Comparison of the Test-Retest Reliability of the Balance Computerized Adaptive Test and a Computerized Posturography Instrument in Patients With Stroke

Archives of Physical Medicine and Rehabilitation journal homepage: www.archives-pmr.org Archives of Physical Medicine and Rehabilitation 2014;-:-------

ORIGINAL ARTICLE

Comparison of the Test-Retest Reliability of the Balance Computerized Adaptive Test and a Computerized Posturography Instrument in Patients With Stroke Chia-Hsin Chen, MD, PhD,a,b,c,* Shih-Feng Lin, BS,a,* Wan-Hui Yu, MS,d Jau-Hong Lin, PhD,e,f Hao-Ling Chen, PhD,d Ching-Lin Hsieh, PhDd From the aDepartment of Physical Medicine and Rehabilitation, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung; bDepartment of Physical Medicine and Rehabilitation, Kaohsiung Medical University Hospital, Kaohsiung; cDepartment of Physical Medicine and Rehabilitation, School of Medicine, Graduate Institute of Neuroscience, College of Medicine, Kaohsiung Medical University, Kaohsiung; dSchool of Occupational Therapy, College of Medicine, National Taiwan University, Taipei City; eDepartment of Physical Therapy, College of Health Science, Kaohsiung Medical University, Kaohsiung; and fDepartment and Graduate Institute of Neurology, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan. *Chen and Lin contributed equally to this work.

Abstract Objective: To compare the test-retest reliabilities of the scores of the Balance Computerized Adaptive Test (CAT) and the Biodex Balance System in patients with stroke. Design: A repeated-measures design (at a 1-wk interval) was used to examine the test-retest reliabilities of the scores of the Balance CAT and the Biodex Balance System. Setting: One rehabilitation unit in a local hospital. Participants: Patients (NZ50) with stroke for more than 6 months and undergoing outpatient rehabilitation completed the Balance CAT and the eyes open (EO)/closed (EC) tests, but only 17 patients finished the Limit of Stability (LOS) test because they were unable to reach all the targets. Interventions: Not applicable. Main Outcome Measures: The Balance CAT and 2 computerized tests of the Biodex Balance System, namely the EO/EC test and the LOS, were used to evaluate balance function. Results: The test-retest reliabilities of the scores of the Balance CAT (Pearson rZ.92, minimal detectable change [MDC] percentZ12.8%) was excellent. Those of the EO/EC and LOS tests were poor to good (Pearson rZ.56e.85, MDC%Z50.8%e126.9%). Conclusions: The test-retest reliabilities of the scores of the Balance CAT were sufficient for assessing balance function in patients with stroke. Moreover, the test-retest reliabilities of the scores of the Balance CAT, one of the functional balance measures, were superior to those of the Biodex Balance System, 1 type of computerized posturography instrument. Therefore, the Balance CAT may be a more reliable measure for clinicians and researchers to use in assessing the balance function of patients with stroke for more than 6 months. Archives of Physical Medicine and Rehabilitation 2014;-:------ª 2014 by the American Congress of Rehabilitation Medicine

A deficit in balance is a common problem in patients with stroke. This deficit may seriously impair their function in activities of Supported by Kaohsiung Municipal Ta-Tung Hospital in Taiwan (grant no. KMTTH-100-026). Disclosures: none.

daily living such as walking and stair climbing, thereby increasing the risk of falling.1-4 Improving balance function is an important goal of rehabilitation.5 Through the use of reliable functional balance measurements, clinicians can effectively assess and monitor the balance function of patients with stroke.6,7

0003-9993/14/$36 - see front matter ª 2014 by the American Congress of Rehabilitation Medicine http://dx.doi.org/10.1016/j.apmr.2014.03.005

2

C-H Chen et al

Functional balance measures, which assess balance function through functional tasks, are one type of balance measure frequently used in clinical and research settings. These functional balance measures, such as the Berg Balance Scale and Postural Assessment Scale for Stroke Patients (PASS), have practical advantages, including ease of use and low cost.8 Unfortunately, few of them can achieve both efficiency and precision, both of which are needed in busy clinics. Computerized adaptive testing has been suggested to meet this need.9,10 The Balance Computerized Adaptive Test (CAT) has been developed to efficiently and precisely assess balance function in patients with stroke through a computerized adaptive testing system.11 The Balance CAT chooses items tailored to an individual patient and skips items that are too easy or too difficult for that patient. This advantage allows clinicians to efficiently assess balance function when time is limited. Some psychometric properties of the scores of the Balance CAT in inpatients with stroke receiving rehabilitation, namely the responsiveness and predictive validity, have been verified.12 However, the test-retest reliability of the Balance CAT in patients with stroke has not been examined. Apart from functional balance scales, another type of balance measure is computerized posturography instruments, such as the Smart Balance Master System (SBM) and Biodex Balance System.a This type of balance measure has been developed to objectively quantify standing balance function with high resolution and thus help clinicians precisely monitor the changes in patients’ balance function.13,14 Objective quantification with high resolution is thought to be helpful for yielding better test-retest reliability (eg, test-retest reliability). Thus, the test-retest reliabilities of the scores of computerized posturography instruments in patients with stroke might be better than those of functional balance scales. Recently, the Biodex Balance System, which is less expensive and more portable than the SBM, has frequently been used in clinical or sports rehabilitation settings.15 However, since no studies have examined the test-retest reliabilities of the scores of the Biodex Balance System in patients with stroke, its utility in such patients is limited. Although the Biodex Balance System and the Balance CAT measure balance function in different ways, both are frequently used to represent and monitor the balance function of patients in clinics and research. Before these balance measures are used, their test-retest reliabilities must be confirmed, compared head to head, or both. Therefore, the purposes of the study were to compare the test-retest reliabilities and random measurement errors (ie, minimal detectable change [MDC]) of the Biodex Balance System and Balance CAT in patients with chronic stroke. The results should be useful for both clinicians and researchers in selecting an appropriate balance measure for patients with chronic stroke. We hypothesized that the test-retest reliabilities of the scores of the Biodex Balance System, which can objectively quantify balance function with high resolution, would be superior to that of a functional balance measure, namely the Balance CAT.

List of abbreviations: CAT EC EO LOS MDC PASS SBM SEM

Computerized Adaptive Test eyes closed eyes open limit of stability minimal detectable change Postural Assessment Scale for Stroke Patients Smart Balance Master System standard error of measurement

Methods Participants We recruited patients undergoing outpatient rehabilitation in an urban rehabilitation center, which provides typical stroke rehabilitation. The following criteria were used to determine whether patients could be included in this study: (1) diagnosis of cerebral hemorrhage or cerebral infarction; (2) stroke onset >6 months before the start of the study (ie, chronic patients); (3) ability to stand without assistive devices for 30 seconds; (4) ability to follow instructions to complete the assessments; (5) no visual impairment; and (6) absence of other neuromusculoskeletal diseases (eg, Parkinsonism, osteoarthritis) that could affect a patient’s balance function. Patients who met the inclusion criteria were invited by a trained physical therapist (S.-F.L.) to join our study. All the invited patients participated in our study. Informed consent for participation was obtained from the participants personally or by proxy. Patients who had previously been trained in using the Biodex Balance System were excluded in this study. This study was approved by the institutional review boards of local ethics committees.

Procedures Demographic data were collected from all participants before testing to determine the general characteristics of the sample population. All subjects were evaluated initially using the Balance CAT and the Biodex Balance System and were then reassessed after 1-week intervals. For each subject, both measures in the testretest sessions were administered by the physical therapist (S.-F.L.) in a counterbalanced order within 24 hours. To ensure that the subjects’ conditions were stable during the study period, changes in their medical conditions between the 2 assessments were recorded. We excluded 13 subjects who developed recurrent strokes or other medical conditions that might result in poorer performance in balance during the test-retest period.

Measures Balance CAT The Balance CAT is a computerized adaptive test developed to evaluate balance function in patients with stroke.11 This measure contains a pool of 34 easily administered items that can be administered by a clinician using a personal digital device connected to the Internet. The resulting scores, transformed using item response theory, range from 0 to 10. The Balance CAT chooses items tailored to an individual patient’s balance function and skips items that are too easy or too difficult for the patient. For example, if a patient cannot stand independently, the system will not select a more difficult task (eg, stand on 1 leg) for further testing. Instead, the system will select an easier task (eg, sit without support). In general, the Balance CAT takes only 4 items (completed within 2min on average) to obtain reliable and valid balance assessments of patients with stroke.11,12 For all the testing with the Balance CAT, the subjects wore their own shoes. The feasibility of the Balance CAT in patients with stroke, in terms of sufficient validity and responsiveness, has been previously determined in inpatients with stroke.11,12 Biodex Balance System The Biodex Balance System, a commercially available device, is used to quantify and train the ability of standing balance in www.archives-pmr.org

Reliability of Balance Computerized Adaptive Test and Biodex

3

patients with pathologic conditions.16 In this study, the subjects were instructed to stand independently on the platform without footwear, keeping their feet apart at a distance equal to the width of the shoulders and their arms by their sides, without support. In the Biodex Balance System, the Clinical Test of Sensory Integration and the Balance and Limit of Stability (LOS) tests were used. Since patients with stroke may have poor balance function, only the eyes open/eyes closed (EO/EC) tests on a firm surface were used in the present study because of safety considerations. For each trial of each condition, subjects were encouraged to maintain a stable upright posture for 30 seconds. Two trials were performed for each test condition, with resting periods of 10 seconds between trials and 5 minutes between conditions. The measures of the EO/EC test were the stability index (EOstability/EC-stability) and the sway index (EO-sway/EC-sway). High scores in the stability index and the sway index indicate poor balance.16 For the LOS test, subjects were instructed to shift their center of gravity to move a ball-shaped cursor from the center target to a blinking target and back as quickly and with as little deviation as possible, keeping their body in a straight line, with the ankle joint as the primary axis of rotation. Three trials were performed, with resting periods of 3 minutes between trials. The target placement was preset by the Biodex Balance System at 50% of the theoretical LOS, based on the subject’s height. Before testing, 1 practice trial was performed. The measures of the LOS test were the duration to reach all targets (LOS-time) and the overall directional control score (LOS-overall control). Higher scores in overall directional control and lower scores in duration indicate better dynamic postural control.16 For each condition, the average of trial data was used for subsequent analysis.

Data analysis Data were analyzed with the IBM SPSS Statistics 19.0 program.b To investigate the test-retest reliability of the Balance CAT and the Biodex Balance System, the Pearson correlation coefficient was used. A Pearson r value >.75 indicates good reliability (>.90, excellent); .50 to .75, moderate to good reliability; and <.50, poor reliability.17 The Pearson r values of both balance systems were also compared using Modified Pearson-Filon Statistics.18 A paired t test was used to examine whether systematic bias existed between repeated measurements. The MDC was calculated from the standard error of measurement (SEM) to quantify the test-retest reproducibility of the measures.19,20 The MDC, with a confidence level of 95%, was calculated by using the following formulas21: SEMZSDall testing scores


pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1  rtestretest

pffiffiffi MDCZ1:96
2
SEM

A total of 50 patients with chronic stroke were recruited in this study. All participants completed the Balance CAT and the EO/EC test twice, at about 1 week apart. However, of the 50 patients, 33 could not finish the LOS test because they were unable to reach all the targets. Therefore, the results of the LOS test were reported based on the data from the 17 remaining patients. The demographic information, clinical characteristics, and functional statuses of the recruited subjects are shown in table 1. Because the functional statuses of the 17 subjects who completed the LOS test were different from those who did not (nZ33) (see table 1), results of the Balance CAT and the EO/EC were divided into 2 parts (nZ17 and NZ50) (table 2).

Data for subjects who completed Balance CAT and EO/EC tests of the Biodex Balance System The mean value of the Balance CAT was about 7 for the first and second assessments, indicating that the subjects in the study, on average, could walk with minimal support and independently transfer between sitting and standing. The Pearson r for the Balance CAT was .92, indicating excellent test-retest reliability. For the Biodex Balance System, the Pearson r ranged from .57 to .85 (see table 2), indicating moderate to good test-retest reliability. The Pearson r values of the Balance CAT were significantly higher than those of the Biodex Balance System (P<.05) (see table 2), except for the comparisons with EO-sway and LOS-overall control of the Biodex Balance System. For the Balance CAT, the paired t test showed no significant difference between test-retest assessments. However, for the Biodex Balance System, a significant difference was found in the EO-sway item (PZ.04). The differences in scores are plotted against the mean scores of the 2 repeated measurements for the Balance CAT and Biodex Balance System for each participant in figure 1. The limits of agreement ratio of the Balance CAT was smaller, indicating that the test-retest reliabilities of the scores of the Balance CAT were higher than those of the Biodex Balance System. The MDC of the Balance CAT was .91 points. The MDC% of the Balance CAT was 12.8%, indicating limited random measurement error. However, the MDC% of all the variables obtained from the Biodex Balance System ranged from 50.5% to 126.9% (see table 2), indicating a substantial to large random measurement error.

Data for subjects who completed all the tests ð1Þ ð2Þ

In these formulas, SDall testing scores is the SD of all scores of 2 sessions’ assessments, and rtest-retest is the coefficient of the testretest reliability derived from the Pearson r in the study. The multiplier of O2 refers to the additional uncertainty introduced by the inclusion of scores from 2 separate assessments. The MDC can also be expressed as a percentage of the parameter’s grand mean (MDC%), which is independent of the units of measurement. An MDC% >30% is considered poor, one of 10% to 30% is considered acceptable, and an MDC% <10% is considered excellent.22 We also used Bland-Altman plots with 95% limits of agreement to visualize the agreement between 2 repeated measurements.23 www.archives-pmr.org

Results

For subjects who completed the LOS tests (nZ17), poor-tomoderate and excellent test-retest reliabilities were found in the Biodex Balance System and Balance CAT, respectively (see table 2). Results of MDC and MDC% also indicate a limited random measurement error for the Balance CAT and a substantial to large one for the Biodex Balance System.

Discussion This study is the first to establish the test-retest reliability, including the Pearson r, paired t test, and MDC values, of the Balance CAT and the Biodex Balance System in patients with chronic stroke. The Pearson r of the Balance CAT was significantly higher than the Pearson r values of all the tests in the Biodex Balance System, except for 2 comparisons (ie, the

4

C-H Chen et al Table 1

Demographics, clinical characteristics, and functional status of the subjects in the study

Characteristic Demographics Age (y) Sex (M/F) Clinical characteristics Stroke type Cerebral hemorrhage Cerebral infarction Side of hemiplegia (R/L) Period of onset to initial evaluation (mo) Functional status Brunnstrom stage Upper extremity Lower extremity Modified Ashworth Scale Upper extremity Lower extremity Barthel Index

NZ50

nZ33

nZ17

P*

61.810.9 (33e84) 31/19

64.19.4 (49e84) 23/10

57.311.9 (33e80) 8/9

.04y .11

13 37 26/24 29.431.7 (6e150)

10 23 19/14 32.535.5 (6e150)

3 14 7/10 23.421.0 (7e65)

.21 .35

3.91.1 (2e6) 4.50.7 (3e6)

3.71.1 (2e6) 4.40.7 (3e6)

4.21.0 (2e6) 4.80.6 (4e6)

.17 .05

.27

1.71.1 (0e4) 1.41.0 (0e4) 94.49.6 (55e100)

1.71.2 (0e4) 1.41.1 (0e4) 92.611.1 (55e100)

1.80.9 (0e4) 1.40.7 (0e2) 97.93.9 (90e100)

.76 .95 .02y

NOTE. Values are mean  SD (minimumemaximum), n, or as otherwise indicated. Abbreviations: F, female; L, left; M, male; R, right. * P value from independent-samples t test between subjects who completed the LOS test (nZ17) and those who did not (nZ33). y Indicates significant difference between 2 groups (P<.05).

EO-sway and LOS-overall control of the Biodex Balance System). These 2 comparisons had similar but not significant results. Although the EO-sway of the Biodex Balance System had statistically comparable Pearson r values, systematic bias was found between test-retest assessments in the EO-sway test, suggesting that test-retest reliability was compromised for the score of EOsway of the Biodex Balance System. These results substantially

Table 2 Measures

support the argument that the test-retest reliabilities of the scores of the Balance CAT were higher than those of the Biodex Balance System in patients with chronic stroke, which generally disproves our hypothesis. The measurements of the Biodex Balance System are commonly believed to yield better psychometric properties than those of clinical balance scales. That belief is based on the

Test-retest reliability and the MDC for Balance CAT and Biodex Balance System 1st Assessment

Data for subjects who completed Balance CAT Balance CAT (NZ50) 7.161.12 Biodex Balance System EO-stability (NZ50) 2.211.72 EO-sway (NZ50) 0.990.45 EC-stability (NZ50) 3.122.05 EC-sway (NZ50) 1.520.76 Data for subjects who completed all the tests Balance CAT (nZ17) 7.871.16 Biodex Balance System EO-stability (nZ17) 1.741.14 EO-sway (nZ17) 0.880.42 EC-stability (nZ17) 2.801.25 1.440.78 EC-sway (nZ17) LOS-time (nZ17) 63.7632.3 LOS-overall control 27.7112.3 (nZ17)

2nd Assessment

Paired t test t (P)

Pearson r

and EO/EC of Biodex Balance System 7.161.23 0.10 (.92) .92

Modified Pearson-Filon Statistics ZPF (P) NA

2.041.41 1.070.57 2.992.02 1.671.04

0.85 2.08 0.64 1.73

(.40) (.04) (.53) (.09)

.57 .85 .72 .78

4.53 1.58 3.27 2.58

8.031.22

1.29 (.22)

.91

NA

0.05 1.29 0.36 0.93 0.09 0.63

.64 .75 .41 .55 .56 .67

1.99 1.43 2.86 2.35 2.44 1.87

1.751.27 0.980.50 2.931.80 1.600.83 63.0635.18 29.2412.41

(.96) (.22) (.72) (.37) (.93) (.54)

(<.001) (.13) (.01) (.02)

(.07) (.17) (.02) (.04) (.04) (.09)

SEM

MDC (MDC%)

0.33

0.91 (12.8)

0.97 0.19 1.00 0.41

2.70 0.52 2.76 1.14

0.35

0.98 (12.3)

0.67 0.21 1.04 0.51 22.06 7.00

1.86 0.58 2.88 1.42 61.15 19.41

(126.9) (50.5) (90.4) (71.2)

(106.8) (62.0) (100.6) (93.4) (96.4) (68.2)

NOTE. Values are mean  SD or as otherwise indicated. Abbreviations: EC-stability, stability index (degree) of EC condition; EC-sway, sway index (degree) of EC condition; EO-stability, stability index (degree) of EO condition; EO-sway, sway index (degree) of EO condition; LOS-time, time duration (s) of LOS test; LOS-overall control, overall directional control (%) of LOS test; NA, not applicable; ZPF, z score of modified Pearson-Filon statistic.

www.archives-pmr.org

5

Balance CAT (N=50)

1.5

Mean+1.96*SD=0.97

1 0.5

Mean=0.01

0 -0.5 -1

Mean-1.96*SD=-0.96

-1.5 0

6 4

Mean+1.96*SD=2.60

2 Mean=-0.17

0 -2

Mean-1.96*SD=-2.94

-4 -6 0

2 4 6 Mean of Test-Retest Assessments

8

Stability Index of Eye Close Condition of BBS (N=50) 6 4

Mean+1.96*SD=2.66

2 Mean=-0.13

0 -2

Mean-1.96*SD=-2.92

-4 -6 0

2 4 6 8 Mean of Test-Retest Assessments

10

Duration of Limit of Stability Test of BBS (N=17)

100

Mean+1.96*SD=61.37 50

0

Mean=-0.71

-50

-100

Mean-1.96*SD=-62.78 0

50 100 150 Mean of Test-Retest Assessments

10

Sway Index of Eye Open Condition of BBS (N=50) 1 Mean+1.96*SD=0.65 0.5 Mean=0.08 0

-0.5

Difference Between Test-Retest Assessments

Stability Index of Eye Open Condition of BBS (N=50)

Difference Between Test-Retest Assessments

2 4 6 8 Mean of Test-Retest Assessments

Difference Between Test-Retest Assessments

Difference Between Test-Retest Assessments

Difference Between Test-Retest Assessments

Difference Between Test-Retest Assessments

Difference Between Test-Retest Assessments

Reliability of Balance Computerized Adaptive Test and Biodex

-1 0

Mean-1.96*SD=-0.48

0.5 1 1.5 2 2.5 Mean of Test-Retest Assessments

3

Sway Index of Eye Close Condition of BBS (N=50) 2 Mean+1.96*SD=1.35 1 Mean=0.15

0 -1

Mean-1.96*SD=-1.05

-2 0

1 2 3 4 Mean of Test-Retest Assessments

5

Overall Control of Limit of Stability Test of BBS (N=17) 30 20

Mean+1.96*SD=21.09

10 Mean=1.53

0 -10 -20 -30

Mean-1.96*SD=-18.04 0

10 20 30 40 50 Mean of Test-Retest Assessments

60

Fig 1 Bland-Altman method for plotting the difference scores against the mean scores of each pairing of the Balance CAT (A) and the BBS (BeG). Note: Not all the data points can be seen if the data overlapped. Limits of agreement ratio (range of limits of agreement divided by mean scores of all assessments): Balance CAT, .27; EO-stability, 2.61; EO-sway, 1.10; EC-stability, 1.83; EC-sway, 1.50; LOS-time, 1.96; LOS-overall control, 1.37. Abbreviation: BBS, Biodex Balance System.

www.archives-pmr.org

6 sensitivity (resolution) of the computerized posturography instrument (ie, the Biodex Balance System) to patient performance. Thus, our findings support the statement that the selection of a measure among competing measures should be based on empirical evidence of psychometric properties (eg, the test-retest reliability in this study).24 Therefore, the test-retest reliabilities of the scores of the Biodex Balance System in patients with chronic stroke may not be superior, or even equal, to that of functional balance scales (eg, the Balance CAT). This finding is in agreement with those of previous studies.25,26 Chien et al26 compared the test-retest reliability of the SBM, another type of computerized posturography instrument, with that of the PASS in patients with mild stroke. They found that the test-retest reliability of the functional balance scale, namely the PASS, was superior to that of the SBM. It seems that the Balance CAT is more suitable than the Biodex Balance System for repeated assessments of balance function in patients with chronic stroke in clinical and research settings. Moreover, the Balance CAT is more cost-effective than the Biodex Balance System, which may improve its utility in clinics. The results on the MDC and MDC% also supported the finding that the test-retest reliabilities of the scores of the Balance CAT was better than that of the Biodex Balance System in patients with chronic stroke. The MDC, a statistical parameter, can be used as a threshold to help users distinguish real change from measurement error for an individual patient. The MDC is different from the minimal clinically important change, which is defined as the smallest value of clinical meaningful improvement perceived as beneficial to patients. In other words, if the value of change between 2 measurements in separate sessions for an individual patient is greater than the MDC value, it can be concluded that the patient has achieved real improvement (beyond measurement error) in the specific characteristic assessed by the measure, such as balance function in the present study, but it does not guarantee clinically meaningful improvement.20,27 Thus, the MDC and MDC% of a measure are useful for comparing the amount of random measurement error between competing measures in both clinical and research settings. To our knowledge, our study is the first to report the MDC and MDC% of the EO/EC test and LOS tests of the Biodex Balance System in patients with chronic stroke. The MDC of the Balance CAT was .91 points (total score of Balance CAT, 0e10), which means that a change in score measured by the Balance CAT of >.91 points indicates real improvement for an individual patient. The MDC% of the Balance CAT was close to 10%, indicating a nearly excellent level of measurement error. However, the MDC% of the Biodex Balance System ranged from 50.5% to 126.9%, which is greatly inferior to that of the Balance CAT. Pickerill and Harter28 examined the test-retest reliability of the most difficult level of the LOS test of the Biodex Balance System in healthy young adults. They found that the MDC% was 52.2% for directional control, and that for test duration was 43.7%.28 A similar result was also found in our study for directional control of the LOS test. However, the MDC% for test duration of the LOS test was greater in the present study than in their study. The greater MDC% may have been due to the greater variation of the test duration used by the patients with stroke. Because the MDC% was much greater than the acceptable criterion (30%), caution must be taken in assessing the change in balance function in patients with chronic stroke when the Biodex Balance System is used. Only 34% of the subjects completed the LOS test, possibly because of the great demands (challenges) of the tests, given the modest balance function of our subject group. The LOS test of the

C-H Chen et al Biodex Balance System is a continuous test using a randomized order of targets, without any breaks in the test between different subtests of the targets reached. If subjects cannot move their center of gravity to all the targets in different directions, the data are not recorded by the system. These tasks may have been too difficult for patients with chronic stroke to achieve consistent performances. In a previous study,26 a higher completion rate (ie, 75%) of the LOS test in the SBM was found in patients with chronic stroke. Instead of being recorded continuously during the subject’s performance, data of the LOS test for the SBM were separately recorded for each direction, which may have been easier than in the Biodex Balance System. In addition, our results also showed that the patients with chronic stroke who had higher clinical/functional status (ie, higher scores on the Barthel Index) were more likely to complete the LOS test of the Biodex Balance System. Therefore, the Biodex Balance System should be used with caution by clinicians and researchers when the LOS test is applied to patients with chronic stroke.

Study limitations This study had 3 limitations. First, the items of the Balance CAT were rated by a rater, which may have incurred recall bias. Such bias might have inflated the test-retest reliabilities of the scores of the Balance CAT. Second, patients who could not stand independently for 30 seconds were excluded from the study because of safety considerations. In addition, some data for the LOS test were missing. These factors may limit the generalization of the findings. The results of the test-retest reliability of this study are not suitable for generalization to patients with severe balance dysfunction. Third, other psychometric properties, such as responsiveness and the practicability of the Biodex Balance System in patients with chronic stroke, need to be studied further. Apart from the mode of measuring global balance function, the Biodex Balance System provides a useful platform for investigating the neuromuscular mechanisms of postural control (eg, specific balance functions interacting with visual information) or for developing training protocols of postural control for athletes, the elderly, and patients with disabilities.29-33 Therefore, different balance measures may be chosen for different purposes.

Conclusions The test-retest reliabilities of the scores of the Balance CAT was sufficient for assessing balance function in patients with stroke. Moreover, the test-retest reliabilities of the scores of the Balance CAT, one of the functional balance measures, were superior to those of the Biodex Balance System, one type of computerized posturography instrument in patients with chronic stroke. Therefore, the Balance CAT may be a more reliable measure for clinicians and researchers to use in assessing the balance function of patients with stroke for more than 6 months.

Suppliers a. Biodex Medical Systems, 20 Ramsey Rd, Shirley, NY 11967. b. IBM Corp, 1 New Orchard Rd, Armonk, NY 10504-1722.

Keywords Postural balance; Psychometrics; Rehabilitation; Stroke www.archives-pmr.org

Reliability of Balance Computerized Adaptive Test and Biodex

Corresponding author Hao-Ling Chen, PhD, School of Occupational Therapy, National Taiwan University, 4F, No.17, Xuzhou Rd, Zhongzheng Dist, Taipei City 100, Taiwan (R.O.C.) E-mail address: hlchen@ntu. edu.tw.

References 1. Bohannon RW, Leary KM. Standing balance and function over the course of acute rehabilitation. Arch Phys Med Rehabil 1995;76:994-6. 2. Nichols DS, Miller L, Colby LA, Pease WS. Sitting balance: its relation to function in individuals with hemiparesis. Arch Phys Med Rehabil 1996;77:865-9. 3. Juneja G, Czyrny JJ, Linn RT. Admission balance and outcomes of patients admitted for acute inpatient rehabilitation. Am J Phys Med Rehabil 1998;77:388-93. 4. Tyson SF, Hanley M, Chillala J, Selley AB, Tallis RC. The relationship between balance, disability, and recovery after stroke: predictive validity of the Brunel Balance Assessment. Neurorehabil Neural Repair 2007;21:341-6. 5. Bobath B. Adult hemiplegia: evaluation and treatment. London: William Heinemann Medical Book; 1990. 6. Wade D. Measurement in neurological rehabilitation. Oxford: Oxford Univ Pr; 1992. 7. Hsieh CL, Sheu CF, Hsueh IP, Wang CH. Trunk control as an early predictor of comprehensive activities of daily living function in stroke patients. Stroke 2002;33:2626-30. 8. Mao HF, Hsueh IP, Tang PF, Sheu CF, Hsieh CL. Analysis and comparison of the psychometric properties of three balance measures for stroke patients. Stroke 2002;33:1022-7. 9. Jette AM, Haley SM. Contemporary measurement techniques for rehabilitation outcomes assessment. J Rehabil Med 2005;37:339-45. 10. Cella D, Gershon R, Lai JS, Choi S. The future of outcomes measurement: item banking, tailored short-forms, and computerized adaptive assessment. Qual Life Res 2007;16:133-41. 11. Hsueh IP, Chen JH, Wang CH, et al. Development of a computerized adaptive test for assessing balance function in patients with stroke. Phys Ther 2010;90:1336-44. 12. Yu WH, Hsueh IP, Hou WH, Wang YH, Hsieh CL. A comparison of responsiveness and predictive validity of two balance measures in patients with stroke. J Rehabil Med 2012;44:176-80. 13. Berg KO, Maki BE, Williams JI, Holliday PJ, Wooddauphinee SL. Clinical and laboratory measures of postural balance in an elderly population. Arch Phys Med Rehabil 1992;73:1073-80. 14. Whitney SL, Poole JL, Cass SP. A review of balance instruments for older adults. Am J Occup Ther 1998;52:666-71. 15. Hinman MR. Factors affecting reliability of the Biodex Balance System: a summary of four studies. J Sport Rehabil 2000;9:240-52.

www.archives-pmr.org

7 16. Biodex Balance System operation and service manual #950-444. Shirley; Biodex Medical Systems; 2009. 17. Portney LG, Watkins MP. Foundations of clinical research: applications to practice. Upper Saddle River: Pearson Prentice Hall; 2009. 18. Raghunathan TE, Rosenthal R, Rubin DB. Comparing correlated but nonoverlapping correlations. Psychol Methods 1996;1:178-83. 19. Schreuders TAR, Roebroeck ME, Goumans J, van Nieuwenhuijzen JF, Stijnen TH, Stam HJ. Measurement error in grip and pinch force measurements in patients with hand injuries. Phys Ther 2003;83: 806-15. 20. Haley SM, Fragala-Pinkham MA. Interpreting change scores of tests and measures used in physical therapy. Phys Ther 2006;86:735-43. 21. Lu WS, Wang CH, Lin JH, Sheu CF, Hsieh CL. The minimal detectable change of the simplified stroke rehabilitation assessment of movement measure. J Rehabil Med 2008;40:615-9. 22. Huang SL, Hsieh CL, Wu RM, Tai CH, Lin CH, Lu WS. Minimal detectable change of the Timed “Up & Go” test and the Dynamic Gait Index in people with Parkinson disease. Phys Ther 2011;91:114-21. 23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135-60. 24. Hobart JC, Lamping DL, Freeman JA, et al. Evidence-based measurementdwhich disability scale for neurologic rehabilitation? Neurology 2001;57:639-44. 25. Liston RA, Brouwer BJ. Reliability and validity of measures obtained from stroke patients using the Balance Master. Arch Phys Med Rehabil 1996;77:425-30. 26. Chien CW, Hu MH, Tang PF, Sheu CF, Hsieh CL. A comparison of psychometric properties of the Smart Balance Master System and the Postural Assessment Scale for stroke in people who have had mild stroke. Arch Phys Med Rehabil 2007;88:374-80. 27. Jette AM, Tao W, Norweg A, Haley S. Interpreting rehabilitation outcome measurements. J Rehabil Med 2007;39:585-90. 28. Pickerill ML, Harter RA. Validity and reliability of limits-of-stability testing: a comparison of 2 postural stability evaluation devices. J Athl Training 2011;46:600-6. 29. Schmitz R, Arnold B. Intertester and intratester reliability of a dynamic balance protocol using the Biodex stability system. J Sport Rehabil 1998;7:95-101. 30. Mattacola CG, Perrin DH, Gansneder BM, Gieck JH, Saliba EN, McCue FC. Strength, functional outcome, and postural stability after anterior cruciate ligament reconstruction. J Athl Training 2002;37: 262-8. 31. Paterno MV, Myer GD, Ford KR, Hewett TE. Neuromuscular training improves single-limb stability in young female athletes. J Orthop Sports Phys Ther 2004;34:305-16. 32. Akbari M, Karimi H, Farahini H, Faghihzadeh S. Balance problems after unilateral lateral ankle sprains. J Rehabil Res Dev 2006;43: 819-23. 33. Cakar E, Durmus O, Tekin L, Dincer U, Kiralp MZ. The ankle-foot orthosis improves balance and reduces fall risk of chronic spastic hemiparetic patients. Eur J Phys Rehabil Med 2010;46:363-8.