- Email: [email protected]
Validity and Reliability of Assessment Tools for Measuring Unsupported Sitting in People With a Spinal Cord Injury
1571
ORIGINAL ARTICLE
Validity and Reliability of Assessment Tools for Measuring Unsupported Sitting in People With a Spinal Cord Injury Claire L. Boswell-Ruys, BAppSc (Physioth), Daina L. Sturnieks, PhD, Lisa A. Harvey, PhD, Catherine Sherrington, PhD, James W. Middleton, MBBS, PhD, Stephen R. Lord, PhD, DSc ABSTRACT. Boswell-Ruys CL, Sturnieks DL, Harvey LA, Sherrington C, Middleton JW, Lord SR. Validity and reliability of assessment tools for measuring unsupported sitting in people with a spinal cord injury. Arch Phys Med Rehabil 2009;90: 1571-7. Objectives: To develop simple tests to assess the abilities of people with spinal cord injury (SCI) to sit unsupported and to assess the construct validity and test-retest reliability of these tests. Design: Cross-sectional comparisons, convenience sample. Setting: Biomechanical laboratory. Participants: People (N⫽30) with SCI between the C6 and the L2 level of 2 months to 37 years duration before assessment. The sample was stratified by impairment level (at T8) and time since injury (1y postinjury). Interventions: Not applicable. Main Outcome Measures: On 2 separate occasions, participants performed tests that measured the distance of upperbody sway and maximal torso leaning, errors made during a coordinated stability task, timed dressing/undressing of the upper body and alternating arm reaching, and percentage change in seated upper body/arm reaching. Results: All tests showed good construct validity in that they distinguished between participants with higher (C6-T7) and lower (T8-L2) level impairments (P⬍.05) and between participants with acute (ⱕ1y) and chronic (⬎1y) lesions (P⬍.05). The tests also showed good to excellent test-retest reliability (intraclass correlation coeffiecient3,1 range, .51–.91). Conclusions: These simple and quick-to-administer tests have both construct validity and test-retest reliability. They would be appropriate for research and clinical purposes to quantify the abilities of people with SCI to sit unsupported. Key Words: Outcome assessment (health care); Posture; Rehabilitation; Spinal cord injuries. © 2009 by the American Congress of Rehabilitation Medicine
From the Prince of Wales Medical Research Institute, University of New South Wales, Randwick (Boswell-Ruys, Sturnieks, Sherrington, Lord); Rehabilitation Studies Unit and the University of Sydney, Sydney (Harvey, Middleton); and the George Institute for International Health, University of Sydney, Sydney (Sherrington), Australia. Presented to the Australian and New Zealand Spinal Cord Society, Melbourne, Australia, November 16 –18, 2006; and the World Confederation for Physical Therapy, Vancouver, Canada, June 2– 6, 2007. Supported by the New South Wales Premier’s Spinal Cord Injury Grant Program and the National Health and Medical Research Council of Australia. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Correspondence to Claire Boswell-Ruys, BAppSc (Physioth), Prince of Wales Medical Research Institute, Barker St, Randwick, NSW, 2031, Australia, e-mail: [email protected]. Reprints are not available from the author. 0003-9993/09/9009-00579$36.00/0 doi:10.1016/j.apmr.2009.02.016
HE ABILITY TO SIT unsupported in the able-bodied popT ulation requires the coordinated use of the whole body, the lower limbs, the trunk, the arms, and the head along with inputs from the sensory systems.1-3 Because of paralysis and sensory loss, people with a SCI have an impaired ability to sit unsupported. In general, the extent of this impairment depends on the neurologic level of injury and time since injury. Unsupported sitting is important for people with SCI because it increases independence with ADLs. Therapists invest significant time training patients with SCI to sit unsupported. Typically, training involves challenging patients to reach and move in different directions while seated. However, there are no standardized training guidelines, and it is not known whether such training strategies are effective. As a first step toward investigating these issues, it is important to devise valid and reliable ways of assessing the abilities of people with SCI to sit unsupported. Unsupported sitting is difficult to assess because it is complex; involves ongoing postural adjustments4; and, as an integral component of all daily activities, cannot be separated from the environment in which it is performed.5 In the laboratory, force plate transducers,6-10 piezo-resistive pressure systems,11 and motion-analysis systems12-15 have been used to assess and model unsupported sitting. This equipment is costly and not appropriate for use in clinical settings. The mFRT, adapted from the stroke population,16 has been used in people with SCI.17,18 It has been shown to be reliable. However, the mFRT is limited to only measuring a person’s ability to reach forward, and so it does not fully reflect a person’s ability to sit unsupported and move his/her center of mass to reach in different planes. There is a need for reliable and valid tests to comprehensively assess unsupported sitting ability in people with SCI, which encompass a range of tasks that have relevance to ADLs. The objectives of this study were (1) to devise a battery of simple, inexpensive, and portable tests suitable for use in the clinical setting that assess salient aspects of unsupported sitting; (2) to examine the reliability of these tests in people with SCI; and (3) to assess the correlation of the measures and determine the validity of the tests by assessing their ability to detect differences in neurologic level of SCI and time since injury.
List of Abbreviations ADLs AP ASIA ICC mFRT SCI
activities of daily living anteroposterior American Spinal Injury Association intraclass correlation coefficient modified Functional Reach Test spinal cord injury
Arch Phys Med Rehabil Vol 90, September 2009
1572
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
METHODS Participants Thirty people with SCI participated in this study (24 men, 6 women). The participants were between 18 and 66 years of age (mean ⫾ SD, 35⫾11y). Time since injury varied from 2 months to 37 years (mean ⫾ SD, 9⫾12y). Participants were included if they were able to sit out of bed for 2 hours and were free of sacral pressure areas. Participants were excluded if they had external bracing or injuries affecting their ability to lift their arms above their head. The neurologic level of injury ranged from C6 to L2 according to the International Standards for the Neurological Classification of Spinal Cord Injury.19 Twenty-one participants had complete (ASIA grade A) lesions, with the remainder having varying degrees of incomplete lesions (3 with ASIA grade B, 5 with ASIA grade C, 1 with ASIA grade D impairments). Participants were categorized into 2 groups based on their level of injury, signifying participants with none or some abdominal muscle innervation, a welldocumented factor identifying the ability of people with SCI to sit unsupported.8,12,20 The high-level group (n⫽14) included people with neurologic levels between C6 and T7, all of whom had ASIA grade A lesions. The remaining participants were in the low-level group (n⫽16), having neurologic levels between T8 and L2 with ASIA A to D impairment grades. Participants were also categorized into 2 groups depending on the time since injury; namely, people who sustained their injury up to 1 year before testing (acute: n⫽10) and people who sustained their injury more than 1 year before testing (chronic: n⫽20). One year postinjury is most widely considered to be classified as chronic as neurorecovery plateaus at approximately 12 months postinjury.21 It can be observed clinically that people with a chronic injury perform better on tasks than people with an acute injury for a given level of injury. There were no differences in the distributions of ASIA motor and sensory scores between the acute and chronic groups (Pⱖ.05). All participants provided informed consent before study participation. Testing Procedures The definition of unsupported sitting used in this study was the ability of a person to sit upright without using the hands or an external support. Participants sat on a flat seat without a backrest. The seat was adjusted so that the hips, knees, and ankles were at 90° with the feet resting on the floor (short sitting). The posterior aspect of the sacrum was positioned in line with the central posterior margin of the seat. The seat was designed to mimic the low treatment plinths commonly found in physiotherapy gymnasiums. A harness was used to ensure the participants’ safety and provide support during the rest periods between each test. The harness was connected to an overhead beam via a force transducer to ensure that the participants were not using the harness for support during the testing. If a force was registered, the test was stopped and recommenced from the beginning. Participants performed a battery of 6 tests on 2 occasions separated by a median (interquartile) of 7 days (3–13d). The order of test administration was the same for all participants, but participants were provided with rest periods as needed and between each test to minimize any fatigue that may have affected the performance of the tests. The second assessment was performed without reference to the first which had been placed in a sealed opaque envelope. Two authors were present at every session, but all testing and measurements were performed by the same examiner. Arch Phys Med Rehabil Vol 90, September 2009
The development of the tests was primarily modeled on tests used in the elderly and other neurologic populations.16,17,20,22-24 Those tests were modified to reflect situations when a person is required to maintain his/her center of mass over his/her base of support while sitting unsupported. These situations included the performance of slow controlled movements, fast alternating movements, and functional tasks. The 6 tests we developed are as follows. Upper-body sway. This test measured the abilities of participants to sit unsupported and remain as still as possible for 30 seconds. The Lord sway meter, which was originally designed to measure total body sway while attempting to stand still, was used to measure upper-body sway.22 It consisted of a 40-cm hinged rod that was fastened by a firm belt to the participants’ chest at the level of the axilla and extended in a horizontal plane from the body. For this test, the rod was positioned to extend in the posterior direction. A ball-point pen mounted vertically at the end of the rod recorded the movements of the upper body on a sheet of graph paper that was fixed to the top of a height-adjustable table. The tip of the pen was placed on the graph paper when the participants had commenced unsupported sitting. The resultant trace was measured in terms of 3 components: maximal lateral displacement, maximal AP displacement and the total length of the sway path (number of square mm traversed by the pen). Small displacement and length indicated better performance. The test was performed 3 times and the mean derived. Maximal balance range. This test was adapted from a standing test.22 Participants were asked to lean as far forward as they could without falling and then return to their starting position. They were then asked to lean back as far as possible without falling and then return to their starting position. The maximal AP distance traversed was measured by using the sway meter described earlier. However, the rod was positioned to extend in the anterior plane for visual feedback. The pen attached to the end of the rod recorded the anterior and posterior movements of the participants on a sheet of graph paper fastened to the top of a height-adjustable table. The participants had 2 attempts at the test, with the longer distance taken as the test result. The score recorded was the maximal AP distance moved. A long distance was considered a better performance. This score was corrected for body height (score ⫻ mean height/participants’ height) measured from the center of the sway meter strap to the top of the seat. Coordinated stability. The coordinated stability test was also adapted from a standing test.22 It measured participants’ abilities to adjust their sitting posture in a steady and coordinated way when near, or at the limits of, their postural equilibrium. The sway meter was again attached to the participants with the rod extending in the anterior plane for visual feedback. The participants were asked to adjust their posture by bending or rotating the upper body so that the sway meter pen tip followed and remained within a convoluted track. Two tracks were tested: the original version from the standing test (test A) and a simpler version (test B). Each track was marked on a piece of paper attached to the top of a height-adjustable table. To complete test A without errors, the sway meter pen had to remain within the track, which was 1.5-cm wide, and the participant had to be capable of adjusting the position of the pen 25cm laterally and 18cm in the AP plane. Test B comprised a less convoluted track and required smaller displacements of the pen (14cm laterally and 12cm anterioposteriorly). In both test versions, a total error score was calculated by summing (1) the number of occasions the pen failed to stay within the path (1 error point), (2) the number of track corners cut (3 error points), and (3) the number of times the participants
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
used their hands for support (5 points). Participants attempted each test twice, with the lesser score (better performance) taken as the test result. Alternating reach test. This test measured the abilities of participants to tap a table 8 times as fast as possible by using alternate arms on each repetition. This test was similar to item 12 of the Berg Balance Scale.23 A table was positioned at the height of the participants’ axillas with the closest edge an arm’s length away. The participants were tested under 2 conditions: supported, with their nonmoving hand grasping their thigh, and unsupported, with their nonmoving hand beside their body (shoulder at 0° and elbow at 90° flexion). The time taken to perform the tests was recorded with a stopwatch. A short time indicated a better performance. Seated reach distance. This test measured the participants’ abilities to reach in different directions as far as possible without falling. This test was similar to the one used in the stroke population.24 A large table was positioned with its closest edge in line with the greater trochanters of the participants and at the height of their iliac crests. A semicircle was cut out of the table to accommodate the participants’ abdomens. The table was covered with a large paper sheet with 5 predrawn lines. The directions of the lines were (1) lateral right (3 o’clock), (2) lateral left (9 o’clock), (3) 45° right (1.30 o’clock), (4) 45° left (10.30 o’clock), and (5) forwards (12 o’clock). A marker pen was taped into the thumb web spaces of the participants’ 2 hands. Without holding on for support, the participants were asked to reach sideways with the right hand as far as possible to make a light mark across the lateral right line before returning to the starting position. This was then performed with the left hand on the lateral left line. This procedure was repeated for each direction; those on the right side of the body with the right hand and those on the left side of the body with the left hand. The greatest reach distance for each direction was measured from the point of confluence of the lines. Arm length was measured from the acromion process to the position of the pen in the thumb web space. The seated reach distance was calculated as a proportion of the length of the participants’ arms (greatest reach distance/arm length); a high proportion indicated a better performance. A perpendicular measure from the top of the seat to the acromion process was used to correct the score for body height (score ⫻ mean height/participants’ height). T-shirt test. This test measured the time taken for participants to put on and take off a t-shirt. The test was designed to be similar to that reported by Chen et al20 but was performed in short sitting as described previously. A table was positioned with its closest edge in line with the participants’ knees and at the height of their iliac crests. A pullover t-shirt was spread out flat on the table face down. Standardized t-shirts were supplied and were 1 size larger than the participants would normally wear. No harness was worn during this test, but a research assistant was ready to support the participants if they were at risk of falling. The participants were required to put on and take off the t-shirt, resting between each maneuver. The test was repeated twice, with the average times calculated for each component (on, off, and total time). Short times indicated a better performance. Statistical Analysis ICC3,1 were calculated to evaluate the test-retest reliability of the 6 tests. The ICC values were interpreted according to a rating system suggested by Shrout and Fleiss25 (⬎.75 excellent reliability, .40 –.75 fair to good reliability, ⬍.40 poor reliability). Independent sample t tests were used to assess differences in tests between the participants with high- and low-level
1573
lesions and participants with acute and chronic injuries. Discriminant function analyses were used to determine the best combination of independent and significant tests for discriminating between injury level (low/high) and injury time (acute/ chronic) status. Spearman correlations were computed to assess the associations between the test scores and ASIA motor and ASIA sensory scores. Pearson correlation coefficients were also calculated to assess the correlation of the test measures. Any test measures with right-skewed distributions were log10 transformed in all analyses. Despite the multiple comparisons made, P values of less than .05 were considered statistically significant (ie, not adjusted to Bonferroni). This approach was taken because of the exploratory nature of this relatively small study for which such adjustments may increase type II errors.26 Analyses were performed by using SPSS for Windows.27,a RESULTS Test-Retest Reliability All participants performed all tests on the 2 testing occasions. The 6 tests were found to be reliable with the ICC3,1 values presented in table 1. The ICCs were excellent for the upper-body sway total length, alternating reach, maximal balance range, coordinated stability, seated reach, and t-shirt tests (ICC range, .79 –.91; 95% confidence interval, .60 –.96). The upper-body sway test AP and lateral range components showed fair to good reliability (ICCs⫽.51 and .72, respectively). Because the tests were deemed reliable, they were considered suitable for further validity analysis. Test Validity Construct validity was assessed in relation to ASIA scores, lesion level, and time since injury. The majority of the tests were significantly correlated with ASIA motor and ASIA sensory scores (table 2). The seated reach test had significant correlations with ASIA motor scores only. The upper-body sway measures poorly correlated with both ASIA motor and sensory scores (see table 2). The upper-body sway (AP and total length), maximal balance range, coordinated stability, and alternating reach tests had significant correlations with subject height (see table 2). This indicated that taller subjects performed more poorly in these tests. Results for the participants categorized in terms of lesion level and time since injury are shown in table 3. Participants with high lesions (C6-T7) performed significantly worse than participants with low lesions (T8-12) in the alternating reach, coordinated stability and tshirt tests, and half the reach directions in the seated reach test. Participants with an acute injury performed significantly worse than participants with a chronic SCI in at least 1 component of each test. The discriminant function analysis for level of injury identified the upper-body sway test total length and supported alternating reach test as being able to discriminate between high and low level impairments. The 2-variable model had a Wilk lambda of .68 (P⫽.006) and a canonical correlation for the discriminant function of .56 and correctly classified 90% of participants into high and low impairment groups, with high sensitivity (86%) and specificity (94%). The standardized canonical correlation coefficients were –.74 for upper-body sway total length and 1.10 for supported alternating reach. For time since injury, upper-body sway total length and seated reach 45° to the right discriminated between acute and chronic injury groups (Wilk ⫽.60, P⫽.001 and canonical correlation⫽.64). The standardized canonical correlation coefficients were –.67 for upper-body sway total length and 0.75 for seated reach 45° Arch Phys Med Rehabil Vol 90, September 2009
1574
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 1: Mean ⴞ SD Test and Retest Scores for the Unsupported Sitting Tests With Corresponding ICCs Test
Upper-body sway (mm) Total length* AP* Lateral* Maximal balance range (cm) Total range* Coordinated stability (score) Test A Test B Alternating reach (s) Supported* Unsupported* Seated reach distance (% arms length) Lateral right Lateral left 45° right 45° left Forward right Forward left T-shirt test (s) On* Off* Total*
Test 1
Test 2
ICC3,1 (95% CI)
90⫾89 18⫾12 15⫾8
104⫾102 24⫾20 14⫾9
0.79 (0.60–0.89) 0.72 (0.49–0.86) 0.51 (0.18–0.73)
23.1⫾14.4
24.2⫾15.5
0.90 (0.81–0.95)
26.6⫾17.5 11.0⫾8.4
21.2⫾15.6 9.8⫾8.9
0.83 (0.67–0.92) 0.86 (0.72–0.93)
4.9⫾1.3 5.5⫾2.4
4.7⫾1.5 5.4⫾2.7
0.82 (0.65–0.91) 0.88 (0.75–0.94)
109⫾10 107⫾10 109⫾13 105⫾15 98⫾22 96⫾24
110⫾10 107⫾11 108⫾12 106⫾14 95⫾29 97⫾23
0.80 (0.62–0.90) 0.86 (0.72–0.91) 0.83 (0.67–0.91) 0.82 (0.66–0.91) 0.87 (0.75–0.94) 0.89 (0.79–0.95)
16.5⫾13.9 9.4⫾9.2 25.6⫾22.8
14.9⫾13.1 8.0⫾6.9 22.9⫾18.7
0.91 (0.82–0.96) 0.85 (0.72–0.93) 0.89 (0.78–0.95)
NOTE. High scores in the sway, coordinated stability, alternating reach, and t-shirt tests and low scores in the maximal balance range and seated reach distance tests indicate impaired performance. Abbreviation: CI, confidence interval. *Natural logs used for statistical testing because of positively skewed data.
to the right. This model correctly classified 87% of participants into acute and chronic groups, again with high sensitivity (80%) and specificity (90%). Correlation of the Test Measures Table 4 presents the associations among the tests. There were significant correlations among all of the tests (P⬍.05) with the exception of some components of the upper-body Table 2: Correlations Between the Test Measures and ASIA Motor and ASIA Sensory Scores and Participant Height ASIA Unsupported Sitting Test
Motor
Sensory
Height
Upper-body sway total length Upper-body sway AP Upper-body sway lateral Maximal balance range Coordinated stability test A Coordinated stability test B Alternating reach supported Alternating reach unsupported Seated reach lateral right Seated reach lateral left Seated reach 45° right Seated reach 45° left Seated reach forward right Seated reach forward left T-shirt on T-shirt off T-shirt total
–.17 –.14 –.33 .37* –.56* –.58* –.51* –.57* .26 .22 .30 .34 .41* .47* –.75* –.62* –.73*
–.05 –.02 –.33 .46* –.47* –.55* –.58* –.54* .29 .42* .40* .43* .44* .48* –.63* –.57* –.61*
.50* .61* .33 –.42* .44* .24 .52* .54* –.36 –.32 –.35 –.32 –.22 –.19 .27 .19 .23
*P⬍.05.
Arch Phys Med Rehabil Vol 90, September 2009
sway test. The strongest correlations were between the coordinated stability test A and the put-on component of the t-shirt test (r⫽.79, P⬍.01), the coordinated stability test B and the supported alternating reach test (r⫽.72, P⬍.01), and all of the seated reach directions with the maximal balance range test (r⫽.69 –.80, P⬍.01). As expected, there were also significant correlations among the components of each test (r⫽.59 –.98, P⬍.01). The strongest correlations were evident within the alternating reach (r⫽.93, P⬍.01) and t-shirt tests (r⫽.86 –.96, P⬍.01). DISCUSSION The 6 tests used in this study were developed to measure unsupported sitting ability in people with SCI. For independence and safety in daily activities, it is necessary for a person with an SCI to be able to maintain his/her center of mass over his/her base of support as he/she performs slow controlled movements, fast alternating movements, and functional tasks. The tests developed encompass these aspects of daily living. The 6 tests proved to have construct validity in that they could discriminate between people with acute and chronic injuries and people with higher and lower lesions. The tests also showed good to excellent test-retest reliability. This study showed that tests of controlled unsupported sitting (coordinated stability test) and functional tasks (t-shirt test), in addition to reaching ability (seated reach distance and alternating reach tests), were able to discriminate between people with lesions above or below the level of T8 in univariate analyses. People with a lesion below the T8 impairment level have intact abdominal muscle innervation and so would reasonably be expected to perform better in unsupported sitting tests than people with higher lesions. Previous work17,20 has found that reaching tests can distinguish a person’s neurologic level. The discriminant function analysis identified supported
1575
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 3: Mean ⴞ SD of Groups Based on Time Since Injury and Level of Injury Time Since Injury Test
Upper-body sway (mm) Total length AP Lateral Maximal balance range (cm) Total distance Coordinated stability (score) Test A Test B Alternating reach (s) Supported Unsupported Seated reach distance (% arm length) Lateral right Lateral left 45° right 45° left Forward right Forward left T-shirt test (s) On Off Total ASIA score Motor Sensory
Level of Injury
Acute (n⫽10)
Chronic (n⫽20)
High C6-T7 (n⫽14)
Low T8-T12 (n⫽16)
149⫾129 24⫾11 19⫾12
61⫾38* 15⫾11* 14⫾6
76⫾49 18⫾13 15⫾5
103⫾113 18⫾11 16⫾11
12.8⫾4.9
29.1⫾16.5*
17.2⫾9.3
29.3⫾18.2
44.2⫾2.2 16.3⫾9.4
21.1⫾14.3* 11.4⫾13.7
36.9⫾15.8 18.5⫾15.5
22.3⫾20.8* 8.5⫾8.0*
5.6⫾1.4 6.5⫾2.5
5.0⫾2.0 5.0⫾2.3*
6.1⫾2.2 6.8⫾2.8
4.4⫾0.9* 4.4⫾1.4*
100⫾9 100⫾12 97⫾11 91⫾14 81⫾15 79⫾16
116⫾18* 113⫾17* 117⫾20* 114⫾19* 108⫾27* 107⫾28*
106⫾15 102⫾13 103⫾16 98⫾17 88⫾20 85⫾19
115⫾18 114⫾18 116⫾21 113⫾21* 109⫾28* 108⫾30*
24.6⫾16.6 14.0⫾12.0 38.6⫾27.5
12.4⫾10.6* 7.1⫾6.7* 19.0⫾17.2*
23.8⫾15.7 13.2⫾9.6 36.3⫾25.4
10.1⫾8.2* 6.1⫾7.8* 16.2⫾15.6*
51⫾5 133⫾44
53⫾12 124⫾44
47⫾7 86⫾15
57⫾10* 162⫾25*
*P⬍.05.
alternating reach and upper-body sway as independent and significant predictors of injury level groups. Interestingly, upper-body sway did not correlate with ASIA motor or sensory scores but did provide unique predictive ability in the multivariate model, despite not being significantly associated with neurologic level in a univariate analysis. This suggests that only after adjusting for reaching ability, the association between upper-body sway and neurologic level becomes apparent. Combined, these 2 complementary unsupported sitting measures had both high sensitivity (86%) and specificity (94%) with regard to classifying people based on SCI level. The validity of the tests was shown further by the significant associations with ASIA scores. People with higher ASIA scores were able to reach further than those with a lower score, similar to the findings of Lynch et al.17 The more available muscle activity and preserved sensation that a person has, the faster he/she is able to perform time-related tests and tests that require greater postural control. All tests were able to distinguish between people with acute and chronic SCI. People with chronic lesions had faster performance times, greater stability, and a longer reach. With increasing time, people with SCI adapt to their impairments, know their abilities better, and are more comfortable with challenging themselves to their limits. Thus, time since injury also comprises a good measure for evaluating the discriminatory ability of assessment measures. Upper-body sway and one of the reach measures (seated reach 45° to the right) were identified as independent predictors of time since injury. As with neurologic level, complementary measures of seated postural control and reaching ability would accurately discriminate between people with acute and chronic SCI, showing 80% sensitivity and 90% specificity.
Participants with a longer trunk had increased sway, took longer to complete tasks, and were unable to reach as far while sitting unsupported. This finding supports clinical observations and compliments the work of Chen et al.20 People with SCI have an upward shift of their center of gravity by approximately 5% of their body length compared with people without paralysis,28 which increases the degree of difficulty in maintaining their center of gravity over their base of support. Furthermore, people with longer bodies have a higher center of gravity.29 To maintain this higher center of gravity over their base of support requires the recruitment of distal musculature, which may be impaired. Thus, when assessing a person’s unsupported sitting, the height of the person is an important consideration. Study Limitations Currently, there are no tests that comprehensively measure the unsupported sitting ability of people with SCI; therefore, this trial was exploratory in nature. Although this is the major study strength, it also highlights certain limitations. In validating the test measures, there was a sound rationale for categorizing participants with regard to level of injury (presence or absence of abdominal muscle innervation), but we acknowledge that the criterion of 1 year for the time since injury measure was somewhat arbitrary. We also acknowledge that while being of sufficient size for measuring test-retest reliability, the sample was at the margin for assessing the validity of the balance measures, particularly in relation to multivariate analysis.30 Thus, the discriminant function models identified here require verification in larger external samples. Finally, because we chose not to adjust P values to avoid type II Arch Phys Med Rehabil Vol 90, September 2009
1576
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 4: Correlations Among the Test Measures Upper-Body Sway
MBR
Coordinated Stability
Alternating Reach
Seated Reach Distance
T-Shirt Test
Unsupported Sitting Test
Length
AP
Lateral
Distance
Test A
Test B
Support
Unsupp
Lat R
Lat L
45° R
45° L
Fwd R
Fwd L
On
Off
Sway AP Sway Lat MBR distance Co Stab A Co Stab B AR Support AR Unsupp SR Lat R SR Lat L SR 45° R SR 45° L SR Fwd R SR Fwd L T-shirt on T-shirt off T-shirt total
.85* .75* –.31 .47* .44* .48* .51* –.32 –.35 –.36 –.37* –.62 –.22 .37* .12 .29
.59 –.32 .48* .55* .50* .43* –.49* –.42* –.47* –.43* –.30 –.27 –.38* .17 .31
–.25 .36 .56* .46* .44* –.25 –.27 –.26 –.31 –.25 –.24 .41* .13 .32
–.58* –.45* –.58* –.51* .74* .69* .78* .74* .80* .79* –.55* –.55* –.55*
.72* .67* .69* –.56* –.48* –.62* –.61* –.67* –.67* .79* .62* .71*
.72* .49* –.64* –.55* –.54* –.45* –.36 –.36 .57* .34 .49*
.93* –.54* –.65* –.59* –.55* –.48* –.44* .63* .50* .56*
–.36 –.55* –.60* –.63* –.63* –.57* .65* .48* .55*
.78* .82* .75* .63* .62* –.54* –.46* –.53*
.73* .84* .64* .64* –.56* –.48* –.54*
.91* .87* .85* –.65* –.65* –.67*
.87* .87* –.65* –.63* –.67*
.98* –.66* –.63* –.65*
–.64* –.66* –.67*
.86* .96*
.94*
Abbreviations: AR, alternating reach; Co Stab, coordinated stability; Fwd, forward; L, left; Lat, lateral; MBR, maximal balance range; R, right; SR, seated reach distance; Support, supported; Sway, upper-body sway; Unsupp, unsupported. *P⬍.05.
errors,26 it is possible that some of the significant univariate associations revealed may have occurred by chance. Based on the validity and reliability findings of this trial, the following tests are recommended as a minimum set for comprehensively assessing unsupported sitting in people with SCI: upper-body sway test total length, seated reach distance (45° to the left and/or right), supported alternating reach test, coordinated stability test A, and t-shirt put-on test. These tests could be routinely used in clinical settings because each takes less than 3 minutes to administer and requires a minimal amount of equipment. Each test should be used to assess a person’s performance against himself/herself until further research is performed that may allow comparison to the SCI population as a whole. Additional research should determine the responsiveness of each test to ensure they also detect meaningful change over time. CONCLUSIONS A battery of tests was developed to measure unsupported sitting in people with SCI. The tests are simple, reliable, and quick to administer. They would be suitable for use in further research and the clinical setting. 1.
2.
3.
4. 5. 6.
References Dean CM, Shepherd RB, Adams R. Sitting balance I: trunk-arm coordination and the contribution of the lower limbs during selfpaced reaching in sitting. Gait Posture 1999;10:135-46. Dean CM, Shepherd RB, Adams R. Sitting balance II: reach direction and thigh support affect the contribution of the lower limbs when reaching beyond arm’s length in sitting. Gait Posture 1999;10:147-53. Crosbie J, Shepherd RB, Squire TJ. Postural and voluntary movement during reaching in sitting: the role of the lower limbs. J Human Mov Studies 1995;28:103-26. Huxham FE, Goldie PA, Patla AE. Theoretical considerations in balance assessment. Aust J Physiother 2001;47:89-100. Carr JH, Shepherd RB. A motor relearning programme for stroke. 2nd ed. Oxford: Butterworth-Heinemann; 1987. Aissaoui R, Boucher C, Bourbonnais D, Lactoste M, Dansereau J. Effect of seat cushion on dynamic stability in sitting during a
Arch Phys Med Rehabil Vol 90, September 2009
7.
8.
9.
10.
11.
12.
13. 14.
15.
16. 17.
18. 19.
reaching task in wheelcahir users with paraplegia. Arch Phys Med Rehab 2001;82:274-81. Shirado O, Kawase M, Minami A, Strax TE. Quantitative evaluation of long sitting in paraplegic patients with spinal cord injury. Arch Phys Med Rehab 2004;85:1251-6. Seelen HA, Potten YJ, Huson A, Spaans F, Reulen JP. Impaired balance control in paraplegic subjects. J Electromyogr Kinesiol 1997;7:149-60. Janssen-Potten YJ, Seelen HA, Drukker J, Spaans F, Drost MR. The effect of footrests on sitting balance in paraplegic subjects. Arch Phys Med Rehab 2002;83:642-8. Kamper D, Parnianpour M, Barin K, Adams T, Linden M, Hemami H. Postural stability of wheelchair users exposed to sustained, external perturbations. J Rehabil Res Dev 1999;36:121-32. Karatas GK, Tosun AK, Kanatl U. Center-of-pressure displacement during postural changes in relation to pressure ulcers in spinal cord-injured patients. Am J Phys Med Rehab 2008;87: 177-82. Bernard PL, Peruchon E, Micallef JP, Hertog C, Rabischong P. Balance and stabilization capability of paraplegic wheelchair athletes. J Rehabil Res Dev 1994;31:287-96. Allison G, Singer K. Assisted reach and transfers in individuals with tetraplegia: towards a solution. Spinal Cord 1997;35:217-22. Curtis K, Kindlin C, Reich K, White D. Functional reach in wheelchair users: the effects of trunk and lower extremity stabilization. Arch Phys Med Rehab 1995;76:360-7. Kukke SN, Triolo RJ. The effects of trunk stimulation on bimanual seated workspace. IEEE Trans Neur Syst Rehab Eng 2004; 12:177-85. Duncan PW, Weiner DK, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol 1990;45:M192-7. Lynch SM, Leahy P, Barker SP. Reliability of measurements obtained with a modified functional reach test in subjects with spinal cord injury. Phys Ther 1998;78:128-33. Bolin I, Bodin P, Kreuter M. Sitting position—posture and performance in C5-C6 tetraplegia. Spinal Cord 2000;38:425-34. Maynard F Jr, Bracken M, Creasey G, et al. International standards for neurological and functional classification of spinal cord injury. Spinal Cord 1997;35:266-74.
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
20. Chen CL, Yeung KT, Bih LI, Wang CH, Chen MI, Chien JC. The relationship between sitting stability and functional performance in patients with paraplegia. Arch Phys Med Rehab 2003;84:1276-81. 21. Burns AS, Ditunno JF. Establishing prognosis and maximizing functional outcomes after spinal cord injury: a review of current and future directions in rehabilitation management. Spine 2001; 26(24 Suppl):S137-45. 22. Lord SR, Menz HB, Tiedemann A. A physiological profile approach to falls risk assessment and prevention. Phys Ther 2003; 83:237-52. 23. Berg K, Wood-Dulphinee S, Williams J, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can 1989;41:304-11. 24. Dean CM, Shepherd RB. Task-related training improves performance of seated reaching tasks after stroke: a randomized controlled trial. Stroke 1997;28:722-8.
1577
25. Shrout P, Fleiss J. Intraclass correlations: uses in assessing rater reliability. Psych Bull 1979;86:420-8. 26. Perneger T. What’s wrong with Bonferroni adjustments. BMJ 1998;316:1236-8. 27. SPSS. SPSS for Windows base system user’s guide. Chicago: SPSS Inc; 1993. 28. Duval-Beaupere G, Robain G. Upward displacement of the centre of gravity in paraplegic patients. Paraplegia 1991;29:309-17. 29. Page R. The position and dependence on weight and height of the centre of gravity of the young adult male. Ergonomics 1974;17: 603-12. 30. Concato J, Feinstein A, Holford T. The risk of determining risk with multivariable models. Ann Int Med 1993;118:201-10. Supplier a. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Arch Phys Med Rehabil Vol 90, September 2009
ORIGINAL ARTICLE
Validity and Reliability of Assessment Tools for Measuring Unsupported Sitting in People With a Spinal Cord Injury Claire L. Boswell-Ruys, BAppSc (Physioth), Daina L. Sturnieks, PhD, Lisa A. Harvey, PhD, Catherine Sherrington, PhD, James W. Middleton, MBBS, PhD, Stephen R. Lord, PhD, DSc ABSTRACT. Boswell-Ruys CL, Sturnieks DL, Harvey LA, Sherrington C, Middleton JW, Lord SR. Validity and reliability of assessment tools for measuring unsupported sitting in people with a spinal cord injury. Arch Phys Med Rehabil 2009;90: 1571-7. Objectives: To develop simple tests to assess the abilities of people with spinal cord injury (SCI) to sit unsupported and to assess the construct validity and test-retest reliability of these tests. Design: Cross-sectional comparisons, convenience sample. Setting: Biomechanical laboratory. Participants: People (N⫽30) with SCI between the C6 and the L2 level of 2 months to 37 years duration before assessment. The sample was stratified by impairment level (at T8) and time since injury (1y postinjury). Interventions: Not applicable. Main Outcome Measures: On 2 separate occasions, participants performed tests that measured the distance of upperbody sway and maximal torso leaning, errors made during a coordinated stability task, timed dressing/undressing of the upper body and alternating arm reaching, and percentage change in seated upper body/arm reaching. Results: All tests showed good construct validity in that they distinguished between participants with higher (C6-T7) and lower (T8-L2) level impairments (P⬍.05) and between participants with acute (ⱕ1y) and chronic (⬎1y) lesions (P⬍.05). The tests also showed good to excellent test-retest reliability (intraclass correlation coeffiecient3,1 range, .51–.91). Conclusions: These simple and quick-to-administer tests have both construct validity and test-retest reliability. They would be appropriate for research and clinical purposes to quantify the abilities of people with SCI to sit unsupported. Key Words: Outcome assessment (health care); Posture; Rehabilitation; Spinal cord injuries. © 2009 by the American Congress of Rehabilitation Medicine
From the Prince of Wales Medical Research Institute, University of New South Wales, Randwick (Boswell-Ruys, Sturnieks, Sherrington, Lord); Rehabilitation Studies Unit and the University of Sydney, Sydney (Harvey, Middleton); and the George Institute for International Health, University of Sydney, Sydney (Sherrington), Australia. Presented to the Australian and New Zealand Spinal Cord Society, Melbourne, Australia, November 16 –18, 2006; and the World Confederation for Physical Therapy, Vancouver, Canada, June 2– 6, 2007. Supported by the New South Wales Premier’s Spinal Cord Injury Grant Program and the National Health and Medical Research Council of Australia. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Correspondence to Claire Boswell-Ruys, BAppSc (Physioth), Prince of Wales Medical Research Institute, Barker St, Randwick, NSW, 2031, Australia, e-mail: [email protected]. Reprints are not available from the author. 0003-9993/09/9009-00579$36.00/0 doi:10.1016/j.apmr.2009.02.016
HE ABILITY TO SIT unsupported in the able-bodied popT ulation requires the coordinated use of the whole body, the lower limbs, the trunk, the arms, and the head along with inputs from the sensory systems.1-3 Because of paralysis and sensory loss, people with a SCI have an impaired ability to sit unsupported. In general, the extent of this impairment depends on the neurologic level of injury and time since injury. Unsupported sitting is important for people with SCI because it increases independence with ADLs. Therapists invest significant time training patients with SCI to sit unsupported. Typically, training involves challenging patients to reach and move in different directions while seated. However, there are no standardized training guidelines, and it is not known whether such training strategies are effective. As a first step toward investigating these issues, it is important to devise valid and reliable ways of assessing the abilities of people with SCI to sit unsupported. Unsupported sitting is difficult to assess because it is complex; involves ongoing postural adjustments4; and, as an integral component of all daily activities, cannot be separated from the environment in which it is performed.5 In the laboratory, force plate transducers,6-10 piezo-resistive pressure systems,11 and motion-analysis systems12-15 have been used to assess and model unsupported sitting. This equipment is costly and not appropriate for use in clinical settings. The mFRT, adapted from the stroke population,16 has been used in people with SCI.17,18 It has been shown to be reliable. However, the mFRT is limited to only measuring a person’s ability to reach forward, and so it does not fully reflect a person’s ability to sit unsupported and move his/her center of mass to reach in different planes. There is a need for reliable and valid tests to comprehensively assess unsupported sitting ability in people with SCI, which encompass a range of tasks that have relevance to ADLs. The objectives of this study were (1) to devise a battery of simple, inexpensive, and portable tests suitable for use in the clinical setting that assess salient aspects of unsupported sitting; (2) to examine the reliability of these tests in people with SCI; and (3) to assess the correlation of the measures and determine the validity of the tests by assessing their ability to detect differences in neurologic level of SCI and time since injury.
List of Abbreviations ADLs AP ASIA ICC mFRT SCI
activities of daily living anteroposterior American Spinal Injury Association intraclass correlation coefficient modified Functional Reach Test spinal cord injury
Arch Phys Med Rehabil Vol 90, September 2009
1572
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
METHODS Participants Thirty people with SCI participated in this study (24 men, 6 women). The participants were between 18 and 66 years of age (mean ⫾ SD, 35⫾11y). Time since injury varied from 2 months to 37 years (mean ⫾ SD, 9⫾12y). Participants were included if they were able to sit out of bed for 2 hours and were free of sacral pressure areas. Participants were excluded if they had external bracing or injuries affecting their ability to lift their arms above their head. The neurologic level of injury ranged from C6 to L2 according to the International Standards for the Neurological Classification of Spinal Cord Injury.19 Twenty-one participants had complete (ASIA grade A) lesions, with the remainder having varying degrees of incomplete lesions (3 with ASIA grade B, 5 with ASIA grade C, 1 with ASIA grade D impairments). Participants were categorized into 2 groups based on their level of injury, signifying participants with none or some abdominal muscle innervation, a welldocumented factor identifying the ability of people with SCI to sit unsupported.8,12,20 The high-level group (n⫽14) included people with neurologic levels between C6 and T7, all of whom had ASIA grade A lesions. The remaining participants were in the low-level group (n⫽16), having neurologic levels between T8 and L2 with ASIA A to D impairment grades. Participants were also categorized into 2 groups depending on the time since injury; namely, people who sustained their injury up to 1 year before testing (acute: n⫽10) and people who sustained their injury more than 1 year before testing (chronic: n⫽20). One year postinjury is most widely considered to be classified as chronic as neurorecovery plateaus at approximately 12 months postinjury.21 It can be observed clinically that people with a chronic injury perform better on tasks than people with an acute injury for a given level of injury. There were no differences in the distributions of ASIA motor and sensory scores between the acute and chronic groups (Pⱖ.05). All participants provided informed consent before study participation. Testing Procedures The definition of unsupported sitting used in this study was the ability of a person to sit upright without using the hands or an external support. Participants sat on a flat seat without a backrest. The seat was adjusted so that the hips, knees, and ankles were at 90° with the feet resting on the floor (short sitting). The posterior aspect of the sacrum was positioned in line with the central posterior margin of the seat. The seat was designed to mimic the low treatment plinths commonly found in physiotherapy gymnasiums. A harness was used to ensure the participants’ safety and provide support during the rest periods between each test. The harness was connected to an overhead beam via a force transducer to ensure that the participants were not using the harness for support during the testing. If a force was registered, the test was stopped and recommenced from the beginning. Participants performed a battery of 6 tests on 2 occasions separated by a median (interquartile) of 7 days (3–13d). The order of test administration was the same for all participants, but participants were provided with rest periods as needed and between each test to minimize any fatigue that may have affected the performance of the tests. The second assessment was performed without reference to the first which had been placed in a sealed opaque envelope. Two authors were present at every session, but all testing and measurements were performed by the same examiner. Arch Phys Med Rehabil Vol 90, September 2009
The development of the tests was primarily modeled on tests used in the elderly and other neurologic populations.16,17,20,22-24 Those tests were modified to reflect situations when a person is required to maintain his/her center of mass over his/her base of support while sitting unsupported. These situations included the performance of slow controlled movements, fast alternating movements, and functional tasks. The 6 tests we developed are as follows. Upper-body sway. This test measured the abilities of participants to sit unsupported and remain as still as possible for 30 seconds. The Lord sway meter, which was originally designed to measure total body sway while attempting to stand still, was used to measure upper-body sway.22 It consisted of a 40-cm hinged rod that was fastened by a firm belt to the participants’ chest at the level of the axilla and extended in a horizontal plane from the body. For this test, the rod was positioned to extend in the posterior direction. A ball-point pen mounted vertically at the end of the rod recorded the movements of the upper body on a sheet of graph paper that was fixed to the top of a height-adjustable table. The tip of the pen was placed on the graph paper when the participants had commenced unsupported sitting. The resultant trace was measured in terms of 3 components: maximal lateral displacement, maximal AP displacement and the total length of the sway path (number of square mm traversed by the pen). Small displacement and length indicated better performance. The test was performed 3 times and the mean derived. Maximal balance range. This test was adapted from a standing test.22 Participants were asked to lean as far forward as they could without falling and then return to their starting position. They were then asked to lean back as far as possible without falling and then return to their starting position. The maximal AP distance traversed was measured by using the sway meter described earlier. However, the rod was positioned to extend in the anterior plane for visual feedback. The pen attached to the end of the rod recorded the anterior and posterior movements of the participants on a sheet of graph paper fastened to the top of a height-adjustable table. The participants had 2 attempts at the test, with the longer distance taken as the test result. The score recorded was the maximal AP distance moved. A long distance was considered a better performance. This score was corrected for body height (score ⫻ mean height/participants’ height) measured from the center of the sway meter strap to the top of the seat. Coordinated stability. The coordinated stability test was also adapted from a standing test.22 It measured participants’ abilities to adjust their sitting posture in a steady and coordinated way when near, or at the limits of, their postural equilibrium. The sway meter was again attached to the participants with the rod extending in the anterior plane for visual feedback. The participants were asked to adjust their posture by bending or rotating the upper body so that the sway meter pen tip followed and remained within a convoluted track. Two tracks were tested: the original version from the standing test (test A) and a simpler version (test B). Each track was marked on a piece of paper attached to the top of a height-adjustable table. To complete test A without errors, the sway meter pen had to remain within the track, which was 1.5-cm wide, and the participant had to be capable of adjusting the position of the pen 25cm laterally and 18cm in the AP plane. Test B comprised a less convoluted track and required smaller displacements of the pen (14cm laterally and 12cm anterioposteriorly). In both test versions, a total error score was calculated by summing (1) the number of occasions the pen failed to stay within the path (1 error point), (2) the number of track corners cut (3 error points), and (3) the number of times the participants
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
used their hands for support (5 points). Participants attempted each test twice, with the lesser score (better performance) taken as the test result. Alternating reach test. This test measured the abilities of participants to tap a table 8 times as fast as possible by using alternate arms on each repetition. This test was similar to item 12 of the Berg Balance Scale.23 A table was positioned at the height of the participants’ axillas with the closest edge an arm’s length away. The participants were tested under 2 conditions: supported, with their nonmoving hand grasping their thigh, and unsupported, with their nonmoving hand beside their body (shoulder at 0° and elbow at 90° flexion). The time taken to perform the tests was recorded with a stopwatch. A short time indicated a better performance. Seated reach distance. This test measured the participants’ abilities to reach in different directions as far as possible without falling. This test was similar to the one used in the stroke population.24 A large table was positioned with its closest edge in line with the greater trochanters of the participants and at the height of their iliac crests. A semicircle was cut out of the table to accommodate the participants’ abdomens. The table was covered with a large paper sheet with 5 predrawn lines. The directions of the lines were (1) lateral right (3 o’clock), (2) lateral left (9 o’clock), (3) 45° right (1.30 o’clock), (4) 45° left (10.30 o’clock), and (5) forwards (12 o’clock). A marker pen was taped into the thumb web spaces of the participants’ 2 hands. Without holding on for support, the participants were asked to reach sideways with the right hand as far as possible to make a light mark across the lateral right line before returning to the starting position. This was then performed with the left hand on the lateral left line. This procedure was repeated for each direction; those on the right side of the body with the right hand and those on the left side of the body with the left hand. The greatest reach distance for each direction was measured from the point of confluence of the lines. Arm length was measured from the acromion process to the position of the pen in the thumb web space. The seated reach distance was calculated as a proportion of the length of the participants’ arms (greatest reach distance/arm length); a high proportion indicated a better performance. A perpendicular measure from the top of the seat to the acromion process was used to correct the score for body height (score ⫻ mean height/participants’ height). T-shirt test. This test measured the time taken for participants to put on and take off a t-shirt. The test was designed to be similar to that reported by Chen et al20 but was performed in short sitting as described previously. A table was positioned with its closest edge in line with the participants’ knees and at the height of their iliac crests. A pullover t-shirt was spread out flat on the table face down. Standardized t-shirts were supplied and were 1 size larger than the participants would normally wear. No harness was worn during this test, but a research assistant was ready to support the participants if they were at risk of falling. The participants were required to put on and take off the t-shirt, resting between each maneuver. The test was repeated twice, with the average times calculated for each component (on, off, and total time). Short times indicated a better performance. Statistical Analysis ICC3,1 were calculated to evaluate the test-retest reliability of the 6 tests. The ICC values were interpreted according to a rating system suggested by Shrout and Fleiss25 (⬎.75 excellent reliability, .40 –.75 fair to good reliability, ⬍.40 poor reliability). Independent sample t tests were used to assess differences in tests between the participants with high- and low-level
1573
lesions and participants with acute and chronic injuries. Discriminant function analyses were used to determine the best combination of independent and significant tests for discriminating between injury level (low/high) and injury time (acute/ chronic) status. Spearman correlations were computed to assess the associations between the test scores and ASIA motor and ASIA sensory scores. Pearson correlation coefficients were also calculated to assess the correlation of the test measures. Any test measures with right-skewed distributions were log10 transformed in all analyses. Despite the multiple comparisons made, P values of less than .05 were considered statistically significant (ie, not adjusted to Bonferroni). This approach was taken because of the exploratory nature of this relatively small study for which such adjustments may increase type II errors.26 Analyses were performed by using SPSS for Windows.27,a RESULTS Test-Retest Reliability All participants performed all tests on the 2 testing occasions. The 6 tests were found to be reliable with the ICC3,1 values presented in table 1. The ICCs were excellent for the upper-body sway total length, alternating reach, maximal balance range, coordinated stability, seated reach, and t-shirt tests (ICC range, .79 –.91; 95% confidence interval, .60 –.96). The upper-body sway test AP and lateral range components showed fair to good reliability (ICCs⫽.51 and .72, respectively). Because the tests were deemed reliable, they were considered suitable for further validity analysis. Test Validity Construct validity was assessed in relation to ASIA scores, lesion level, and time since injury. The majority of the tests were significantly correlated with ASIA motor and ASIA sensory scores (table 2). The seated reach test had significant correlations with ASIA motor scores only. The upper-body sway measures poorly correlated with both ASIA motor and sensory scores (see table 2). The upper-body sway (AP and total length), maximal balance range, coordinated stability, and alternating reach tests had significant correlations with subject height (see table 2). This indicated that taller subjects performed more poorly in these tests. Results for the participants categorized in terms of lesion level and time since injury are shown in table 3. Participants with high lesions (C6-T7) performed significantly worse than participants with low lesions (T8-12) in the alternating reach, coordinated stability and tshirt tests, and half the reach directions in the seated reach test. Participants with an acute injury performed significantly worse than participants with a chronic SCI in at least 1 component of each test. The discriminant function analysis for level of injury identified the upper-body sway test total length and supported alternating reach test as being able to discriminate between high and low level impairments. The 2-variable model had a Wilk lambda of .68 (P⫽.006) and a canonical correlation for the discriminant function of .56 and correctly classified 90% of participants into high and low impairment groups, with high sensitivity (86%) and specificity (94%). The standardized canonical correlation coefficients were –.74 for upper-body sway total length and 1.10 for supported alternating reach. For time since injury, upper-body sway total length and seated reach 45° to the right discriminated between acute and chronic injury groups (Wilk ⫽.60, P⫽.001 and canonical correlation⫽.64). The standardized canonical correlation coefficients were –.67 for upper-body sway total length and 0.75 for seated reach 45° Arch Phys Med Rehabil Vol 90, September 2009
1574
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 1: Mean ⴞ SD Test and Retest Scores for the Unsupported Sitting Tests With Corresponding ICCs Test
Upper-body sway (mm) Total length* AP* Lateral* Maximal balance range (cm) Total range* Coordinated stability (score) Test A Test B Alternating reach (s) Supported* Unsupported* Seated reach distance (% arms length) Lateral right Lateral left 45° right 45° left Forward right Forward left T-shirt test (s) On* Off* Total*
Test 1
Test 2
ICC3,1 (95% CI)
90⫾89 18⫾12 15⫾8
104⫾102 24⫾20 14⫾9
0.79 (0.60–0.89) 0.72 (0.49–0.86) 0.51 (0.18–0.73)
23.1⫾14.4
24.2⫾15.5
0.90 (0.81–0.95)
26.6⫾17.5 11.0⫾8.4
21.2⫾15.6 9.8⫾8.9
0.83 (0.67–0.92) 0.86 (0.72–0.93)
4.9⫾1.3 5.5⫾2.4
4.7⫾1.5 5.4⫾2.7
0.82 (0.65–0.91) 0.88 (0.75–0.94)
109⫾10 107⫾10 109⫾13 105⫾15 98⫾22 96⫾24
110⫾10 107⫾11 108⫾12 106⫾14 95⫾29 97⫾23
0.80 (0.62–0.90) 0.86 (0.72–0.91) 0.83 (0.67–0.91) 0.82 (0.66–0.91) 0.87 (0.75–0.94) 0.89 (0.79–0.95)
16.5⫾13.9 9.4⫾9.2 25.6⫾22.8
14.9⫾13.1 8.0⫾6.9 22.9⫾18.7
0.91 (0.82–0.96) 0.85 (0.72–0.93) 0.89 (0.78–0.95)
NOTE. High scores in the sway, coordinated stability, alternating reach, and t-shirt tests and low scores in the maximal balance range and seated reach distance tests indicate impaired performance. Abbreviation: CI, confidence interval. *Natural logs used for statistical testing because of positively skewed data.
to the right. This model correctly classified 87% of participants into acute and chronic groups, again with high sensitivity (80%) and specificity (90%). Correlation of the Test Measures Table 4 presents the associations among the tests. There were significant correlations among all of the tests (P⬍.05) with the exception of some components of the upper-body Table 2: Correlations Between the Test Measures and ASIA Motor and ASIA Sensory Scores and Participant Height ASIA Unsupported Sitting Test
Motor
Sensory
Height
Upper-body sway total length Upper-body sway AP Upper-body sway lateral Maximal balance range Coordinated stability test A Coordinated stability test B Alternating reach supported Alternating reach unsupported Seated reach lateral right Seated reach lateral left Seated reach 45° right Seated reach 45° left Seated reach forward right Seated reach forward left T-shirt on T-shirt off T-shirt total
–.17 –.14 –.33 .37* –.56* –.58* –.51* –.57* .26 .22 .30 .34 .41* .47* –.75* –.62* –.73*
–.05 –.02 –.33 .46* –.47* –.55* –.58* –.54* .29 .42* .40* .43* .44* .48* –.63* –.57* –.61*
.50* .61* .33 –.42* .44* .24 .52* .54* –.36 –.32 –.35 –.32 –.22 –.19 .27 .19 .23
*P⬍.05.
Arch Phys Med Rehabil Vol 90, September 2009
sway test. The strongest correlations were between the coordinated stability test A and the put-on component of the t-shirt test (r⫽.79, P⬍.01), the coordinated stability test B and the supported alternating reach test (r⫽.72, P⬍.01), and all of the seated reach directions with the maximal balance range test (r⫽.69 –.80, P⬍.01). As expected, there were also significant correlations among the components of each test (r⫽.59 –.98, P⬍.01). The strongest correlations were evident within the alternating reach (r⫽.93, P⬍.01) and t-shirt tests (r⫽.86 –.96, P⬍.01). DISCUSSION The 6 tests used in this study were developed to measure unsupported sitting ability in people with SCI. For independence and safety in daily activities, it is necessary for a person with an SCI to be able to maintain his/her center of mass over his/her base of support as he/she performs slow controlled movements, fast alternating movements, and functional tasks. The tests developed encompass these aspects of daily living. The 6 tests proved to have construct validity in that they could discriminate between people with acute and chronic injuries and people with higher and lower lesions. The tests also showed good to excellent test-retest reliability. This study showed that tests of controlled unsupported sitting (coordinated stability test) and functional tasks (t-shirt test), in addition to reaching ability (seated reach distance and alternating reach tests), were able to discriminate between people with lesions above or below the level of T8 in univariate analyses. People with a lesion below the T8 impairment level have intact abdominal muscle innervation and so would reasonably be expected to perform better in unsupported sitting tests than people with higher lesions. Previous work17,20 has found that reaching tests can distinguish a person’s neurologic level. The discriminant function analysis identified supported
1575
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 3: Mean ⴞ SD of Groups Based on Time Since Injury and Level of Injury Time Since Injury Test
Upper-body sway (mm) Total length AP Lateral Maximal balance range (cm) Total distance Coordinated stability (score) Test A Test B Alternating reach (s) Supported Unsupported Seated reach distance (% arm length) Lateral right Lateral left 45° right 45° left Forward right Forward left T-shirt test (s) On Off Total ASIA score Motor Sensory
Level of Injury
Acute (n⫽10)
Chronic (n⫽20)
High C6-T7 (n⫽14)
Low T8-T12 (n⫽16)
149⫾129 24⫾11 19⫾12
61⫾38* 15⫾11* 14⫾6
76⫾49 18⫾13 15⫾5
103⫾113 18⫾11 16⫾11
12.8⫾4.9
29.1⫾16.5*
17.2⫾9.3
29.3⫾18.2
44.2⫾2.2 16.3⫾9.4
21.1⫾14.3* 11.4⫾13.7
36.9⫾15.8 18.5⫾15.5
22.3⫾20.8* 8.5⫾8.0*
5.6⫾1.4 6.5⫾2.5
5.0⫾2.0 5.0⫾2.3*
6.1⫾2.2 6.8⫾2.8
4.4⫾0.9* 4.4⫾1.4*
100⫾9 100⫾12 97⫾11 91⫾14 81⫾15 79⫾16
116⫾18* 113⫾17* 117⫾20* 114⫾19* 108⫾27* 107⫾28*
106⫾15 102⫾13 103⫾16 98⫾17 88⫾20 85⫾19
115⫾18 114⫾18 116⫾21 113⫾21* 109⫾28* 108⫾30*
24.6⫾16.6 14.0⫾12.0 38.6⫾27.5
12.4⫾10.6* 7.1⫾6.7* 19.0⫾17.2*
23.8⫾15.7 13.2⫾9.6 36.3⫾25.4
10.1⫾8.2* 6.1⫾7.8* 16.2⫾15.6*
51⫾5 133⫾44
53⫾12 124⫾44
47⫾7 86⫾15
57⫾10* 162⫾25*
*P⬍.05.
alternating reach and upper-body sway as independent and significant predictors of injury level groups. Interestingly, upper-body sway did not correlate with ASIA motor or sensory scores but did provide unique predictive ability in the multivariate model, despite not being significantly associated with neurologic level in a univariate analysis. This suggests that only after adjusting for reaching ability, the association between upper-body sway and neurologic level becomes apparent. Combined, these 2 complementary unsupported sitting measures had both high sensitivity (86%) and specificity (94%) with regard to classifying people based on SCI level. The validity of the tests was shown further by the significant associations with ASIA scores. People with higher ASIA scores were able to reach further than those with a lower score, similar to the findings of Lynch et al.17 The more available muscle activity and preserved sensation that a person has, the faster he/she is able to perform time-related tests and tests that require greater postural control. All tests were able to distinguish between people with acute and chronic SCI. People with chronic lesions had faster performance times, greater stability, and a longer reach. With increasing time, people with SCI adapt to their impairments, know their abilities better, and are more comfortable with challenging themselves to their limits. Thus, time since injury also comprises a good measure for evaluating the discriminatory ability of assessment measures. Upper-body sway and one of the reach measures (seated reach 45° to the right) were identified as independent predictors of time since injury. As with neurologic level, complementary measures of seated postural control and reaching ability would accurately discriminate between people with acute and chronic SCI, showing 80% sensitivity and 90% specificity.
Participants with a longer trunk had increased sway, took longer to complete tasks, and were unable to reach as far while sitting unsupported. This finding supports clinical observations and compliments the work of Chen et al.20 People with SCI have an upward shift of their center of gravity by approximately 5% of their body length compared with people without paralysis,28 which increases the degree of difficulty in maintaining their center of gravity over their base of support. Furthermore, people with longer bodies have a higher center of gravity.29 To maintain this higher center of gravity over their base of support requires the recruitment of distal musculature, which may be impaired. Thus, when assessing a person’s unsupported sitting, the height of the person is an important consideration. Study Limitations Currently, there are no tests that comprehensively measure the unsupported sitting ability of people with SCI; therefore, this trial was exploratory in nature. Although this is the major study strength, it also highlights certain limitations. In validating the test measures, there was a sound rationale for categorizing participants with regard to level of injury (presence or absence of abdominal muscle innervation), but we acknowledge that the criterion of 1 year for the time since injury measure was somewhat arbitrary. We also acknowledge that while being of sufficient size for measuring test-retest reliability, the sample was at the margin for assessing the validity of the balance measures, particularly in relation to multivariate analysis.30 Thus, the discriminant function models identified here require verification in larger external samples. Finally, because we chose not to adjust P values to avoid type II Arch Phys Med Rehabil Vol 90, September 2009
1576
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys Table 4: Correlations Among the Test Measures Upper-Body Sway
MBR
Coordinated Stability
Alternating Reach
Seated Reach Distance
T-Shirt Test
Unsupported Sitting Test
Length
AP
Lateral
Distance
Test A
Test B
Support
Unsupp
Lat R
Lat L
45° R
45° L
Fwd R
Fwd L
On
Off
Sway AP Sway Lat MBR distance Co Stab A Co Stab B AR Support AR Unsupp SR Lat R SR Lat L SR 45° R SR 45° L SR Fwd R SR Fwd L T-shirt on T-shirt off T-shirt total
.85* .75* –.31 .47* .44* .48* .51* –.32 –.35 –.36 –.37* –.62 –.22 .37* .12 .29
.59 –.32 .48* .55* .50* .43* –.49* –.42* –.47* –.43* –.30 –.27 –.38* .17 .31
–.25 .36 .56* .46* .44* –.25 –.27 –.26 –.31 –.25 –.24 .41* .13 .32
–.58* –.45* –.58* –.51* .74* .69* .78* .74* .80* .79* –.55* –.55* –.55*
.72* .67* .69* –.56* –.48* –.62* –.61* –.67* –.67* .79* .62* .71*
.72* .49* –.64* –.55* –.54* –.45* –.36 –.36 .57* .34 .49*
.93* –.54* –.65* –.59* –.55* –.48* –.44* .63* .50* .56*
–.36 –.55* –.60* –.63* –.63* –.57* .65* .48* .55*
.78* .82* .75* .63* .62* –.54* –.46* –.53*
.73* .84* .64* .64* –.56* –.48* –.54*
.91* .87* .85* –.65* –.65* –.67*
.87* .87* –.65* –.63* –.67*
.98* –.66* –.63* –.65*
–.64* –.66* –.67*
.86* .96*
.94*
Abbreviations: AR, alternating reach; Co Stab, coordinated stability; Fwd, forward; L, left; Lat, lateral; MBR, maximal balance range; R, right; SR, seated reach distance; Support, supported; Sway, upper-body sway; Unsupp, unsupported. *P⬍.05.
errors,26 it is possible that some of the significant univariate associations revealed may have occurred by chance. Based on the validity and reliability findings of this trial, the following tests are recommended as a minimum set for comprehensively assessing unsupported sitting in people with SCI: upper-body sway test total length, seated reach distance (45° to the left and/or right), supported alternating reach test, coordinated stability test A, and t-shirt put-on test. These tests could be routinely used in clinical settings because each takes less than 3 minutes to administer and requires a minimal amount of equipment. Each test should be used to assess a person’s performance against himself/herself until further research is performed that may allow comparison to the SCI population as a whole. Additional research should determine the responsiveness of each test to ensure they also detect meaningful change over time. CONCLUSIONS A battery of tests was developed to measure unsupported sitting in people with SCI. The tests are simple, reliable, and quick to administer. They would be suitable for use in further research and the clinical setting. 1.
2.
3.
4. 5. 6.
References Dean CM, Shepherd RB, Adams R. Sitting balance I: trunk-arm coordination and the contribution of the lower limbs during selfpaced reaching in sitting. Gait Posture 1999;10:135-46. Dean CM, Shepherd RB, Adams R. Sitting balance II: reach direction and thigh support affect the contribution of the lower limbs when reaching beyond arm’s length in sitting. Gait Posture 1999;10:147-53. Crosbie J, Shepherd RB, Squire TJ. Postural and voluntary movement during reaching in sitting: the role of the lower limbs. J Human Mov Studies 1995;28:103-26. Huxham FE, Goldie PA, Patla AE. Theoretical considerations in balance assessment. Aust J Physiother 2001;47:89-100. Carr JH, Shepherd RB. A motor relearning programme for stroke. 2nd ed. Oxford: Butterworth-Heinemann; 1987. Aissaoui R, Boucher C, Bourbonnais D, Lactoste M, Dansereau J. Effect of seat cushion on dynamic stability in sitting during a
Arch Phys Med Rehabil Vol 90, September 2009
7.
8.
9.
10.
11.
12.
13. 14.
15.
16. 17.
18. 19.
reaching task in wheelcahir users with paraplegia. Arch Phys Med Rehab 2001;82:274-81. Shirado O, Kawase M, Minami A, Strax TE. Quantitative evaluation of long sitting in paraplegic patients with spinal cord injury. Arch Phys Med Rehab 2004;85:1251-6. Seelen HA, Potten YJ, Huson A, Spaans F, Reulen JP. Impaired balance control in paraplegic subjects. J Electromyogr Kinesiol 1997;7:149-60. Janssen-Potten YJ, Seelen HA, Drukker J, Spaans F, Drost MR. The effect of footrests on sitting balance in paraplegic subjects. Arch Phys Med Rehab 2002;83:642-8. Kamper D, Parnianpour M, Barin K, Adams T, Linden M, Hemami H. Postural stability of wheelchair users exposed to sustained, external perturbations. J Rehabil Res Dev 1999;36:121-32. Karatas GK, Tosun AK, Kanatl U. Center-of-pressure displacement during postural changes in relation to pressure ulcers in spinal cord-injured patients. Am J Phys Med Rehab 2008;87: 177-82. Bernard PL, Peruchon E, Micallef JP, Hertog C, Rabischong P. Balance and stabilization capability of paraplegic wheelchair athletes. J Rehabil Res Dev 1994;31:287-96. Allison G, Singer K. Assisted reach and transfers in individuals with tetraplegia: towards a solution. Spinal Cord 1997;35:217-22. Curtis K, Kindlin C, Reich K, White D. Functional reach in wheelchair users: the effects of trunk and lower extremity stabilization. Arch Phys Med Rehab 1995;76:360-7. Kukke SN, Triolo RJ. The effects of trunk stimulation on bimanual seated workspace. IEEE Trans Neur Syst Rehab Eng 2004; 12:177-85. Duncan PW, Weiner DK, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol 1990;45:M192-7. Lynch SM, Leahy P, Barker SP. Reliability of measurements obtained with a modified functional reach test in subjects with spinal cord injury. Phys Ther 1998;78:128-33. Bolin I, Bodin P, Kreuter M. Sitting position—posture and performance in C5-C6 tetraplegia. Spinal Cord 2000;38:425-34. Maynard F Jr, Bracken M, Creasey G, et al. International standards for neurological and functional classification of spinal cord injury. Spinal Cord 1997;35:266-74.
ASSESSMENT OF SITTING IN SPINAL CORD INJURY, Boswell-Ruys
20. Chen CL, Yeung KT, Bih LI, Wang CH, Chen MI, Chien JC. The relationship between sitting stability and functional performance in patients with paraplegia. Arch Phys Med Rehab 2003;84:1276-81. 21. Burns AS, Ditunno JF. Establishing prognosis and maximizing functional outcomes after spinal cord injury: a review of current and future directions in rehabilitation management. Spine 2001; 26(24 Suppl):S137-45. 22. Lord SR, Menz HB, Tiedemann A. A physiological profile approach to falls risk assessment and prevention. Phys Ther 2003; 83:237-52. 23. Berg K, Wood-Dulphinee S, Williams J, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can 1989;41:304-11. 24. Dean CM, Shepherd RB. Task-related training improves performance of seated reaching tasks after stroke: a randomized controlled trial. Stroke 1997;28:722-8.
1577
25. Shrout P, Fleiss J. Intraclass correlations: uses in assessing rater reliability. Psych Bull 1979;86:420-8. 26. Perneger T. What’s wrong with Bonferroni adjustments. BMJ 1998;316:1236-8. 27. SPSS. SPSS for Windows base system user’s guide. Chicago: SPSS Inc; 1993. 28. Duval-Beaupere G, Robain G. Upward displacement of the centre of gravity in paraplegic patients. Paraplegia 1991;29:309-17. 29. Page R. The position and dependence on weight and height of the centre of gravity of the young adult male. Ergonomics 1974;17: 603-12. 30. Concato J, Feinstein A, Holford T. The risk of determining risk with multivariable models. Ann Int Med 1993;118:201-10. Supplier a. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Arch Phys Med Rehabil Vol 90, September 2009