The kinematics of upper extremity reaching: a reliability study on people with and without shoulder impingement syndrome

Physical impairments to the upper extremity can significantly affect the ability to perform daily life activities [1]. The evaluation of the motor performance using kinematic data is often performed in order to establish the impact of the physical impairments on function [2, 3]. The tasks chosen to evaluate motor performance should reflect the movement deficits characteristic of the target disorder and present an appropriate challenge for the spectrum of patients who would be evaluated. For example, people with shoulder impingement syndrome (SIS) present alterations in the movement of the scapula (increased or decreased scapular posterior tilting and lateral rotation) [4–6], humeral head (superior displacement with respect of the glenoid) [7] and clavicle (increased elevation and retraction) [6, 8] during arm elevation. These movement deficits are most likely associated with a reduction of the subacromial space [9, 10] leading to impingement of the subacromial structures [3]. The subacromial space reduction is more pronounced while performing arm elevation in the frontal plane as compared to the sagittal plane, thus leading to higher demands for shoulder control [11]. Therefore, the frontal plane is an optimal test position to appropriately challenge this target population [8].

Ideally, tasks used to evaluate motor performance should reflect activities of daily living. Most daily life activities require coordinated multi-joint movement of the upper extremity that balances stability and mobility while optimizing a goal oriented movement pattern. Unfortunately, the evaluation of the motor performance in people with SIS has been confined to simple arm elevation movements. Therefore, the upper extremity of people with SIS needs to be evaluated during multiarticular functional tasks in order to better represent the motor performance required during daily life activities [8].

Functional tasks, such as reaching out/pointing to targets and reaching for objects, have been used to evaluate the motor performance of people with physical impairments. For example, kinematic analysis of reaching has been used as an evaluative measure of upper extremity motor performance in people with shoulder dysfunctions [12] and in people after a stroke [13, 14]. Two groups have evaluated the reliability of such reaching tasks and have reported favourable results [12, 14]. Lin et al. [12] evaluated the within session reliability of shoulder girdle kinematics during functional tasks, including reaching tasks. They found intraclass correlation coefficients (ICC) values ranging from 0.73 to 0.99 and standard error of measurement (SEM) equal to or less than 2.5° for subjects with and without shoulder disorders. Wagner et al. [14] evaluated the within session reliability of forward reaching tasks in subjects with hemiparesis after stroke. Depending on the kinematic variable and the demands of the motor task, they observed ICC ranging from 0.04 to 0.99, and minimal detectable change ranging from 7.4% to 98,9%.

A reaching task that evaluates impairment characteristics of people with SIS was developed to evaluate the motor performance of this population [8]. However, the psychometric properties for evaluating longitudinal change in motor performance have not been established for this reaching task. The objective of this study was to characterize the reproducibility (ICC and SEM) and minimal detectable change (MDC) for reaching kinematics in people with and without SIS. A second objective was to evaluate the impact of the number of trials on reproducibility.

Within the same session, the reliability of scores based on the first two trials of upper extremity total joint excursion was good to very good for both patients and asymptomatic subjects and very good when using either the first five or last five trials (Figures 3 and 4; Table 1). There was a trend towards higher reliability coefficients when within session reliability was based on five trials. However, the ICC was only significantly higher when using the last five trials compared to the first two trials for trunk rotation and shoulder elevation for the healthy group (Figure 3). There were no significant differences in the level of reliability between the subjects with and without SIS, except for clavicular protraction/retraction for which the SEM were smaller in the healthy group (Figure 4).

The intrasession reliability was very good for end-reach position when using either the first two trials, first five trials or last five trials. However, for some joints, the 95%CI was larger when using the first two trials. Significant differences were only observed for elbow flexion/extension in the healthy group, where the SEM were lower when using the last five trials compared to the first two or five trials (Figure 4). Globally, the SEM and MDC were smaller for the last five trials compared to the first two or five trials. There were no significant differences between the subjects with and without SIS, except for wrist flexion/extension for which the SEM were smaller in the healthy group (Figure 4).

Between session (test-retest), reliability of upper extremity total joint excursion was good to very good in healthy subjects when using the mean of the first two or five trials, and very good when using the mean of the ten trials (Figure 5 and 6; Table 2). As for the intersession reliability for the upper extremity position at the end of reaching, the intersession reliability was again good to very good when using the mean of the first two or five trials, and very good when using the mean of the ten trials (Figure 5 and 6; Table 2). For most of the joints, the range of the 95%CI and the MDC were smaller when using all ten trials, especially for the position at the end of reaching. However, it did not result in significant differences in reliability for both the total joint excursion and the joint position at the end of reaching.

This study established that kinematic parameters of a functional task can be consistently performed by patients with SIS or asymptomatic subjects within a single session and by asymptomatic subjects across a test-retest period. It was also established that MDC could be used as a benchmark for clinically relevant differences when evaluating a patient following an intervention.

The intra- and intersession reliability statistics indicate that ideally, five to ten trials should be performed depending if intra- or intersession evaluations are performed. If only intrasession evaluation is needed, either the first or last five trials should be used to establish stable results. As for intersession reliability, using the mean of 10 trials did not result in better reliability coefficients but did define smaller absolute error estimates so that smaller minimal detectable change could be identified. The majority of time required for this testing is for the set up procedures (positioning of the markers, probing bony landmarks, calibration). The actual reaching task is fairly quick and easy for most participants. Therefore, an additional five repetitions to reduce test-retest measurement error is worthwhile. However, if patients experience high level of pain during the reaching task, then fewer repetitions should be performed. If fewer repetitions are performed, then changes in time have to be analyzed in light of higher absolute error estimates. Lin et al. [12] also found similar findings for reaching tasks in that the measurement error decreases when the number of trials increased.

The coordination of the multiple degrees of freedom available during the reaching tasks could explain the need for multiple trials in order to have good repeatability with acceptable confident intervals. Reaching implies the coordination of numerous upper extremity joints. Coordination of movement is the process of mastering the redundant degrees of freedom in order to have a controllable system [1]. As a result, a given movement can be performed using different kinematic and kinetic patterns [24]. By providing a standardized starting position and a target for the final position, a standardized task goal was created. Although providing a target for the final position may seem structured, many functional movements like taking objects off of shelves or pressing an elevator button require reaching with a specific end target position. Furthermore, the subjects retained a large amount of control and potential variability in how that goal is executed. Since this specific targeted task is new to the subject, several trials may need to be executed before the subject develops a consistent strategy and becomes comfortable with the execution of the task. In the present study only three practice trials were performed before the recording. The data suggest that additional practice trials might improve repeatability. The fact that a target was provided may explain why end-reach position was more consistent than total joint excursion. Different combinations of kinematic and kinetic patterns can be used to get to the target. However, these patterns result in a final joint position that is more constant across trials.

Reliability within sessions was not statistically different between patients and healthy subjects, although there was a trend for the healthy subjects to be more consistent. Pain may have interfered to some extent with the consistency of the performance of patients with SIS. The plane chosen to execute the task and the position of the target were selected to challenge patients with SIS, so this effect was expected. The pain level was evaluated after each trial for the subjects with SIS using the Present Pain Index [25]. The mean pain level was low at 0.9 out of 5 (Standard deviation = 0.8), but all the subjects experienced pain during the evaluation session. Patients experiencing pain can be expected to try different movement strategies in order to reduce the pain arising from impingement of the subacromial structures under the coracoacromial arch.

To determine whether change in the motor performance is meaningful, the MDC values can be used [26]. For example, if the same patient who had a total excursion in clavicular elevation of 14° on the initial evaluation has an end-reach position in clavicular elevation of 8.5° during reassessment 2 weeks later, the clinician will be able to state confidently that the patient has demonstrated statistically meaningful improvement because the change of 5.5° is greater than the MDC value (4.9°). In fact, in a previous study, a difference of 6.3° was observed for total excursion in clavicular elevation between a subgroup of subjects with SIS and healthy subjects [8]. Following a single session of rehabilitation, a reduction of 2.8° in clavicular excursion was observed. The change was statistically significant, but still not meaningful. MDC can be used to set short to mid term measurable treatment goals.

A number of limitations should be considered when interpreting the results. First, the current proposed method cannot be generalized to reaching in other planes of movement since the reliability has only been tested for a target located in the frontal plane. Second, the intersession reliability was only evaluated in a subgroup of healthy subjects. This could explain the large variations in the 95%CI obtained for the between session measurements and the magnitude of the MDC for some variables (mostly the more distal joints-elbow and wrist). Finally, a skin-based method, which involves digitizing bony landmarks and infrared-emitting diodes, was used for measuring the upper-extremity kinematics. Therefore, skin motion artefacts may affect data accuracy[12]

The reaching task proposed to evaluate the motor performance of people with SIS was found reliable in people with and without SIS. Furthermore, the minimal difference necessary to infer a meaningful change in motor performance was determined, indicating that relatively small changes in task performance can be interpreted as a change in motor performance. This reaching task may now be used to characterize changes in the motor performance of people with SIS. In future studies, it will be important to analyze the effects of medical/rehabilitation interventions on the motor performance in light of the minimal detectable change.