Mini Balance Evaluation Systems Test

Balance Disorders: 

(Godi, et al, 2013; n = 93 patients with balance disorders, mean age = 66.2(13.2) years, 53 female/40 male, Dx: n = 3 Parkinson’s, n = 25 hemiplegia, n = 6 MS, n = 5 vestibular disorders, n = 4 neoplasm CNS, n = 6 neuromuscular disease, n = 8 hereditary ataxia, n = 8 poly neuropathy, n = 6 age-related balance disorders) 

• SEM = 1.26 (1.01-1.65)

Balance Disorders: 

(Godi,et al, 2013) 

• Clinically meaningful change is improvement of 4 points (out of 28 total)

Balance Disorders:

(Godi, et al, 2013; n = 32 patients with balance disorders, mean age = 67.3 (13.5) years, 19 female/13male, Dx: n = 8 Parkinson’s, n = 7 hemiplegia, n = 3 vestibular disorders, n = 4 age-related balance disorders, 10 with “other neuro" Dx) 

• Test-retest 1 to 3 days after baseline assessment. Excellent test-retest reliability ICC = 0.96 (0.94-0.99) 

Balance Disorders:  

(Godi, et al., 2013; n = 32 patients, sample per above) 

• Excellent interrater reliability ICC = 0.98 (0.97-0.99) 

Balance Disorders:

(Franchignoni, et al. 2010; n = 115, mean age = 62.7 (16) years, 62 female/53 male, Dx: n = 21 Parkinson’s, n = 15 neuromuscular disease, n = 14 hereditary ataxia, n = 11 MS, n = 22 hemiparesis, n = 7 peripheral vestibular disorders, n = 6 TBI, n = 10 nonspecific age-related balance disorders, n = 3 encephalopathy, n = 3 cervical myelopathy, n = 2 CNS neoplasm) 

• Factor and Rasch analysis performed to identify optimal test psychometrics for item selection for Mini BESTest. High item separation reliability r = 0.98, Item separation index = 7.35, and Person separation reliability = 0.86 without item redundancy. 

(Godi, et al., 2013; n = 93 patients with balance disorders, mean age = 66.2 (13.2)years, 53 female/40male, Dx: n = 3 Parkinson’s, n = 25 hemiplegia, n = 6 MS, n = 5 vestibular disorders, n = 4 neoplasm CNS, n = 6 neuromuscular disease, n = 8 hereditary ataxia, n = 8 poly neuropathy, n = 6 age-related balance disorders} 

• Excellent internal consistency at baseline testing and follow-up assessment: Cronbach alpha = 0.90 and 0.91 respectively

Balance Disorders:

(Godi, et al., 2013) 

• Excellent concurrent validity of Mini-BESTest with Berg Balance Scale r = 0.85 (CI 0.78-0.90); 
• Excellent validity between Mini-BESTest and mean global rating of change score in rehabilitation r = 0.72 (CI 0.61-0.81)

Adult with Balance Deficits

(Franchignoni, et al., 2010) 

• Internal construct validity supported based on hierarchy of item difficulty (Rasch analysis. Item difficulty ranged from -4 to +2.5 logits)

Adults with Balance Deficits

(Franchignoni, et al., 2010; n = 115 with diverse neurological diagnoses, 53 men, mean age 62.7 ± 16 years, diagnoses included hemiparesis, Parkinson disease, neuromuscular diseases, hereditary ataxia, multiple sclerosis, peripheral vestibular disorders, traumatic brain injury, cervical myelopathy) 

• High content validity since individual items are part of well know balance batteries, such as Berg Balance Scale, Clinical Test of Sensory Integration in Balance, Dynamic Gait Index and TUG. 
• Factor analysis selected 24 out of the 36 original BESTest items likely to represent dynamic balance; Rasch analysis was used to improve rating categories and to delete 10 items that were misfitting or showed local dependency.  The model was verified using confirmatory factor analysis

• Supported based on balance performance items that examine the construct of dynamic balance.

Balance Disorders:

(Godi, et al., 2013) 

• No evidence of floor or ceiling effect in mixed neurologic cohort at baseline mean score = 12.8 (6.9); or after treatment in rehabilitation center mean score = 15.8 (6.9); 2 participants (2.1%) reached Mini BESTest top score.

Balance Disorders : 

(Godi, et al., 2012; n = 93 patients with balance disorders) 

• Subjects had 10 physical therapy treatment sessions for balance disorders. 
• SEM = 1.26 (1.01-1.65); MDC = 3.5 ; MCID = 4 pts; Out of 40 subjects who global rating of change was > 3 (moderate to large improvement) 38/40 (95%) showed Mini-BESTest changes score > 4 pts 
• Excellent ability to identify those subjects with clinically meaningful improvement: AUC = 0.92 (0.84-0.97), sensitivity = 94%, specificity = 81%, discriminative accuracy to detect moderate improvement in balance = 79%
• Discriminatory accuracy excellent, area under ROC curve 0.92, to determine a moderate to large global rating of change improvement.

Chronic Stroke 

(Tsang et al 2013, n = 106 people with stroke, mean age 57.1 ± 11.0 years, 73 men, 33 women and 48 controls, mean age 60.2 ± 9.3 years, 28 men and 20 women)

• Scores of ≤ 17.5 identified those with a history of falling, sensitivity 64%, specificity 64% AUC 0.64

Chronic Stroke

(Tsang et al, 2013)

• Excellent for total score ICC = 0.96, Poor to excellent reliability for individual items kappa = 0.37 to 1.0

Chronic Stroke 

(Tsang et al 2013)

• Excellent for total scores ICC = 0.96, poor to excellent reliability for individual items kappa = 0.36 to 1.0

Chronic Stroke 

(Tsang et al 2013)

• Excellent Cronbach’s alpha = 0.89, 0.93, 0.94 for each rater

Chronic Stroke

(Tsang et al 2013) 

• Excellent correlation with Berg Balance Scale r = 0.83
• Adequate correlation with Functional reach test r = 0.55
• Excellent correlation with one leg stand on paretic side r = 0.83
• Adequate correlation with one leg stand on non-paretic side r = 0.54
• Excellent correlation with Timed “Up & Go” r = -0.82

(Bergstrom et al 2012, n = 9, mean age 78.4 years, range 66-90 years, median 17 months post CVA.  Swedish version)

• Excellent correlation with Berg Balance Scale r = 0.94
• Excellent correlation with Timed “Up & Go” r = -0.89

Chronic Stroke 

(Tsang et al 2013)

• The mini-BESTest discriminates between those with a history of stroke and healthy control subjects (p < 0.001)
• The mini-BESTest discriminates between subjects with chronic stroke with and without a history of falling (p = 0.03)

Chronic Stroke 

(Tsang et al 2013)

• Skewness = - 0.81
• Floor effect: 0 participants with lowest score
• Ceiling effect: 0.9% participants with highest score

Parkinson’s Disease: 

(Leddy et al, 2011, subset of 24 subjects, MDS-UPDRS = 71 ± 21.9, disease duration = 6.9 ± 3.38; Hoehn and Yahr stage 1 – 2 participants, stage 2 – 11 participants, stage 2.5 – 6 participants, stage 3 – 3 participants, stage 4 – 2 participants, 21% were classified as fallers.)

• SEM calculated = 6.16% or 1.99 points

Parkinson’s Disease: 

(Leddy et al, 2011)

• MDC calculated = 17.1% or 5.52 points

Parkinson’s Disease:

(Leddy, et al, 2011; total sample n = 80, mean age = 68.2 (9.3)years, mean H & Y = 2.4 (0.64), 31% were fallers) 

• Preferred Fall risk cut score 63% (20/32 points total) had adequate ability to identify fallers (sensitivity = 0.88, specificity = 0.78). To maximize sensitivity and LR-, a cut score of 72% (23/32) was identified (sensitivity = 0.96, specificity = 0.47) 

(Duncan, Leddy, & Earhart, 2013; n =56 with idiopathic PD, mean age = 69.5 years, 32 male/23 female, H&Y stage (2 = 21, 2.5 = 25, 3 = 9, 4 = 1) mean stage 2.4 (0.5)) 

• Using a cut score of 16/32 points, the Mini-BESTest demonstrated adequate ability to predict 6-month prospective fallers (AUC = 0.80, sensitivity = 0.75, specificity = 0.79, LR+ = 3.57, LR- = 0.32). Mini-BESTest was superior to gait measures to detect fall risk in PD cohort. 

(Duncan,et al, 2012; Baseline n = 80 PD, Six-month evaluation n = 51, mean age = 67.5 (8.8), years post diagnosis 7.7(3.9), H & Y stage 2.4 (0.6), UPDRS 37.8 (13.1), 27% fallers; 12-month evaluation n = 40, mean age 67.3 (9.5), years post diagnosis 7.2 (4.1), H&Y stage 2.3 (0.6), UPDRS 39.3 (13.3), 37% fallers) 

Based on a Cut Score < 20/32 points total (63%); 

• Adequate ability to predict fallers at 6 month prospectively (AUC = 0.87, sensitivity = 0.86, specificity = 0.78, LR+ = 3.97, LR- = 0.18) 
• Adequate ability to predict fallers at 12 months prospectively (AUC = 0.77, sensitivity = 0.62, specificity = 0.74, LR+ = 2.37, LR- = 0.52) 

(Duncan, et al, 2013; n = 80 with idiopathic PD, mean age = 68.2 (9.7), mean MDS-UPDRS 41.3 (14.7), H & Y stage [1 = 4, 2 = 27, 2.5 = 30, 3= 13, 4= 6]) 

Based on a recommended Cut score < 20/32 points to detect fallers: 

• Adequate ability to identify fallers based on retrospective fall report (AUC = 0.86, sensitivity = 0.88, specificity = 0.78, LR+ = 4.03, LR- = 0.15} 
• Adequate ability to predictive fallers based on 6 –month prospective falls (AUC = 0.87, sensitivity = 0.86, specificity = 0.78, LR+ = 3.97, LR- = 0.18); and for 12 month prospective falls (AUC = 0.77, sensitivity = 0.62, specificity = 0.74, LR+ = 2.37, LR- = 0.52) 

(Mak and Auyeung 2013, n = 110, non-recurrent fallers n=86, 34 female, mean age 63.5± 9.3; recurrent fallers n=24, 10 female, mean age 62.2 ± 7.5)

• Scores of ≤ 19 identified recurrent fallers; sensitivity 79%, specificity 67% AUC 0.75

(King et al 2012)

• Scores of ≤ 21 differentiate those with and without postural response deficits with a sensitivity of 89%, specificity 81%

Parkinson’s Disease:

(Leddy et al, 2011; subset of subjects n = 24, MDS-UPDRS = 71 (21.9), disease duration mean 6.9 (3.38), 21% fallers; H & Y stages [1 = 2, 2 = 11, 2.5 = 6, 3 = 3, 4 = 2], 21% n = 5 fallers) 

• Excellent test retest reliability (ICC = 0.92)

Parkinson’s Disease:

(Leddy, et al., 2011; subset of subjects, n = 15 MDS-UPDRS = 74.2 (18.6), disease duration = 6.8 years (3.26), H & Y stages (1 = 2, 2 = 7, 2.5 = 3, 3 = 2, 4 = 1), 20% fallers) 

• Excellent Inter-rater reliability (ICC = 0.91)

Parkinson's Disease:

(King, et al., 2012; n = 97, mean age 65.6 (7.1, time since diagnosis mean = 6.5 (5), H & Y mean = 2.3 (0.6) stages 1-4, UPDRS mean = 31.6 (11.2), 59 male/38 female.) 

• Excellent concurrent validity between Mini-BESTest and Berg (r = 0.79); 
• Adequate correlation between Mini-BESTest and UPDRS-disease severity (r = -0.51). 

(Leddy, et al., 2011; total sample n = 80, mean age = 68.2 (9.3) years, mean H & Y = 2.4 (0.64), 31% were fallers) 

• Mini-BESTest has excellent correlation with BESTest (r = 0.955) 

(Duncan, et al., 2013; n = 80 with idiopathic PD, mean age = 68.2 (9.7), mean MDS-UPDRS 41.3 (14.7), H & Y stage [1 = 4, 2 = 27, 2.5 = 30, 3 = 13, 4 = 6]) 

• Excellent concurrent validity between the Mini-BESTest and Brief BESTest (r = 0.94) 

(McNeely, et al., 2012; n = 22, mean age = 71.3 (7.6), 13 male/9 female, MMSE mean 28 (1.8), disease duration mean 7.0 (4.2) years, mean MED_UPDRS = 39.1 (9.2) off meds, 25.3 (6.9) on meds, mean H & Y stage 2.2 (0.3) 

• Excellent concurrent validity between Mini-BESTest and Berg Balance Scale (r = 0.83); 
• Adequate correlation between Mini-BESTest and Activity Specific Balance Confidence Scale ABC (r = 0.66). 

(McNeely, et al., 2011; n = 23 mean age = 62 (9), mean disease duration = 15 (6) years) 

• Adequate correlation between Mini-BESTest scores and UPDRS postural stability item (item 3.12) (r = -0.44), p < 0.001

(Duncan, et al., 2013; total n = 53, Off meds group n = 28: mean age = 70 (7.4) years, 15 males/13 females, H &Y stage 2 = 8, 2.5 = 15, 3 = 5, 3 fallers & 5 gait freezers; On meds group n = 53, mean age = 68 (8.5) years, 31 males/22 females, H&Y 1 = 2, 2 = 28, 2.5 = 16, 3 = 2, 4 = 5, 11 fallers and 17 gait freezers) 

• Excellent concurrent validity between Mini-BESTest and Four square step test (r = -0.65) p < 0.001  
• Mini-Best had adequate predictive validity to discriminate fallers and nonfallers with PD (AUC = 0.80, sensitivity = 0.88, specificity = 0.78); superior to 4SSS for identifying fall risk in PD. 

(Duncan, Leddy & Earhart, 2011; n = 80 with idiopathic PD, H & Y stages I-IV, age > 40 years, no details on sample) 

• Excellent correlation between Mini-BESTest and Five times sit to stand test (r = -0.71 p < 0.001)
• Multiple regression revealed that Mini-BESTest and Nine Hole peg test explained 53% of variance in FTSTS

(Bergstrom et al 2012, n = 9 with mild to moderate Parkinson disease, mean age 60.3 years, age range 46-85 years, 8 female, Hoehn and Yahr stages I-III)

• Excellent correlation with  Berg Balance Scale r = 0.94
• Excellent correlation with Timed “Up & Go” r = -0.81
• Poor correlation with Falls Efficacy Scale r = 0.26

(Mak and Auyeung 2013)

• The mini-BESTest score was a significant predictor of recurrent falls using multivariate logistic regression.

Parkinson’s Disease:

(Leddy, et al., 2011; total sample n = 80, mean age = 68.2 (9.3) years, mean H & Y = 2.4 (0.64), 31% were fallers) 

• Significant difference between mean Mini BESTest score of fallers= 14.3 (6.2) and nonfallers = 22.9 (5.5) in PD cohort (average difference between groups of 27%)
• Total mini-BESTest scores discriminated between fallers and non-fallers AUC = 0.86
• No specific section of the BESTest or mini-BESTest captured the primary balance deficit

(King, et al., 2012; n = 97 see above) 

• Excellent ability to detect those PD patients with balance deficits based on H & Y stages 1 - 2 versus 3 - 4; AUC = 0.91; Cut off score to distinguish those with and without balance deficits < 21/28 total pts (sensitivity = 89%, specificity = 81%) Mini- BESTest was superior to Berg in discriminating disease severity in PD cohort based on H & Y stage.

Parkinson's Disease

(King, et al., 2012; n = 97, see above) 

• No evidence of ceiling effect, normal distribution of scores in PD cohort with varied H&Y stages, disease severity.
• Skewness -0.93 

(McNeely, et al., 2012; n = 22, see above) 

• No ceiling effect was seen in Mini BESTest scores for PD cohort as compared to ceiling effect seen in Berg scores.

(Bergstom et al., 2012)

• no subjects scored the maximum or minimu on the mini-BESTest