549032
research-article2014
CRE0010.1177/0269215514549032Clinical RehabilitationChen et al.
CLINICAL REHABILITATION
Article
A prospective study of the responsiveness of the original and the short form Berg Balance Scale in people with stroke
Clinical Rehabilitation 2015, Vol. 29(5) 468–476 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0269215514549032 cre.sagepub.com
Kuan-Lin Chen1,2, Yei-Tai Chou3,4, Wan-Hui Yu5, Cheng-Te Chen6, Ching-Lin Shih7 and Ching-Lin Hsieh5,8
Abstract Objective: The study aim was to examine the responsiveness of the Berg Balance Scale (BBS) and that of its short form (SFBBS) at both the individual person level and the group level. Design: A repeated-measurements design. Setting: Hospital and home setting. Participants: Patients with stroke. Results: Totals of 226, 202, and 168 patients with stroke were assessed with the BBS at 14, 30, and 90 days after stroke, respectively. The SFBBS data were extracted from the patients’ responses on the BBS. At the group level, the BBS and the SFBBS had sufficient and similar responsiveness. For the Rasch scores, the effect sizes of the three change scores for the BBS and the SFBBS, respectively, had similar ranges between 0.38 and 0.88 and between 0.39 and 0.85, respectively. The standardized response means of the three change scores for the BBS and the SFBBS ranged from 0.74 to 1.33 and from 0.72 to 1.13, respectively. At the individual person level, the BBS detected significant balance improvement in about twice as many patients as the SFBBS detected. Conclusion: The responsiveness of the BBS at the individual person level was better than that of the SFBBS in patients with stroke. The BBS is recommended as an outcome measure to better detect changes in individual patients. Keywords Individual-person-level responsiveness, Rasch analysis, short form Berg Balance Scale, stroke, significance of change, outcome measurement Received: 14 April 2014; accepted: 04 August 2014 1School
of Occupational Therapy, College of Medicine, National Cheng Kung University, Tainan, Taiwan 2Department of Physical Medicine and Rehabilitation, National Cheng Kung University Hospital, Tainan, Taiwan 3Department of Psychology, National Chung Cheng University, Chiayi, Taiwan 4Research Center for Psychological and Educational Testing, National Taiwan Normal University, Taipei, Taiwan 5School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan
6Center
for Teacher Education, National Tsing Hua University, Hsinchu, Taiwan 7Center for Teacher Education, National Sun Yat-Sen University, Kaohsiung, Taiwan 8Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital, Taipei, Taiwan. Corresponding author: Ching-Lin Hsieh, School of Occupational Therapy, College of Medicine, National Taiwan University, 4F, No 17, Xuzhou Rd, Taipei 100, Taiwan. Email: [email protected]
Downloaded from cre.sagepub.com at NORTHERN ARIZONA UNIVERSITY on May 31, 2015
469
Chen et al.
Introduction The responsiveness of an outcome measure is vital to detecting changes over time.1 The ability to quantify changes is an important aspect of most clinical assessment tools. Comparisons of the responsiveness of competing measures (e.g., short forms vs. long forms or newly-developed measures vs. legacy measures) can guide the selection of measures for clinicians and researchers. Several studies using group-based responsiveness indices show that the well-known Barthel Index and Functional Independence Measure have similar levels of responsiveness.2-6 However, groupbased indices of responsiveness may be limited and can be misleading.7 Hobart et al.7 recently compared the responsiveness of the two measures at both the group and individual person levels. Their results show that the group-level responsiveness of the Barthel Index is similar to that of the Functional Independence Measure, as found in several previous studies.4,8-10 However, the Barthel Index is less responsive than the Functional Independence Measure at the individual person level.7 This issue is critical and should be brought to light to help clinicians and researchers select a responsive outcome measure among competing measures. The Berg Balance Scale, a well-known measure for assessing an elderly person’s balance function,11 is generally considered to be the gold standard of balance measures.12 Because the Berg Balance Scale is time-consuming and has inconsistent rating systems across items, its short form was developed to improve the utility.8 The short form has 7 items and a 3-level rating scale (0-2-4). The short form has been recommended for clinicians because it is easier and faster to administer than the original and it is psychometrically similar to the original, including its responsiveness.8 However, the responsiveness of the Berg Balance Scale has been compared to that of the short form only at the group level. No such comparison has been performed at the individual person level. Without empirical evidence on the individual person level responsiveness, we cannot determine whether the capacity of the short form to detect change in an individual patient with stroke is
similar to that of the original in both clinical and research settings. The aim of this study was to examine the responsiveness of the Berg Balance Scale and that of its short form at both the individual person level and the group level. We hypothesized that the Berg Balance Scale would have better responsiveness than the its short form because the Berg Balance Scale has more items and more response levels (14 items with 5 levels) than does the short form (7 items with 3 levels).
Method Study participants The data were retrieved from a prospective cohort study, “The Quality of Life After Stroke Study in Taiwan”.13 In this study, patients with stroke were recruited from the National Taiwan University Hospital between December 1999 and May 2000. Patients were included if they met the following criteria: (1) diagnosis of cerebral hemorrhage (431), cerebral infarction (434), or other (430, 432, 433, 436, 437) based on the International Classification of Diseases, ninth revision, clinical modification codes; (2) first onset of cerebrovascular accident; (3) stroke onset within 14 days before hospital admission; (4) ability to follow verbal instructions; and (5) ability to give informed consent personally or by proxy. The clinical diagnosis of stroke was confirmed by neuroimaging examination (CT/MRI). Subjects were excluded if they suffered from another stroke or other major disease(s) during the follow-up periods.
Procedure The patients were assessed with the Berg Balance Scale at 14, 30, and 90 days after stroke onset. All assessments were administered by a trained occupational therapist at a hospital or the patients’ home.
Measures The Berg Balance Scale 14,15 was designed to measure functional balance performance among older
Downloaded from cre.sagepub.com at NORTHERN ARIZONA UNIVERSITY on May 31, 2015
470
Clinical Rehabilitation 29(5)
people with impairment in balance function. The Berg Balance Scale has sound psychometric properties in patients with stroke11,13,14,16 The short form of the Berg Balance Scale was derived from the original to improve the utility. The short form has 7 items with three levels (0-2-4) collapsed from the original five levels (0–1–2–3–4). The selected items were those with the highest internal consistency and the greatest responsiveness.8 The total score of the short form ranges from 0 to 28. The short form has satisfactory psychometric properties that are similar to those of the original.8
Data analysis The data of the Berg Balance Scale and its short form were organized to form a pooled data set respectively for each of three time points. Rasch analyses were used to determine if the first pooled data set (obtained at 14 days after stroke) fit the requirements of the partial credit model. The Rasch measurements of the persons and the items were estimated by WINSTEPS.17 The item measurements of the Berg Balance Scale and the short form on the second and third time points were anchored as those on the first time point to maintain the measurement scale. The Rasch score and its corresponding standard error for each individual participant on each time point were further used to examine the individual person level responsiveness of both measures. Descriptive statistical analysis was conducted to characterize the demographic and clinical properties of the participants. Effect size, standardized response mean, and paired t-test were used to examine the responsiveness of the Berg Balance Scale and the short form at the group level. All statistical analyses were performed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). The effect size is a measure of change obtained by dividing the mean change between baseline and follow-up measurements by the standard deviation (SD) of the baseline measurement.18 The standardized response mean is the mean change in scores between two measurements divided by the SD of those changes in scores.19 Effect size and standardized response mean values of 0.20, 0.50, and 0.80
were considered to show small, moderate, and large responsiveness, respectively.18,20 The paired t-test was used to determine the statistical significance of the change in scores.21 The level of statistical significance was set at p ≤ .05. To further compare the group level responsiveness of the Berg Balance Scale to that of the short form, the 95% confidence intervals (CIs) of the differences in effect size and standardized response mean of both measures were constructed with 10,000 bootstrap samples. If the value zero was included in the 95% CIs, the effect size or standardized response mean values between the Berg Balance Scale and the short form were not significantly different. The significance of change for each patient was used for analysis of the responsiveness of both measures at the individual person level, as proposed by Hobart et al.7 Significance of change is a measure of change obtained by dividing the size of the change score for an individual patient (e.g., change scores of 14 days and 30 days after stroke onset) by the standard error of the difference, which is the square root of the sum of the two squared standard error values (e.g., the standard error of the difference for a person between 14 days and 30 days after stroke onset is the square root of the sum of the two squared standard error of the score at 14 days and 30 days after stroke). Significance of change was used for analyzing what the magnitude of a patient’s change was in standard error of the difference units and whether the change was statistically significant at the individual person level. The significance of change is a continuous index and is useful for quantifying each patient’s change score. Paired-t test was then used to compare the significance of change of the Berg Balance Scale to that of the short form. In addition, according to the size and direction of the change score, the significance of change for each patient was categorized into one of the five groups: (1) significance of change ≥1.96 (significant improvement), (2) 0< significance of change
research-article2014
CRE0010.1177/0269215514549032Clinical RehabilitationChen et al.
CLINICAL REHABILITATION
Article
A prospective study of the responsiveness of the original and the short form Berg Balance Scale in people with stroke
Clinical Rehabilitation 2015, Vol. 29(5) 468–476 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0269215514549032 cre.sagepub.com
Kuan-Lin Chen1,2, Yei-Tai Chou3,4, Wan-Hui Yu5, Cheng-Te Chen6, Ching-Lin Shih7 and Ching-Lin Hsieh5,8
Abstract Objective: The study aim was to examine the responsiveness of the Berg Balance Scale (BBS) and that of its short form (SFBBS) at both the individual person level and the group level. Design: A repeated-measurements design. Setting: Hospital and home setting. Participants: Patients with stroke. Results: Totals of 226, 202, and 168 patients with stroke were assessed with the BBS at 14, 30, and 90 days after stroke, respectively. The SFBBS data were extracted from the patients’ responses on the BBS. At the group level, the BBS and the SFBBS had sufficient and similar responsiveness. For the Rasch scores, the effect sizes of the three change scores for the BBS and the SFBBS, respectively, had similar ranges between 0.38 and 0.88 and between 0.39 and 0.85, respectively. The standardized response means of the three change scores for the BBS and the SFBBS ranged from 0.74 to 1.33 and from 0.72 to 1.13, respectively. At the individual person level, the BBS detected significant balance improvement in about twice as many patients as the SFBBS detected. Conclusion: The responsiveness of the BBS at the individual person level was better than that of the SFBBS in patients with stroke. The BBS is recommended as an outcome measure to better detect changes in individual patients. Keywords Individual-person-level responsiveness, Rasch analysis, short form Berg Balance Scale, stroke, significance of change, outcome measurement Received: 14 April 2014; accepted: 04 August 2014 1School
of Occupational Therapy, College of Medicine, National Cheng Kung University, Tainan, Taiwan 2Department of Physical Medicine and Rehabilitation, National Cheng Kung University Hospital, Tainan, Taiwan 3Department of Psychology, National Chung Cheng University, Chiayi, Taiwan 4Research Center for Psychological and Educational Testing, National Taiwan Normal University, Taipei, Taiwan 5School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan
6Center
for Teacher Education, National Tsing Hua University, Hsinchu, Taiwan 7Center for Teacher Education, National Sun Yat-Sen University, Kaohsiung, Taiwan 8Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital, Taipei, Taiwan. Corresponding author: Ching-Lin Hsieh, School of Occupational Therapy, College of Medicine, National Taiwan University, 4F, No 17, Xuzhou Rd, Taipei 100, Taiwan. Email: [email protected]
Downloaded from cre.sagepub.com at NORTHERN ARIZONA UNIVERSITY on May 31, 2015
469
Chen et al.
Introduction The responsiveness of an outcome measure is vital to detecting changes over time.1 The ability to quantify changes is an important aspect of most clinical assessment tools. Comparisons of the responsiveness of competing measures (e.g., short forms vs. long forms or newly-developed measures vs. legacy measures) can guide the selection of measures for clinicians and researchers. Several studies using group-based responsiveness indices show that the well-known Barthel Index and Functional Independence Measure have similar levels of responsiveness.2-6 However, groupbased indices of responsiveness may be limited and can be misleading.7 Hobart et al.7 recently compared the responsiveness of the two measures at both the group and individual person levels. Their results show that the group-level responsiveness of the Barthel Index is similar to that of the Functional Independence Measure, as found in several previous studies.4,8-10 However, the Barthel Index is less responsive than the Functional Independence Measure at the individual person level.7 This issue is critical and should be brought to light to help clinicians and researchers select a responsive outcome measure among competing measures. The Berg Balance Scale, a well-known measure for assessing an elderly person’s balance function,11 is generally considered to be the gold standard of balance measures.12 Because the Berg Balance Scale is time-consuming and has inconsistent rating systems across items, its short form was developed to improve the utility.8 The short form has 7 items and a 3-level rating scale (0-2-4). The short form has been recommended for clinicians because it is easier and faster to administer than the original and it is psychometrically similar to the original, including its responsiveness.8 However, the responsiveness of the Berg Balance Scale has been compared to that of the short form only at the group level. No such comparison has been performed at the individual person level. Without empirical evidence on the individual person level responsiveness, we cannot determine whether the capacity of the short form to detect change in an individual patient with stroke is
similar to that of the original in both clinical and research settings. The aim of this study was to examine the responsiveness of the Berg Balance Scale and that of its short form at both the individual person level and the group level. We hypothesized that the Berg Balance Scale would have better responsiveness than the its short form because the Berg Balance Scale has more items and more response levels (14 items with 5 levels) than does the short form (7 items with 3 levels).
Method Study participants The data were retrieved from a prospective cohort study, “The Quality of Life After Stroke Study in Taiwan”.13 In this study, patients with stroke were recruited from the National Taiwan University Hospital between December 1999 and May 2000. Patients were included if they met the following criteria: (1) diagnosis of cerebral hemorrhage (431), cerebral infarction (434), or other (430, 432, 433, 436, 437) based on the International Classification of Diseases, ninth revision, clinical modification codes; (2) first onset of cerebrovascular accident; (3) stroke onset within 14 days before hospital admission; (4) ability to follow verbal instructions; and (5) ability to give informed consent personally or by proxy. The clinical diagnosis of stroke was confirmed by neuroimaging examination (CT/MRI). Subjects were excluded if they suffered from another stroke or other major disease(s) during the follow-up periods.
Procedure The patients were assessed with the Berg Balance Scale at 14, 30, and 90 days after stroke onset. All assessments were administered by a trained occupational therapist at a hospital or the patients’ home.
Measures The Berg Balance Scale 14,15 was designed to measure functional balance performance among older
Downloaded from cre.sagepub.com at NORTHERN ARIZONA UNIVERSITY on May 31, 2015
470
Clinical Rehabilitation 29(5)
people with impairment in balance function. The Berg Balance Scale has sound psychometric properties in patients with stroke11,13,14,16 The short form of the Berg Balance Scale was derived from the original to improve the utility. The short form has 7 items with three levels (0-2-4) collapsed from the original five levels (0–1–2–3–4). The selected items were those with the highest internal consistency and the greatest responsiveness.8 The total score of the short form ranges from 0 to 28. The short form has satisfactory psychometric properties that are similar to those of the original.8
Data analysis The data of the Berg Balance Scale and its short form were organized to form a pooled data set respectively for each of three time points. Rasch analyses were used to determine if the first pooled data set (obtained at 14 days after stroke) fit the requirements of the partial credit model. The Rasch measurements of the persons and the items were estimated by WINSTEPS.17 The item measurements of the Berg Balance Scale and the short form on the second and third time points were anchored as those on the first time point to maintain the measurement scale. The Rasch score and its corresponding standard error for each individual participant on each time point were further used to examine the individual person level responsiveness of both measures. Descriptive statistical analysis was conducted to characterize the demographic and clinical properties of the participants. Effect size, standardized response mean, and paired t-test were used to examine the responsiveness of the Berg Balance Scale and the short form at the group level. All statistical analyses were performed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). The effect size is a measure of change obtained by dividing the mean change between baseline and follow-up measurements by the standard deviation (SD) of the baseline measurement.18 The standardized response mean is the mean change in scores between two measurements divided by the SD of those changes in scores.19 Effect size and standardized response mean values of 0.20, 0.50, and 0.80
were considered to show small, moderate, and large responsiveness, respectively.18,20 The paired t-test was used to determine the statistical significance of the change in scores.21 The level of statistical significance was set at p ≤ .05. To further compare the group level responsiveness of the Berg Balance Scale to that of the short form, the 95% confidence intervals (CIs) of the differences in effect size and standardized response mean of both measures were constructed with 10,000 bootstrap samples. If the value zero was included in the 95% CIs, the effect size or standardized response mean values between the Berg Balance Scale and the short form were not significantly different. The significance of change for each patient was used for analysis of the responsiveness of both measures at the individual person level, as proposed by Hobart et al.7 Significance of change is a measure of change obtained by dividing the size of the change score for an individual patient (e.g., change scores of 14 days and 30 days after stroke onset) by the standard error of the difference, which is the square root of the sum of the two squared standard error values (e.g., the standard error of the difference for a person between 14 days and 30 days after stroke onset is the square root of the sum of the two squared standard error of the score at 14 days and 30 days after stroke). Significance of change was used for analyzing what the magnitude of a patient’s change was in standard error of the difference units and whether the change was statistically significant at the individual person level. The significance of change is a continuous index and is useful for quantifying each patient’s change score. Paired-t test was then used to compare the significance of change of the Berg Balance Scale to that of the short form. In addition, according to the size and direction of the change score, the significance of change for each patient was categorized into one of the five groups: (1) significance of change ≥1.96 (significant improvement), (2) 0< significance of change
