5…”