Validity and Reliability of Wearable Motion Sensors for Clinical Assessment of Shoulder Function in Brachial Plexus Birth Injury

Abstract

The modified Mallet scale (MMS) is commonly used to grade shoulder function in brachial plexus birth injury (BPBI) but has limited sensitivity and cannot grade scapulothoracic and glenohumeral mobility. This study aims to evaluate if the addition of a wearable inertial movement unit (IMU) system could improve clinical assessment based on MMS. The system validity was analyzed with simultaneous measurements with the IMU system and an optical camera system in three asymptomatic individuals. Test-retest and interrater reliability were analyzed in nine asymptomatic individuals and six BPBI patients. IMUs were placed on the upper arm, forearm, scapula, and thorax. Peak angles, range of motion, and average joint angular speed in the shoulder, scapulothoracic, glenohumeral, and elbow joints were analyzed during mobility assessments and MMS tasks. In the validity tests, clusters of reflective markers were placed on the sensors. The validity was high with an error standard deviation below 3.6°. Intraclass correlation coefficients showed that 90.3% of the 69 outcome scores showed good-to-excellent test-retest reliability, and 41% of the scores gave significant differences between BPBI patients and controls with good-to-excellent test-retest reliability. The interrater reliability was moderate to excellent, implying that standardization is important if the patient is followed-up longitudinally.

Keywords: brachial plexus birth injury; clinical evaluation; inertial movement unit; kinematic analysis; scapula movement; shoulder function.

Figure 1

(A–C). Illustrates the IMU placements used in Part I (validity) and Part II (reliability). The reflective markers used in Part II were placed either on an orthoplastic shell with the sensor in the center of the shell (upper and lower arms, (A) or directly on the sensor’s front and sides (scapula (B) and thorax (C)). The local coordinate systems, marked with white arrows (B,C), were all defined so that after calibration/sensor alignment, they were oriented in the same way.

Figure 2

Examples of data from one test person during four of the tasks included in the modified Mallet score MMS3 hand to neck (A), MMS4 hand to spine (B), MMS5 hand to mouth (C), MMS6 internal rotation (D) as simultaneously measured with the reference system (red line) and the IMU system (thick blue line). The segment helical angle was calculated for scapula (upper row), upper arm (middle row) and forearm (bottom row).

Figure 3

Bland–Altmann plots illustrate system agreements of the IMU system with the reference system. The helical angle was calculated for each segment, scapula (A), upper arm (B) and forearm (C). Angular errors were calculated for tasks involving large shoulder movements in one plane (upper row, shoulder flexion−extension, MMS1 global abduction and MMS2 global external rotation), for elbow mobility tasks (middle row, elbow flexion−extension and forearm pronation–supination); tasks involving movement in both elbow and shoulder (bottom row, MMS3 hand to neck, Mallet MMS4 hand to spine, MMS5 hand to mouth and MMS6 internal rotation). The mean error is illustrated with a blue line, and the error 95% confidence intervals are marked with red lines.

Figure 4

Outcome measures from the assessment of (A) shoulder and (B,C) elbow mobility. Group mean and standard error of mean are illustrated (BPBI: red left bar, control: blue right bar). ICC from test–retest reliability and p-values from t-tests are shown in the upper right corner of each subplot and are highlighted green for outcome scores with both good test–retest reliability (ICC > 0.75) and a significant group difference (p < 0.05).

Figure 5

Outcome measures from the assessment of two tasks from the modified Mallet scale (MMS); (A) global abduction and (B) global external rotation. Group mean and standard error of mean are illustrated (BPBI: red left bar, control: blue right bar). ICC from test–retest reliability and p-values from t-tests are shown in the upper right corner of each subplot and are highlighted green for outcome scores with both good test–retest reliability (ICC > 0.75) and a significant group difference (p < 0.05).

Figure 6

Outcome measures from the assessment of two tasks from the modified Mallet scale (MMS); (A) hand to neck and (B) hand to spine. Group mean and standard error of mean are illustrated (BPBI: red left bar, control: blue right bar). ICC from test–retest reliability and p-values from t-tests are shown in the upper right corner of each subplot and are highlighted green for outcome scores with both good test–retest reliability (ICC > 0.75) and a significant group difference (p < 0.05).

Figure 7

Outcome measures from the assessment of two tasks from the modified Mallet scale (MMS); (A) hand to mouth and (B) internal rotation. Group mean and standard error of mean are illustrated (BPBI: red left bar, control: blue right bar). ICC from test–retest reliability and p-values from t-tests are shown in the upper right corner of each subplot and are highlighted green for outcome scores with both good test–retest reliability (ICC > 0.75) and a significant group difference (p < 0.05).