Abstract
Background The protractor method is a proposed clinical assessment tool, the first to measure vertical scapular position, that directly compares scapular and spinal landmarks. This tool has the potential to reliably and accurately measure excessive scapular elevation or depression.
Objective The purpose of this study was to determine reliability and validity of the protractor method to measure resting scapular position.
Design An interrater and intratester reliability and validity study was conducted.
Methods Testing was conducted on the same day by 2 physical therapists who were blinded to each other's results. The vertical distances between the spinous process of C7 and the superior margin of the medial aspect of the spine of the scapula (C7 method) and the spinous process of T8 and the inferior angle of the scapula (T8 method) were palpated and measured on the symptomatic shoulder in 34 people with current shoulder pain using the protractor method. Measurements were compared with 2-dimensional camera analysis to assess validity.
Results For intertester reliability, the standard error of measure, minimal detectable change, and intraclass correlation coefficient were 6.3 mm, 17.3 mm, and .78, respectively, for the C7 method and 5.7 mm, 15.7 mm, and .82, respectively, for the T8 method. For intratester reliability, the standard error of measure, minimal detectable change, and intraclass correlation coefficient were <0.9 mm, <2.5 mm, and .99, respectively. For validity, significant correlations (r) and mean differences were .83 and 10.1 mm, respectively, for the C7 method and .92 and 2.2 mm, respectively, for the T8 method.
Limitation The results of this study are limited to static measurement of the scapula in one plane.
Conclusion Both protractor methods were shown to have good reliability and acceptable validity, with the T8 method demonstrating superior clinical utility. The clinical use of the T8 method is recommended for measurement of excessive resting scapular elevation or depression.
The 1-year worldwide prevalence of neck and shoulder pain has been reported as ranging from 16.7% to 75.1% and 4.7% to 46.7%, respectively.1,2 When taken together, these symptoms have been shown to be the most common causes of musculoskeletal pain, resulting in large costs to the health sector, as well as indirect costs through work disability.3 In order to decrease these costs and to provide comprehensive rehabilitation for these patients, clinicians first need to be able to reliably and accurately measure factors related to shoulder dysfunction, such as resting scapular position, which is considered an important factor in the assessment of these conditions.4,5
Scapular position has been linked to rotator cuff pathology, impingement, and shoulder dysfunction, with particular reference to the potential impact of scapular elevation position on shoulder pain.6–9 Increased scapular elevation position has been described as being more likely to be found in people with full-thickness rotator cuff tears than in people who are asymptomatic.9 Lin et al7 found greater elevation of the scapula with a functional task in individuals with shoulder dysfunction compared with people without dysfunction. People with impingement also were found to have higher scapular elevation with arm elevation compared with people who were asymptomatic.6,8 Although there is a greater body of evidence for the potential involvement of scapular elevation in shoulder pathologies, scapular depression also has been associated with shoulder and neck pain, such as with thoracic outlet syndrome. A study by Azevedo et al10 showed that healthy young individuals with a depressed scapula had significantly lower pressure pain thresholds in the upper trapezius muscle compared with individuals with normal scapular position.10 In summary, clinicians should expect that there will be a difference between symptomatic and asymptomatic shoulders on examination of the scapula. Additionally, the nondominant scapula may be elevated in relation to the dominant side, by approximately 5 mm.11 Therefore, it would seem important to reliably and accurately measure scapular elevation or depression position.
Longitudinal studies also have shown that that scapular position or dyskinesis is important and can be predictive of shoulder pain,12–14 although not all studies concur.15,16 However, these longitudinal studies have not incorporated the measurement of elevation. Due to conflicting evidence, establishing a reliable and valid means of measuring scapular elevation is needed for further large prospective studies to help clarify the causal relationship of scapular position to pain.
Several laboratory methods have been used to measure scapular position, including radiography,17 Moiré topography,18 infrared/visual spectrum motion analysis system,19 electromagnetic tracking systems,20 and electromechanical digitizers.21 Although these methods have been shown to have high precision, they are not practical in clinical use.17 Previous authors have utilized a number of different clinical tools, such as the Palpation Meter (Performance Attainment Associates, St Paul, Minnesota),22 goniometer,23 inclinometer,24 sliding caliper,25 string measurement,26 and tape measure.25,27 Angular measurement tools, such as a goniometer or inclinometer, cannot adequately determine a linear distance, such as elevation of the scapula. Additionally, linear measurement tools, such as a tape measure, sliding caliper, string measurement, or the Palpation Meter cannot provide information on elevation or depression of the scapula, as there are no easily identifiable and consistently placed bony landmarks directly above or below prominent scapular landmarks.
The Lennie method11 has been the only measure to examine the elevation or depression position of the scapula. However, this method is limited because it is does not determine whether the scapula is elevated or depressed relative to a set spinal landmark, as it is purely a comparison with the opposite side. As both scapulae could potentially sit in elevation or depression, assessment that compares side-to-side position may minimize or exclude absolute changes in scapular position. An important aspect of measuring scapular position is to measure the vertical distance between a scapular landmark and a spinal landmark. This measurement would potentially provide a repeatable and valid measure of position, as the position of the opposite scapula may vary, but a vertebral body such as C7 would remain relatively unchanged. It appears that there are no previous studies that have determined the clinical utility of a clinical method that examines scapular elevation or depression in a resting position compared with a fixed spinal landmark.
Therefore, research is needed to determine a reliable and accurate method of measuring scapular elevation and depression position. The protractor method could be a useful tool in measuring elevation and depression of the scapula, both clinically and in future research. It is a compact and low-cost measurement tool that we developed using a metal right-angle protractor. The protractor method has the potential to enhance physical therapists' ability to assess and monitor neck and shoulder pain, with the overarching aim of improving outcomes for these conditions.
Although there may be a multitude of proposed methods for validating a measurement tool such as this, there has yet to be an agreed-on reference standard method of measuring scapular position in relation to spinal landmarks. The use of real-time ultrasound may be a potential method of measurement, but the length or width of the probe cannot measure the distance between the spine and the scapula in one field of view. Plain film radiographs are subject to magnification, true distortion,28 and scapular shadow29 errors. Thoracic rotation away from or toward the x-ray film may cause parallax error and, therefore, cannot be used as a reference standard. Computerized tomography, as described by Ho et al,30 exposes people to unethically high levels of radiation and was not appropriate for this study. Closed magnetic resonance imaging (MRI) utilizes a supine patient position, which does not provide a participant position similar to that used clinically when assessing the scapula. Three-dimensional (3D) motion analysis was not used in this study due to its questionable accuracy in the measurement of the scapula relating to excessive skin movement, with the lack of ethical feasibility for the alternative of using bony pins to identify landmarks. As the clinical measure used in this study was a 2-dimensional (2D) measure, 2D camera equipment was chosen as the validating reference standard.
The primary aim of this study was to evaluate the intertester reliability and validity of a protractor method of measuring resting elevation or depression of the scapula relative to a spinal landmark. The secondary aim was to determine which of the 2 sets of bony landmarks used (C7 or T8 method) is the most reliable and accurate measurement of scapular position for clinical use.
Method
Study Design
This was a reliability and validity study consisting of 2 parts: (1) an intertester and intratester reliability assessment using 2 investigators (A.O., R.K) and conducted in 1 day using the protractor method and (2) an assessment of the validity of this method compared with 2D analysis.
All testers were qualified physical therapists with more than 4 years' experience each in musculoskeletal physical therapy diagnosis and patient management. Both investigators underwent identical familiarization and practice trials of the techniques on 5 participants over approximately 5 hours (pilot test). Examiner training included testers observing each other and discussion of reasons for differences between the testers in the pilot test until an agreed-on method was determined. Pilot testing confirmed which landmarks to use during the study. Despite using the same standardized procedures, results from the pilot testing indicated that there were differences in determining the exact location of the landmarks (especially for the spine of scapula) between testers. To limit the effect of these differences on the validity between the 2 measurement methods in the present study, the testers identified the appropriate landmarks for each participant from which to use the protractor method, and each tester then photographed the participant with his or her own identified landmarks for later 2D analysis.
Participants
Participants with current shoulder pain were recruited through community-based physical therapy practices, local radio advertisement, and flyers in local community centers and sports clubs. Testing was carried out on the symptomatic side, with dominance recorded (Tab. 1). Participants were included if they were over 18 years of age, had a body mass index (BMI) of less than 30 kg/m2, and had current shoulder pain that was elicited by shoulder movement but not provoked by cervical or thoracic movements. Prospective participants were excluded from the study if they had “pins and needles” or numbness in any distribution in the upper limb, pain above the level of C7 or below the elbow, current thoracic pain, or suspected glenohumeral joint instability on screening.
Participant Characteristics (N=34)a
A priori analysis suggested that a minimum of 34 participants were needed to achieve 80% power to detect a significant difference of 1 cm (SD=2.2) between the protractor method and the 2D camera analysis at an alpha level of .05.31
Procedure
Pretesting.
Prior to testing, all participants provided written informed consent. Participants completed a questionnaire, which inquired about date of birth, hand dominance, and shoulder pathology. They also completed the Disability of Arm, Shoulder and Hand (DASH) questionnaire.32 Height and weight were measured by the same tester for each participant. Tester and method order were randomized prior to testing, with each test carried out in a closed-curtained cubicle in the testing area to ensure blinding between testers. Participants wore attire that completely exposed the posterior thoracic region while covering the front of the torso. Participants were seated on an adjustable stool, in normal relaxed sitting, with their arms by their side and thumbs facing forward while looking forward at a mark on the wall fixed at eye level. During pilot testing, we identified the potential for postural fatigue during testing. Such fatigue could lead to participants being unable to maintain an erect upright posture throughout testing, possibly leading to a participant sitting in a different posture at the beginning of testing from when seated for the final measurement taken by the second tester later in the study. Such changes in posture would lead to significant differences in results for both reliability and validity, as distances between landmarks would differ with posture. As a result, participants were provided with breaks between measurements and were consistently asked to sit in their own relaxed sitting posture prior to each measurement being taken.
Opsite (Smith & Nephew Medical Ltd, Hull, United Kingdom) (5 cm) was applied to each participant in the general area of the landmarks that were palpated. The use of Opsite ensured easy removal of marks on the skin, thus enhancing tester blinding, as no rubbing marks or incompletely removed marks were left on the skin.
Testing.
For each participant, the first tester performed the pretesting procedures, identified landmarks, used the protractor method, and then removed landmarks. The first tester then repeated the landmark identification and protractor method twice more before taking a photograph of his own marked landmarks. The second tester repeated this process, except for the pretesting procedures.
Landmark identification.
Landmarks included for palpation in this study were the spinous processes of C7 and T8, the superior margin of the medial aspect of the spine of the scapula, and the inferior angle of the scapula. Previous literature has indicated that C7 and T8 are valid landmarks.33 Haneline et al34 determined that the mean spinal level corresponding to the inferior angle of the scapula was midway between the T8–T9 interspace and the upper body of T9. Identification of the spinous process of C7 was determined as the most caudal cervical spinous process that did not disappear on cervical extension,33,35 and the tester then counted down the vertebra to identify the spinous process for T8. For all landmarks (except the superior border of the scapula), the most inferior portion of the bone was palpated, then soft tissue was released and a horizontal line placed along the most inferior aspect of the landmark. Next, a vertical line was drawn through the most central point of the landmark. These 2 lines created a crosshatch, and the intersection of these lines was used as a single point for measurement. For the C7 method, bony landmarks marked by the testers comprised the most inferior aspect of the spinous process of C7 and the superior margin of the medial border of the spine of the scapula (Fig. 1). For the T8 method, the most inferior aspect of the spinous process of T8 was marked, as well as the inferior angle of the scapula (Fig. 2).
Photograph showing the alignment of the protractor for the C7 method. The dotted line indicates the vertical distance between the superior margin of the medial aspect of the root of the spine of the scapula (lower crosshatch) and the inferior aspect of the spinous process of C7 (upper crosshatch).
Photograph showing the alignment of the protractor for the T8 method. The white line indicates the vertical distance between the inferior angle of the scapula (upper crosshatch) and the inferior aspect of the spinous process of T8 (lower crosshatch).
Protractor method.
For both methods, a metal right-angle protractor with a 20-cm side length and a scale that began from the edge of the metal was used (Empire carpenter square, Mukwonago, Wisconsin). A protractor that has at least 16-cm side length was required, as previous data indicated that the inferior angle of the scapula is, on average, 10.1 cm (range=6.7–15.3 cm) laterally from the spine.36 We also attached 2 simple spirit levels to the protractor to ensure that level measurements were assessed. For both methods, a vertical distance was identified between landmarks using the protractor. For the C7 method, the protractor was aligned so that the bottom of the protractor lay horizontally between the root of the spine of the scapula and the spine. The vertical distance was read from the bottom of the protractor to C7 (dotted line in Fig. 1). The tester recorded this measurement 3 times before moving to the next method. For the T8 method, the protractor was again placed horizontally, aligning between the inferior angle of the scapula and the spine. The vertical distance was read from the bottom of the protractor to T8 (white line in Fig. 2). Measurements were recorded in the same manner as for C7, with mean measurements recorded.
2D analysis.
Participants were instructed to sit in the same manner as previously described. Data were collected using a digital camera (Exilim EX-ZR800, Casio, Tokyo, Japan). Scapular position was examined in posterior view only, with all landmarks visible.
Data Processing and Reduction
Two-dimensional analysis was performed with Siliconcoach Live software (The Tarn Group Ltd, Dunedin, New Zealand) using the anatomical landmarks described above and the Siliconcoach Live digital vertical measurement tool. This analysis was done by a third tester (L.M.), who was blinded to all other results. Tester randomization order was maintained, with the C7 measurement taken first. Photographs were viewed in full screen mode to increase pixel accuracy. The vertical measurement tool was used in smallest width line, with contrasting color to allow for greatest accuracy.
Data Analysis
Statistical processing was performed using IBM SPSS for Windows version 22.0 (IBM Corp, Armonk, New York) and the level of significance was set at α=.05. Means and standard deviations for the demographic and protractor method data were calculated.
To assess reliability, standard errors of measurement (SEMs), minimal detectable changes (MDCs), intraclass correlation coefficients (ICCs), and paired-sample t tests were ascertained for the average tester scores. The SEM were calculated from the square root of the mean square error term of a repeated-measures analysis of variance as previously described.37 Minimal detectable change was calculated using the formula (1.96 × √2 × SEM).37 The ICCs with corresponding confidence intervals were calculated using the 2,k model for intertester reliability and the 2,1 model for intratester reliability. For interpretation of the ICCs, reliability coefficients less than .50 indicated poor reliability, values ranging from .50 to .75 indicated moderate reliability, values ranging from .75 to .90 indicated good reliability, and values greater than .90 indicated excellent reliability.25
Validity was examined for agreement between the protractor methods and 2D camera analysis by calculating correlations, the mean difference (with limits of agreement), and SEMs between these measurements. Bland-Altman plots also were constructed. The Bland-Altman plots are used to calculate the mean difference between 2 methods of measurement, with 95% limits of agreement as the mean difference.38 The smaller the range between the 95% limits of agreement, the better the agreement.39
Results
The demographic data of the participants (N=34) are presented in Table 1. There were a greater number of male participants than female participants, with most individuals reporting shoulder pain in their dominant side. The participants' ages ranged from 23 to 71 years, with length of time of shoulder pain ranging from 1 month to 50 years. The DASH questionnaire scores ranged from 0.83 to 39.16, indicating mild to moderate levels of disability.
Using the protractor method, the average values for the vertical distance between C7 and the superior aspect of the spine of the scapula were 56.6 and 57.3 mm for tester 1 and tester 2, respectively. The 2D analysis revealed average values for this vertical distance of 46.5 mm for tester 1 and 46.6 mm for tester 2. Using the protractor method, the average values for the vertical distance between T8 and the inferior angle of the scapula were 31.9 and 30.0 mm for tester 1 and tester 2, respectively. Two-dimensional analysis revealed average values for this vertical distance of 34.1 mm for tester 1 and 32.4 mm for tester 2.
Reliability
Intertester reliability revealed SEMs 6.3 mm or less, MDCs were less than 17.5 mm, and ICCs were .78 or greater (Tab. 2), with greater reliability indicated for the T8 method. There were no significant differences between testers. Intratester reliability revealed SEMs were 0.9 mm or less, MDCs were 2.5 mm or less, and ICCs were .99 (Tab. 2).
Intertester and Intratester Reliability Statistics for the C7 and T8 Protractor Methoda
Validity
Examination of validity revealed significant correlations (P<.001) of .92 or less, with greater validity indicated for the T8 method (Tab. 3). On average, the difference in vertical distance between the C7 protractor method and the 2D analysis was greater than the difference for the T8 method (Tab. 3, Figs. 3A and 3B). For validity, SEMs were less than 4.2 mm for both methods, with a smaller SEM for the T8 method. Bland-Altman plots graphically demonstrated the larger variances and mean difference for C7 in comparison with T8 (Figs. 3A and 3B).
Validity Statistics for the Protractor Method Versus Photo 2D Analysis for Tester 1 and Tester 2 for Both C7 and T8a
Bland-Altman plots for (A) C7, illustrating the difference between photo 2-dimensional (2D) analysis and the protractor method for tester 1, and (B) T8, illustrating the difference between photo 2D analysis and the protractor method for tester 1.
Discussion
To our knowledge, this is the first study to examine the absolute vertical measurement of scapular position relative to a spinal landmark with a clinically feasible tool. Results indicated good levels of intertester reliability for both methods, with no significant differences between testers. However, the results for the T8 method showed slightly better intertester reliability, as indicated by more favorable ICC, SEM, and MDC values. Intratester reliability was excellent, and little difference between the methods was evident due to the very high reliability. In addition to being reliable, both protractor methods were shown to be valid measurements of resting scapular position, with high and significant correlations (Tab. 3) between the tester and the reference standard of digital analysis. The T8 method also showed slightly better results for validity, with a higher correlation (Tab. 3) and smaller limits of agreement (Tab. 3, Fig. 3) between the tester and digital analysis, compared with the C7 method.
With an intertester SEM of 5.7 mm (as reported for T8), clinicians can be 68% confident that differences greater than this value are real differences and are not due to intertester error.40 As the SEM may be considered a within-subject standard deviation,41 it is likely that 68% of scores will be within 1 standard deviation or SEM of a normally distributed sample.42–44 Previous research has shown that people who are symptomatic have 19 to 23 mm more scapular elevation than those who are asymptomatic.7,8 As the results for reliability and validity in this study identified a SEM of less than 5.7 mm for the T8 method, this finding may indicate that the method may accurately determine the difference between people who are symptomatic and those who are asymptomatic, particularly if measured by only one tester.
The MDC accounts for 2 error components when 2 measurements are being taken, such as in a test-retest situation (eg, preinterevention and postintervention), by incorporating the square root of 2.37 The formula for MDC used in this statistical analysis utilized a z score of 1.96 to allow for a 95% confidence interval.37,44 As a clinical example, if the MDC for the T8 protractor method measured before treatment by one clinician is 25 mm and the MDC measured after treatment by a separate clinician is 35 mm, this difference would likely be error, as the intertester MDC for the T8 protractor method at 15.3 mm is larger than the 10-mm test-retest change. If the posttreatment value is 40.5 mm, it is not likely to be error, as the MDC of 15.3 mm is smaller than the test-retest change of 15.5 mm. If preintervention and postintervention measurements were taken by one tester, the protractor method would be able to pick up an even smaller difference of 2.5 mm. To reliably and accurately measure scapular position using the T8 method, clinicians require a right-angled metal protractor with a spirit bubble fixed on each arm. A spirit bubble on each arm is necessary for the measurement of the left and right sides.
The results of this study may indicate that the T8 method is a superior measure compared with the C7 method, which may be a result of a relative lack of curvature in the thoracic spine at the level of T8, making it easier to rest the protractor on the individual for measurement. For the C7 method, a substantial cervical lordosis or thoracic kyphosis caused the upper margin of the protractor to rest against the occiput, and not the cervical spine. Subsequently, the protractor would be positioned posteriorly to the skin marking, introducing potential parallax error. Additionally, greater consistency between testers when identifying the inferior angle of the scapula compared with the medial aspect of the spine of the scapula was evident. As a result, the mean value for the T8 method was less than 2.2 mm below the reference standard of digital analysis, whereas mean value for the C7 method was greater than 10 mm above the reference standard (Tab. 3, Fig. 3).
Previous authors have reported poor levels of intertester reliability in palpation of thoracic landmarks.45 Additionally, the pilot study demonstrated that differences between testers for landmark identification would substantially affect validity. Therefore, to ensure that the validity of the protractor method was determined and not the validity of palpation, the protractor method was compared with a digital 2D image using the testers' own landmarks.
Although da Costa et al22 also described a measurement of vertical distance between a spinal landmark and scapular landmark, the measurement assessed the distance between C7 and the posterior angle of the acromion following a diagonal line. The protractor method, however, assesses a true measure of vertical distance between landmarks at a right angle to a horizontal plane. The current study demonstrated a lower SEM (5.67 mm versus 6.4 mm) and a higher ICC (.82 versus .76) than reported by da Costa et al,22 indicating a more reliable measure of elevation or depression of the scapula. Other authors23 also have found that a diagonal measure of scapular elevation demonstrates poor reliability and validity.
Although other methods have been shown to be effective in measuring scapular position against the opposite side,11 such a method could potentially be limited due to the potential variance in position of the opposite scapula. Differences in scapular position between dominant and nondominant sides can be usual in people who are pain-free.11,46,47 Therefore, a rehabilitation goal for scapular dyskinesis that aims for scapular symmetry may not be appropriate, and measurement methods need to be designed without comparison between scapulae. Additionally, both shoulders of individuals who are symptomatic have been shown to have an elevated scapular position compared with individuals who are asymptomatic.8,19 During humeral elevation. Inappropriate neuromuscular strategies may potentially affect both shoulders,46,48,49 altering the position of the asymptomatic scapula. Using spinal landmarks to measure scapular position offers a more robust method of measurement than a comparison to the opposite scapula. Therefore, it is suggested that the protractor method is a superior method of measuring scapular position, as it is a measure of resting scapular position compared to a fixed spinal landmark, particularly as scapular position in relation to the spine may vary considerably among individuals.50
Previous research on asymptomatic and symptomatic groups has identified that the greatest variances in scapular elevation or depression were identified above 120 degrees of humeral elevation.6 The current study utilized only measurements in a resting position, with no measurements of scapular position taken in higher ranges or through shoulder range. Given the lack of prior research regarding the utility of the protractor method in measuring scapular elevation during motion and prior research that indicates higher positions are important, future research is needed to assess protractor method utility in higher ranges or through range of motion.
Future studies to determine normative values, pathological threshold, and minimal clinically important difference (MCID) for the T8 method would be useful. Determining normative values for the T8 method would allow appropriate rehabilitation goal setting that does not rely on within-patient comparisons between scapulae. Normative values also can be established for different athletic populations such as tennis players and swimmers, who may have a different healthy scapular position from that of the remainder of the population. Future studies also could help to indicate whether a measurable pathological threshold value actually exists for scapular elevation or depression. Determining a protractor-measured pathological threshold could potentially provide a specific outcome value for rehabilitation goal setting. Additionally, determining a protractor-measured MCID would help clinicians to determine whether their interventions are helpful to patients. A comparison between the MDC given in the present study and a protractor-measured MCID would allow clinicians to determine whether the protractor method is sufficiently sensitive to detect worthwhile changes.
In conclusion, both the C7 and T8 protractor methods were shown to have moderate intertester reliability, with the T8 method demonstrating superior reliability and excellent validity. We recommend the use of the T8 method for measurement of resting scapular position, although further research is needed to assess the validity of this method throughout movement or in different scapular or glenohumeral positions and planes.
Footnotes
All authors provided concept/idea/research design, writing, data collection and analysis, participants, and consultation (including review of manuscript before submission). Dr McKenna provided project management, facilities/equipment, institutional liaisons, and administrative support. The authors acknowledge Guy Anza and Kerry Higgins for assistance with providing equipment and thank Curtin University Radio for their help with recruitment. They also acknowledge all of the participants in this study and thank them for their time in contributing to the study.
The Curtin University Human Research Ethics Committee approved this study, and all rights of the individual were protected.
- Received March 13, 2015.
- Accepted August 28, 2015.
- © 2016 American Physical Therapy Association