Reseach Article

An Attempt to Improve Accuracy and Reliability of Shoulder Active Range of Motion Goniometric Measurement: A Randomized Controlled Trial

Jake Irwin*, Kimberly M Richards, Michael S Johnson, Derek Clark and Stephen N Housley

Department of Physical Therapy, Byrdine F. Lewis School of Nursing and Health Professions, Georgia State University, USA

Received date: 07 Aug 2017; Accepted date: 08 Oct 2017; Published date: 12 Oct 2017.

*Corresponding author: Jake Irwin, Byrdine F. Lewis School of Nursing and Health Professions, Georgia State University, USA, Tel: 404-413- 1212; E-mail: jirwin5@gsu.edu
Abstract

This study aims to demonstrate a more reliable method for measurement of shoulder AROM is available and should be researched further as these measurements are performed consistently in clinical settings with poor quantity of evidence to support their ability to produce consistent results. We performed a study in order to determine the reliability of a novel method of measurement of shoulder active range of motion, using body contours to align the goniometer, and compare it to the reliability of the traditional method, which utilizes clinician determined landmarks and imaginary lines. The mechanism for this improvement is based on the concept that body contours are not going to appreciably change during an episode of care, while a clinician’s assessment of bony landmarks is more variable especially when different providers measure the same patient. We enrolled a convenience sample of 22 healthy volunteers to test the hypothesis that the inter- and intra-rater reliability of joint range of motion assessment will be improved with the use of this new technique when compared to the traditional method of measurement of shoulder range of motion.We utilized two groups of examiners taking AROM measurements on each shoulder of our 22 volunteers using both methods of measurement followed by a repeat assessment performed 2 weeks after the initial assessment. The study was limited by our use of a convenience sample of young, healthy volunteers. It is also limited by having too few examiners to determine the statistical significance of reliability. Our study resulted in improved inter and intra rater reliability using the novel method, and justifies further research into improving the usefulness of goniometry via the use of body contours to align the stationary and movement arms instead of relying on a clinician to determine bony landmarks on the wide variety of individual body types assessed in a clinical environment. This finding may also be useful in a healthcare marketplace transitioning towards an increased need to demonstrate objective improvements in order to justify payment.

Introduction

In an increasingly competitive healthcare marketplace, physical therapists are obligated to prove that our services produce reliably measureable results. Objective assessments, such as range of motion measurements, are an important tool for the physical therapist to determine dysfunction and to measure improvement. Physical therapists regularly utilize goniometric measurement with this type of goniometer in orthopedic clinical practice. However, there is a paucity of evidence to support the use or reliability of traditional methods. The traditional method also relies heavily on the examiner determining proper alignment of the goniometer via bony prominences and imaginary lines i.e. perpendicular to the floor. The reliability of our measurements is essential to their use in clinical practice, however it has been poorly represented in the research available to date [1-3]. The present study aims to employ a novel method to measure shoulder flexion and abduction and demonstrate improved inter and intra-rater reliability over the traditional measurement technique. We enrolled a convenience sample of 22 healthy volunteers to test the hypothesis that the inter- and intra-rater reliability of joint range of motion assessment will increase with the use of the novel method versus the traditional method commonly used in physical therapy practice today. If this is the case, studies that follow may assess the use of a non-traditional methods of measuring range of motion at different joints throughout the human body. Improving reliability of our objective measurements may also be of value in a changing healthcare marketplace in which more emphasis is placed on outcomes, and reliable techniques will be required in order to achieve payment goals.

Methods

Subjects

A convenience sample of 22 participants was recruited from the Georgia State University Doctor of Physical Therapy Program. The study was approved by the Institutional Review Board at Georgia State University and Informed Consent was obtained from all participants. The participants were eligible for the study if they were between 23 and 33 years of age, they could actively move their upper extremity into full shoulder flexion and abduction and they consented to take part in the study. The exclusion criteria for our study was any participant who had a current shoulder pathology that was being treated by a physician.

The study staff included two licensed physical therapists with 7 and 13 years of clinical experience in the assessment and treatment of orthopedic conditions as well as 2 current students in the Georgia State University Doctor of Physical Therapy Program. The staff was split into two groups, with a licensed PT and a student assigned to each group.

Design

Joint Measurements: All traditional goniometer measurements were taken using a standard 6-inch clear plastic rulongmeter goniometer with 360 degree head. This tool was utilized due to its common use in clinical scenarios and its cost effectiveness compared to computer based or electronic goniometers that would not apply to as large a clinical population as the plastic version. The examiners recorded the measurements for each motion found from each method in one degree increments. For both the traditional and novel measurements, the participant was asked to perform the motion to end range, and the goniometer was placed according to the standards for the respective measurement technique. When performing the standard measurement for shoulder flexion, the goniometer was placed according to traditional bony landmarks. The stationary arm was aligned with the midline of the thorax, while the movement arm was aligned with the lateral shaft of the humerus in line with the lateral epicondyle. The axis of the goniometer was placed at the greater tuberosity of the humerus [4]. To measure shoulder abduction utilizing the traditional method, the stationary arm was aligned perpendicular to the floor, the axis at the anterior aspect of the middle of the head of the humerus as assessed by the examiner and the movement arm aligned to the midline of the anterior aspect of the shaft of the humerus [4]. For the measurement of shoulder flexion utilizing the novel method, the stationary arm was placed on the lateral aspect of the midline of the thorax, the movement arm of the goniometer was placed midline on the posterior aspect of the humerus, and the axis in the axilla, and the [figure 1]. Pressure was applied to both arms such that the soft tissue was impressed upon until a firm end feel was interpreted by the examiner. For measurement of shoulder abduction the stationary arm was pressed into the lateral aspect of the midline of the thorax, the movement arm was pressed into the lateral aspect of the midline of the shaft of the humerus, and axis was pressed into the axilla [figure 2]. Pressure was applied as before such that the examiner pushed into soft tissue until a firm end feel was sensed.

Study participants performed a warm up routine prior to being measured that included 10 repetitions of shoulder pendulums in both directions on each arm in order to prevent prior activity level from affecting the ROM measured. Measurements of bilateral shoulder flexion and abduction were then obtained in standing. The traditional goniometric technique was employed for the measurement of flexion and abduction on one shoulder of the participant, followed by novel goniometric measurements of flexion and abduction on the same shoulder of the participant. The goniometer was placed by the student member of each team, and the number was read and recorded by the therapist member of the team. The same procedure was then performed on the other shoulder. Each participant was first measured by one set of examiners, then the participant proceeded to the next station for measurement by the second set of examiners, repeating

Figure 1: Novel method for measuring shoulder flexion
Figure 2: Novel method for measuring shoulder abduction

the same procedure. Each participant then returned in 2 weeks, and the measurements were repeated following the same parameters. All of the participants were assigned a number and their measurements were recorded next to the number they had been assigned.

The active range of motion obtained by each examiner for each style of measurement of shoulder flexion and abduction from the first trial were compared to those from the second trial in order to determine intra-rater reliability. The measurements obtained from each method on both dates were then compared between examiners in order to determine interrater reliability for each technique. Those findings were then analyzed to determine if the novel method demonstrated greater reliability than the traditional method.

Data Analysis: An initial descriptive analysis was carried out to calculate measures of central tendency, standard deviation (SD) and 95% confidence interval (CI) for both raters, across all movements, methods of measurement, and sessions.

Intra-rater Reliability: Intra-rater reliability was assessed by calculating the Intra class Correlation Coefficient (ICC), using a two-way mixed effects model1, for both raters across all movements and methods. ICC values were classified as poor (< 0.4), fair (0.4-0.59), good (0.6-.74), or excellent (0.75-1.00)2. A paired sample t-test was used to compare the mean difference between the first and second session.

Interrater Reliability: Interrater reliability was assessed by calculating the ICC, using a two-way mixed effects model1, between the first measurements obtained by 2 independent raters for both methods and movements. A paired sample t-test was used to compare the mean difference between the first sessions for 2 independent raters.

SEM: Inter-trial inconsistency from high inter-subject variability may artificially inflate the ICC value. However, the standard error of measurements (SEM) is not affected by inter-subject variability. Cdonsequently, the SEM is reported in concurrence with ICC. SEM was calculated using the formula SEM=SD√(1-r).

Criterion Validity: Criterion validity was estimated with ICC to determine if both standard and novel methods of measurement produced comparable results across all comparisons. ICC was chosen over a Pearson’s correlation coefficient due to the advantage of controlling for association bias. All analyses were run using SPSS Statistics Version 22.0 (SPSS Inc. Armonk, NY USA). The alpha level was set at 0.05. All statistical analyses were performed by an author not involved in data collection (S.N.H.).

Results

The participants comprised of 10males and 12 females, mean age (25.2). The ICC values for the intra-rater reliability for the measurements of flexion and abduction obtained with the standard goniometric procedure ranged from 0.585 and 0.842 [Table 1]. The ICC values for the intra-rater reliability for measurements of flexion and abduction obtained with the novel goniometric procedure ranged from 0.647 and 0.851 [Table 1].

The ICC values for the inter-rater reliability for the measurements flexion and abduction obtained with the standard goniometric procedure ranged from 0.695 and 0.781 [Table 2]. The ICC values for inter-rater reliability for the novel goniometric procedure ranged from 0.803 and 0.811 [Table 2].

Discussion

This study was performed in order to examine the reliability of standard goniometry for shoulder flexion and abduction using the traditional and widely accepted bony landmarks, and comparing those results to that of a novel method utilizing body contour in an effort to standardize the method and improve clinical reliability. The use of bony landmarks requires increased clinical skill and is subject to greater variability due to the wide range of body types that clinicians will be presented with in the clinic. Utilizing body contours decreases reliance on arbitrary assessments of location of said landmarks and allows improved reliability of measurement. This improvement in reliability adds clinical significance to the objective findings when evaluating and re-assessing patient progress and dysfunction. When assessing the traditional method, the inter class correlation values (ICC) for inter-rater reliability for Rater A flexion met the criterion for a good rating, and for abduction met the criteria for an excellent rating. The ICC values for inter-rater reliability for Rater B flexion met the criterion for a fair rating, and abduction met the criterion for a rating of good. The inter-rater reliability only met the criterion for a rating of good for both shoulder flexion and abduction utilizing the traditional method [Table 3].

Utilizing the novel approach to measuring shoulder flexion and abduction range of motion by utilizing the body’s natural contours, we found the inter-rater reliability met the criterion to classify both flexion and abduction for Rater A as good, and both flexion and abduction for Rater B as excellent [Table 3]. Our results indicate that the inter-rater reliability of our examiners is better for the novel approach than for the traditional method of measurement of both shoulder flexion and abduction.

INTRARATER

Standard

Pre

Post

ICC

95% CI Up

95% CI Low

Rater A

FLX

153.4565

152.3913

0.78

0.604

0.878

ABD

158.1957

157.4783

0.842

0.715

0.913

Novel

Flx

151.7826

152.9565

0.781

0.608

0.879

ABD

157.9348

165.0435

0.783

0.338

0.907

Rater B

Standard

FLX

158.4565

154.2609

0.585

0.179

0.782

ABD

162.0217

165.5217

0.778

0.588

0.879

Novel

Flx

155

153.7826

0.828

0.69

0.904

ABD

165.2826

164.6304

0.851

0.73

0.917

Table 1:Intra-rater reliability for the measurements of flexion, and abduction

INTERRATER

Raters A-B

Standard Pre

IRW

RIC

ICC

95% CI Up

95% CI Low

FLX

153.4565

158.4565

0.638

0.233

0.817

ABD

158.1957

162.0217

0.781

0.6

0.879

Standard Post

FLX

152.3913

154.2609

0.679

0.425

0.822

ABD

157.4783

165.5217

0.767

0.022

0.916

Raters A-B

Novel Pre

FLX

151.7826

155

0.811

0.535

0.911

ABD

157.9348

165.2826

0.805

0.224

0.926

Novel Post

FLX

152.9565

153.7826

0.764

0.574

0.869

ABD

165.0435

164.6304

0.88

0.784

0.934

Table 2:Inter-rater reliability for the measurements flexion and abduction

Intra-Rater

Rater A

Flexion

Abduction

Standard

Good

Excellent

Novel

Good

Good

Rater B

Flexion

Abduction

Standard

Fair

Good

Novel

Excellent

Excellent

Inter-Rater

Raters A-B

Flexion

Abduction

Standard

Good

Good

Novel

Excellent

Excellent

Table 3:Ratings for each method

Pooled Bland-Altman plots were generated from the data for both groups of testers to evaluate the validity of measurements. There was a significant difference in means for novel and standard as well as significant proportional bias for flexion resulting in low criterion validity. The information was also interpreted to show non-significant differences and low proportional bias for abduction thus presenting as relatively valid.

In the relatively few other studies available on the subject [1-3], the standard method of measurement for shoulder flexion and abduction have found excellent inter-rater reliability. While the measurements from our study ranged from fair to excellent for inter-rater reliability, and good for intra-rater reliability for the traditional measurement technique. However, our findings do not support classification of the traditional method as having excellent reliability. Future research should be performed to validate the reliability of both the traditional and our novel goniometric technique for the measurement of active shoulder flexion and abduction. Future research on both methods to determine accuracy compared to a gold standard is also in order. As standard goniometry is consistently utilized in orthopedic clinical settings, and the current literature is significantly lacking in research regarding the reliability of the traditional technique, a novel method is suggested here. This method utilizes body contour in an effort to obtain more reliable findings, and this study finds the novel method to be more reliable than the traditional approach. While new technologies (electronic goniometers, computer analysis) are forthcoming, the most common approach continues to be manual use of plastic goniometers therefore that is the mechanism examined in this study. Future research should include a greater number of examiners in order to determine statistical significance.

Possible limitations to be considered for this study include the utilization of participants that were obtained as a convenience sample of college students without current shoulder pathology. While this was sufficient for the purposes of comparing intra-rater and inter-rater reliability, it would be beneficial for future research to include a larger population of both symptomatic and asymptomatic participants. Range of motion measurements are typically obtained over a period of weeks to months in a clinical setting, allowing for a greater degree of normalization of measurements to occur, while in this study the results were obtained on only two different dates spaced two weeks apart.

In an effort to blind the study and reduce bias during measurements, a two person team approach was employed, with one tester placing the goniometer on the participant and then having the second tester read and record the measurement. This method decreased intra-rater reliability due to multiple testers attempting to obtain one measurement. In a typical clinical setting, the treating clinician would perform all the steps in taking a ROM measurement, thus giving a more clinically accurate intra-rater reliability.

During the second round of measurements, examination group A used a different goniometer than was used in the first round of measurements. This change in devices between tests increases external validity by making use of a different piece of equipment, which likely occurs in a clinical environment, but reduces internal validity by allowing for the possibility of variation between tools.

Two different methods for reading the goniometers were utilized between the two groups of testers. One group placed the goniometer into position, then removed the device from the participant in order to read and record the measurement. The other group took the reading withthe goniometer in place on the participant. The variance between these methods will decrease inter-rater reliability, but improves external validity and generalization to practice due to the variation in individual clinicians’ technique.

This study is the first of its kind to examine our novel method of measuring active shoulder flexion and abduction using the body’s natural contours. We have demonstrated good validity, and improved reliability for our novel technique over the traditional method of measuring shoulder abduction. Clinically, our novel body contour method is a reliable and valid alternative to the standard for measurement technique for abduction.

Acknowledgement

Georgia State University College of Nursing and Health Professions

References
  • Kolber M, Hanney W (2012) The reliability and concurrent validity of shoulder mobility measurements using a digital inclinometer and goniometer: a technical report. Int J Sports PhysTher 7(3): 306-313. 
  •      
  • Boone DC, Azen SP, Lin CM, Spence C, Baron C, et al. (1978) Reliability of goniometric measurements.  Phys Ther. 58(11): 1355-1360
  • Riddle DL, Rothstein JM, Lamb RL (1987) Goniometric Reliability in a Clinical Setting: Shoulder Measurement. Phys Ther 67(5): 688-673
  • Shrout PE, Fleiss JL (1979) Intraclass correlations: uses in assessing rater reliability. Psychol bull86(2): 420-428.
  • Hallgren KA (2012) Computing Inter-Rater Reliability for Observational Data: An Overview and Tutorial. Tutor Quant Methods Psychol8(1): 23-34.