Research Report

Is the BESTest at Its Best? A SuggestedBrief Version Based on InterraterReliability, Validity, InternalConsistency, and TheoreticalConstructParminder K. Padgett, Jesse V. Jacobs, Susan L Kasser

B a c k g r o u n d . The Balance Evaluation Systems Test (BESTest) and Mini-BESTestare clinical examinations of balance impairment, but the tests are lengthy and theMini-BESTest is theoretically inconsistent with the BESTest.

Objec t ive . The purpose of this study was to generate an alternative version of theBESTest that is valid, reliable, time efficient, and founded upon the same theoreticalunderpinnings as the original test.

D e s i g n . This was a cross-sectional study.

M e t h o d s . Three raters evaluated 20 people with and without a neurologicaldiagnosis. Test items with the highest item-section correlations defined the newBrief-BESTest. The vaUdity of the BESTest, the Mini-BESTest, and the new Brief-BESTest to identify people with or without a neurological diagnosis was compared.Ihterrater reliability of the test versions w as evaluated by intraclass correlationcoefficients. Validity was further investigated by determining the ability of eachversion of the examination to identify the fall status of a second cohort of 26 peoplewith and without multiple sclerosis.

Results . Items of hip abductor strength, functional reach, one-leg stance, lateralpush-and-release, standing on foam with eyes closed, and the Timed "Up & Go" Testdefined the Brief-BESTest. Intraclass correlation coefficients for all examination ver-sions were greater than .98. The accuracy of identifying people from the first cohortwith or without a neurological diagnosis was 78% for the BESTest versus 72% for theMini-BESTest or Brief-BESTest. The sensitivity to fallers from the second cohort was100% for the Brief-BESTest, 71% for the Mini-BESTest, and 86% for the BESTest, andall versions exhibited specificity of 95% to 100% to identify nonfallers.

Limitat ions . Eurther testing is needed to improve the generalizability of findings.

C o n c l u s i o n s . Although preliminary, the Brief-BESTest demonstrated reliabilitycomparable to that of the Mini-BESTest and potentially superior sensitivity whilerequiring half the items of the Mini-BESTest and representing all theoretically basedsections of the original BESTest.

P.K. Padgett, DPT, Department ofRehabilitation and Movement Sci-ence, University of Vermont, Burl-ington, Vermont.

J.V. Jacobs, RhD, Department ofRehabilitation and Movement Sci-ence, University of Vermont, 305Rowell BIdg, 106 Carrigan Dr, Bur-lington, VT 05405 (USA). Addressall correspondence to Dr Jacobsat: jjacobs@uvm.edu.

S.L. Kasser, PhD, Department ofRehabilitation and Movement Sci-ence, university of Vermont.

[Padgett PK, Jacobs JV, Kasser SL.Is the BESTest at its best? A sug-gested brief version based oninterrater reliability, validity, inter-nal consistency, and theoreticalconstruct. Phys Ther. 2012;92:1197-1207.]

© 2012 American Physical TherapyAssociation

Published Ahead of Print:June 7, 2012

Accepted: May 30, 2012Submitted: February 12, 2012

Post a Rapid Response tothis article at:ptjournal.apta.org

September 2012 Volume 92 Number 9 Physical Therapy • 1197

The Brief-BESTest: A Suggested Brief Version of the BESTest

I njuries resulting from falls contrib-ute to decreased health statusand increased mortality, particu-

larly for individuals of advanced ageor with chronic disease.' In addition,falls are linked to a reduction in over-all ftinctioning and to early admis-sion to long-term care facilities.^-'Balance impairments, which oftenlead to injurious falls, can be quanti-fied in a clinical setting in orderto direct therapeutic rehabilitationaimed at mitigating an individual'sspecific impairments and minimizingthe risk of falls.

To aid in balance assessment andtherapeutic prescription, several reli-able clinical tools have been devel-oped.'•<''-'' Although widely used,most of these assessments provide ameasure of stability based on a singlecontext of balance impairment.'''-^Research, however, has demonstratedthat postural impairments may beevident across several contexts ofbehavior. • •"' A recently developedbalance assessment tool, the BalanceEvaluation Systems Test (BESTest),examines balance performance in6 specific contexts (or systems, astermed by the test's developers) ofpostural control: mechanical con-straints, limits of stability, anticipa-tory postural adjustments, posturalresponses to an induced loss of bal-ance, sensory orientation, and gait.s'In some cases, balance control maybe compromised by a single balancesystem or subset of systems. TheBESTest allows for the identificationof specific balance systems respon-sible for poor balance performanceand, therefore, can help direct clini-cal interventions.5

The BESTest has been found to bereliable across raters evaluating acohort of individuals with and with-out various neurological diagnoses,and its validity was initially con-firmed on the basis that BESTestscores correlate with reportedscores of balance confidence.'^"

BESTest scores also have been vali-dated to differentiate people withand without fibromyalgia, chronicobstructive pulmonar)' disease, andmultiple sclerosis (MS).'^-''' In addi-tion, the BESTest exhibited hightest-retest and interrater reliabilitywhen used to evaluate participantswith Parkinson disease, and BESTestscores were sensitive to these partic-ipants' prospective or retrospectivefall reports."-''^ This initial literatureshows promise for the BESTest, butits cUnical feasibility is extremelylimited due to the time required tocomplete all 36 items.

To address potential limitations ofthe BESTest's redundancy and lengthytest duration, Franchignoni et aV^identified a subset of the originalBESTest items (ie, the Mini-BESTest),which consists of 16 items that canbe administered in' approximately 15to 20 minutes. The Mini-BESTest hasbeen reported to be just as reliableand capable of identifying fall statusas the BESTest for individuals withParkinson disease.'^ In addition, theMini-BESTest exhibited superiorpsychometric properties comparedwith the Berg Balance Scale for iden-tifying motor impairments in peoplewith Parkinson disease.'?

Although the reduced time to admin-ister the Mini-BESTest renders theexamination more efficient than theoriginal BESTest, anecdotal reportssuggest this assessment remains toolengthy, given increasing constraintson patient contact time in the clinic.In addition, although the Rasch anal-ysis used to define the test offers apowerftil technique to generate anexamination consisting of nonredun-dant items that measure a correlatedconstruct and represent a range ofdifficulty to prevent ceiling orfloor effects, the result was con-trary to the theoretical basis of theBESTest. Specifically, the items defin-ing the Mini-BESTest represent asingular construct (termed "dynamic

balance" by the authors) identifiedby the Rasch analysis, but excludeditems related to mechanical con-straints and to limits of stability.

The construction of the Mini-BESTestthus implies that postural controlrepresents a single construct andthat a clinical assessment need onlyto evaluate this construct, asopposed to the original BESTest,which sought a global assessment ofmultiple constructs that influencebalance impairment. Multiple con-structs influence postural controland may be important for a broaderclinical assessment of balanceimpairment intended for use acrossdiverse clinical populations. Specifi-cally, the Mini-BESTest's lack ofitems assessing mechanical con-straints or limits of stability couldinhibit its sensitivity when appliedto people with musculoskeletalimpairment or impaired limits of sta-bility. Furthermore, without assess-ing these contexts of postural con-trol, clinicians would be uninformedto direct interventions on the basisof these impairments. Indeed, theexisting literature on the BESTestsuggests that mechanical constraintsor limits of stability differentiategroups with and without clinicalhealth conditions or groups witb andwithout a faU history.s'.' -M.iâ Alter-native methods, such as classical testtheory, therefore, may offer anotherapproach to shortening the BESTestbased on its original theoreticalunderpinnings, thereby advancingthe clinical goal of generating a time-efficient balance assessment acrossseveral influential constructs of pos-tural control for use across multipleclinical populations.

This preliminary study, therefore,evaluated the internal consistency ofitems in each section of the BESTestand used item-total correlations toidentify each section's most repre-sentative item. Each section's mostrepresentative item then was

1198 • PhysicalTherapy Volume 92 Number 9 September 2012

The Brief-BESTest: A Suggested Brief Version of the BESTest

included in a new Brief-BESTestexamination. We evaluated each testversion for interrater reliability andits validity to identify individualswith and without a neurologicaldiagnosis. We ftirther tested eachexamination version's validity byevaluating its ability to identify thereported fall history of people withor without MS. This second cohortwas chosen because lower-limbstrength and postural performanceat the limits of stability have beenfound to represent significant predic-tors of future falls in people with MSand because the BESTest subsectionson mechanical constraints and limitsof stability are significantly affectedin people with MS."'•2° Thus, evalu-ating the different examination ver-sions' ability to identify fall status inpeople with and without MS wouldprovide insight into w^hether retain-ing an evaluation of mechanical con-straints and limits of stability isimportant to the utility of the Brief-BESTest versus the Mini-BESTest. Wehypothesized that the Brief-BESTestwould exhibit psychometric proper-ties comparable or superior to thoseof the Mini-BESTest, but with feweritems and with representation fromevery section of the original BESTest.

MethodParticipantsTw enty participants were includedin the first cohort (Tab. 1). Recruit-ment occurred with the intent toinclude participants with a widerange of balance abilities. Partici-pants were included if they wereable to stand independently andambulate 6.1 m (20 ft) with or with-out an assistive device, and werewilling to complete the BESTest(45-60 minutes). No other criteriawere applied. Five of the 20 partici-pants in the first cohort reported atleast 1 fall in the previous 2 months,for a total of 7 falls (range=0-2). Aspart of a larger study,'''•2' thus repre-senting a secondary analysis withinthis study, the second cohort

Table 1.Participant Characteristics of the First Cohort"

Participant No.

1

2

3

4

S

6

7

8

9

10

11

12

13

14

15

15

17

18

19

20

Sex

F

F

F

M

F

M

M

M

M

M

F

M

M

M

F

F

M

M

M

F

Age (y)

60

51

56

75

70

63

83

60

68

63

57

64

75

60

57

64

65

69

51

61

Neurologicai Diagnosis

N/A

N/A

N/A

N/A

Cerebrovascular accident (stroke)

Parkinson disease

Essential tremor

N/A

Parkinson disease

Multiple sclerosis

Multiple sclerosis

Parkinson disease

N/A

Peripheral neuropathy secondary to diabetes

Multiple sclerosis

N/A

Multiple sclerosis

Parkinson disease

N/A

N/A

° N/A=not applicable, F=female, M=male.

included 13 people with MS (8women and 5 men; mean age=50years, range=31-64) and 13 peoplewithout MS (8 women and 5 men;mean age=50 years, range=31-66).People with MS were recruited byadvertisement in the local chapter ofthe National MS Society and wereincluded if they: (1) had neurologist-diagnosed MS, (2) had an ExpandedDisability Status Scale score of lessthan 6, and (3) had no uncorrectedhearing or visual impairments. Peo-ple without MS were recruited byadvertising within the local com-munity and were included if they:(1) had no self-reported neurologi-cal, musculoskeletal, or psycliiatricdisorders; (2) had no uncorrectedhearing or visual impairments; and(3) were matched to the individualswith MS according to sex, similarheight and weight, and within 2years of age. The disease severity of

the participants with MS rangedfrom 0 to 4.5 on the Expanded Dis-ability Status Scale. Seven of the par-ticipants with MS in cohort 2reported at least one fall in the pre-vious 3 months, for a total of 18reported falls (range=0-6 falls; 5participants reported multiple falls).The participants without MSreported no faUs. All individuals gavewritten informed consent to partici-pate in the study.

InstrumentThe BESTest consists of 36 itemsgrouped into 6 specific postural con-trol systems: biomechanical con-straints, stability limits and vertical-ity, anticipatory postural adjustments,postural responses to external per-turbations, sensory orientation dur-ing stance, and stability in gait. Eachitem is scored based on a 4-level ordi-nal scale from 0 to 3. A score of 0

September 2012 Volume 92 Number 9 Physical Therapy • 1199

The Brief-BESTest: A Suggested Brief Version of the BESTest

indicates failure or inabuity to com-plete the task, and a score of 3 indi-cates successful completion of thetask according to all scoring criteria.As such, total scores range from 0 to108, with subsection totals rangingfrom 0 to 15-21 (depending on thenumber of items in the respectivesubsection).

The Mini-BESTest is a subset of 14tasks (16 items due to bilateralassessment) from sections of theBESTest related to anticipatory pos-tural adjustments, reactive posturalresponses, sensory orientation, andstability in gait. The Mini-BESTest'sitems are scored on a 3-level ordinalscale from 0 to 2.

TrainingThe raters were a student in a Doctorof Physical Therapy program and 2doctorate researchers with expertisein the postural control of individualswith balance disorders. Raters pre-pared to administer the BESTest byreviewing the w ritten version of thetest and viewing the accompanyingDVD provided by the test developer.In addition, one of the BESTest'soriginal developers provided a2-hour training session to the raters.The raters practiced administeringand scoring the BESTest on studentand community volunteers. Thesepractice sessions allowed all raters tobecome familiar and comfortablewith the implementation and scoringof the BESTest prior to the studies.

Data CollectionFollowing the training sessions, thefull 36-item BESTest w as adminis-tered to all participants, regardless ofcohort, although all 3 raters concur-rently rated only the first cohort. Thespace was organized to facilitatetransitions from one item to the next

. in order to minimize fatigue andmobility requirements. Five rest peri-ods were offered at regular intervals,and participants were instructed torequest additional rest if needed.

For the first cohort, 1 of the 3 ratersadministered the test while that rateror another rater served as a spotterfor the participant during task per-formance in order to minimize therisk of falling. All raters indepen-dently scored the test for each par-ticipant. One rater evaluated the par-ticipants' BESTest performance forthe second cohort.

Data AnalysisThe data from the first cohort wereanalyzed for internal consistency,validity, and interrater reliabilityusing PASW version 18 (SPSS Inc,Chicago, Illinois). First, Cronbachalpha and item-total correlationswere generated for each BESTest sec-tion and rater. The item with thehighest average item-total correla-tion for each section was selectedfor inclusion in a new Brief-BESTest.If the item assessed a lateralizedbehavior with a companion item thatassesses behavior on the other side(ie, items that are performed on theleft and right sides), both items wereincluded in the Brief-BESTest.

After establishing the Brief-BESTest,total scores were calculated for eachversion of the BESTtest. For the Mini-BESTest, total scores were generatedby transforming scores from theBESTest's 4-point ordinal scale tothe Mini-BESTest's 3-point scale.Item-total correlations and Cronbachalpha also w ere reported for theBrief-BESTest and Mini-BESTest toconfirm internal consistency andeach item's contribution to therespective examination's total score.Interrater reliability of each testversion was analyzed using 2-way,mixed-model intraclass correlationcoefficients (ICC) testing for abso-lute agreement. Validity was initiallyassessed from the data of the firstcohort by single-variable logisticregression models to determine thesensitivity, specificity, overall accu-racy, and positive and negative like-lihood ratios of each rater's total

BESTest, Mini-BESTest, and Brief-BESTest scores to identify partici-pants with or without a netirologicaldiagnosis. Similar logistic regressionmodels were used to determine thesensitivity, specificity, overall accu-racy, and positive and negative like-lihood ratios of the second cohort'sscores for each examination versionto identify participants w ith orwithout a reported fall history (ie,whether they reported experiencingat least 1 fall in the previous 3months).

ResultsInternal Consistency andItem-Total CorrelationsThe average Cronbach alpha coeffi-cients for each section of theBESTest were .839, .621, .874, .863,.813, and .920 for mechanical con-straints, limits of stability and verti-cality, anticipatory postural adjust-ments, postural responses, sensoryorientation, and gait, respectively.The items with the highest item-totalcorrelation coefficients to theirrespective section totals were hipabduction, forward reaching, single-leg stance, lateral compensatorystepping, standing with eyes closedon foam, and the Timed "Up & Go"Test (Eig. 1). These items, therefore,defined the Brief-BESTest. It shouldbe noted that single-leg stance wasselected as the representative itemfor the section on anticipatory pos-tural adjustments based on an item-total correlation of .805, which wasjust slightly higher than the value of.800 elicited by the rise-to-toes item.

Cronbach alpha and item-total corre-lations for the Mini-BESTest andBrief-BESTest are identified in Table2. Cronbach alpha was higher forthe Mini-BESTest than for the Brief-BESTest in both the first and secondcohorts, but both versions exhibitedvalues above .85. On average, theitem-total correlations were .732 forthe Mini-BESTest and .737 for theBrief-BESTest in the first cohort and

1200 • PhysicalTherapy Volume 92 Number 9 September 2012

The Brief-BESTest: A Suggested Brief Version of the BESTest

c

(UO

uo

.9-

.8-

.6-

15 .5-

o 4

•'S

Oí 2

. 1 -

MechanicalConstraints

çê£?

Stability Limits/Verticality

AnticipatoryPostural

Adjustments

PosturalResponses

SensoryOrientation

Stability inGait

a iai

CO

o < ..ï '^

o

Figure 1.Item-total correlation coefficients for the items of the Balance Evaluation Systems Test (BESTest) to their respective section totals. Theblack bars represent the 8 items that were included in the Brief-BESTest on the basis of having the highest item-total correlations.

.617 for both the Mini-BESTest andBrief-BESTest in the second cohort.

Interrater Reliability and ValidityAll 3 versions of the examination(score distributions identified inFig. 2) exhibited very strong levels ofinterrater reliability: ICC (95% confi-dence interval [CI])=.985 (.959-.994) for the BESTest, .995 (.988-.998) for the Mini-BESTest, and .994(.986-.997) for the Brief-BESTest.Total BESTest, Mini-BESTest, andBrief-BESTest scores significantly dif-ferentiated people from the firstcohort with and without diagnosedneurological disorders or injuries(Tab. 3). BESTest scores w^ere more

sensitive than the mini or brief ver-sions to identify people with neuro-logical disorders, whereas levels ofspecificity were similar among allversions of the examination. Therelative sensitivity and specificity ofthe Mini-BESTest versus the Brief-BESTest depended on the rater, butthe Brief-BESTest's average sensitiv-ity and specificity were 3% higherand 4% lower, respectively, thanthe Mini-BESTest's sensitivity andspecificity.

For the second cohort of 26 individ-uals with and without MS, the Brief-BESTest was 100% accurate in iden-tifying whether the participants

reported no falls or at least 1 fall inthe previous 3 months. The Mini-BESTest and original BESTest alsoprovided high levels of specificity forpeople without a faU history, butexhibited lower sensitivities for peo-ple wiûi a fall history than the Brief-BESTest (Tab. 4).

DiscussionThe results support our hypothesisthat the Brief-BESTest, defined fromitems with the highest item-totalcorrelations, exhibits psychometricproperties comparable or superior tothose of the Mini-BESTest. Althoughthe original BESTest better identifiedpeople with a neurological diagnosis

September 2012 Volume 92 Number 9 Physical Therapy • 1201

The Brief-BESTest: A Suggested Brief Version of the BESTest

Table 2.Internal Consistency of the Mini-BESTest and Brief-BESTest Based on Cronbach Alphaand Item-Total Correlations

Item

Item-Total Correlation Coefficients

Cohort 1 Cohort 2

Mini-BESTest (Cronbach alpha=.949 for cohort 1 and .90S for cohort 2)

Sit-to-stand

Rise-to-toes

Left single-leg stance

Right single-leg stance

Forward compensatory stepping

Backward compensatory stepping

Leftward compensatory stepping

Rightward compensatory stepping

Standing with eyes open on firm surface

Standing with eyes closed on foam surface

Incline stance

Change gait speed

Gait with head turns

Gait with pivot turn

Gait over obstacle

Dual-task Timed "Up & Go" Test

.626

.753

.653

.742

.807

.631

.875

.834

.434

.684

.725

.699

.858

.591

.699

.647

Excluded due to lack of variability

.740 -

.765

.611

.813

.746

.888

.495

.534

.627

.630

.569

.837

.591

.124

Brief-BESTest (Cronbacn alpha=.917 for cohort 1 and .856 for cohort 2)

Hip abduction

Forward functional reach

Left single-leg stance

Right single-leg stance

Leftward compensatory stepping

Rightward compensatory stepping

Standing with eyes closed on foam surface

Timed "Up & Go" Test

.755

.330

.708

.760

.904

.820

.820

.799

.579

.495

.866

.810

.605

.619

.401

.562

compared with either abbreviatedversion of the examination, the Brief-BESTest and Mini-BESTest exhibitedvery similar levels of overall accu-racy. The Brief-BESTest offered thehighest sensitivity and overall accu-racy to identify people with andwithout MS who reported at least 1fall in the previous 3 months. In addi-tion, all 3 versions of the examina-tion exhibited ver>' high levels ofinterrater reliability. Thus, the Brief-BESTest offers an even more abbre-viated alternative to the Mini-BESTestwith similar or superior psychomet-ric properties.

The primary objective of this studywas to shorten the BESTest in a the-oretically consistent manner, withsecondary objectives to provide pre-liminary comparisons of each ver-sion's psychometric properties. Thepreliminary validity analysis to iden-tify people with and without diag-nosed neurological disorders is notintended to suggest that any versionof the BESTest would be used todiagnose the existence of neurologi-cal disorders. The analysis wasinstead chosen to demonstrate eachexamination version's ability to dif-ferentiate people with and without a

neurological diagnosis regardless ofpathology, thereby supporting theiruse as a generalized balance assess-ment across the sampled clinicaldiagnoses.

The 8 selected items of the Brief-BESTest also exhibit face validity forrepresenting each context of bal-ance impairment beyond providingthe highest item-total correlationsfor their respective contexts ofimpairment. In previously publishedresearch, each item has been shownto discriminate groups with andwithout neurological diagnoses, toassociate with or predict falls orfractures, or to associate with eco-logical tasks or limited participationin activities of daily living.2°'22-4o i^addition, the Brief-BESTest includesmany assessments reported to bemost frequently executed by physi-cal therapists (eg, one-leg stance,functional reach, and Timed "Up &Go" Test).-*' Thus, the Brief-BESTestitems appear to provide valid repre-sentative assessments for each con-text of balance impairment assessedby the original BESTest.

Both the Brief-BESTest and Mini-BESTest eUcited a Cronbach alphaof greater than .85, regardless ofcohort, although this measure ofinternal consistency was higher forthe Mini-BESTest. This finding is notsurprising, as the Mini-BESTest w asderived based on a Rasch analysisdesigned to identify items repre-senting a single construct, and ouranalysis confirms its internal consis-tency. In contrast, the Brief-BESTestwas defined from the items moststrongly representative of each ofthe BESTest's section scores that rep-resent different contexts of posturalcontrol. The Brief-BESTest's itemsappear similarly associated to thetotal score as those of the Mini-BESTest are to its total score. Thus,although including items from all 6contexts of postural control exam-ined by the original BESTest, the

1202 • Physical Therapy Volume 92 Number 9 September 2012

The Brief-BESTest: A Suggested Brief Version of the BESTest

selected examination items appearappropriate for assessing balanceimpairment.

The analysis of fall history on peoplewith and without MS demonstrateseither the importance of the Brief-BESTest's retention of all 6 contextsof postural control or that the Mini-BESTest includes additional itemsthat diminish its sensitivity to falls.Although further testing is necessaryto identify which is true, measuresof lower-limb strength and the abilityto maintain balance at the limits ofstability are associated with falling inpeople with MS, and the 2 BESTestsubsections are significantly differ-ent between people with and with-out MS.i-' o Thus, although theimproved sensitivity of the Brief-BESTest may have been due to theremoval of insensitive items ratherthan the retention of items related tomechanical constraints or limits ofstability, including relevant contextsof balance impairment in, peoplewith MS likely contributed to itscombined sensitivity and specificityto fall history.

Thé Mini-BESTest and BESTest havepreviously been reported to elicit86% and 84% accuracy, respectively,in identifying fallers and nonfallerswith Parkinson disease."' Thus, ourresults suggest all versions of theBESTest could provide similar orhigher levels of accuracy to identifythe fall status of people with MS. Alarger study is needed to confirm thisfinding, and ftirther testing acrossmultiple patient populations remainsnecessary to determine each test ver-sion's relative capability to serve as afalls screening tool.

When evaluating the internal consis-tency of the BESTest, items with par-tictilarly low correlations to theirsection totals were generally thosewith little variability across partici-pants and that exhibited ceilingeffects (eg, base of support, standing

16-,

14-

• | i o -

t 8-

I ^2 -

0

16-

14-

12-

10-

8 -

6-

4 -

2 -

0

16-

14-

I2 -

0

Original BESTest

n With Neurologic Diagnosis• No Neurologic Diagnosis

n II n nO'

Total Score (% Possible)

Mini-BESTest

O With Neurologic Diagnosis• No Neurologic Diagnosis

Total Score (% Possible)

Brief-BESTest

D With Neurologic Diagnosis• No Neurologic Diagnosis

n n n n n tl

Total Score (% Possible)

Figure 2.Frequency histograms of scores (percentage of possible maximum) from the first andsecond cohorts on the original Balance Evaluation Systems Test (BESTest) (top), Mini-BESTest (middle), and the proposed Brief-BESTest (bottom). With the combinedcohorts, scores represent 24 individuals with a neurological diagnosis (white bars) and22 without a neurological diagnosis (black bars).

September 2012 Volume 92 Number 9 PhysicalTherapy • 1203

The Brief BESTest: A Suggested Brief Version of the BESTest

Table 3.Ability of the BESTest, Mini-BESTest, and Brief-BESTest to Differentiate People With and Without a Neurological Diagnosis forEach Rater

Rater No.

BESTest total score

1

2

3

Mini-BESTest total score

1

2

3

Brief-BESTest total score

1

2

3

Mean (9S% Confidence intervai)Percent Totai Score

Croup WithoutNeuroiogicai

Diagnosis

95(91-98)

89.(87-91)

92 (88-97)

90 (86-95)

90 (85-94)

90 (84-96)

91 (83-100)

84 (80-87)

90 (82-97)

Croup WithNeuroiogicai

Diagnosis

74(61-88)

72 (59-84)

73(61-86)

64(46-81)

61 (43-78)

59(41-77)

59(39-79)

57 (38-76)

61 (43-80)

Regression Modei) ^ Vaiue(P Vaiue)

13.60 (.00023)

11.61 (.00066)

10.37 (.0013)

11.54 (.00068)

10.07 (.0015)

13.09 (.00030)

10.04 (.0015)

9.72 (.0018)

9.50 (.0021)

% Sensitivity,% Specificity,% Accuracy

82, 67, 75

82, 78, 80

82, 78, 80

64, 78, 70

64, 78, 70

64, 89, 75

64, 78, 70

64, 89, 75

74, 67, 70

Positive, NegativeLiiteiihood Ratios

2.48, 0.27

3.73, 0.23 •

3.73, 0.23

2.91, 0.46

2.91, 0.46

5.82, 0.40

2.91, 0.46

5.82, 0.40

2.24, 0.39

Table 4.Ability of the BESTest, Mini-BESTest, and Brief-BESTest to Differentiate People With and Without a Self-Reported Recent FallHistory From the Second Cohort

Test Version

Brief-BESTest

Mini-BESTest

BESTest

Mean (95% Confidence intervai)Percent Totai Score

Croup WithFail History

59 (45-73)

68 (52-83)

77 (66-87)

Croup WithoutFaii History

95 (92-98)

94(91-96)

96 (94-98)

Regression Modei) ^ Vaiue(P Value)

30.29 (<.OOOO1)

20.61 (<.OOOO1)

24.47 (<.OOOO1)

% Sensitivity,% Specificity,% Accuracy

100, 100, 100

71, 100, 92

86, 95, 92

Positive, NegativeLiiceiihood Ratios

Infinity, 0.00

Infinity, 0.29

1 7.20, 0.46

on a firm surface, feet-in-placeresponses). In addition, although theassessments of verticaUty and stabil-ity limits are included in the samesection, those on verticality did notcorrelate well with their sectionscores, nor did they correlate well inan exploratory analysis with the sen-sory orientation scores (not shown).The interrater reliability of the verti-cality items also was reportedly lowin previous research,» suggestingthese items do not provide ver)'meaningful contributions to theexamination.

Unexpectedly, the dual-task Timed"Up & Go" Test provided the lowestitem-total correlations for the gait

section despite the Timed "Up & Go"Test providing the highest correla-tions. One potential reason may bethat the examination was challeng-ing for most participants, with dual-task costs on either walking speedor counting being evident for mostparticipants. Instructional standard-ization for attentional focus also maybe an important factor,'' as the par-ticipants could have differentiallyprioritized either task. Given that thescoring of this item diminishes basedon impaired performance of eitheror both tasks and our itemwise inter-rater reliability for the item was ade-quate (Kendall W= .76), this was nota likely cause of low item-total cor-relations. Alternatively, the addition

of a dual task may represent anothersystem of impairment that repre-sents cognitive-motor interaction. Itw ould be of interest to evaluate dual-task analogs to multiple tasks (reach-ing, one-leg stance, stance on firmand foam surfaces, and the Timed"Up & Go" Test) in order to deter-mine whether dual-task impairmentrepresents a unique context ofimpairment.

ConclusionsGiven the economics of clinical eval-uation allowing an extremely limitedamount of patient-clinician contacttime, it becomes even more impera-tive to develop an efficient examina-tion. At the same time, the presenta-

1204 • Physical Therapy Volume 92 Number 9 September 2012

The Brief-BESTest: A Suggested Brief Version of the BESTest

tion of balance impairment ismultifactorial, and the consequencesof impaired balance are deserving ofadequate assessment. This study con-firms the validity and reliability ofboth the BESTest and Mini-BESTestfor raters in research and raters intraining as physical therapists, butalso provides initial support for theBrief-BESTest. The Brief-BESTest elic-ited equally high levels of interraterreliability as the existing versions ofthe examination, its ability to differ-entiate individuals w ith and w^ithoutdiagnosed neurologic disorders orinjuries w as similar to that of theMini-BESTest, and its ability to differ-entiate people with and without MSbased on fall history was superiorto that of either existing version.These psychometric properties arebalanced by the most clinically feasi-ble combination of only 8 scoreditems (compared with the 16 or 36items of the other versions) whileremaining theoretically intact withrepresentative items from all 6 sys-tems of postural control assessed bythe original BESTest. The findings,however, are preliminary, represent-ing a modest number of participantsin both cohorts. Eurther use andexamination of the Brief-BESTest areneeded to confirm its validity andgeneralizability, as well as to identifyclinically useful cutoff scores.

All authors provided concept/idea/researchdesign, writing, and data collection andanalysis. Dr Jacobs provided project man-agement. Dr Padgett and Dr Kasser providedstudy participants. Dr Jacobs and Dr Kasserprovided facilities/equipment. The authorsacknowledge Dr Fay Horak for providingtraining.

These data were presented as a poster andabstract at the Joint World Congress of theInternational Society for Posture & GaitResearch (ISPCR) and Gait & Mental Func-tion; June 24-28, 2012; Trondheim, Norway.

The study was funded by the University ofVermont's Department of Rehabilitation andMovement Science.

DOI: 10.2S22/ptj.201200S6

References1 Centers for Disease Control and Preven-

tion, National Center for Injur)' Preventionand Control. Web-based Injur>' StatisticsQuer)' and Reporting System (WISQARS).2005. Available at: http://www.cdc.gov/ncipc/wisqars. Accessed March 24, 2010.

2 Bro\vn AP. Reducing falls in elderly peo-ple: a review of exercise Interventions.Physiother Theory Pract. 1999; 15:59-68.

3 Tinetti ME. Performance-oriented assess-ment of mobilit)' problems in elderlypatients. / Am Geriatries Soc. 1986;34:119-126.

4 Stevens JA, Corso PS, Finkelstein EA, MillerTR. The costs of fatal and nonfatal fallsamong older adults. Inj Prev. 2006;12:290-295.

5 inetti ME, Williams CS. Falls, injuries dueto falls, and the risk of admission to a nurs-ing home. N Engl J Med. 1997;337:1279-1284.

6 Tyson SF. How to measure balance in clin-ical practice: a systematic review of thepsychometrics and clinical utilit)' of mea-sures of balance activity for neurologicalconditions. Clin Rehabil. 2009;23:824-840.

7 Mancini M, Horak FB. The relevance ofclinical balance assessment tools to differ-entiate balance deficits. EurJ Phys Reha-bil Med. 2010;46:239-248.

8 Yelnik, A, Bonan I. Clinical tools for assess-ing balance disorders. Clin Neurophysiol.2008;38:439-445.

9 Horak FB, Wrisley DM, Frank J. The Bal-ance Evaluation Systems Test (BESTest) todifferentiate balance deficits. Phys TTher.2009:89:484-498.

10 Shumway-Cook A, WooUcott MH. MotorControl: Theoiy and Praetieal Applica-tions. Baltimore, MD: Williams & Wilkins;1995:119-140.

11 Opheim A, Jahnsen R, Olsson E, StanghelleJK. Balance in relation to walking deterio-ration in adults with spastic bilateral cere-bral pals)'. Phys Ther. 2012;92:279-288.

12 Jones KD, Horak FB, Winters-Stone K,et al. Fibromyalgia is associated withimpaired balance and falls./ Clin Rheuma-tol. 2009;15:l6-21.

13 Beauchamp MK, Sibley KM, Lakhani B,et al. Impairments in systems underlyingcontrol of balance in COPD. Chest. 2011Nov 23 [Epub ahead of print].

14 Jacobs JV, Kasser SL. Balance Impairmentin people with multiple sclerosis: prelim-inary evidence for the Balance EvaluationSystems Test. Gait Posture. 2012 Apr 21[Epub ahead of print).

15 Leddy AL, Crowner BE, Earhart GM. Func-tional Gait Assessment and Balance Evalu-ation System Test: reliabiUt)', validity, sen-sitivity, and specificity for identifyingindividuals with Parkinson disease whofall. Phys Ther. 2011;91:102-113.

16 Leddy AL, Crowner BE, Earhart GM. Utility'of the Mini-BESTest, BESTest, and BESTestsections for balance assessments in indi-viduals with Parkinson disease. / NeurolPhys Ther. 2011;35:90-97.

17 Duncan RP, Leddy AL, Cavanaugh JT, et al.Accuracy of fall prediction' in Parkinsondisease: six-month and 12-month prospec-tive analyses. Parkinsons Dis. 2012;2012:237673.

18 Franchignoni FB, Horak FB, Godi M, et al.Using psychometric techniques toimprove the Balance Evaluations SystemsTest: the Mini-BESTest. / Rehabil Med.2O1O;42:323-331.

19 King LA, Priest KC, Salarian A, et al. Com-paring the Mini-BESTest with the BergBalance Scale to evaluate balance disor-ders in Parkinson's disease. ParkinsonsDis. 2012;2012:3754l9.

20 Kasser SL, Jacobs JV, Foley JT, et al. A pro-spective evaluation of balance, gait, andstrength to predict falling in women withmultiple sclerosis. Areh Phys Med Reha-bil. 20ll;92:1840-1846.

21 Jacobs JV, Kasser SL. Effects of dual taskingon the postural performance of peoplewith and without multiple sclerosis: a pilotstudy. J Neurol. 2012;259:l 166-1176.

22 Hilliard MJ, Martinez KM, Janssen I, et al.Lateral balance factors predict future fallsin community-living older adults. ArchPhys Med Rehabil. 2008;89:1708-1713.

23 Nadollek H, Brauer S, Isles R. Outcomesafter trans-tibial amputation: the relation-ship between quiet stance abilit)', strengthof hip abductor muscles and gait. Phys-iother Res Int. 2002;7:203-2l4.

24 Jenkins ME, Johnson AM, Holmes JD, et al.Predictive validity of the UPDRS posturalstability score, and the Functional ReachTest, when compared with ecologicallyvalid reaching tasks. Parkinsonism RelatDisord. 2010;l6:409-4ll.

25 Weiner DK, Duncan PW, Chandler J,Studenski SA. Functional reach: a markerof physical frailt)'. / Am Geriatr Soeiety.1992;40:203-207.

26 Frzovic D, Morris ME, Vowels L. Clinicaltests of standing balance: performance ofpersons with multiple sclerosis. Areh PhysMed Rehabil. 2000;81:215-221.

27 Mak MK, Pang MY. Parkinsonian singlefallers versus recurrent fallers: differentfall characteristics and clinical features.J Neurol. 2010;257:1543-1551

28 Muir SW, Berg K, Chesworth B, et al. Bal-ance impairment as a risk factor for fallsin communit)'-dwelling older adults whoare high functioning: a prospective study.Phys Ther. 2010;90:338-347.

29 Komatsu T, Jung Kim K, Kaminai T, et al.Clinical factors as predictors of the risk offalls and subsequent bone fractures due toosteoporosis in postmenopausal women./ Bone Miner Metab. 2006;24:4l9 - 424.

30 Jacobs JV, Horak FB, Tran VK, Nutt JG.Multiple baliuice tests improve the assess-ment of postural stabilit)' in subjects withParkinson's disease. / Neurol NeurosurgPsyehiatiy. 2006;77:322-326.

31 Maki BE, Edmondstone MA, McIlroy WE.Age-related differences in laterallydirected compensator)' stepping behavior./ Gerontol A Biol Sei Med Sei. 2000;55:M270-M277.

September 2012 Volume 92 Number 9 PhysicalTherapy • 120S

The Brief-BESTest: A Suggested Brief Version of the BESTest

32 Mansfield A, Peters AL, Liu BA, Maki BE. 36 Buatois S, Guegtien R, Gatichard GC, et al. 40 Dite W, Connor HJ. Curtis HC. ClinicalEffect of a perturbation-based balance Posturography and risk of recurrent falls in identification of multiple fall risk eariytraining program on compensator)' step- healthy non-institutionalized persons aged itfter unilateral transtibial amputation,ping and grasping reactions in older over 65. Gerontology. 2006;52:i45-i52. Arch Phys Med Rehabil. 2007,88.109-114.

37 Cohen HS, Kimball KT. Usefulness of 4 l Sibley KM, Straus SE, IiHiess EL, et al. Bal-some current balance tests for identifying anee assessment practices and use of stan-

33 King LA, Horak FB. Lateral stepping for individuals with diseciuilibrium due to ves- dardized balance measures among Ontariopostural correction in Parkinson's disease. tibular impairments./Ke5íííi/?e.s. 2008;18: physical therapists. Phys Ther. 2011;91:Arch Phj'S Med Rehabil. 2008-89.492-499. 295-303. 1583-1591.

34 Winters-Stone KM, Torgrimson B, Honik 38 Herman T, Giladi N, Hausdorff JM. Prop- 42 Abernethy B. Dual-task methodolog)' andFB, et al. Identifying factors associated erties of the "timed up and go" test: more motor skills research: some methodologi-with falls in postmenopausal breast cancer than meets the eye. Gerontology. 2011; cai constraints. Journal of Human Move-survivors: a multi-disciplinary approach. 37:203-210. ment Studies. 1988,14:101-152.Arch Phys Med Rehabil. 2011;92:646-652. 3^, ^^„^^^ j c,^^„g „ diamond B, et al. The

35 Frenklach A, Louie S, Koop MM, Bronte- Timed Up and Go test: predicting falls inStewart H. Excessive postural sway and ALS. Amyotroph Lateral Scler. 2007;8:the risk of falls at different stages of Par- 292-295.kinson's disease. Mov Disord. 2009;24:377-385.

1206 • Physical Therapy Volume 92 Number 9 September 2012

The Brief BESTest: A Suggested Brief Version of the BESTest

Appendix.Scoring Form for the Brief Balance Evaluation Systems Test (Brief-BESTest).°

Brief Balance Evaluation Systems Test Patient/Subject:Examiner:Date:

General Note: "instability" is defined as using more than an ankle strategy to maintain balance (eg, a hip strategy isused).

Section 1. Biomechanical Constraints

Item 1 : Hip/Trunk Lateral Strength"Rest fingertips in my hands while you lift your leg to the side

and hold, keep trunk vertical. You will hold for 10 s."Count 10 s, watch for straight knee; if they use moderate torce

on your hands, score as "without keeping trunk vertical."

(3) Normal (10 s with trunk vertical)(2) Mild (10 s without trunk vertical)(1) Moderate (1 hip abducts with trunk vertical)(0) Severe (neither hip, 10 s and vertical or not

vertical)

Section ¡I. Stabiiity Limits

Item 2: Functional Reach Forward ."Stand normally; lift both arms straight in front of you; reach

as far forward as you can with arms parallel to the rulerwithout lifting your heels."

2 attemptsObserve that patient does not lift heels, rotate trunk, or

protract scapula.Watch for vertical initial alignment. Record best reach.

(3) >32 cm (12.5 in)(2) 16.5-32 cm (6.5-12.5 in)(1) <16.5 cm (6.5 in)(0) No measurable lean (or must be caught)

Trial 1 (cm or in)

Trial 2 (cm or in)

Section Iii. Transitions-Anticipatory Posturai Adjustment

Items 3 and 4: Stand on One Leg-Left and Right"Look ahead; hands must stay on hips; bend one leg behind

you; stand on 1 leg as long as you can for up to 30 s. Donot let your lifted leg touch the other leg."

Allow 2 attempts, record best attempt; record time up to 30 s(stop time if hands off hips or leg on floor or leg touchessupporting leg).

(3) Normal (stable >20 s)(2) Trunk motion OR 10-20 s(1) Stand 2-10 s(0) Unable

Left

Seconds

Right

Seconds

Section iV. Reactive Posturai Response

Items 5 and 6: Compensatory Stepping-Lateral, Left and Right"Stand with feet nearly together; lean into my hands; 1 will

remove my hands; do whatever necessary to keep balance,trying to take 1 step."

Note: Stand next to and behind participant. Place hand ongreater trochanter and brace yourself to hold the person'sweight shifted to supported leg.

(3) Recovers with 1 side/crossover step(2) Several steps to recover independently(1) Steps but needs assist to prevent fall(0) No step OR falls

Left

Right

Section V. Sensory Orientation

Item 7: Stance With Eyes Closed, on Foam Surface"Stand on foam with your eyes closed, your hands on your

hips, and your feet close but not touching. Start by lookingstraight ahead, and 1 will start timing when you close youreyes. Stay as stable as possible and try to keep your eyesclosed for the entire time. The goal is 30 s."

Two trials, if necessary. Patient must step off foam betweentrials.

Section VI. Stabiiity in Cait

Item 8: Timed "Up St Go" Test"When 1 say 'go,' stand up and walk quickly but safely to the

tape, turn, and walk back and sit in chair."Start with back against chair, stop timing when buttocks hit

the chair; chair should have arms to push from, if necessary.Imbalance might include trips or lateral/backward stumblesor crossovers.

(3) 30 s stable(2) 30 s unstable(1)<30s(0) Unable

(3) Fast, <11 s, good balance(2) Slow, >11 s, good balance(1) Fast, <11 s, imbalance(0) Slow, >11 s, imbalance

TOTAL:

Trial 1 (s)

Trial 2 (s)

Time (s)

The scoring form for the Brief-BESTest examination may not be used or reproduced without written permission of the authors.

September 2012 Volume 92 Number 9 Physical Therapy • 1207

Copyright of Physical Therapy is the property of American Physical Therapy Association and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.