Reliability and Validity of a Clinical Assessment Tool for Measuring Scapular Motion in All 3 Anatomical Planes

A convenience sample generated a total of 67 inquiries from the general population in Lexington, Kentucky. All volunteers were screened for eligibility based on the following inclusion criteria: age between 18 and 99 years, willingness to attend 2 testing sessions separated by at least 24 hours, the ability to lift the right upper extremity to at least 120° in the scapular plane, and no self-reported medical restrictions relating to the upper extremity or spine at the time of the study. An a priori power analysis conducted using nQuery software (version 8.1; Statistical Solutions Ltd) indicated a sample size of 60 participants would have 90% power to detect a difference in means of 3° in scapular motion and minimize the chance of type II error.

We identified and enrolled 60 participants (29 women, 31 men; age = 30 ± 14 years, height = 1.73 ± 0.10 m, mass = 75.32 ± 16.90 kg) who met the inclusion criteria. All participants completed 2 testing sessions, with an average time between sessions of 9 days. All participants provided written informed consent, and the study protocol was approved by the University of Kentucky Institutional Review Board (No. 43537).

The EasyAngle electric goniometer (Meloq AB) was used to perform clinical measurements of scapular motion (Figure 1). Before data collection, an upright polyvinyl chloride (PVC) pole was placed at 30° anterior to the frontal plane relative to the participant's sitting location, marking the scapular plane. Participants were instructed to actively raise the upper extremity with the wrist touching the PVC pole until they reached 120°, as confirmed using a standard goniometer. When the individual reached 120° of upper extremity elevation, a quick-grip mini bar clamp (Irwin Tools) was used to mark and physically limit 120° of upper extremity elevation on the PVC pole (Figure 2). An I-beam square bubble level (model 7724; Johnson Level & Tool Manufacturing Co, Inc) was used to calibrate the electric goniometer for measurements taken in the sagittal plane as described.

Three-dimensional motion capture was recorded using a Nexus (Vicon Motion Systems Ltd) 14-camera, high-speed, infrared, video-based, optical, motion-capture system. Raw marker trajectory data were stored and reconstructed using Nexus software (version 2.9; Vicon Motion Systems Ltd). Reconstructed kinematic data were exported and analyzed using Visual3D (version 9 Professional; C-Motion, Inc).

The clinical measurements using the electric goniometer were recorded independently by each examiner during upper extremity elevation in the frontal, transverse, and sagittal planes. To facilitate consistency of the clinical measurements between examiners and to accommodate the placement of the retroreflective markers used for 3D optical motion capture, a standard operating procedure was implemented. For each anatomical plane, the calibration technique and specific location for the electric goniometer based on several scapular landmarks followed a standard procedure. To measure scapular motion in the frontal plane, we calibrated the electric goniometer to the floor directly beneath participants to represent 0°. The electric goniometer was placed on the scapular spine at one-third of the distance between the root of the scapular spine and the posterior acromial angle, as measured and marked using a cloth tape measure (Figure 1A), and oriented posteriorly (Figure 1B). To measure scapular motion in the transverse plane, the electric goniometer was calibrated using a perpendicular edge of a floor tile beneath participants to represent 0°. The electric goniometer was placed at the same location on the scapular spine as described for frontal-plane motion but oriented superiorly (Figure 1C). To record scapular motion in the sagittal plane, we calibrated the electric goniometer to the vertical I-beam square level to represent 0°, placed on the most prominent portion of the medial scapular border and oriented laterally (Figure 1D).

All participants began each trial seated in an upright position on a 35-cm-tall stool with their feet flat on the floor. The motion of upper extremity elevation was explained and demonstrated. They were able to practice the motion several times and ask questions before data collection. To begin each trial, the examiner applied the electric goniometer to the specified scapular landmark and instructed the participant to assume an upright and relaxed sitting posture. The scapular rest position was recorded, and then the participant was prompted to perform the desired condition. After completing active movement, the individual held the final position for several seconds while the examiner measured the end scapular position. Total excursion values were calculated by subtracting the initial scapular position (rest) from the final scapular position (end) after motion was completed. Three trials of active upper extremity elevation were recorded for each scapular condition, totaling 9 trials for data collection. Constant pressure and contact were maintained with the scapular landmark during each movement. The order of anatomical planes was randomized before testing, and the same order was used on both test days. Clinical measurements of scapular motion were interpreted following the guidelines of the International Society of Biomechanics¹⁸ : positive scapular motion in the frontal, transverse, and sagittal planes is identified as downward rotation, internal rotation, and posterior tilt, respectively.

On 1 day of testing, 3D optical motion capture was recorded simultaneously with the clinical measurements. Surface reflective markers were attached to the participant using 2-sided tape following the procedures outlined by Chu et al⁸  in a validation study of the marker-based motion-capture model of scapular motion (Figure 3). A scapular acromial marker cluster (AMC) was created using a rigid triangular body and was applied to the posterior acromial process and medial to the posterior acromial calibration marker (Figure 1B). Recording of scapular motion using an AMC has shown excellent within-session reliability (ICC range = 0.90–0.98) and a standard error of measurement (SEM) of 2.25° for active upper extremity elevation, protraction, and retraction and has been validated against criterion standard techniques, such as dynamic radiography.^8,19  The raw kinematic camera data were collected at 200 Hz and smoothed using a low-pass Butterworth filter with a cutoff frequency of 6 Hz. Joint coordinate systems and segment parameters for the trunk, pelvis, and scapula were oriented with the x axis pointed anteriorly, the y axis oriented superiorly, and the z axis oriented laterally (Figure 3).¹⁸  A Euler rotation sequence for scapular motion in the frontal and transverse planes was resolved as Y, X, Z and calculated relative to the thorax per the guidelines of the International Society of Biomechanics.¹⁸ 

We applied a test-retest design to examine the intrarater reliability of each examiner between testing sessions and the interrater reliability of both examiners in the same testing session for clinical measurements recorded using the electric goniometer. Both the intrarater and interrater reliability of scapular measurements recorded during rest and excursion for each anatomical plane were assessed via ICC (2,3) using the average of 3 trials of motion. We interpreted the ICCs as poor (<0.5), moderate (0.5–0.75), good (0.76–0.90), or excellent (>0.90) reliability.²⁰  Measurement precision was determined by calculating the SEM and the minimal detectable change score at the 90% CI.²¹ 

The criterion validity of the electric goniometer versus the reference standard of the 3D optical motion-capture system for measuring total scapular excursion in each anatomical plane was determined using several approaches. First, a paired t test was used to compare the average excursion of 3 trials of motion between the electric goniometer and the 3D optical motion-capture system. We set the α level a priori at ≤.05, although a Bonferroni correction was applied to account for the 3 comparisons of each condition in each plane. This correction reduced the α level to ≤.017. Second, the root mean square error (RMSE) was calculated to determine the error associated with the electric goniometer versus the 3D optical motion-capture system for each condition. Third, we created Bland-Altman plots to examine the average difference and limits of agreement (LOAs) between the electric goniometer and 3D optical motion-capture system. The LOAs were calculated by multiplying the SD of the average difference by 1.96 to observe the 95% CIs.²²  Validity was determined by observing the RMSE values: <5° indicated strong, 5° to 10° indicated moderate, and >10° indicated poor validity.^8,23  Analysis of the Bland-Altman plots revealed a systematic average difference of −7° between scapular excursions recorded using the electric goniometer and the 3D optical motion-capture system for scapular motion measured in the frontal plane. Therefore, a correction factor of +7° was applied to the clinical data for mean scapular excursion in the frontal plane.

We observed moderate to good intrarater reliability for determining the scapular rest position and scapular excursion between testing sessions (Table 1). Interrater reliability was good to excellent for measuring the scapular rest position and moderate for measuring the scapular excursion during a testing session (Table 2).

The validity results are presented in Table 3. Bland-Altman plots are provided in Supplemental Figures 1 through 3. We found differences between the mean scapular excursions recorded using the electric goniometer and 3D optical motion-capture system for the frontal (P < .001), transverse (P = .015), and sagittal (P < .001) planes. The RMSE ranged from 7° to 10°, and the average difference ranged from −7° to 4°.

We investigated the reliability and validity of a novel electric goniometer for measuring scapular motion in each anatomical plane during upper extremity elevation. The intrarater and interrater reliability of the clinical scapular measurements was investigated across days and between examiners and addressed the criterion validity of measurements recorded using the clinical assessment device compared with the reference standard of 3D optical motion capture. Our results indicated that the electric goniometer was a reliable device for measuring the scapular rest position and total excursion in each anatomical plane when a standard operating procedure was used. Furthermore, the electric goniometer had moderate validity for measuring scapular excursions in all 3 anatomical planes in a clinical setting.

Before data collection, we hypothesized that the measurements recorded using the electric goniometer would not exceed 10° of error compared with those recorded using the 3D optical motion-capture system. Although the resultant P values demonstrated differences among mean values for scapular excursion in each anatomical plane, the comparison of means alone was not sufficient for a complete validity analysis.²⁴  Therefore, we used a multistep approach to assess validity using statistics, such as RMSE, average difference, and LOA.²²  The threshold of RMSE was rooted in the notion that 10° of error would exceed both the SEM and minimal detectable change score at 90% CI, such that error >10° would indicate an invalid measurement of scapular motion. Additionally, previous researchers^8,25,26  found that RMSE values >10° indicated inaccurate measures of true scapular motion. In our study, the RMSE values were ≤10° for all planes of motion. Furthermore, the average difference between the electric goniometer and the 3D motion-capture system ranged from −7° to 4° across the 3 anatomical planes. Taken together, these results suggest that the electric goniometer was capable of measuring scapular motion in each anatomical plane during upper extremity elevation with a moderate degree of accuracy.

The RMSE associated with upper extremity elevation in the frontal plane highlights a limitation of using a 3D optical motion-capture system with an AMC to capture scapular motion. The AMC represents the scapula, and its motion was recorded using the 14-camera 3D optical motion-capture system to represent scapular movement. Placement of the AMC on the posterior acromion was difficult and restricted access to the scapular spine. As shown in Figures 1B and 1C, placement of the electric goniometer was limited to the medial aspect of the scapular spine due to the AMC's position on the acromion. Therefore, the correction factor was applied to frontal-plane data, the plane of motion most affected by the AMC. The correction reduced the RMSE value from 10° to 7° and increased the associated P value to .96, illustrating no difference between the electric goniometer and the 3D optical motion-capture system when measuring scapular motion in the frontal plane during upper extremity elevation.

We observed a similar error between the measurement methods during motion in the sagittal plane. Given the difference between methods despite an RMSE of 9° and an average difference of 4°, we suspect that accessory motion from spinal flexion and extension contributed to the overall differences in scapular measurement. Although participants were orally instructed to not move their spines during each trial and were closely observed during testing, it was not possible to eliminate the inherent motion of the spine. This concept highlights a limitation of calibrating the electric goniometer to a standalone vertical surface (I-beam square level). To overcome this limitation in the future, we suggest calibrating the electric goniometer to the participant's spine before measuring sagittal-plane motion. This adjustment in calibration will ideally capture the inherent trunk position of the participant and account for any initial spinal offset in the sagittal plane.

Although each measurement recorded by the electric goniometer introduced a specific limitation, the comparison between mean excursion values recorded by the device and those from previous research is encouraging. Specifically, the average scapular external rotation in our study in the transverse plane (−8°) was identical to the average scapular external rotation that Chu et al⁸  recorded using the AMC (−8°) and similar to the value noted by McClure et al³  using bone pins (−6°). In addition, the average total excursion of scapular posterior tilt recorded using the electric goniometer (18°) agreed with the average excursion that Ludewig et al²  identified using intracortical measurement techniques (18°). These comparisons with earlier findings using criterion standard measurement techniques demonstrate promising capabilities for scapular measurement in each anatomical plane during upper extremity elevation.

A strength of our study was the examination of both the intrarater and interrater reliability. Previous investigations of the reliability of clinical measurement of scapular motion are limited. Reliability values have been described^23,27  for measurements recorded by a single examiner between 2 testing sessions on the same day, whereas in other works,^12,28  the authors did not provide the time between testing sessions. We examined both the intrarater reliability of the electric goniometer across 2 testing sessions and the interrater reliability within a single testing session. The average rest position and average excursion values from the 3 trials of motion were analyzed for reliability. To minimize the risk of error between measurement techniques, we used standardized placement procedures. Our results are consistent with those of previous research²⁹  on digital goniometer measurement in finding a higher intrarater than interrater reliability, even when a standard procedure was used.

The electric goniometer was reliable for determining both scapular rest position (ICC [2,3] range = 0.692–0.874) and total excursion (ICC [2,3] range = 0.628–0.790) across an average of 9 days. The associated error was less when we measured rest position (SEM = 3°) than scapular excursions (SEM range = 2°–4°). The decrease in ICC values and increase in SEM between the rest and excursion measurements could be linked to variations in individuals' movement patterns across days. Furthermore, our same-day interrater reliability of scapular rest position (ICC [2,3] range = 0.833–0.912) across all 3 anatomical planes was higher than that reported by Watson et al²⁷  (ICC range = 0.21–0.52). Reliability for total scapular excursion was also higher in our study (ICC [2,3] range = 0.545–0.724) than in the study of Watson et al²⁷  (ICC = 0.23) during upper extremity elevation. These results demonstrating increased interrater reliability support the concept that the electric goniometer is reliable when used by multiple raters in a single testing session.

This study had limitations. The electric goniometer offers a surface-based assessment, which is affected by soft tissue obstruction and movement. As reported in the literature,^17,19,30  scapular clinical assessment is limited by the presence of soft tissue and skin movement artifact. This limitation was apparent during the measurement of frontal-plane scapular motion during upper extremity elevation; a correction factor was necessary to account for an average difference of −7°. In addition, measurement of the scapula during sagittal-plane motion may be inhibited by the posterior bunching of soft tissues during active movement. Conversely, measuring scapular motion in the sagittal plane is challenging because of difficulty palpating the scapula as it wraps around the thorax, making the prominent bony aspects on the scapula difficult to discern. As a result, we believe that either teaching videos or hands-on training using the device before implementation in practice or research may be necessary. Another limitation was the sample population. Our participants were asymptomatic and did not have any pathologic shoulder conditions at the time of the study. This highlights a constraint to the clinical application of the electric goniometer as a screening tool versus a diagnostic device based on our work. Future researchers should include patients with pathologic shoulder conditions to determine if scapular motion measured using the electric goniometer can discriminate between healthy and pathologic states.

Ultimately, the results of this investigation demonstrated that the IMU-equipped electric goniometer was a reliable and moderately valid device to measure scapular motion in each anatomical plane during upper extremity elevation in a healthy population. The degree of error associated with the device when measuring scapular-motion excursions depended on soft tissue and palpation restrictions. We recommend using a clear and defined standard operating procedure when scapular measurements are taken by more than 1 examiner. These findings provide evidence of a clinically portable and consistent device for objectively measuring scapular motion in the clinical setting.

We acknowledge Rui Chen and Denise Conway for the donation of the EasyAngle (Meloq AB) electric goniometer. We did not receive any financial or monetary gain from Meloq AB.

This work was funded in part by grant No. 1718MGP02 from the National Athletic Trainers' Association Research & Education Foundation and in part by grant No. UL1TR001998 from the National Center for Advancing Translational Sciences of the National Institutes of Health to the University of Kentucky Center for Clinical and Translational Science. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the University of Kentucky.