Berg Balance Scale (BBS)


Berg, Wood-Dauphinee, Williams, and Maki (1992), and Berg, Wood-Dauphinee, and Williams (1995) examined both the validity and reliability of the BBS. Following these psychometric studies, current research on the BBS has focused mainly on comparing the psychometrics of the BBS to other balance measures (eg. Mao, Hsueh, Tang, Sheu, & Hsieh, 2002) and on testing the psychometrics of the short form of the BBS, the BBS-3P (Chou et al., 2006).

Floor/Ceiling Effects

In a study by Mao et al. (2002), three common balance scales (BBS, the Balance subscale of the Fugl-Meyer Assessment, and the Postural Assessment Scale for Stroke Patients) were compared. A significant floor effect was detected in the BBS and the balance subscale of the Fugl- Meyer 14 days after stroke onset. One possible explanation for this effect is that the least demanding item in these two tests is to sit independently. As patients with severe impairments may be unable to do any of the other activities – for example, stand on one foot, step up on a stool etc. – even as they improve, the result is a floor effect (the score does not show change for those with severe impairments). Thus, in those with severe impairments, you may wish to consider using another scale such as the Postural Assessment Scale for Stroke Patients, which is applicable to all stroke patients. The BBS also had a significant ceiling effect at 90 and 180 days after stroke onset for those with higher-level function, suggesting that the BBS may not be able to discriminate balance function after 90 days, the point at which people typically reintegrate into leisure and community activities.


Internal consistency :
Berg et al. (1995) conducted a study to assess the internal consistency of the BBS in both elderly long-term care residents and patients with stroke. The BBS was administered to elderly residents (n=113) at baseline, and at 3, 6 and 9 months, and to patients with stroke (n=70) at 2, 4, 6 and 12 weeks post-stroke onset. At each evaluation, Cronbach’s alphas were greater than 0.83 and 0.97 for the elderly residents and patients with stroke respectively, showing that the BBS has excellent internal consistency.

Inter-rater and intra-rater:
Berg et al. (1995) also assessed inter-rater reliability . Therapists administered the BBS to 35 patients with stroke within 24 hours of the independent rater. Similarly, caregivers were asked to test the elderly residents within one week of the independent rater. To assess intra-rater reliability , 18 residents and 6 stroke patients were assessed one week apart by the same rater. There was agreement between the raters (ICC = 0.98) and the same rater was consistent at two points in time (ICC = 0.97).

Stevenson (2001) examined inter-rater reliability of the BBS among 48 patients with stroke assessed by two different raters at initial assessment (T1) and second assessment within 24 hours (T2). Agreement between T1 and T2 data was excellent (ICC = 0.92).

Flansbjer, Blom & Brogardh (2012) conducted an intra-rater test-retest reproducibility study whereby 50 patients with chronic stroke were assessed with the BBS on 2 occasions, 7 days apart, by one rater. Test-retest reliability was excellent (ICC=0.88). The mean difference of test scores, measured by the Bland and Altman technique, was high and positive (d=0.72), indicating a learning effect.
Note: When performing a Bland and Altman analysis, a mean difference close to zero indicates higher agreement between measurements.

Content validity:
The items were selected based on interviews with 12 geriatric clients and 10 professionals. The list of items was revised following a pretest of all preliminary items.

Construct validity:
Convergent/Discriminant validity:
Seventy acute stroke clients were tested on the BBS, the Barthel Index, and the balance subscale of the Fugl-Meyer Assessment at 4, 6 and 12 weeks post-stroke. Correlations between the BBS and the Barthel Index were excellent (ranging from r = 0.80 to r=0.94, and correlations between the BBS and the balance subscale of the Fugl-Meyer ranged from adequate to excellent (ranging from r = 0.62 to r = 0.94) (Berg et al., 1992).

BBS scores were also reported to correlate with the Functional Independence Measure (r = 0.57 to 0.70) (Juneja, Czyrny, & Linn, 1998) and r = 0.76 (Wee, Bagg, & Palepu, 1999).

Known Groups:
Stevenson (2001) examined the known groups validity of the BBS among 48 patients with acute stroke, using Dunn’s method. Patients were grouped according to Functional Ambulation Category (FAC) scores: ASSIST (FAC score ? 2, requiring physical assistance, n=16), SBA (FAC score = 3, requiring stand-by assistance, n=17) or INDEP (FAC score ? 4, independently ambulant, n=15). There was a significant difference between the INDEP and ASSIST groups (Q = 4.47, p<0.05) and the INDEP and SBA groups (Q = 3.07, p<0.05), but not between the SBA and ASSIST groups.

Criterion Validity:
Concurrent validity:
Flansbjer, Blom & Brogardh (2012) reported excellent relationships between the BBS and the Single-Leg Stance (SLS) in 50 patients with chronic stroke (r=0.65 – 0.79, p<0.001), using Pearson product moment correlation coefficients.

Liston and Brouwer (1996) showed that BBS scores related to dynamic Balance Master measures (r ? 0.48) in 20 ambulatory hemiparetic subjects.

Mao et al. (2002) reported excellent relationships between BBS scores and the balance subscale of the Fugl-Meyer (r = 0.90 to 0.92), and Postural Assessment Scale for Stroke Patients (r = 0.92 to 0.95) at 4 assessment times (14, 30, 90, and 180 days post-stroke).

Tyson and DeSouza (2004) tested the concurrent validities of the sitting section of the Motor Assessment Scale , the Berg Balance Scale and Rivermead Mobility Index using Spearman’s rho and found that BBS scores correlated with the the appropriate comparator tests (r = 0.32 to 0.74), except the weight shift test and step-up tests which did not form significant relationship with Berg Balance Scale (r = 0.26 and 0.19 respectively).

Smith, Hembree, and Thompson (2004) found that the BBS correlated with Functional Reach (r = 0.78). When the relationship between the two measures for subjects with similar motor impairments were examined (based on the four categories of stroke severity from the motor section of the Fugl-Meyer as suggested by Duncan et al. (1992)), correlations differed according to patient severity, with the lowest correlation being for those with moderately severe motor impairment (a score of 36-55 on the Fugl-Meyer; r = 0.24), and the highest correlation for those with moderate motor impairments (a score of 56-79 on the FM; r = 0.80).

Thirty-one elderly clients were measured on the BBS, and on lab measures of postural sway and clinical measures of balance and mobility including the Tinetti Balance Subscale, the Barthel mobility subscale, and the Timed Up and Go Test . Postural sway correlated adequately with the BBS (r = -0.55), and clinical measure correlations ranged from poor (r = -0.46) to excellent (r = -0.67) and were negatively correlated (note: low scores on postural sway and clinical measures indicate normal function, whereas a high score on the BBS indicates normal function, resulting in a negative correlation). Correlation with the Tinetti Balance Subscale was excellent (r = 0.91), and the BBS adequately correlated with the Barthel mobility subscale (r = 0.67). Correlation with the Timed Up and Go Test was excellent and negative (r = -0.76), meaning that a low score on the Timed Up and Go Test (a low score suggests normal functioning) corresponds to a high score on the BBS (a high score indicates balance is intact) (Berg et al., 1992).

In this same study, correlations between scores on the BBS and ratings of 113 residents of a home for the elderly and their caregivers ranged from poor to adequate (elderly: r=0.39 to r=0.41; caregivers: r=0.47 to r=0.61) (Berg et al., 1992).

Predictive validity:
One hundred thirteen elderly individuals were followed for 12 months, and were classified as having 0, 1, and > 2 falls during that time. The relative risk of falling over the next 12 months was 2.7 times more likely in patients who obtained a BBS score < 45 (Berg et al., 1992).

Admission BBS was adequately predictive of length of stay (LOS) in rehabilitation unit (r = -0.39) (this negative relationship suggests that a higher BBS score results in a shorter length of stay) (Juneja et al., 1998).

In Mao et al. (2002), the predictive validity of the BBS was assessed by comparing the results of the BBS at 14, 30, and 90 days after stroke with that of the Motor Assessment Scale at 180 days after stroke by use of Spearman’s correlation coefficient. The scores of the BBS at the earlier 3 days after stroke points were highly correlated with the MAS scores on evaluations on 180 days after stroke (r > 0.8), indicating excellent predictive validity.

Fulk, Reynolds, Mondal & Deutsch (2010) examined the predictive validity of the 6MWT and other widely used clinical measures (FMA LE, self-selected gait-speed, SIS and BBS) in 19 patients with stroke. The BBS was found to not be a significant predictor of mean steps per day (r = 0.54; P = 0.016). Although gait speed and balance were related to walking activity, only the 6MWT was found to be a predictor of community ambulation in patients with stroke.


Flansbjer, Blom & Brogardh (2012) examined the responsiveness of the BBS among 50 patients with chronic stroke who were assessed on 2 occasions, 7 days apart. The standard error of measurement (SEM), i.e. the smallest change that indicates a real improvement for a group of individuals, was 3%. The smallest real difference (SRD) for a single individual was 8%.

Stevenson (2001) examined the responsiveness of the BBS among 48 patients with acute stroke assessed by different raters over three intervals – initial assessment (T1), second assessment within 24 hours (T2), and third assessment after approximately 1 to 2 weeks of intervention (T3). Patients were categorized into one of three groups according to Functional Ambulation Category (FAC) scores: ASSIST (FAC score ≤ 2, requiring physical assistance); SBA (FAC score = 3, requiring stand-by assistance); and INDEP (FAC score ≤ 4, independently ambulant). All groups demonstrated statistically significant increases in BBS performance from T1 to T3 (Wilcoxon Signed Rank Test, W = 77.4 – 106, p ≤ 0.002), but not from T1 to T2. Minimal Detectable Change (MDC) from T1 to T2 = 5.8 (90% CI) indicating that a minimum absolute change score of 6 points represents change in a patient’s BBS performance when assessed by two different raters within 24 hours (all patients ± 6; INDEP group ± 6; SBA group ± 5; ASSIST group ± 7).

Mao et al. (2002) assessed the responsiveness of the BBS, the Balance subscale of the Fugl-Meyer Test, and the Postural Assessment Scale for Stroke Patients by calculating effect size (ES) (dividing the mean change scores by the standard deviation of the change score in the same subjects). The ES showed that the BBS was moderately responsive in detecting changes before 90 days after stroke. The ES for the BBS were greatest in the interval between 14 and 30 days (0.80) and diminished the further one moved through time from the stroke event (30 to 90 days after stroke, ES = 0.69). The ES for the BBS was considered poor at 90-100 days after stroke (ES = 0.40). The changes in the BBS at each stage were significant.

To determine whether the responsiveness of the measures varied depending on the initial stroke-induced deficits, patients were stratified into 1 of the following 3 groups on the basis of their Balance subscale of the Fugl-Meyer Test scores: 0 to 35, severe; 36 to 79, moderate; and 80, mild. The responsiveness of the BBS at different stages for subjects with different levels of stroke severity (ES = 0.21) suggested that the BBS is generally sensitive to change over time after a stroke. However, the BBS was found to be less responsive than the Balance subscale of the Fugl-Meyer Test and Postural Assessment Scale for Stroke Patients for severe patients with stroke, at 14 to 30 days after stroke. The reason for this finding might be that the BBS was not originally designed for patients with stroke, and only one item of the scale assesses balance ability in the sitting position. Because sitting balance is one of the first postures to be restored after a stroke, it seems that the BBS is lacking items to detect change in patients who are unable to stand independently.

Wood-Dauphinee, Berg, Bravo, and Williams (1997) reported an ES of 0.66 for initial 6-week post-stroke evaluation period, ES = 0.25 for 6-12 weeks post-stroke, and an overall ES = 0.97.

Salbach et al. (2001) used standardized response mean (SRM = mean change/standard deviation of change) to estimate the responsiveness of the 5-metre walk test, the 10-metre walk test, the BBS, the Barthel Index , the Stroke Rehabilitation Assessment of Movement, and the Timed Up and Go in 50 subjects with residual gait deficits after a first-time stroke. The SRM from 8-38 days post-stroke for the BBS was 1.04. The BBS was rated as the second most responsive measure (the 5-metre walk test was the most responsive measure) and was recommended for use in patients who have suffered a severe stroke.

English, Hillier, Stiller, and Warden-Flood (2006) investigated the sensitivity of gait speed, the BBS and the Motor Assessment Scale in 78 subjects receiving inpatient rehabilitation following a first or recurrent stroke to detect change over time. Subjects were assessed within one week of admission and one week of discharge. The BBS was sensitive to change (only two patients showed no change) and demonstrated a large ES (d = 1.01).

  • Benaim, C., Pérennou, D. A., Villy, J., Rousseaux, M., Pelissier J. Y. (1999). Validation of a Standardized Assessment of Postural Control in Stroke Patients: The Postural Assessment Scale for Stroke Patients (PASS). Stroke, 30, 1862-1868.
  • Berg, K.O., Wood-Dauphinee, S., Williams, J. L., Maki, B. (1989). Measuring balance in the elderly: Validation of an instrument. Physiotherapy Canada, 41(6), 304-311.
  • Berg, K., Wood-Dauphinee, S. L., Williams, J. I., Maki, B. E. (1992). Measuring balance in the elderly: Validation of an instrument. Canadian Journal of Public Health, 83(S2), S7-S11.
  • Berg, K., Wood-Dauphinee, S. L., Williams, J. I. (1995). The Balance Scale: reliability assessment with elderly residents and patients with an acute stroke. Sc and J Rehabil Med, 27(1), 27-36.
  • Chou, C. Y., Chien, C. W., Hsueh, I. P., Sheu, C. F., Wang, C. H., Hsieh, C. L. (2006). Developing a Short Form of the Berg Balance Scale for People With Stroke. Physical Therapy, 86(2), 195-204.
  • Juneja, J., Czyrny, J. J., Linn, R. T. (1998). Admission balance and outcomes of patients admitted for acute inpatient rehabilitation. Am J Phys Med Rehabil, 77, 388-393.
  • Liston, R., Brouwer, B. J. (1996). Reliability and validity of measures obtained from stroke patients using the balance master. Arch Phys Med Rehabil, 77, 425-430.
  • English, C. K., Hillier, S. L., Stiller, K., Warden-Flood, A. (2006). The sensitivity of three commonly used outcome measures to detect change amongst patients receiving inpatient rehabilitation following stroke. Clin Rehabil, 20(1), 52-55.
  • Fulk, G. D., Reynolds, C., Mondal, S., & Deutsch, J. E. (2010). Predicting home and community walking activity in people with stroke. Arch Phys Med Rehabil, 91, 1582-1586.
  • Flansbjer, U-B., Blom, J., & Brogardh, C. (2012). The reproducibility of Berg Balance Scale and the Single-Leg Stance in chronic stroke and the relationship between the two tests. Physical Medicine & Rehabilitation, 4(3), 165-170.
  • Mao, H. F., Hsueh, I. P., Tang, P. F., Sheu, C. F., Hsieh, C. L. (2002). Analysis and comparison of the psychometric properties of three balance measures for stroke patients. Stroke, 33, 1022.
  • Miyamoto, S. T., Lombardi, I. J., Berg, K. O., Ramos, L. R., Natour, J. (2004). Brazilian version of the Berg balance scale. Braz J Med Biol Res, 37(9), 1411-1421.
  • Salbach, N. M., Mayo, N. E., Higgins, J., Ahmed, S., Finch, L. E., Richards, C. L. (2001). Responsiveness and predictability of gait speed and other disability measures in acute stroke. Arch Phys Med Rehabil, 82, 1204-1212.
  • Scherfer, E., Bohls, C., Freiberger, E., Heise, K. F., Hogan, D. (2006). Berg-Balance-Scale – German Version – Translation of a standardized instrument for the assessment of balance and risk of falling. Physioscience, 2, 59-66.
  • Smith, P. S., Hembree, J. A., Thompson, M. E. (2004). Berg Balance Scale and Functional Reach: Determining the best clinical tool for individuals post acute stroke. Clin Rehabil, 18, 811-818.
  • Stevenson, T.J. (2001). Detecting change in patients with stroke using the Berg Balance Scale. Australian Journal of Physiotherapy, 47, 29-38.
  • Tyson, S. F., De Souza, L. H. (2004). Reliability and validity of functional balance tests post stroke. Clin Rehabil, 18, 916-923.
  • Wee, J. Y. M., Bagg, S. D., Palepu, A. (1999). The Berg Balance Scale as a predictor of length of stay and discharge in an acute stroke rehabilitation setting. Arch Phys Med Rehabil, 80, 448-452.
  • Wood-Dauphinee, S., Berg, K. O., Bravo, G., Williams, J. L. (1997). The Balance Scale: responsiveness to clinically meaningful changes. Can J Rehab, 10, 35-50.