VALIDATION OF A NEW METHOD FOR ASSESSING SCAPULAR ANTERIOR‐POSTERIOR TILT

Abstract

Background:

Electromagnetic tracking systems have enabled some investigators and clinicians to measure tri‐planar scapular motion; yet, they are not practical and affordable options for all clinicians. Currently, the ability to affordably quantify scapular motion is limited to monitoring only the motion of scapular upward rotation, with use of a digital inclinometer.

Hypothesis/Purpose:

The objective of this study was to determine the criterion‐related validity of a modified digital inclinometer when used to measure the motion of scapular anterior‐posterior (AP) tilt.

Materials & Methods:

Thirteen volunteers, free from any history of shoulder injury, reported for a single testing session. Each subject underwent a brief shoulder and posture examination in order to confirm the absence of pathology. Subjects actively performed clinically relevant amounts of humeral elevation in the scapular plane while in a seated position. An electromagnetic tracking system (Ascension Technology, Burlington, VT) and a modified inclinometer (Pro 360, Baseline®, Fabrication Enterprises, White Plains, NY) were used to acquire scapular AP tilt over the same shoulder motions. Criterion‐related validity was determined using Pearson Product Moment correlations.

Results:

Correlation analyses revealed significant moderate to good associations (r = 0.63 to 0.86, p < 0.01) between scapular AP tilt measures obtained with a digital inclinometer and an electromagnetic tracking system.

Conclusions

A modified digital inclinometer is a moderately valid device to use for the quantification of scapular AP tilt. Further study is warranted to establish reliability and to validate use of the device in patients with shoulder injury or pathology. The modified inclinometer expands the clinician's ability to quantify scapular kinematic motion during the clinical evaluation and rehabilitation process.

Level of Evidence:

Level 3

INTRODUCTION

The use of electromagnetic tracking devices has allowed researchers and clinicians to obtain in‐vivo measures of scapular motion, expanding clinical understanding of scapular tri‐planar motion.^1‐3 In addition to gaining an understanding of how scapular rotations contribute to overall shoulder motion, investigators and clinicians are becoming more aware of the scapular movement patterns exhibited by various patients populations,⁴ including those with adhesive capsulitis,^5,6 multi‐directional instability,⁷ and rotator cuff pathologies,^8‐12 when performing scapular plane shoulder elevation. Patients with adhesive capsulitis appear to initiate scapular upward rotation earlier ⁵ and tend to exhibit greater amounts of upward rotation⁶ in comparison to healthy subjects. Conversely, subjects with multidirectional instability seem to exhibit altered scapulohumeral rhythm,¹³ including reductions in scapular upward rotation and increases in scapular medial tilt.^7,14 Evidence suggests that patients with rotator cuff pathology or shoulder imp‐ingement tend to present with alterations in all three components of scapular motion, including reductions in scapular upward rotation, decreased posterior tilt and increased medial scapular rotation.^8‐10,15 Increases in scapular and clavicular elevation, scapular upward rotation and posterior tilting in impingement patients have also been noted; yet, it has been suggested these altered kinematics may be compensations for pain.^16,17 Unfortunately, while this scapular kinematic data is available to clinicians and could be used to support clinical intervention, most clinicians are unable to monitor and quantify tri‐planar scapular motions due to the cost and lack of accessibility to the necessary equipment.

Various attempts have been made to establish clinically accessible methods that are valid and reliable for qualitatively ^18‐21 and quantitatively ^22‐27 evaluating scapular kinematics. Many of the quantitative methods have addressed the scapula's relative position, displacement and posture, and in doing so have not addressed the scapula's three‐dimensional orientation. Although techniques like the double square^28,29 and the lateral scapular slide test^25,27 have been used in clinically based studies, the validity and the clinical usefulness of these examination procedures is questionable.^26,30,31 Few studies have presented clinically accessible means of quantifying scapular angular orientation. ^22,32 Sobush et al. established the validity and reliability of the Lennie test, which examined resting scapular upward rotation position.²² However, Johnson et al³² were the first to develop a reliable and valid means of quantifying scapular upward rotation using a modified digital inclinometer to measure the scapula's angular orientation at varying degrees of humeral elevation. The inclinometer was deemed reliable and valid for assessing upward rotation in both healthy subjects and those with shoulder pathology, noting good to excellent validity (r = 0.59 – 0.92) when compared to measures obtained with an electromagnetic tracking system.³² The reliability and validity of other inclinometers have also been established for making similar upward rotation measures.^33,34 Use of the inclinometer is dependent on the clinician's ability to align the inclinometer with the spine of the scapula in order to measure scapular upward rotation associated with varying amounts of humeral elevation. The alignment results in the inclinometer being positioned relative to a perpendicular axis that is directed orthogonally to the scapula approximating the axis about which scapular upward and downward rotation occurs.³² Other authors ^34‐39 have also explored scapular upward rotation using this clinically useful technique in the shoulders of healthy individuals and symptomatic and asymptomatic overhead athletes.

Ludewig and Reynolds⁴ recently described the role that scapular anterior‐posterior (AP) tilt plays in overall scapular function, its contributions to humeral elevation, and more specifically its connection to shoulder pathologies. The accessory contributions and variable nature of medial‐lateral (ML) scapular tilt were also noted in the review. Presently, aside from the visually based observation studies designed to rate the orientation and angular motions of the scapula^18‐21,40 and those that assess overall shoulder girdle/scapular position³⁰ there are no studies that specifically measure AP or ML tilt of the scapula. Unfortunately, without the use of an electromagnetic tracking device most clinicians lack the means to quantitatively assess these significant components of scapulothoracic function, most notably AP tilt. By expanding use of a gravity‐dependent digital inclinometer to include measures of scapular AP tilt, the authors' hope to enhance the evaluative capabilities for the shoulder, which will facilitate clinicians' abilities to monitor the outcomes of their rehabilitative efforts relative to scapular neuromuscular control. While the reliability of the inclinometer to measure scapula AP tilt has been established,⁴¹ no efforts have been made to validate this instrument for measuring AP tilt. Therefore, the objective of this study was to establish the criterion‐related validity of a modified digital inclinometer when used to measure the motion of scapular AP tilt. The authors' hypothesized that a modified digital inclinometer would exhibit strong criterion‐related validity when compared to measures obtained from an electromagnetic tracking system.

MATERIALS & METHODS

Subjects

A sample of thirteen, healthy college students (21.46 + 1.13 years; 1.76 + 0.11 meters; 76.19 + 12.57 kg; 8 males, 5 females; 11 right‐handed, 2 left‐handed) volunteered to participate in the study. Pilot data was collected as part of a separate study and a power analysis was performed using the free statistical package G*Power 3⁴² to determine the sample size. The study was powered to 80%, using effects sizes ranging from r = 0.76 ‐ 0.85 and α = 0.05, resulting in a sample size of 13 participants. To be included subjects had to be between the ages of 18 and 40 years old, free from any previously diagnosed upper extremity, neck or back injury. Individuals were excluded if they 1) presented with any neurological conditions that may result in muscle weakness and consequent decreased upper extremity range of motion, including cervical disc disease or stenosis, stroke, or brachial plexus injury, 2) had experienced any trauma or injury to their dominant or non‐dominant shoulder in the last 6 months, 3) had surgery involving the neck, back, or either upper extremity, 4) suffered from rheumatoid arthritis, and/or 5) were potentially pregnant. Individuals with allergies to tape adhesives were also excluded from participating in the study.

Subjects read and signed an informed consent document approved by the University's Institutional Review Board prior to data collection. Subjects completed a questionnaire designed to ascertain demographic data and health history in order to ensure no history of neck, spine or upper extremity injury. The primary investigator performed a brief physical screen in order to determine the presence of any current shoulder injury or pathology. The primary investigator was a certified athletic trainer with 14 years of clinical experience at the time of the study. The screen involved measures of shoulder range of motion, strength, and a subset of evaluative special tests to ensure the absence of previous or current injury.⁴³ Subjects presenting with no shoulder injury history, range of motion and strength measures that were consistent bilaterally, and the absence of pain and abnormal findings for the orthopedic special tests were included. Active and passive shoulder ranges of motion (flexion, abduction, medial/lateral rotation) were assessed using standard goniometric measurement techniques.⁴⁴ A single trial was performed for each motion. Similarly, shoulder strength (flexion, abduction, medial/lateral rotation) was assessed using standard manual muscle testing techniques.⁴⁵ A handheld dynamometer (ErgoFet 300, Hogan Health Industries, West Jordan, UT) was utilized to obtain strength measures for each muscle group. Two break tests were performed for each muscle group and the mean of the two trials was calculated for each muscle group strength assessment. (Table 1)

Table 1.

Subject demographic data and physical characteristics.

Demographics
Age (years)		21.46 ± 1.13
Height (meters)		1.76 ± 0.11
Weight (kg)		76.18 ± 12.57
Body Composition (%)		16.55 ± 5.62
BMI (kg/m2)		22.79 ± 6.99
Gender		8 males, 5 females
Hand Dominance		11 right, 2 left
Scapula Posture		7 normal scapula posture
		5 medial border prominence
		1 inferior angle prominence
Range of Motion (deg)
		Active	Passive
Flexion	D	176.85 ± 6.23°	179.85 ± 0.55°
	ND	176.54 ± 7.05°	179.23 ± 1.96°
Abduction	D	179.31 ± 1.70 °	179.85 ± 0.38°
	ND	179.31 ± 1.70°	179.23 ± 0.28°
Laternal Rotation	D	107.15 ± 7.71°	129.46 ± 12.38°
	ND	101.92 ± 10.81°	117.54 ± 11.45°
Medial Rotation	D	53.00 ± 12.40°	61.38 ± 13.54°
	ND	56.15 ± 9.01°	66.85 ± 11.44°
Normalized Strength (%)
Flexion	D	19.29 ± 4.80
	ND	19.18 ± 4.26
Abduction	D	18.53 ± 3.90
	ND	17.93 ± 4.72
Medial Rotation	D	17.25 ± 4.43
	ND	17.43 ± 3.56
Lateral Rotation	D	12.54 ± 3.37
	ND	12.72 ± 2.39

Means and standard deviations are provided for each characteristic. BMI=body mass index; D=dominant; ND=non‐dominant. Strength measures were obtained using a hand held dynamometer (ErgoFet 300, Hogan Health Industries, West Jordan, UT). All strength measures were normalized to the body weight of the subject and are recorded as a percentage of body weight.

To help characterize the physical presentation of the subjects, each underwent a brief postural and scapular movement screen as described by Kibler et al.¹⁹ The position and orientation of the scapula was observed with subjects standing and when performing humeral elevation, starting with their hands at their sides and moving their arms overhead. Noteworthy scapular positions included 1) excessive superior placement or translation, 2) inferior angle prominence, 3) medial border prominence and/or normal scapular appearance.¹⁹ Although some subjects presented with dyskinetic scapular movement patterns, all shoulders were classified as non‐pathologic due to the lack of previous or current shoulder injury. (Table 1)

Instrumentation

A digital inclinometer (Pro 360, Baseline®, Fabrication Enterprises, White Plains, NY) was used to assess scapular AP tilt during humeral elevation trials. A series of modifications were made to the inclinometer to make it suitable for measuring AP tilt. Specially designed wooden feet were used with the inclinometer in order to appropriately align the inclinometer with scapula bony landmarks. Each “foot” measured 5.3cm × 1.0cm × 1.0 cm and included a blunted end used to align the scapular bony landmarks. A custom‐made plate measuring 7.0cm × 20.5cm (Lexan*, SABIC Innovative Plastics, Pittsfield, MA) was attached to the inclinometer, allowing for adjustable spacing of the “feet” (1 cm increments). (Figures 1A & B) Each “foot” was attached to a 3.5cm corner brace and affixed to the Lexan plate using standard hardware. The modifications allowed for alignment with a vertically oriented axis passing through the intersection of the spine and medial border of the scapula and the inferior angle of the scapula. Using palpation, the investigator was able to align the inclinometer with these two points (Figure 2). Strong intrarater reliability (ICC_3,2 = 0.93 – 0.99) using this device was established for the primary investigator as part of a separate study.⁴¹

Figure 1.

(A) Modified digital inclinometer, (B) Orientation of inclinometer relative to scapula. The digital inclinometer (Pro 360, Baseline®, Fabrication Enterprises, White Plains, NY) was oriented vertically, in plane with the dorsum of the scapula, and was aligned with the inferior angle of the scapula and the intersection of the scapular spine and medial border.

Figure 2.

Scapular alignment points for the inclinometer.

Three dimensional scapular kinematic data for validation of the digital inclinometer were collected using an electromagnetic tracking system (Ascension Technology, Burlington, VT), and Motion Monitor software (Innovative Sports Training, Chicago, IL). The system's short‐range transmitter was rigidly fixed to metal free scaffolding to minimize interference with the system. Three sensors were used to track motion of the trunk, humerus and scapula. The trunk sensor was affixed to the subject's cervical spine (proximal to the spinous process of C₇) using double sided tape and reinforced with 1” athletic tape. The scapula sensor was mounted on the acromion process in a similar fashion; while the humeral sensor was strapped proximally on the humerus with the manufacturer provided strapping system. (Figure 3) Using twelve digitized landmarks, local anatomic and reference coordinate systems approved by the International Society of Biomechanics were constructed in order to calculate position, orientation and displacement of the scapula and humerus.⁴⁶ Kinematic data were collected at 100 Hz.

Figure 3.

Subject testing set up with the electromagnetic tracking system.

The accuracy of the electromagnetic tracking system and the reliability of the testing procedures have been established for the laboratory in which this study was conducted and have been published previously.⁴⁷ For system accuracy, the root mean squared error for linear position, angular orientation and linear displacement are 5.3mm, 0.23°, and 3.1mm, respectively in this laboratory. While values for linear displacement have not been provided by the manufacturer or in the literature, the position and orientation values documented by the laboratory are similar to previously reported values.⁴⁸ The ICC _(3,1) values for intrasession reliability ranged from 0.812 to 0.999 for glenohumeral motion (elevation, plane, rotation) and 0.983 to 0.999 for scapular motion (AP tilt, ML tilt, upward/downward rotation). Validation data using the tracking system in the laboratory suggested that the equipment is valid for capturing human motion (r = 0.85 to 0.99); however, the system routinely under‐represented the amount of humeral elevation being achieved.⁴⁷ Similar errors have been noted previously in the literature.⁴⁹ Therefore, when reporting findings for this study, relative values as opposed to absolute values of motion were reported.

Patient Preparation and Data Collection

Subjects were seated on a plastic stool one arm's length from a stationary screen. Subjects were seated to minimize compensatory changes of the lower extremity and trunk, which could impact shoulder biomechanics. The subjects were asked to move their hands to selected points along the screen that corresponded to specific shoulder ranges of motion (rest, 30°, 45°, 60°, 90°, 120°) in the scapular plane. Shoulder ranges of motion were monitored with a digital inclinometer and hand placement was marked on the screen with respect to the specific shoulder ranges of motion, to ensure repeatable arm placement throughout the data collection. Data collection occurred in the plane of the scapula. All testing involved the dominant shoulder.

The electromagnetic sensors then were attached using double‐sided tape and straps in order to secure the sensors to the subjects. After affixing the sensors, a separate sensor was used to digitize a series of 12 anatomical landmarks in order to create local anatomic and reference coordinate systems.⁴⁶ The landmarks included, 1) spinous process of C₇, 2) spinous process of T₈, 3) spinous process of T₁₂, 4) jugular notch of the sternum, 5) xiphoid process, 6) coracoid process, 7) AC joint, 8) posterolateral corner of the acromion, 9) inferior angle of the scapula, 10) intersection of medial border and spine of the scapula, 11) medial epicondyle and 12) lateral epicondyles of the humerus.

Testing order was randomized by arm position and was maintained for both inclinometer and electromagnetic tracking system measures. Subjects began with the dominant hand at his/her side and were then asked to move the hand to a selected position and to hold that position while kinematic data were acquired. Three trials were performed for each shoulder position with both the inclinometer and the tracking system (rest, 30°, 45°, 60°, 90°, 120° of humeral elevation). These shoulder positions were selected as they are commonly reported in the shoulder biomechanics literature. Additionally, a single anatomical position trial was captured with both instruments for use during data processing. Each trial lasted 10‐15 seconds, and subjects were provided with 5‐10 seconds rest in between each trial and an additional one minute of rest in between each set of three trials. For inclinometer bony landmark orientation, the primary investigator relied on palpation to identify the position of the inferior angle, medial border and spine of the scapula for each trial. The primary investigator was responsible for all measures taken with the inclinometer, but was blinded while aligning the inclinometer. The inclinometer values were recorded by a laboratory assistant after each trial and the inclinometer was repositioned for each trial. The primary investigator was also blinded to the electromagnetic tracking system data with respect to scapula angular values during each testing session. All trials were completed first with the electromagnetic tracking system, and then followed by the digital inclinometer. A total of six trials were completed at each shoulder position; three using the electromagnetic tracking system and three with the inclinometer. Kinematic data collection using both measurement systems did not occur concurrently within each trial due to the documented distortion of the electromagnetic field caused by the introduction of metal in the testing environment.^50‐52

Data processing

Humeral and scapular kinematics were computed from the MotionStar sensor data using the Euler angle sequences approved by the ISB.⁴⁶ A Butterworth filter with a cutoff frequency of 8 Hz was used to smooth the data.⁵³ Tracings and values associated with humeral elevation positions were observed visually in Motion Monitor to confirm stability of the data. Stability of the data was defined by fluctuations of < 1° for a minimum of 50 consecutive data points (0.5 seconds). The Motion Monitor processed data was then exported into Excel (Microsoft Corporation, Redman, WA) where means for the humeral and scapular kinematics were calculated. The scapular kinematic means were then used for the validity analyses.

The assessment of criterion‐related validity was based upon relative values obtained from the electromagnetic tracking system and inclinometer as opposed to absolute measurements. To compute relative scapular values or the change in scapular orientation from baseline or anatomic position, the scapular values obtained during the single anatomic position trial were subtracted from the values observed at rest, 30°, 45°, 60°, 90°, 120° of humeral elevation for both measurement devices during each of the three trials. The means of the three trials for each subject at each position of elevation were then calculated and used during data analysis. Using this approach to quantify scapular orientation enabled the calculation of the relative change in scapular orientation and accounted for differences in the measurement devices. While the scapular orientations obtained with the inclinometer are based upon the inclinometer's position and gravity, orientations of the scapula obtained with the electromagnetic tracking system are based upon the relative position of the scapula with respect to the thorax.

Data analysis

Means and standard deviations were calculated for the demographic data. Criterion‐related validity was assessed via Pearson Product Moment correlation (r) analyses where orientation data of the inclinometer and the electromagnetic tracking system were compared. Coefficients of determination (r²) were used to evaluate the clinical meaningfulness of the relationships. The upper and lower limits of agreement (+ 1.96 standard deviation) were calculated for the paired mean differences (inclinometer vs. electromagnetic tracking system) in order to examine the relationship between measurement error and the mean of the two tests.⁵⁴ Additionally, intrarater reliability was calculated for the data set using intraclass correlation coeffecients (ICC_3,1). SPSS version 17.0 was used for the statistical analyses and an α = 0.05 was used for all statistical analyses.

RESULTS

The assessment of intrarater reliability for the primary investigator revealed strong repeatability for this data set. Within session reliability at each humeral elevation was found to be excellent with ICC_3,1 ranging from 0.97 to 0.99.

The criterion‐related validity analysis revealed significant moderate to good correlations (r = 0.63 to 0.86, p < 0.01) across several combinations of humeral elevation angles consistent with the shoulder biomechanics literature. (Table 2) The validity measures were considered to be moderate to good based upon the range described by Portney and Watkins where 0.00 to 0.25 represents no or little relationship; 0.25 to 0.50 represents a fair relationship; 0.50 to 0.75 signifies a moderate to good; and values > 0.75 considered good to excellent⁵⁵. An examination of the slope coefficients revealed slopes that were equal to 0.52 + 0.06. The coefficients of determination (r² = 0.40 to 0.74) revealed that 40‐74% of the error variance can be accounted for by common factors involving the two devices. Measures of scapular AP tilt obtained with the inclinometer were consistently greater than those obtained with the electromagnetic tracking system.

Table 2.

Comparisons of anterior‐posterior tilt measures between a modified digital inclinometer and electromagnetic tracking system.

Humeral Elevations	r*	Slope	Intercept	MD (degrees)	95% Limits of Agreement
Rest, 30°, 45°, 60°, 90°, 120°	0.63	0.46	0.41	4.45	‐16.79, 7.90
Rest, 30°, 60°, 90°, 120°	0.67	0.48	0.48	4.32	‐16.66, 8.03
Rest, 60°, 90°, 120°	0.69	0.52	0.32	4.75	‐17.09, 7.58
Rest, 90°, 120°	0.79	0.57	0.66	3.92	‐15.57, 7.72
Rest, 120°	0.86	0.57	0.70	3.66	‐15.85, 8.52

MD=the mean difference between inclinometer and electromagnetic tracking measures, reported in degrees; r=correlation coefficient.

^*All correlations were significant at the p < 0.01 level.

The inclinometer and electromagnetic tracking system AP tilt values reflect a gradual increase in scapular posterior tilt as greater amounts of humeral elevation are achieved. (Table 3) The limits of agreement analysis indicated that the mean of the paired differences for the two devices ranged from 3.67° + 6.21° to 4.75° + 6.29° when looking across all humeral elevation angles for each combination of humeral elevations. The 95% upper and lower limits of agreement resulted in a +12° envelope with respect mean differences for each combination of humeral elevations. (Figure 4) Although one subject's data fell below the lower limit of agreement, the analysis revealed moderate agreement between the two devices as data from all other subjects routinely fell within the limits of agreement.

Table 3.

Anatomical Neutral and Relative Scapular Anterior‐Posterior Orientations.

Active Humeral Elevation Position	Digital Inclinometer Measures		Electromagnetic Tracking System Measures
	Anatomic Neutral	Relative	Anatomic Neutral	Relative
	68.68°		‐11.75°
Rest		0.00°		0.00°
30°		4.45°		1.88°
45°		7.38°		2.29°
60°		9.98°		2.75°
90°		12.00°		7.56°
120°		20.06°		12.73°

Note: Values reflect the mean relative change in scapular AP tilt for all subjects at each humeral elevation position from anatomic neutral for the digital inclinometer and electromagnetic tracking system. The negative anatomic neutral trial for the electromagnetic tracking system reflects an anteriorly tilted scapula. Increases in relative values reflect movement of the scapula to a more posteriorly tilted orientation.

Figure 4.

Bland & Altman plot depicting the 95% limits of agreement between devices for scapular AP tilt measures for the humeral elevation positions rest, 30°, 45°, 60°, 90° & 120°.

DISCUSSION

The comparison between an electromagnetic tracking system and a modified digital inclinometer revealed moderate to good criterion‐related validity for the measurement of scapular AP tilt using the inclinometer. Further analysis also demonstrated that this novel device was able to account for 40%‐74% of the error variance between devices, which was also reflected in the limits of agreement assessment. Although some of the current statistical findings were not as strong as those identified for use of the modified digital inclinometer for measuring scapular upward rotation,³² the results are valuable as they reflect an initial attempt at establishing an affordable and clinically viable means of assessing scapular AP tilt.

The current effort to establish the inclinometer as a valid tool for assessing scapular AP tilt comes at a time when considerable emphasis continues to be placed on the importance of understanding tri‐planar scapular kinematics and their relationships with shoulder pathologies.⁴ During normal humeral elevation a combination of scapular upward rotation, posterior tilt, and lateral rotation occur.^3,4 It has been suggested that limitations in these scapular motions can alter glenohumeral joint kinematics and stability, and ultimately contribute to the development of shoulder pathologies.⁴ Although the involvement of scapular AP tilt is less apparent in patients presenting with multidirectional instability and adhesive capsulitis, studies have noted relative reductions in posterior tilt in patients presenting with shoulder impingement during assessments of humeral elevation.^8‐10,15 In these reports the noted differences were relatively small in comparison to healthy control subjects; however, given the limited space beneath the coracoacromial arch, small variations in scapular kinematics could place the subacromial structures at an increased risk of injury. Interestingly, McClure et al¹⁷ identified increases in scapular posterior tilt and upward rotation when assessing patients with subacromial impingement. Laudner et al¹⁶ also noted increases of posterior tilt and scapular elevation; however, this was observed in patients with internal impingement. Regardless, Laudner et al¹⁶ and McClure et al¹⁷ suggested that these increases in scapular motion were compensatory adjustments made to reduce soft tissue loading associated with the respective forms of impingement. Having validated the inclinometer for use when measuring scapular AP tilt in the present study, clinicians will now be able to readily compare patient findings with the data from these clinically‐based studies and better direct rehabilitation efforts.

The data revealed moderate to good criterion‐related validity when analyzing the data across all subjects, using selected combinations of commonly reported positions of humeral elevation. While an electromagnetic tracking system does allow for static and dynamic in‐vivo scapular kinematic assessment, a static assessment appeared to be more prudent given the manner in which the inclinometer had been modified and would be used clinically. Furthermore, previous work by Johnson et al³² found that the strengths of the relationships between the inclinometer and electromagnetic tracking system were stronger when comparing to the static versus the dynamic electromagnetic tracking system measures. Overall, the slope coefficients associated with the observed relationships suggest positive relationships between measures taken with both devices. The slopes of the humeral elevation combinations fell within the range of 0.46 to 0.57, suggesting that as the ‐modified inclinometer detected 1° of change in scapular AP tilt, the electromagnetic tracking system recorded approximately a 0.50° increase. Although these findings are lower than the findings reported by Johnson et al³², they are consistent with an electromagnetic tracking system validation study that noted under‐reporting of shoulder joint range of motion angles when comparing tracking system data with standard goniometric measures.⁴⁷ Although the mean differences between the devices were higher than desired, the consistency of the slopes, slope intercepts and mean differences can be attributed to systematic differences. The authors would suggest that three factors: differences in measurement systems, palpation error, and morphological variability, may have contributed to the variations in these values. The shear comparison between a 2‐dimensional, gravity‐dependent measurement system, generating Euclidean angles with an electromagnetic tracking system that results in Euler angle data may have consistently contributed to the differences observed between the measurement devices. Similarly, alignment of the inclinometer was based upon the inferior angle and scapular spine, while the tracking system data for the scapula was obtained by placing a sensor on the acromion process and relied on a local coordinate system whose origin was the posterolateral aspect of the acromion.⁴⁶ While both systems are able to generate angular measures, the basis of their calculations may have resulted in some inherent discrepancies in the actual angular measurements. The role of palpation and landmark identification has been examined on multiple fronts as it relates to discrepancies in shoulder kinematic descriptions.^51,52,56,57 Recent work involving computational modeling has shown that shoulder kinematic descriptions are sensitive to select anatomical landmarks and that uncertainty or error in landmark identification can introduce variability into these shoulder kinematic descriptions.⁵⁷ Based upon in‐vivo data collected using an electromagnetic tracking system, de Groot⁵⁶ noted that palpation errors result in approximately 2° kinematic description inaccuracy. However, de Groot also suggests that a greater amount of variability and error is introduced into the description due to inter‐subject morphology.⁵⁶ Others suggest that movement characteristic, subject demographics and the interplay between skin‐based sensors and soft tissue artifact may warrant consideration relative to electromagnetic tracking system accuracy.^49,58 While there may be a design limitation in the inclinometer that does not sufficiently control for morphological variability and that some inherent error in our “gold standard” exists, having one investigator responsible for all study‐related palpations likely helped to minimize systematic error in the kinematic descriptions that could be attributed to both palpation error and inter‐subject morphology.

The data processing and analysis approach used to assess scapular AP tilt was unique when compared to the approach that others^32,33 have used to quantify scapular kinematics. Electromagnetic tracking systems have been established as the gold standard for scapular kinematic assessment; yet, the authors' experience would suggest that examining these relationships would be best served by examining relative scapular orientations over an arc of select humeral elevations as opposed to absolute scapular orientations at the select points of humeral elevation.⁴⁷ Therefore, the decision was made to examine the ability of the inclinometer to track scapular AP tilt across the selected ranges of humeral elevations. While for clinicians it is important to examine how segment or joint motion may change at selected intervals, it is equally important to examine the extent to which the scapula would tilt over an arc of humeral elevation. Likewise, clinicians evaluate discrete skills and comprehensive functional abilities in patients, and although they often attempt to address scapular kinematics at discrete points of humeral elevation, realistically during functional activities the scapula rarely remains fixed. Ultimately, although this study incorporated a unique data analysis approach, the study findings support the use of this clinically available and novel tool/technique for measuring scapular AP tilt.

When considering the values obtained for scapular AP tilt, the measures are consistent with AP tilt values described in previous studies.^3,59,60 Although strong relationships were observed between measurements from the inclinometer and the tracking system, additional analyses indicated that further refinement of the inclinometer may be necessary to enhance the agreement between devices. The coefficients of determination revealed that depending on the combination of humeral elevation angles, 40‐74% of the error variance can be accounted for by factors common to both devices. An assessment of the limits of agreement between the devices further exposed the disparity between the devices as it relates to comparably quantifying scapular AP tilt. Again, inaccuracies in the tracking system's ability to assess shoulder kinematics, the influence of palpation error, and certainly morphological variability could all have contributed to the noted error observed in both systems. Variability associated with scapular medial rotation^3,61 may have contributed to some of the weaker findings as well, particularly during early phases of humeral elevation. It is unclear whether similar issues may have been present during validation of the digital inclinometer for use with scapular upward rotation as this data for the lower ranges of humeral elevation were not presented. Given the noted differences and confounding factors one design modification to consider could include use of a third foot to facilitate inclinometer alignment within the scapular plane. Johnson et al³² used a bubble level to minimize axial rotation of the inclinometer. Unfortunately, given the required positioning of the inclinometer to obtain scapular AP tilt, use of a bubble level was inappropriate. While the primary investigator did attempt to keep the base of the inclinometer oriented with the plane of the scapula during testing, the addition of a third contact point on the scapula may improve inclinometer orientation relative to the scapular plane. Given inter‐subject morphological variability and the displacement of the inferior angles as the scapula achieves greater upward rotation during humeral elevation, a third contact point may enhance the accuracy and precision of scapular AP tilt measures.

One of the limitations of the current study was that it did not include subjects with shoulder pathologies, limiting the generalizability of our validity findings. To fully elucidate the usefulness of the inclinometer in patients with shoulder pathologies additional validation studies are necessary. A comparison between subjects with and without scapular dyskinesis may also be useful. Although many claim a connection between scapular dyskinesis and shoulder pathology, few have been able to establish this relationship. While some subjects were identified as demonstrating dyskinetic scapular movement patterns using a validated instrument,¹⁹ the design of the present study was not appropriate to address this question. Since the initiation of our study Uhl et al²¹ and Ellenbecker et al⁴⁰ have examined the validity and reliability of the dyskinesis evaluation system described by Kibler et al,¹⁹ while others have established the reliability and validity of an entirely new dyskinesis assessment scheme.^18,20 Any attempts to determine the usefulness of the inclinometer in assessing subjects with dyskinesis should take the evolution of these dyskinesis assessment pieces into consideration. Furthermore, although the large limits of agreement envelopes can be partially attributed to differences between measurement devices and a small sample size, further refinement of the device should be considered.

CONCLUSION

The findings of the current study suggest that a modified digital inclinometer is a moderately valid instrument to use when attempting to quantify scapular AP tilt. The current study represents the first to attempt to validate the use of this instrument as a clinically accessible and affordable means of quantifying scapular AP tilt. Expanding the exploration of the validity and reliability of this instrument to measure AP tilt in a patient population is critical. However, the authors' believe that further work in this area will ultimately provide clinicians with an accessible means of quantifying scapular motion both during clinical evaluations and throughout the injury rehabilitation process. Clinicians will when then have another means by which they can monitor and document patient progress with respect to scapular kinematics.