If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Northwestern University Prosthetics-Orthotics Center, Northwestern University Feinberg School of Medicine, Chicago, ILDepartment of Physical Medicine and Rehabilitation, Northwestern University Feinberg School of Medicine, Chicago, IL
Northwestern University Prosthetics-Orthotics Center, Northwestern University Feinberg School of Medicine, Chicago, ILDepartment of Physical Medicine and Rehabilitation, Northwestern University Feinberg School of Medicine, Chicago, IL
Department of Physical Medicine and Rehabilitation, Northwestern University Feinberg School of Medicine, Chicago, ILRehabilitation Institute of Chicago, Chicago, IL
To evaluate the validity and reliability of the Berg Balance Scale (BBS) for use in people with lower-limb amputation.
Design
Cross-sectional study.
Setting
Research laboratory.
Participants
Individuals (N=30; age, 54±12y; 20 men) with unilateral transtibial (n=13), unilateral transfemoral (n=14), or bilateral (n=3) lower-limb amputation of dysvascular (n=7), traumatic (n=14), infectious (n=6), or congenital (n=3) origin.
Interventions
Not applicable.
Main Outcome Measures
BBS, 2-minute walk test, L test, Prosthesis Evaluation Questionnaire-Mobility Subscale, Activities-specific Balance Confidence Scale, and Frenchay Activities Index; self-reported descriptors were also collected, including frequency of prosthesis use, number of falls in 12 months before the visit, fear of falling, and daily mobility aid use.
Results
The BBS had high interrater reliability (intraclass correlation coefficient =.945) and internal consistency (α=.827). Relations between the BBS scores and those of other outcome measures were all statistically significant (P≤.001). Significant group differences in BBS scores were observed for fear of falling (P=.008) and mobility aid use (P<.001), but not for multiple (≥2) falls in the previous 12 months (P=.381). BBS items involving reaching forward, turning 360°, tandem standing, and standing on 1 leg had relatively greater frequencies of lower scores across participants.
Conclusions
The BBS appears to be a valid and reliable clinical instrument for assessing balance in individuals with lower-limb amputation, but it may not be able to discriminate between individuals with greater or lesser fall risk. Limitations in prosthetic motion and control may be responsible for the challenges experienced on items of lower performance. Future studies would be useful to assess the responsiveness of the BBS to interventions aimed at improving balance in individuals with lower-limb amputation.
Falls pose a significant risk to persons with lower-limb amputation (LLA) during and after the rehabilitation process. One study reported that 20.5% of persons with LLA in inpatient rehabilitation experienced ≥1 fall, with 13.6% of the falls occurring while the persons with LLA were standing or walking and 17.9% of fall injuries including bleeding/hemorrhage, lacerations, head trauma, and fractures.
found that more than half (52.4%) of the persons with LLA who wore their prosthesis daily experienced a fall in the prior 12 months and among those who fell, 75% fell more than once and 40.4% suffered a fall injury. Another study reported that 54.9% of persons with LLA experienced falling in the past 12 months, 66.7% wore their prosthesis at the time of the fall, 43.3% fell more than once, and 57% sustained a fall injury.
Falls may be related to both physical and psychological factors. Nearly half (49.2%) of community-dwelling persons with LLA reported fear of falling, causing 76.2% of them to avoid activities. Reduced balance confidence and fear of falling restrict physical mobility, social activity, and community participation
The influence of falling, fear of falling, and balance confidence on prosthetic mobility and social activity among individuals with a lower extremity amputation.
The increased risk of falls in persons with LLA is likely a direct consequence of compromised functional balance, resulting from reduced muscle strength,
Comparative study of the strength of the flexor and extensor muscles of the knee through isokinetic evaluation in normal subjects and patients subjected to trans-tibial amputation.
The effects of walking speed and prosthetic ankle adapters on upper extremity dynamics and stability-related parameters in bilateral transtibial amputee gait.
for producing appropriately timed reactions following a balance disturbance. For persons with LLA, the ability to maintain upright balance during standing minimizes the risk of falls.
Therefore, accurate assessment of balance in persons with LLA can measure rehabilitative progress and identify individuals who may be at risk for falling. Currently, no standardized, comprehensive clinical assessment tool quantifies the balance of persons with LLA. Therefore, clinical assessment of lower-extremity strength and range of motion (ROM)
The influence of falling, fear of falling, and balance confidence on prosthetic mobility and social activity among individuals with a lower extremity amputation.
Functional gait assessment and balance evaluation system test: reliability, validity, sensitivity, and specificity for identifying individuals with Parkinson disease who fall.
Test-retest reliability and minimal detectable change on balance and ambulation tests, the 36-item short-form health survey, and the unified Parkinson disease rating scale in people with parkinsonism.
The administration of the BBS requires 15 to 20 minutes using minimal, inexpensive equipment. Importantly, BBS scores demonstrate association with the likelihood of falling in older adults,
The accuracy of individual Berg Balance Scale items compared with the total Berg score for classifying people with chronic stroke according to fall history.
Functional gait assessment and balance evaluation system test: reliability, validity, sensitivity, and specificity for identifying individuals with Parkinson disease who fall.
Identifying patients with increased risk of falling enables rehabilitation professionals to prescribe interventions, targeted therapies, and/or detailed monitoring.
Already as an accepted clinical measure, the BBS has sufficient validity and reliability to identify the increased risk of falling in individuals with conditions that compromise balance. Thus, the BBS is a potentially appropriate measure to assess balance and fall risk in persons with LLA. In addition, a recent study observed significant differences in the results of certain clinical mobility measures (Timed Up & Go, 180° Turn, and Four Square Step Test) between persons with unilateral transtibial amputations who fell once and those who fell more than once in a 6-month period.
These measures use mobility tasks similar to those included in the BBS, further supporting the potential effectiveness of the BBS to identify persons with LLA who are at increased fall risk.
Although multiple studies have used the BBS to assess balance in persons with LLA of various ages and activity level (ranging from young soccer players to older dysvascular amputees of limited mobility),
Once validated, the BBS may serve as a valuable clinical instrument to assess balance in persons with LLA. Because it examines specific aspects of functional balance (eg, weight shifting and turning), the use of the BBS may provide additional insights into specific balance deficits of persons with LLA. Therefore, the purpose of this study was to assess the validity (convergent validity, known groups discrimination, and floor/ceiling effects) and reliability (interrater and internal consistency) of the BBS for assessing balance in persons with LLA. Given their relationship with balance, it was anticipated that groups with a fear of falling,
The influence of falling, fear of falling, and balance confidence on prosthetic mobility and social activity among individuals with a lower extremity amputation.
The university's institutional review board granted ethical approval for this study. Participants were included if they had experienced unilateral or bilateral lower-limb loss at or proximal to the ankle; used a prosthesis for ambulation with or without an additional mobility aid; did not have an upper-extremity amputation; and had a residual limb in good condition (ie, no open wounds or infections).
To identify the minimum sample size necessary to yield results of sufficient power and confidence, a priori power analyses were conducted using methods by Walter et al
respectively, to test 2 primary study aims: (1) interrater reliability and (2) group discrimination of participants classified retrospectively as fallers. The reliability analysis was based on a study design in which the acceptable reliability coefficient bandwidth was .65 to .85
a) was based on a study design for independent group comparison (2-tailed) of BBS scores between “fallers” and “nonfallers” and using an effect size estimated from a similar study on individuals with Parkinson's disease in which a faller was defined as experiencing ≥2 falls in 12 months prior to data collection.
Setting the critical alpha to .05 and desired power to 80%, reliability and group discrimination power analyses estimated sample sizes of 28 and 22, respectively. A sample size of 30 was deemed appropriate for this study, and data were collected from the target number.
Data collection
After obtaining written informed consent, data collection for each participant occurred during 1 visit. Information on subject age, sex, height, mass, number of LLAs and amputation level(s), and the following self-reported descriptors was obtained: amputation etiology, time since most recent major amputation, frequency (average days per week and hours per day) of prosthesis use, experience using a prosthesis (years), number of falls in the 12 months before the visit, presence of fear of falling, and type of mobility aid(s) used for daily ambulation (eg, community and/or household). A fall was defined as “coming to rest inadvertently on the ground or other level, other than as a consequence of lost consciousness, a violent blow, stroke, or epileptic seizure”
and presence of fear of falling was confirmed if the participant responded affirmatively to the question “Are you afraid of falling?” Subjects were considered to use a mobility aid for daily ambulation irrespective of the type of aid(s) used.
Because there is no criterion standard for quantifying balance in persons with LLA, convergent validity of the BBS was evaluated by collecting information on several clinical outcome measures that assess constructs related to balance and mobility. Following the collection of descriptors, participants completed, in random order, the Activities-specific Balance Confidence (ABC) Scale,
The L test of functional mobility: measurement properties of a modified version of the timed “up & go” test designed for people with lower-limb amputations.
to assess aspects of functional balance and mobility (see table 1 for details). To limit the effects of fatigue, participants had 5 minutes seated rest between each performance test.
Table 1Description of outcome measures
Outcome Measure
Measurement
Data
Description
Time Required for Administration (min)
Validity and Reliability in Persons With LLA (Reference)
ABC Scale
Perceived balance and confidence
Ordinal 0%–100%
16 ADLs
5–10
Test-retest reliability: ICC=.91; Cronbach's α=.93 (Miller et al
14 tasks performed in chronological sequence: (1) sitting-to-standing; (2) standing unsupported; (3) sitting with back unsupported; (4) standing-to-sitting; (5) transfer between seats with and without armrests; (6) standing unsupported with eyes closed; (7) standing unsupported with feet together; (8) reaching forward while standing; (9) picking up an object from the floor; (10) turning to look over left and right shoulders; (11) turning 360°; (12) alternately stepping up on a stool while unsupported; (13) tandem standing while unsupported; (14) standing on one leg.
15–20
Not applicable
L Test
Mobility/balance performance
Time (10th of a second)
Participants rise from a chair without armrests, walk 3m forward, turn right, walk 7m, walk around a floor marker, retrace their path to the chair, and sit down. The participants are timed from the moment their trunk leaves contact with the back of the chair until the trunk regains contact.
1–2 including practice
Intrarater reliability ICC=.97; interrater reliability ICC=.96 (Deathe and Miller
The L test of functional mobility: measurement properties of a modified version of the timed “up & go” test designed for people with lower-limb amputations.
Each participant completed these tests in the same well-lit room while wearing comfortable clothing, shoes, and their everyday prosthesis. To ensure protocol compliance and minimize distraction, talking was discouraged during test administration. Before data collection, participants were requested to refrain from adjusting custom settings on their prostheses (eg, joint impedance or alignment) throughout testing to maintain consistency between tests. Following the demonstration and completion of 1 practice trial (or lap in the case of the 2MWT), the L Test and the 2MWT were performed once at the participants' comfortable walking speed with their usual walking aid, if requested. Subjects received standardized verbal encouragement during the L Test, but not the 2MWT, as described in standardized protocols for validating these tests in persons with LLA.
The L test of functional mobility: measurement properties of a modified version of the timed “up & go” test designed for people with lower-limb amputations.
In the BBS, the self-selected limb placed in front for tandem standing (item 13) and stood upon during single-limb standing (item 14) was recorded for persons with LLA affected unilaterally.
After the completion of the performance tests, participants received 20 minutes of seated rest before a different rater administered the BBS a second time. The order of BBS administration by the 2 raters (rater A and rater B) was systematically randomized: rater A administered the BBS first for odd-numbered participants, and rater B administered the BBS first for even-numbered participants (numbers assigned by order of visit). For each participant, only 1 rater was present during the administration of the BBS, and throughout the study, raters avoided communication about performance. Rater A administered all participants' L Test and 2MWT. Rater A is a biomedical engineer with 6 months' training on the BBS, including multiple practice sessions, and rater B is a prosthetist/orthotist with 4 years' experience using the BBS.
Data analysis
Summary statistics of the participant characteristics and outcome scores were derived to describe the sample and facilitate comparison with the literature. Because of the ordinal nature of the BBS data, nonparametric statistical tests were performed (in SPSSb) for all analyses and the critical alpha was set at .05.
Convergent validity was assessed by quantifying relationships (Spearman rank correlation coefficient [ρ]) between the participants' BBS scores and that of the other outcome measures. It was anticipated that the BBS would share a positive monotonic relationship with the ABC Scale, PEQ-MS, FAI, and 2MWT and a negative relationship with the L Test. The known groups method was used to evaluate the ability of the BBS to discriminate between subjects classified by clinically relevant descriptors that may affect balance performance (ie, presence of fear of falling, amputation level, amputation cause, and mobility aid use for daily ambulation) or be affected by balance performance (ie, multiple [≥2] falls in the 12 months before data collection). It was anticipated that groups with the following characteristics would perform worse on the BBS (ie, reduced scores):
1.
Fear of falling, due to hesitancy to challenge balance
The influence of falling, fear of falling, and balance confidence on prosthetic mobility and social activity among individuals with a lower extremity amputation.
Independent group comparisons were performed using the Mann-Whitney U test. A Bonferroni adjustment was applied to account for the multiple independent group tests and reduce potential for type I error, in which a P value of ≤.01 was considered statistically significant. The percentage of participants who achieved the maximum BBS score of 56 was used to estimate ceiling effects. Only the results of rater A were used for these validity tests.
Internal consistency and interrater reliability were evaluated through the estimation of Cronbach's alpha as a measure of homogeneity and the intraclass correlation coefficient (2,1) as a measure of absolute agreement between rater scores, respectively. The scores from both raters were used for these reliability tests.
The BBS data from rater A (the “reference” rater) were used for validation analyses and presentation of results because this rater also administered the other performance tests and questionnaires. The BBS data from rater B (the “comparison” rater) were used solely for reliability analyses.
Results
Table 2, Table 3 show summary statistics of participants' characteristics and outcome scores, respectively. The BBS demonstrated high convergent validity, because correlations between the BBS scores and those of other outcome measures were all statistically significant with good to excellent strength
(table 4). Results of the known-groups method are summarized in table 5, with significance achieved only between groups with and without fear of falling and those who did and did not use a mobility aid for daily ambulation. It should be noted that 1 data point is missing from the data set on self-reported falls.
Table 2Participant characteristics and proportions
Data of mass without prosthesis represent 28 participants because 2 participants did not feel comfortable having their prosthesis weighed.
85±19 (50–118)
Time since most recent major amputation (y)
21±15 (0.5–55)
Prosthesis usage (days/week×hours/day=hours/week)
86±30 (15–126)
Experience using a prosthesis (y)
18±14 (0.3–52)
Sex, n
Male
20
Female
10
Amputation etiology, n
Dysvascular
7
Traumatic
14
Infection
6
Congenital
3
Amputation level, n
Unilateral transtibial
13
Unilateral transfemoral
14
Bilateral transtibial
2
Bilateral transtibial/transfemoral
1
Mobility aids used for daily ambulation, n
Cane
10
Walker
5
Scooter
4
Manual wheelchair
4
NOTE. The reported number of participants who use a mobility aid includes overlap because some participants reported using more than 1 aid for daily ambulation.
∗ Data of mass without prosthesis represent 28 participants because 2 participants did not feel comfortable having their prosthesis weighed.
The L test of functional mobility: measurement properties of a modified version of the timed “up & go” test designed for people with lower-limb amputations.
All participants attempted each BBS task with varying degrees of performance success. Figure 1 shows a histogram of BBS scores. Because a ceiling effect was quantified by the percentage of participants who achieved the maximum score on the BBS, results suggest a small ceiling effect (10% of the participants achieved the maximum score of 56). However, it is notable that 70% of the participants achieved a score ≥50, indicating a skewed distribution favoring higher scores. The BBS demonstrated high interrater reliability (intraclass correlation coefficient [2,1]=.945) and internal consistency (α=.827 and .826 for rater A and rater B, respectively). Histograms for BBS items (fig 2) reveal that items 8 (reaching forward), 11 (turning 360°), 13 (tandem standing), and 14 (standing on 1 leg) received noticeably greater relative frequency (at least 40%) of scores below the maximum score of 4. During tandem standing (item 13), 44% of the participants affected unilaterally placed their sound limb in front and 100% stood on their sound limb during single-limb standing (item 14). Prosthesis bulk prevented some participants from bringing their feet together for item 7, but this was not penalized when their feet were considered to be “as close as possible.”
Fig 1Histogram of BBS scores by participant frequency as recorded by rater A.
The BBS performance by community-dwelling persons with LLA (BBS score, 51±5) was comparable to that of people with other pathological conditions prone to unsteadiness, including poststroke unilateral hemiparesis (49±6),
Functional gait assessment and balance evaluation system test: reliability, validity, sensitivity, and specificity for identifying individuals with Parkinson disease who fall.
The results suggest that when assessing balance in community-dwelling individuals with LLA, a slight ceiling effect exists (10% of the participants achieved the maximum score of 56), with BBS scores noticeably skewed toward the higher range (see fig 1). The 2 outlier participants who scored in the 30 to 40 range produced 2 of the lowest FAI and ABC Scale scores, indicating minimal participation in activities of daily living and low balance confidence, respectively. Also, they relied on more than 1 type of mobility aid for daily ambulation. Future studies that examine the BBS with respect to rehabilitation interventions and balance improvement may revisit these findings.
As expected, the BBS demonstrated monotonic relationships with the other outcome measures that assess constructs related to balance and mobility, providing evidence that the BBS does assess balance in persons with LLA. Compared with the questionnaires (ABC Scale, PEQ-MS, and FAI), performance tests (the 2MWT and the L Test) showed stronger relationships with the BBS. This was expected, given that these measures, like the BBS, assess performance, rather than the perception of capability and/or activity level. Furthermore, the reliability of the BBS was high, providing confidence that it is a meaningful clinical measure. Importantly, the estimated values of Cronbach's alpha lie safely in the region of acceptable values (0.7–0.9),
suggesting little to no redundancy and sufficient homogeneity across items.
As anticipated, participants performed worse on the BBS if they were classified as having fear of falling; using a mobility aid for daily ambulation; having a unilateral transfemoral amputation as opposed to transtibial; having a dysvascular-related amputation; and having experienced ≥2 falls in the previous 12 months, as opposed to ≤1. However, group differences were statistically significant only for fear of falling and use of a mobility aid. Importantly, differences were minimal and not significant between groups classified by retrospective falls. Compared with applications in people with other balance compromised conditions,
The accuracy of individual Berg Balance Scale items compared with the total Berg score for classifying people with chronic stroke according to fall history.
it is difficult to address the reason(s) and is beyond the scope of this study. In addition, participants were classified as fallers on the basis of self-reported fall history, which may reduce measure accuracy because of recall bias. However, the BBS was also unable to discriminate between groups on the basis of amputation etiology and level, both factors that impact postural control
; therefore, the BBS may not sufficiently challenge persons with LLA to expose these differences. For example, people with unilateral transfemoral amputation may use different strategies to perform item 12 (alternate stepping on a stool while unsupported) than would their transtibial counterparts (eg, circumduction, rather than knee flexion), but may still be capable of completing the task in the same amount of time and thus achieve the same score. A similar issue may apply to persons with LLA due to dysvascular cause because they may attempt to complete the stepping task as quickly as possible to avoid prolonged single-limb stance. Although a valid and reliable outcome measure for this group, further modifications may be appropriate to refine the BBS to expose the effect of specific amputation characteristics on balance performance among persons with LLA.
Participants appeared to have greater difficulty performing items with specific movements most likely constrained by the limited ROM and control of the prosthesis. For example, lack of sagittal plane joint control and ROM may have limited the ability to perform sustained forward reaching (item 8). In addition, most passive prostheses have limited frontal plane ROM and this may have affected the ability to maintain tandem standing (item 13) because this requires some foot eversion and possibly internal rotation. The reduced ability to perform controlled internal rotation may also limit performance during 360° turning on the spot (item 11). In all these tasks, it is reasonable to assume that reduced sensory feedback that facilitates locating the center of pressure under the prosthesis and its relationship with the base of support may also contribute to cautious performance and reduced scores. Systematic studies would be useful to investigate such prosthetic constraints on BBS performance.
On item 14, 40% of the participants scored between 0 and 3, showing inability to maintain single-limb standing for more than 10 seconds. Significantly, 33% of the participants with a unilateral amputation scored between 0 and 3, despite all participants standing on their sound limb. Unaided walking requires some period of single-limb stance, and so these data highlight a crucial balance impairment that may negatively affect safe ambulation and participation in activities of daily living.
This consideration is important for unilateral prosthesis users because the sound limb ideally should be able to facilitate compensatory balance. Rehabilitation strategies that specifically focus on improving single-limb balance (eg, controlled weight shifting) may benefit functional balance of persons with LLA.
Study limitations
Although this study included a wide range of amputation levels and etiologies, the participant cohort consisted of a convenience sample of relatively active persons with LLA. Probably, this composition caused a skewed distribution toward the higher range of BBS scores. This limitation must be addressed when considering the generalizability of the study results. Also, fall data were based on self-reported history and possibly impeded by recall bias; therefore, as an additional form of validation, further studies should evaluate the ability of the BBS to predict future falls.
Importantly, results from this study confirm that the BBS is capable of characterizing balance performance in community-dwelling persons with LLA. However, the responsiveness of the BBS to identify changes in balance performance resulting from therapeutic interventions is unknown. Investigations on the responsiveness of the BBS would be necessary before making recommendations of its clinical use.
Conclusions
Overall, the BBS demonstrated high validity and reliability and therefore appears to be an appropriate clinical measure for assessing balance in community-dwelling persons with LLA. Even though the BBS was able to discriminate those persons with LLA who possess fear of falling and use a mobility aid for daily ambulation, this was not so for those classified as fallers (≥2 falls in the previous 12mo). However, fall data were based on self-reported experience and additional studies would be useful to assess the ability of the BBS to predict falls prospectively. The BBS appears useful in exposing the challenges to balance that persons with LLA face because of constraints on prosthetic ROM and control. Yet, it appears that the current BBS items may not be challenging enough to expose the balance deficits previously identified in community-dwelling individuals with LLA. Furthermore, responsiveness of the BBS is unknown. Further studies are recommended to explore the responsiveness of the BBS, particularly regarding its utility for assessing the effectiveness of therapeutic interventions and prosthetic componentry for improving upright balance, and to refine it further for persons with LLA (eg, Rasch analysis and modification of tasks to increase their challenge to functional balance).
Suppliers
a.
G*Power 3; Institut für Experimentelle Psychologie, Heinrich-Heine-Universität, Gebäude 23.02 und 23.03, Universitätsstr. 1, 40225 Düsseldorf, Germany.
b.
SPSS v20; IBM Corporation, 1 New Orchard Rd, Armonk, NY 10504.
Acknowledgments
We thank R.J. Garrick, PhD, for her assistance with editing the manuscript.
References
Pauley T.
Devlin M.
Heslin K.
Falls sustained during inpatient rehabilitation after lower limb amputation: prevalence and predictors.
Am J Phys Med Rehabil.2006; 85 (quiz 33-5): 521-532
The influence of falling, fear of falling, and balance confidence on prosthetic mobility and social activity among individuals with a lower extremity amputation.
Comparative study of the strength of the flexor and extensor muscles of the knee through isokinetic evaluation in normal subjects and patients subjected to trans-tibial amputation.
The effects of walking speed and prosthetic ankle adapters on upper extremity dynamics and stability-related parameters in bilateral transtibial amputee gait.
Functional gait assessment and balance evaluation system test: reliability, validity, sensitivity, and specificity for identifying individuals with Parkinson disease who fall.
Test-retest reliability and minimal detectable change on balance and ambulation tests, the 36-item short-form health survey, and the unified Parkinson disease rating scale in people with parkinsonism.
The accuracy of individual Berg Balance Scale items compared with the total Berg score for classifying people with chronic stroke according to fall history.
The L test of functional mobility: measurement properties of a modified version of the timed “up & go” test designed for people with lower-limb amputations.
Supported by National Institutes of Health (grant no. 3UL1RR025741 ), National Institute on Disability and Rehabilitation Research (grant nos. H133P110013 and H133E080009 ), and the David Rubin, MD, Enrichment Fund. The opinions contained in this article are those of the authors and do not necessarily reflect those of the US Department of Education or Health and Human Services.
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.
00524-8&id=gr1.jpg){kind=link}
00524-8&id=gr2.jpg){kind=link}